Oceanography Posters

In this work, we introduce the hydrodynamics of the Suez canal in 1-D and 2-D to study the behaviour of the water dynamics after the new suez canal in 2015. The suez canal is one of the most important water ways for the international trade. The environmental impact of the suez canal on the mediatranian sea is too high which drived by the water logon from the red sea with different salinity. indeed, the salinity profile after the new suez canal building became a crucial parameter in the environmental hazard of the new project. and left us with more doupt about the environmental impact of this project on the surronding environment. we provide a multi-displines integration to well understand the salinity profile by using mathematical modeling tools, remote sensing data, and in-situ data for more realastic simulation.

Urmia Lake as a most vital water bodies in Iran, has been shrinking since the late twentieth century and its area has dramatically decreased. To develop and apply any plans to survive the lake, qualitative and quantitative analysis and any modeling, deriving physical information such as volume, area and their changes are very crucial. The objectives of this study were therefore, intended firstly, to study the bathymetry of Urmia Lake with a more satisfactory approach using Landsat- LDCM satellite image and in situ measurement data. The polynomial model was developed to predict the water depth in Urmia Lake area. This model was developed with the input series of reflectance values from blue, green, red and NIR bands in the Landsat- LDCM satellite imagery for Urmia Lake taken on 12 April 2013 of the sampling sites from actual depth measured were taken on the same date. Also, using a large archive of Landsat imagery (TM, ETM+ and LDCM), a counter of equivalent elevation were established for deriving the bathymetry of desiccated areas by mapping the edges of the lake and finally assembled with bathymetry derived from polynomial model. In-situ depth measurements were used to evaluate resultant derived bathymetric map. This comparison shows reasonable agreement between the Landsat-derived depths and those measured in the field with RMSE of 0.27 cm and R²=0.91. The maximum and mean depths measured were 4.9 and 11 m respectively. The maximum depth measured was located at the upper part of the lake.

Secondly, to make stage curves of lake, multi-temporal changes of water body have been derived from Landsat, MODIS and AVHRR satellite image sets since 1972. In this regard, the area of Urmia Lake at different level was estimated base on object oriented and pixel base classification using 78 satellite images. Finally, stage curve (volume-area-elevation relations) was derived from Bathymetry map. These findings will provide valuable information and can be utilized for various water related resource management systems and environmental impact assessments.

Seasonal and Extended range prediction of Indian summer monsoon has been a challenging task for meteorological communities. It is a well established fact that accurate representation of ocean state is very essential for better seasonal as well as short term weather forecast. Any inaccuracy in the representation of upper ocean thermal structure in the ocean initial state will lead to drastic errors in coupled model forecast. The data assimilation techniques have promising role in achieving an accurate ocean state in the context of recent advances in ocean and atmospheric observational networks. The structure of ocean salinity controls the density field and thereby playing a major role in influencing the ocean dynamics. It has been a challenging task to understand the strong intra-seasonal and inter-annual variability of salinity structure in the regions of large fresh water discharge and high precipitation such as Bay of Bengal. Even the well organized assimilation schemes/models have limitations in producing accurate salinity analysis due to the lack of significant salinity observations (both spatially and temporally). Recent advancement in satellite technology has made possible the measurement of sea surface salinity (SSS) with unprecedented spatio-temporal coverage. Aquarius is the recent operational mission which measures the global SSS since 2011. We assimilated Aquarius SSS in the Global Ocean Data Assimilation (GODAS) for the period 2011 to 2014. GODAS is the assimilation module for MOM4 (ocean component of CFS, the forecast model used in India and many other countries). The assimilation scheme in GODAS is 3DVAR. The present study addresses the impact of assimilating Aquarius SSS in improving the ocean analysis and the intra-seasonal and inter-annual variability of salinity in the key regions such as Bay of Bengal. The results are validated with ECCO2 model and in-situ RAMA buoy observations. The results are discussed in this study.

Changes in the phenology of physical and ecological variables associated with climate change are likely to have significant effect on many aspects of the Baltic ecosystems. We apply a set of phenological indicators to multiple environmental variables measured by satellite sensors for 17-35 years to detect possible changes in the seasonality in the Baltic Sea environment. We detect significant temporal changes such as earlier start of the summer season and prolongation of the productive season in multiple variables ranging from basic physical drivers to ecological status indicators. While increasing trends in the absolute values of variables like sea-surface temperature (SST), diffuse attenuation of light (Ked490) and satellite-detected chlorophyll concentration (CHL) are detectable, the corresponding changes in their seasonal cycles are more dramatic. For example, the cumulative sum of 30,000 W m^-2 of surface incoming shortwave irradiance (SIS) was reached 23 days earlier in 2014 compared to the beginning of the time series in 1983. The period of the year with SST of at least 17 ˚C has almost doubled (from 29 days in 1982 to 56 days in 2014), the period with Ked490 over 0.4 m^-1 has increased from about 60 days in 1998 to 240 days in 2013, i.e. quadrupled. The period with satellite-detected CHL of at least 3 mg m^-3 has doubled from approximately 110 days in 1998 to 220 days in 2013. While the timing of both the phytoplankton spring and summer blooms have advanced, the annual CHL maximum that in the 1980s corresponded to the spring diatom bloom in May has now switched to the summer cyanobacteria bloom in July.

Mediterranean being one of the most oligotrophic marine environment offers an ideal test place for this iconoclastic approach where we hypothesize that algae blooms over the surface waters is entirely governed by the wet dust deposition. During the long range transport of desert dust, it has a chance of interacting with cloud water. Upon this contact it has ben shown that the prokaryotes become active and releases oxalate as an osmosolute. This oxalate is used to form iron oxalate under any circumstances and may undergo decarboxylation reaction if the solar light energy is above a threshold level. Decarboxylation reaction results with the formation of reduced iron carbon dioxide and a carbonyl radical. Being radical it may combine with another carbonyl radical and may form yet another oxalate ion that acts as a feedback mechanism. The pH of the rain event further shifts carbon dioxide into its bioavailable bicarbonate form. The collapse of iron bearing mineral further enriches the media with other essential trace elements. It has been also shown that Saharan dust containing rain events further enriched with various essential amino acids. Thus any wet deposition event blessed with desert dust triggers phytoplankton bloom over the surface ocean and this event can be traced with satellite imageries.  The only drawback is the time necessary for phytoplankton concentration to reach above a critical level that can be detected with satellite sensors and this can be reached in 2-3 weeks. This time lag has been and still misleads the scientific community and algae blooms over the surface ocean is still being tried to linked with other factors without any success. Mediterranean offers an ideal place to test our iconoclastic approach and we can demonstrate the impact of wet dust deposition over algae blooms over the surface ocean via various satellite imageries.

 

Making use of the fine resolution of satellite SAR imagery from Envisat, ERS, ALOS-1, and RADARSAT-1,we observe small eddies during the spring and summer months in Lake Superior, which are important in lake mixing and nearshore water quality. During these months there is a thermal gradient between warmer near shore waters and colder offshore waters which enhances cyclonic coastal currents, as identified by sea surface temperature (SST) imagery for Lake Superior from the Advanced Along Track Scanning Radiometer, over the period 1995-2012.  In this study, we will compare the location, diameter, rotational sense, and seasonality of identified eddies in terms of the strength of the thermal gradient and general circulation patterns within the lake to identify possible patterns of eddy formation, evolution, and dissipation. We will compare these results with a circulation model as well.

One of the most meaningful parts composing the problems of monitoring and analysis of the natural environment is the process of obtaining and using of the observation data. Currently there is a huge amount of different types of the observing of the condition of the oceans and seas. Satellite data has have reached the leading position among them taking into account the amount of the provided geophysical information. Technological progress and water remote sensing allows us to obtain information about the main characteristics of the water environment (such as temperature, salinity, flow, level, etc.) with high precision, frequency, time, spatial resolution and a huge coverage area. Unfortunately, this information usually presented on sets of irregular points at asynchronous time moments. That is why processing of the observation data is a very significant process, moreover the quality of interpolation often determines the solution accuracy of the problems in oceanography and other Earth sciences.

This work is devoted to the satellite observation data processing in problems of the Earth water system monitoring, analyzing and forecasting. The satellite observation data on the sea surface temperature of the World Ocean and separate seas are considered. The analysis of the data helped to identify some systematic errors in the data. Therefore methods of additional verification of observations are discussed which are based on statistical methods, allowing not only to consider specific data field, but also to establish common features, typical for the entire set of realizations of temperature field. Numerical experiments are carried out on data verification based on the three sigma rule. The results of numerical calculations for verification of satellite data on the sea surface temperature are presented.

This study was supported by the Russian Foundation for Basic Research (project №15-01-01583, №16-01-00548) and by the Russian Science Foundation (project №14-11-00609).

1. V. I. Agoshkov, S.A. Lebedev, E. I. Parmuzin, Numerical solution of the variational assimilation problem using on-line SST data. Izvestiya, Atmospheric and Oceanic Physics, 2009. Vol. 38, No. 1. Pp. 1-20. 

2. Zalesny V.B., Gusev A.V., Chernobay S.Yu., Aps R., Tamsalu R., Kujala P., Rytkönen J. The Baltic Sea circulation modelling and assessment of marine pollution, Russ. J. Numer. Analysis and Math. Modelling, 2014, V 29, No. 2, pp. 129–138. 

3. N. B. Zakharova, V. I. Agoshkov, E. I. Parmuzin A new interpolation method for observation data obtained from ARGO buoys system. Russian Journal of Numerical Analysis and Mathematical Modelling. 2013. Volume 28, Issue 1, Pp. 67–84.

4. Agoshkov V.I., Parmuzin E.I., Zakharova N.B., Zalesny V.B., Shutyaev V.P., Gusev A.V. Variational assimilation of observation data in the mathematical model of the Baltic Sea dynamics // Russ. J. Numer. Anal. Math. Modelling, 2015, V. 30, No. 4, PP. 203–212.

In this study, the previously developed CWAVE_ENV [Li et al.,2011] algorithm is applied to the full ASAR wave mode dataset (December, 2002 – April, 2012) during the whole mission of ENVISAT to generate an independent sea state dataset.

The CWAVE_ENV algorithm was developed devoting for ASAR wave mode wave data to derive integral wave parameters, including significant wave height (SWH) and mean wave period. Our previous validation based on four-month ASAR wave mode data acquired during December 2006 to March 2007 suggests a very good agreement with in situ buoy measurements with a bias of 0.06 m and an RMSE of 0.70 m. The algorithm now is being applied to the ten-year ASAR wave mode dataset to obtain SWH and mean wave period. The cal/val is also being conducted by comparing with spatially-temporally collocated in situ buoy measurments and cross-over Radar Altimeter measurments.

We think that the product will be a great add-on to the global surface wave measurments. All the retrievals will be produced in the standard Netcdf data format for public access.  

In addition, global wave climate analysis combing the CWAVE_ENV retrieval (for full sea state) and the ESA standard ASAR wave mode level2 product (for swell) will be also presented.

It is acknowledged that utilization of the ASAR wave mode Level1b data is authorized by ESA and the dataset is provided by Ifremer/cersat.

 

Reference:

Li,X.-M., S. Lehner and Th.Bruns (2011), Ocean wave integral parameter measurements using ENVISAT/ASAR wave mode data, IEEE Transactions on geosciences and remote sensing, 49(1).    

Global and regional mean sea level change have been reconstructed by several studies using data from satellite altimetry and tide gauge records. Commonly, principal component analysis (PCA) is used during the process, for example to reduce noise, to remove seasonal variations in the timeseries, or to investigate the different sources of dynamic sea level variability. However, PCA is limited in isolating physical signals and therefore in the ability to separate individual physical modes of sea level variations. The PCA results in modes, which are mathematically orthogonal. This is not necessarily true for real signals, and the resulting PCA modes may therefore be difficult to interpret physically.

Therefore, we apply the independent component analysis (ICA) to decompose sea level change derived from altimetry and tide gauge data. Here, the main hypothesis is that the physically independent signals likely create statistically independent modes of sea level change which can be extracted by ICA. The implementation of ICA includes an initial PCA to identify the orthogonal patterns of interest, which are then rotated to be as independent as possible.

We anticipate that the ICA method better characterizes independent physical processes leading to regional sea level change as it uses higher than second-order moments of the probability density function that exist in sea level records for example derived from satellite altimetry. The physical sources of the different dynamic sea level modes derived by ICA are analyzed. Results are compared with modes of dynamic sea level change simulated with the Finite Element Sea-Ice Ocean Model (FESOM).

The Indian Ocean is the third largest ocean in the world covering approximately 20% of the total ocean area of world. To date, field observations, especially of phytoplankton, are scarce in the Indian Ocean and large areas of the Indian Ocean remains understudied, despite their importance in the controlling global primary production and biogeochemical cycles. In this study optical properties, measured in July-August 2014 during the SO235-OASIS (OrgAnic very short lived Substances and their air sea exchange from the Indian Ocean to the Stratosphere) cruise from Port Louis, Mauritius to Male, Maldives are used to characterise the phytoplankton in this particular ocean. The phytoplankton is characterised by combining the results obtained with different optical measurements (high performance liquid chromatography, spectrophotometry, fast repetition rate fluorometry (FRRf), flow cytometry and hyperspectral radiometry, and satellite remote sensing) and microscopic cell counts. The obtained optical information are then utilised to characterise the community structure and physiological condition of phytoplankton in the Indian Ocean.

From top-of-atmosphere (TOA) observations, atmospheric correction for ocean color inversion aims at distinguishing atmosphere and water contributions. From a methodological point of view, our approach relies on a Bayesian inference using Gaussian Mixture Model prior distributions on reference spectra of aerosol and water reflectance [1]. A reference spectrum for the aerosol characterizes the specific signature of the aerosols on the observed aerosol reflectance. A reference spectrum for the water characterizes the specific signature of chlorophyll-a, suspended particulate matters and colored dissolved organic matters on the observed sea surface reflectance. In our Bayesian inversion scheme, prior distributions of the marine and aerosol variables are set conditionally to the observed values of covariates, typically acquisition geometry acquisition conditions and pre-estimates of the aerosol and water reflectance in the near-infrared part of the spectrum. The numerical inversion exploits a gradient-based optimization from quasi-randomized initializations.

 

We evaluate our estimates of the sea surface reflectance from the MERIS TOA observations. Using the MERMAID radiometric in-situ dataset, we obtain significant improvements in the estimation of the sea surface reflectance, especially for the 412, 442, 490 and 510 nm bands, compared with the standard ESA MEGS algorithm and the a state-of-the-art neural network approach (C2R). The mean gain value on the relative error for the 13 bands between 412 and 885 nm is of 57% compared with MEGS algorithm and 10% compared with the C2R. The water leaving reflectances are used in Ocean Color for the estimation of the chl-a concentration, the colored dissolved organic matters absorption and the suspended particulate matters concentration underlying the potential of such approach to improve the standard level 2 products in coastal areas.  We further discuss the potential of MEETC2 for the incoming OLCI / Sentinel 3 mission that will be launched in December 2015.

 

[1] MEETC2: Ocean Color Atmospheric corrections in coastal complex waters using a Bayesian latent class model and potential for the incoming Sentinel 3 - OLCI mission, in publication at Remote Sensing of Environment.

Index term: 1) Atmospheric corrections. 2) Ocean color. 3) Bayesian inversion. 4) Gaussian Mixture Model.

By methods of remote sensing, the mesoscale meteorological phenomenon, known as Novaya Zemlya bora, downslope windstorms, which leaves pronounced footprints on the synthetic aperture radar (SAR) images of the sea surface, is investigated. In the winter seasons the Kara Sea is typically ice covered, while the Barents Sea has very low ice concentrations. Weather data indicate that during December - February the windiest conditions occur with monthly mean wind speeds of 10-15 m/s mainly having southeasterly direction. Although a number of authors have referred to these high speed winds as a bora, there has been no systematic study of this phenomenon. Regular images from Sentinel-1A allow taking a look on this phenomenon from space and restoring extent, strength and duration of the Novaya Zemlya bora, and even its impact on oceanography of the Eastern Barents Sea. It is established the local bora covers either coastal/open zones of the Barents Sea to the west from Novaya Zemlya up to 150-200 km, or very rarely less area in the Kara Sea that can lead to catastrophic consequences in the coastal waters of the archipelago. Finally it is shown that routinely use of spaceborne SARs, such as Sentinel-1A, Radarsat-2, etc. for detection, monitoring and research of the polar local winds has no an alternative.

Lake Baikal in the Russian Federation, the deepest and unique lake in the world, represents itself specific hydrological and weather regimes as well as different natural phenomena. Among unique those there are internal waves, giant ice rings, local winds, oil and gas seeps, etc. These phenomena leave pronounced footprints both on the lake surface and in the ice cover. Some of them are typical for inland or semi-closed seas, whereas others are specific for large lakes. These phenomena can be imaged, mapped and studied by remote sensing and in particular by combined use of synthetic aperture radar (SAR) and optical images with middle and high resolution. Most of them for the first time were discovered in satellite images. Moreover, routinely acquired SAR images of Sentinel-1A and Landsat-7/8 help scientists studying them from space allowing the extraction of a number of characteristics and useful features. In this paper the most distinct examples are presented and analyzed with purpose to emphasize the uniqueness of Lake Baikal, which is sometimes considered as a marine laboratory. In conclusion, the use of multi-sensor SAR and optical images is the best way to study and monitor different phenomena occurred in large lakes.

The most versatile and promising technology for solving problems of monitoring and analysis of the natural environment is a four-dimensional variational data assimilation of observation data. The development of computational algorithms for the solution of data assimilation problems in geophysical hydrodynamics is important in the contemporary computation and informational science to improve the quality of long-term prediction by using the hydrodynamics sea model. These problems are applied to close and solve in practice the appropriate inverse problems of the geophysical hydrodynamics.

In this work the variational data assimilation problems in the Baltic Sea water area were formulated and studied [1]. We assume, that the unique function which is obtained by observation data processing is the function and we permit that the function is known only on a part of considering area (for example, on a part of the Baltic Sea).

Numerical experiments on restoring the ocean heat flux and obtaining solution of the system (temperature, salinity, velocity, and sea surface height) in the Baltic Sea primitive equation hydrodynamics model [2] with assimilation procedure were carried out. In the calculations we used daily sea surface temperature observation from Danish meteorological Institute, prepared on the basis of measurements of the radiometer (AVHRR, AATSR and AMSRE) and spectroradiometer (SEVIRI and MODIS). The spatial resolution of the model grid with respect to the horizontal variables amounted to 0.0625*0.03125 degree. The results of the numerical experiments are presented.

This study was supported by the Russian Foundation for Basic Research (project №16-01-00548) and project №14-11-00609 by the Russian Science Foundation.

References.

[1]   Agoshkov V.I., Parmuzin E.I., Zakharova N.B., Zalesny V.B., Shutyaev V.P., Gusev A.V. Variational assimilation of observation data in the mathematical model of  the Baltic Sea dynamics // Russ. J. Numer. Anal. Math. Modelling, 2015, V. 30, No. 4, PP. 203–212.

[2]   Zalesny V.B., Gusev A.V., Chernobay S.Yu., Aps R., Tamsalu R., Kujala P., Rytkönen J. The Bal-tic Sea circulation modelling and assessment of marine pollution, Russ. J. Numer. Analysis and Math. Modelling, 2014, V 29, No. 2, pp. 129–138.

Satellite altimetry has revolutionized our understanding of ocean dynamics thanks to finer spatial sampling and global coverage. Nevertheless, coastal data have been flagged as unreliable due to land and calm water interference in the altimeter and radiometer footprints and uncertainty about high frequency tidal and atmospheric forcing. Recent developments in processing are now bringing new possibilities to sea level studies in the coastal zone. In particular, our research has focussed on the retracking, i.e. the on-ground processing that fits a model to the signal received by the satellite in order to increase the precision of the estimated geophysical parameters.

This study will present the three years of work that have led to the design, validation and application of the first global multi-mission coastal-retracked altimetry reprocessing based on the same algorithm, i.e. ALES (Adaptive Leading Edge Subwaveform retracker), whose aim is to retrack with the same precision both open ocean and coastal data.

1. Design

ALES selects part of each returned echo and models it with a classic ‘open ocean’ Brown functional form, by means of least square estimation whose convergence is found through the Nelder-Mead nonlinear optimization technique. By avoiding echoes from bright targets along the trailing edge, it is capable of retrieving the majority of coastal waveform up to 2 to 3 Km from the coasts. By adapting the estimation window to the significant wave height, it preserves the precision of the standard open-ocean data. ALES is currently adapted for Envisat, Jason-1, Jason-2 and AltiKa missions.

2. Validation

A first validation was performed comparing the Total Water Level Envelope (TWLE) from ALES product and from the standard Sensor Geophysical Data Record (SGDR) with the in-situ measurements of tide gauges (TGs) located in the Adriatic Sea and on the Agulhas Bank. TWLE represents the combined effect of ocean tides and atmospheric forcing in addition to the sea level anomaly with respect to the mean sea surface and it is therefore particularly appropriate for the comparison with sea level from the TGs.

ALES retracker led to roughly a half of the analysed tracks showing a marked improvement in correlation with the TG records, with the root mean square (RMS) difference being reduced by a factor of 1.5 for Jason-1 and Jason-2 and over 4 for Envisat in the Adriatic Sea (at the closest point to the TG). ALES succeeds in providing reliable estimations in areas where data were not included in the only present multi-mission coastal altimetry dataset, which did not include retracking (CTOH dataset). 

Subsequently, the Significant Wave Height (SWH) was validated against buoy data in the German Bight, available from the Bundesamt fuer Seeschiffahrt und Hydrographie. The German Bight is particularly challenging for SWH retrieval due to typically low sea state and the presence of tidal flats and islands. A higher correlation is seen using ALES for up to three 1-Hz points toward the coast for each of the tracks (for both Envisat and Jason missions), meaning about 7 to 22 km in terms of spatial improvement. The 1-Hz SWH estimations have a constantly lower standard deviation (std) compared to the original SGDR product. 

3. EXAMPLES OF APPLICATIONS FOR SEA LEVEL STUDIES

The suitability of the new dataset for sea level studies on a regional scale was tested in the North Sea/Baltic Sea transition area. The analysis was focused on the variability of the annual cycle during the Envisat years (2002-2010) and was performed using the reference ESA Sea Level Climate Change Initiative (ESA SL_cci) dataset for comparison, an extensive network of TGs for validation and external datasets (wind and climatology) to discuss the drivers of the seasonality. ALES is able to increase the quantity of 1 Hz along-track data while improving the quality.

The area of study was divided into sub-basin and the harmonic analysis was performed at different distances from the coastline.The results in the North Sea/Baltic Sea transition show that:

* ALES dataset, using state-of-the-art geophysical corrections, is able to estimate accurately the annual cycle up to the coast. Only ALES-based amplitude estimates of the annual cycle are in agreement with TGs within 1 cm in every sub-basin.

* ALES product is able to detect a slope in the amplitude that follows the bathymetry of the Skagerrak sea and correspond to the narrow Norwegian coastal current, which is not well represented in the SL_cci estimates.

The ALES dataset is currently being applied for sea level studies in Indonesia, a hotspot of sea level rise with few reliable in-situ data.

4.DISTRIBUTION

A complete global ALES-retracked dataset is now available free of charge from the Physical Oceanography Distributed Active Archive Center (PO.DAAC) at the NASA Jet Propulsion Laboratory, Pasadena, CA. http://podaac.jpl.nasa.gov. The file structure is the same as standard SGDR products, with the addition of the fields concerning the ALES retracking, i.e. Range and Significant Wave Height. Users are invited to integrate the new retracked fields with state-of-the-art geophysical corrections in order to maximise the impact of the improvements.

The Guadalquivir estuary is one of the largest and most productive estuarine systems of the Western Europe. Located in the southwest coast of the Iberian Peninsula, its waters mix with those of the Gulf of Cadiz (Atlantic Ocean). The dynamics of its turbidity plume are important for the functioning of the estuary and the adjacent coastal region. Recently, several episodes of high turbidity have occurred which habitually produce negative effects on water quality, and thus on economy and society. In addition, high concentrations of suspended particulate matter in waters directly affect or govern numerous water column and benthic processes. In the case of the marine region influenced by the estuary, turbidity is one of the primary threats to the health of this complex and challenging environment. Ten-year ocean color observations between 2002 and 2012 from the MEdium Resolution Imaging Spectrometer (MERIS FR, 300 m) onboard the ENVISAT multispectral platform are used to quantitatively assess the variability of sediment properties in the Guadalquivir estuary and adjacent region. The combination of high-frequency observations from automated sensors and in situ data in addition to synoptic satellite imagery resulted in an excellent complement to better study water quality parameters. The turbid plume has been detected by the backscattering characteristics of the surface waters in the vicinity of the estuary mouth. Remote Sensing Reflectance at 665 wavelength (Rrs665, 1/sr) and turbidity reveal the effects of physical processes governing both the spatial and temporal distribution of the plume. The findings of this work indicate that both respond to seasonal and inter-annual variations of rainfall and mainly, the freshwater discharges from Alcalá del Río dam. These patterns represent a crucial factor governing phytoplankton primary productivity in the pelagic ecosystem of the Eastern Gulf of Cadiz. Furthermore, our results suggest that the plume varies on fortnightly scales, presenting smaller dimensions at neap tides and achieving higher dimensions during and after spring tides. Additionally, an asymmetric behavior of plume signal in connection with the semidiurnal tidal cycle (flood and ebb tide) is also found, with maximum values of Rrs665 at low tide. Both fortnightly and semidiurnal cycles appear to be superimposed on the large time-scale variability, and present throughout the year. This study demonstrates that tidal forcing is one of the important ocean processes that drive significant changes of the turbidity plume in coastal waters, better characterizing its variability and advancing our understanding of sediment dynamics in highly variable estuaries.

Some satellite sensors are deployed to image the ocean in the visible portion of the spectrum. The colour of the sea, or more precisely the spectral water-leaving radiance, gives us information about the concentration of water constituents, e.g., chlorophyll, coloured dissolved organic matter, or suspended mineral particles. The bidirectional nature of the radiance distribution at the water surface and the interaction of this radiance with the air-sea interface depend on the viewing- and sun-geometry and surface waves. Up to now, many atmosphere-ocean radiative transfer models implement wind-depending wave slope statistics by Cox and Munk. Wave facets are basically randomly orientated on a plane level (vertically invariant) and they are independent from each other. This approach inhibits interactions of light with the actual structure of the wave system. Indeed, wave shadows and multiple scattering of reflected light at the surface are important factors that influence the bidirectional reflectance distribution function (BRDF) in particular at low sun zenith angles (of more than 60°). If we consider state-of-the-art wave modelling, including sea state-depending wave elevations and wind-depending wave slope variance, perceptible deviations of the light transmittance and reflectance of the air-water interface occur at low sun in comparison with wind-depending wave slope statistics. Radiance reflectance distributions are presented as function of sea state (classes of significant wave height and mean wave period) and wind speed. Furthermore, effects of light polarization are discussed. The inclusion of appropriate wind and wave data, i.e., wave heights and periods, can help to reduce uncertainties related to the Fresnel-reflecting ocean surface – in particular for large solar zenith angles. This especially regards remote sensing of ocean colour at high latitudes and atmospheric correction, but it could also be applied for example for aerosol retrievals.

Since the first altimeter missions and the improvements performed in the accuracy of sea surface height measurements from 1992 onwards, the importance of global quality assessment of altimeter data has been increasing. Global Cal/Val studies are usually performed by the analysis of internal consistency and cross-comparison between all missions. In this study, the steric and mass contributions to the sea level provided by Argo profiling floats and the Gravity Recovery And Climate Experiment (GRACE) mission respectively are used as independent sources of comparison to analyze the altimetry errors.

Argo profiling floats are spread out over almost the global open ocean since 2004. However, they measure temperature and salinity vertical profiles, providing only the steric contribution to the total sea level content measured by altimeters. The missing mass contribution is derived from the GRACE data set from 2003 onwards.

The comparison is performed with the first objective of detecting global and regional altimeter mean sea level drifts. A second goal is to assess the impact of new altimeter standards (orbit, geophysical corrections, ground processing) and of new version of altimeter merged products such as the 2014 AVISO reprocessing or the Sea Level CCI data set. We also focus our work on sensitivity analyses of the method of comparison to various parameters. In particular, we determine to which extent the altimeter quality assessment is affected by a different pre processing of altimeter data, a sub sampling of the Argo network and a change of the reference depth used to compute Argo dynamic heights.

The South Pacific Convergence Zone (SPCZ), the largest rain band worldwide during austral summer, is important to atmospheric circulation (including cyclone genesis) and ocean circulation. The variability of SPCZ is examined using ocean surface wind products derived from NASA’s QuickSCAT (1999-2009) and ESA’s ASCAT (2007 onward) satellite scatterometers and ERA-Interim atmospheric reanalysis (1981 onward). From these products, indices were developed to represent the SPCZ strength, area, and centroid location. Excellent agreement is found in terms of the temporal variations of the indices derived from the satellites and reanalysis wind products, despite some small differences in the time-mean SPCZ strength. The SPCZ strength, area, and centroid latitude have a dominant seasonal cycle. In contrast, the SPCZ centroid longitude is dominated by intraseasonal variability due to the influence by the Madden-Julian Oscillation. The SPCZ indices are all correlated with El Niño-Southern Oscillation indices. Interannual and intraseasonal variations of SPCZ strength during strong El Niño are approximately twice as large as the respective seasonal variations. SPCZ strength depends more on the intensity of El Niño rather than the central- vs. eastern-Pacific type. The longer ERA-Interim product is also used to examine decadal variations of the SPCZ indices. The change from positive to negative Pacific Decadal Oscillation phase around 1999 results in a westward shift of the SPCZ centroid longitude, much smaller interannual swing in centroid latitude, and a decrease in SPCZ area. This study improves the understanding of the variations of the SPCZ on multiple time scales and reveals the variations of SPCZ strength not reported previously. The diagnostics analyses can be used to evaluate climate models.

We estimate coastal currents in two regions of the Mexican Pacific from along track coastal altimetry and in situ data. Each region presents a different challenge for the 1Hz X-Track coastal altimetry. In southern México, a low salinity buoyancy flow advances poleward while large eddies generated off the Gulfs of Papagayo (Nicaragua), Fonseca (Honduras) and Tehuantepec (southern Mexico) slowly propagate offshore. The passage of eddies has significant influence on the sea-level signal and along-shore velocities. In the along-shore current time series, from 1993 to date and within 50 km of the coast, variability associated to mesoscale is significant but poleward currents are clearly dominant during most winter seasons. We also discuss a time series of geostrophic alongshore flow off the west coast of Baja California. This is a typical Eastern boundary with intense equatorward winds and a coastal upwelling region active in spring and summer. Coastal currents influenced by the wind regime and the Eastern boundary circulation flow equatorward most of the year. During summer a warm poleward coastal current is reported to flow against the climatological winds. We use a one-year time series of velocities measured with an acoustic Doppler current profiler to validate coastal altimetry data. Our results, consistent with hydrographic data and observed currents, show the seasonal occurrence of a warm poleward flow in late summer or early autumn. Finally, we use a combination of gridded and along-track coastal altimetry to discuss the role of mesoescale eddies on the variability of these two coastal regions.

Since the early 80's satellite altimetry, resulted in an abundance of sea surface height measurements. The exploitation of these data sets from past and current satellite missions is crucial to both oceanographic and geodetic applications. This work presents a correlation analysis of Jason-1, Jason-2 and CryoSat-2 Sea Level Anomalies (SLAs) with global and regional climatic indexes that influence the ocean state, while the cyclo-stationarity of the SLAs is examined. Three oscillation indexes have been investigated, as representative of climate-change and seasonal forcing on the sea level. The first one is the well-known Southern Oscillation Index (SOI) corresponding to the ocean response to El Niño/La Niña-Southern Oscillation (ENSO) events. The next index investigated is the North Atlantic Oscillation (NAO) index, which corresponds to the fluctuations in the difference of atmospheric pressure at sea level between the Icelandic low and the Azores high. The last index investigated is the Mediterranean Oscillation Index (MOI) which refers to the fluctuations in the difference of atmospheric pressure at sea level between Algieres and Cairo. The raw data used are SLA values from Jason-1 and Jason-2 for thirteen consecutive years (2002-2014), ENVISAT for seven consecutive years (2003-2009) and SLA values from CryoSat-2 for five consecutive years (2010-2014) over the entire Mediterranean Basin. The possible correlation are investigated in both annual and seasonal scales, while a regional multiple regression and a principal component analysis between the SLAs and oscillation indexes is carried out. The latter intends to model any possible correlations between the Mediterranean sea level variations and the aforementioned global and regional climatic phenomena. Finally, evidence of the sea level cyclo-stationarity in the Mediterranean Sea is deduced from the analysis of empirically derived covariance functions at monthly intervals from the available SLA data.

With all the variety of models used for calculation of primary production from remote sensing data, a choice of the most realistic one remains a non-trivial issue. The use of regional biological parameters additionally increases the degree of correspondence between a model and in-situ observations. In this work, we estimate primary production in the Northeast Atlantic Ocean in 1998 - 2005 years using three frequently used models: two models are based on the remotely measured chlorophyll-a concentration (VGPM and PSM) and one – on the remotely measured coefficient of light absorption by phytoplankton pigments (Aph-PP). The model results are farther compared with in-situ observations of primary production in the area 20⁰ - 51⁰ N and 10⁰ – 40⁰ W.

The primary production models use as the input level 3 Ocean Color data provided by the OC-CCI database (http://www.esa-oceancolour-cci.org). Photosynthetic model parameters are taken from experimental measurements of phytoplankton photosynthetic efficiency under different light conditions, obtained for Northeast Atlantic phytoplankton species.

The results show a close similarity in the patterns of primary production obtained by different models, although the absolute values in different models differ substantially. PSM model is found to describe better the observed seasonal and spatial variability of the primary production in the Northeast Atlantic as compared to the two other models. However, in most of the cases PSM slightly underestimates the production values. Use of regional photosynthetic parameters contributes to closer approximation of in-situ observations by the models

While analyzing sea-level measurements it is assumed that planetary waves have an input in different sorts of oceanic phenomena and processes, though until recent years attempts to determine them in the ocean were limited.

Satellite altimetry data open new horizons in studying of oceanographic fields variability, allows us to continuous monitoring of the sea level and provides with level-based velocity fields. These methods show great perspectives applied to different aspects, in particular, to fishery oceanographic basis and gas and oil extraction.  Basing on the satellite altimetry data frequency spectrum and wave parameters can be acquired.

Spatial structure of low-frequency wave-like motions, observed in the World ocean, can be described as the alternation of cyclonic and anticyclonic eddies and interpreted as the standing-progressive wave:

ξ=Acos(mx)cos(ly)cos(σt-kx-ny)

Where А – amplitude, t – time,   x, y  – rectangular coordinate system values, σ – frequency, k, n  – zonal and meridional wave numbers, m, l – modulation parameters, which are inversely proportional to the cell size. Particles motion trajectories in the proposed model are closed, which is similar to those in eddies.

Even abrupt phase shifts to the opposite values in sea-level oscillations can be easily spotted in such systems.  Model of sea-level perturbations can be displayed as the sum of a short (sum of wavenumbers) and a long (subtraction of wavenumbers) progressive waves. Their parameters can be assessed based on Complex Singular Value Decomposition (CSVD).

Sea-level field contains nodal and amphidromic points and lines with progressive waves propagating between them. At these very points horizontal advection must be the strongest and total transport of water and characteristics such as temperature, salinity, buoyancy, chlorophyll concentration etc. must be intensified. This situation can be observed in the straits and narrownesses near nodal points of the standing-progressive wave and in the middle of the amphidromic zones.

By using this type of wave oscillations in depth-integrated continuity equation, we can easily go to the kinematic standing-progressive wave model and detect the position of momentum and water properties (temperature, salinity, buoyancy, chlorophyll-a concentration etc.) full fluxes zones.

The Japan/East Sea satellite altimetry data (delayed-time sea level anomalies) was obtained from AVISO project (Archiving, Validation and Interpretation of Satellite Oceanographic data), which is the part of altimetry data processing segment SSALTO (Segment Sol multimissions d'ALTimétrie, d'Orbitographie et de localisation précise) run by French Space Agency (CNES-The Centre National d'Études Spatiales) for 1992.10.14 – 2012.02.08 with 7-days lag period. The data was received from Jason-1, TOPEX/Poseidon, ENVISAT, GFO-1, ERS1/2, GEOSAT satellites and released as the product on a regular Mercator grid map with spatial resolution 1/3° x 1/3° using the optimal interpolation. Time series arrays were arranged in a matrix and thoroughly analyzed (www.aviso.altimetry.fr/duacs/). We found the described wave structure in the Japan/East Sea, assessed the degree of stationarity and proposed the phenomenological model of the so-called ‘basin’ waves, which are, basically, the standing-progressive topographic planetary waves, to explain oceanological fields’ spatiotemporal variability in closed seas abyssal basins (e.g. Tsushima/Ulleung Basin). 

In this study, empirical and semianalytical algorithms are developed and compared for optically complex waters to retrieve the diffuse attenuation coefficient of downwelling irradiance (Kd(k)) from satellite data. In the first approach, a band ratio algorithm was used. Various sets of MERIS band ratios were tested to achieve the best estimates for Kd(490) based on the in situ dataset which was measured in Nordic lakes (oligotrophic to eutrophic conditions). In the second approach, Kd(490) was expressed as a function of inherent optical properties which were retrieved from MERIS standard products. The algorithms from both approaches were tested against an independent data set and validated in optically complex coastal waters in the Baltic Sea and in Nordic lakes with high concentrations of coloured dissolved organic matter (0.32.5 m-1). As a result, a combined band ratio algorithm was developed, which provides a promising approach (R2=0.98, RMSE=17%, N=34, p<0.05) for estimating Kd(490) over a wide range of values (0.3– 6.1 m-1).

The methods were tested on MERIS data, however due to the similar spectral characterization of the Ocean Land and Colour Instrument (OLCI) sensor on board Sentinel-3; we can assume that the approaches used here will be applicable on OLCI data as well.

The wind climate of the Mediterranean Sea has been estimated from atmospheric modelling (Cavaleri 2005, Lavignini et al. 2006) and QuikSCAT (Furevik et al. 2011). The latter shows the Aegean Sea as a promising area for offshore wind power development. According to the Hellenic Wind Energy Association (HWEA), the sites of particular interest for offshore wind energy are located close to the mainland and islands in the Aegean Sea. Wind farm developers aim to select local areas with favorable wind conditions to optimize the annual energy production and the economic profit. In the Aegean Sea, where the spatial variations in wind speed are very high, accurate resource mapping is of great importance as the produced wind power is proportional to the cubed wind speed.  It is challenging to model the wind resource and it is costly to measure from the ground at every place of interest.  Maps based on Synthetic Aperture Radar (SAR) are expected to prove valuable for the exploitation of the excellent wind resource of the Aegean Sea, to the benefit of the national economy. High-resolution SAR satellite data bring new information for pre-feasibility for instance at the policy planning level.

For accurate wind resource mapping from satellite it is necessary to collect many images to reduce the uncertainty. The 10-year Envisat ASAR archive has been used for wind resource mapping. Wind maps from satellite are retrieved at 10 m. DTU Wind Energy has developed a method for extrapolation of winds to turbine hub heights at around 100 m using a combination of satellite wind fields and the long-term climate of atmospheric stability from the mesoscale model (Badger et al. 2016). The result of the mean wind speed at hub-height for the Aegean Sea is shown in Figure 1. The map shows the stability dependent winds (SDW).

It is planned to combine the Envisat wind fields with Sentinel-1a and Sentinel-1b wind fields to further detail the offshore wind resource within the New European Wind Atlas. The work is in progress. Sentinel-1a images are processed at DTU Wind Energy near-real-time and we are updating our wind resource software. A service-based on satellite SAR-derived winds for wind resource estimation is available at DTU Wind Energy.

The project was supported by ESA ResGrow and satellite data from ESA Envisat and Copernicus Sentinel-1.

Badger et al. 2016 abstract (same issue)

Cavalari 2005 The wind and wave atlas of the Mediterranean Sea – the calibration phase. Advances in Geosciences, 2, 255–257.

Furevik, BR, Sempreviva, AM, Cavaleri, L, Lefèvre, J-M & Transerici, C 2011, 'Eight years of wind measurements from scatterometer for wind resource mapping in the Mediterranean Sea' Wind Energy, vol 14, no. 3, pp. 355-372.

Lavignini et al. 2006 Offshore Wind Climatology over the Mediterranean Basin. Wind Energ. 2006; 9:251–266

The geodynamic features in Chile turn this region particularly important for studies in Geophysics, Geodesy and Geology. Principally due to convergence zone between the Nazca and South American tectonic plates, and important events associated to this natural condition, i.e. Earthquakes, Tsunamis, and volcanism. These phenomena cause deformation in the crust of the earth, and derived information is fundamental to understand the interior processes of the Earth, that typically geophysics describe by geophysical inversion. In Chile, study of the long-term vertical deformation in the Ocean-Continent interface is of special interest due to THE NEARBY inter plate convergence zone generated principally by the Nazca and South American PLATES. Taking the difference between the satellite altimetry data (ALT) that gives the absolute sea level variation and the Tide Gauge observations (TG) that record the relative sea level variation, we obtain the absolute vertical crustal motion of the TG sites. We consider 13 TG stations administered by the Hydrographic and Oceanographic Service of the Chilean Navy (SHOA) distributed along the Chilean coast and Chiloé island. Furthermore, altimeter measurements from the multi-satellite missions (Topex 1994-2002, Jason-1 2002-2009 and Jason-2 2009-2015) of the coast in a time period between 1994 and 2015 obtained from Radar Altimeter Database System (RADS) - TU Delft, are included. The altimetry data has been corrected for tropospheric and ionospheric effects, solid earth tide, ocean tide and sea state bias. Barometer inverse correction in altimetry data was not corrected as it is consistent with TG observations. The ALT/TG vertical deformations are compared with vertical GPS measurements from 11 continuous station, which cover a data record period of over four year. The GPS time series solutions were calculated in ITRF08 geodetic reference frame with BERNESE 5.2 GNSS software and the methodology and models that establish the state of the art in GPS data processing. The co-seismic deformations generated by some major earthquakes (e.g. Maule Earthquake) during the study period, were removed based on a catalog of location, depth and magnitude information provided by the National Seismological Centre of the University of Chile. In general, the deformations calculated in the Chilean continental coast show subsidence of the earth's crust up to 11 mm/year, and 67 mm/year on the island of Chiloe. Comparison between GPS and ALT/TG time series show good consistency in trend. Regarding absolute sea level change in the Chilean coast, this presents an increase of 8.5 mm/year in the North (Arica) and is decreasing southbound; 1.6 mm/year in Puerto Williams station. A preliminary investigation of long-term tidal data and altimetry measurements reveals that sea level in the Northern Chilean coast is rising significantly faster than global average rates (3mm/year). However, a higher density and long period of TG and continuous GPS stations is required to confirm our results and conclusions.

The Advanced SCATterometer (ASCAT) instruments on the METOP satellites have been providing operational ocean surface wind and land surface soil moisture data services since 2006, as part of the EUMETSAT Polar System Program (EPS). The third instrument in the series is expected to be launched in 2018 on board METOP-C.

Looking towards the past, the ASCAT mission follows up and indeed bases its heritage on another very successful C-band ESA scatterometer science mission on board the satellites ERS1 and ERS2. Looking towards the future, a follow-on for ASCAT is already being developed and planned for launch in 2023 on board the METOP Second Generation (SG) platforms. In the context of the EPS-SG program, this new C-band scatterometer is expected to further continue the operational scatterometer data services provided by ASCAT and to further expand the European C-band scatterometer data record started by ERS1 in 1991.

The METOP-SG scatterometer is a real aperture C-band (5.355 GHz) side looking radar with a fixed measurement geometry comprising 6 fan beams, similar to ASCAT. This paper further describes its heritage and specification, highlighting the main technological innovations towards providing improved stability, coverage, resolution, as well as the added HV polarisation measurement capability to facilitate improved measurement of very high winds. 

An approach of automatic seasonal adjustment and trend extraction based on the singular spectrum analysis technique has been used to investigate the Mediterranean mean sea level time series from the TOPEX and Jason series of satellite radar altimeters (1993-2012).

The results indicate that the Mediterranean mean sea level is dominated by several harmonic components. The annual signal is particularly strong and almost covers 73.62 % of the original sea level series variation whiles its amplitude is about 15 cm.

The extracted trend shows that the Mediterranean mean sea level has risen by about 60 mm during the last twenty years. Large non-seasonal fluctuations (related to North Atlantic Oscillation climatic phenomenon) were revealed during both winters 2009-2010 and 2010-2011.

Applying the least squares linear regression analysis to the extracted trend of adjusted Mediterranean Sea level series, provides a rate of 2.44 ± 0.4 mm yr^-1 between 1993 and 2012.

Water dynamics in the straits between the North and Baltic Seas during two major Baltic inflows occurred in January 1993 and 2003, based on satellite altimetry data, are investigated. It was noted that before the Baltic inflow occurs surge of water mass to the east coast of the North Sea, and the level difference between the two seas, is about 60 cm. The main research is on the low-frequency fluctuations of sea level, its wave characteristics, by wavelet - and frequency-directional spectral analysis reveals the wave nature of the mechanisms leading to the major Baltic inflow. A comparison of the empirical characteristic of the obtained low-frequency waves with the theoretical dispersion relations for the gradient-vorticity waves showed that in a period of Major Baltic Inflow in the Danish Straits fluctuations are identified as baroclinic Rossby waves. The analysis of cyclonic activity in the northern hemisphere showed that during Inflow observed persisted cyclones, which prove the possibility of resonance mechanism of occurrence of major Baltic inflow.

The use of remote sensing has increased greatly in recent years however in the case of Ireland its use is not well established yet, especially in the marine field. Takinginto account that fisheries and aquaculture plays an important economic and social role in this area, a database of Sea Surface Temperature (SST) from 1982 to 2014 and Chlorophyll-a (Chl-a) from 1998 to 2013 have been generated. For the generation of these sub-products the NOAA Optimum Interpolation (OI) SST V2 and the chlorophyll-a from ESA CCI Ocean Colour Project have been used. Annual and monthly climatologies for SST and Chla have been produced and the new sub-products were exported in broadly used format such as GeoTiff, ascii or NetCDF. Moreover, basic statistical parameters have been extracted for the ICES Divisions and three main areas: West of Scotland, Celtic Sea and Irish Sea, in response to the end-users requirements. The analysis of the SST time series revealed a difference of ~2°C between the West of Scotland and the Celtic Sea areas. Moreover, an increasing trend was detected for all the areas; this trend seems to be more pronounced in the northern waters (~0.5°C in the last 10 years). The analysis of Chl-a showed a clear difference between coastal and open waters with the highest concentrations located close to the UK coast and Irish Sea. In spite of the cloud coverage in the area, the monthly climatology showed a Chl-a concentration peak in spring with a second small peak in autumn with the exception of Irish Sea where no pattern was recognised. The results presented here are preliminary and more detailed analyses are being carried out. The products and data generated will be made available in the near future through Marine Institute downstream data services.

The inflow of Atlantic Water (AW) into the Nordic Seas is of fundamental importance for the Arctic, European and global climate. The West Spitsbergen Current (WSC) is a main source of heat for the Arctic, greatly affecting its sea ice conditions and air temperatures. The aim of this study is to analyze the vertical structure of the current and reconstruct it along its path in 2000-2013 using a synergy of data: i) in situ measurements collected along the WSC by the Institute Oceanology of Polish Academy of Sciences (IOPAS) regularly every July in 2000-2013, ii) satellite altimetry, and iii) an information about the WSC vertical structure simulated in the NEMO (1/12°) numerical model in the same time period.

The preliminary results will be presented based on the in situ 1-year long (2009-2010) mooring measurements, absolute geostrophic currents obtained using CTD data and combined with VMADCP and LADCP and geostrophic currents calculated using satellite altimetry. The relationship between the in situ and satellite observations will be discussed and used to reconstruct the current. The vertical structure of the WSC will be presented and discussed. Finally, the model simulations will be used to estimate the error of the method used for the reconstruction.

Defining the local natural background of suspended particles and biological production is needed for several development and maintenance activities in harbour areas. The Central Dredging Association (CEDA) has noted that the impacts of dredging-induced turbidity should be evaluated as an increase above background level, not as absolute values. Environmental limits should be based on the resilience of the local ecosystem, while accounting for natural fluctuations in turbidity level. (CEDA Information Paper 2011).

Implementation of remote sensing in monitoring the background level of suspended and dissolved particles near the harbor areas can give much wider understanding of ongoing processes in marine environment compared to other methods. We used MERIS data with 300m resolution from years 2006-2011 to analyse the spatial distribution of total suspended matter (TSM), Chlorophyll a (Chl a) and coloured dissolved organic matter (CDOM) concentrations in the vicinity of 8 largest ports located along the Estonian coastline. MERIS images were processed using the Case-2 Regeional  water quality  processor available in BEAM software. The validation results of MERIS products in Estonian harbour areas were analysed  in our earlier work (Raag at al 2014).

The analysis of monthly mean TSM for years 2006-2011 reviled that lowest TSM and lowest seasonal variability from 0.72 to 2.8 mg/L was near the ports of Tallinn and Muuga that locate at the costline of central part of Gulf of Finland. Near harbours that locate at the entrance and in the eastern part of the Gulf of Finland the seasonal variability was more pronounced. The seasonal variability of TSM was largest near port Virtsu located in Moonsund, where TSM concentrations reached 14 mg/L, in April and October. while summer minimum was 2.2 mg/L in August. The highest average TSM concentration (from 9.8 to 24.3 mg/L) throughout the season was observed near Pärnu harbour located in the Gulf of Riga. Also, extremely high monthly mean CDOM was observed near port Pärnu with maximum value of 2 m^-1  in April and minimum value of 1 m^-1 in June. The lowest averaged monthly mean CDOM (form 0.08 to 0.18 m^-1) was near ports Tallinn and Muuga that located at the coastline of central Gulf of Finland.  

The seasonal dynamics of Chl a was similar in seven ports out of eight: maximum in April and minimum in summer  followed by less pronounced maximum in autumn.  Near port Lehtma which is open to the central Baltic waters the autumn maximum of Chl a was higher than the spring maximum. The highest monthly mean  values of Chl a were near Pärnu  form 21 mg/m³ in July  to 37 mg/m³ in April. Relatively high monthly mean values of Chl a (from 13 mg/m³ 21 mg/m³) throughout the season was also observed near port Sillamäe which is the most eastern port in Gulf of Finland belonging to  Estonia. In summer  the Chl a was < 10 mg/m³ in vicinity of other analysed port areas.

Analysis reviled that the variability of natural background TSM, CDOM and Chl a concentration near harbour areas differ greatly along the Estonian coastline what is about  1200 km long. In further port development plans before the dredging operations the satellite based average maps of TSM concentrations can be taken as natural background and the ecosystem resilience level can be set according local variability of TSM.  In addition the background level of CDOM and Chl a broadens the understanding of local ecosystem.

 

References

Raag, Laura, Liis Sipelgas, and Rivo Uiboupin. "Analysis of natural background and dredging-induced changes in TSM concentration from MERIS images near commercial harbours in the Estonian coastal sea." International Journal of Remote Sensing 35.18 (2014): 6764-6780.

 

 

Earth observation satellites have proven through the years that they are capable of recording changes on ecological parameters in the environment and pollution tracking on cases. But radionuclide dispersion cannot be directly detected using satellite remote sensing systems. However the levels of radionuclides in the marine environment, especially the levels for the soluble ones, are associated with other physical, chemical, biological and geological parameters of the natural environment.

Considering this attribute, a program concept has been developed to utilize sea parameters like sea surface salinity (SSS), sea surface temperature (SST) and ocean color. The challenge of the study is the establishment of potential relations between satellite measurements of the marine environment and field radiological measurements. Such potential relations are expected to lead to an innovative tool based on remote sensing data and in situ ¹³⁷Cs measurements for the remote radioactivity detection of the Greek marine ecosystem both for routine control and emergency recordings. Besides, chemical pollutants could be also remotely detected by the tool.

The presented results are an effort of the tool’s development. Satellite data series are acquired from SMOS, Terra and Aqua satellites and in particular, measurements from MIRAS and MODIS instruments correspondingly (SMOS images provided by ESA in the frame of the joint ESA, NOA, NCSR“D” project entitled AOSMOS.4681). These measurements include: sea surface temperature (SST), chlorophyll A (Chla), particulate organic carbon (Poc), particulate inorganic carbon (Poc), Cdom index (Cdom), instantaneous photosynthetically available radiation (IPar) and daily photosynthetically available radiation (Par) (Sykioti et al, 2014). Furthermore, salinity and sea surface temperature in situ measurements were acquired from the HCMR Poseidon system. The ¹³⁷Cs measurements were derived from sea water sample processing using an ammonium molybdophosphate (AMP) radioanalytical pre-concentration method (Florou et al, 2010, Florou et al, 2014) in the environmental radioactivity laboratory of NCSR ”D”.

The first step of the tool’s creation was to establish a spatial database in a GIS environment using all available measurements spanning from December 2011 to September 2015. Other steps include the interpolation and integration of the data for the study of the spatial and temporal variation especially for the area of Souda, North Crete, which has been chosen as a test site for a point to point comparison for all the time span of measurements.

Following up, the correlation between the satellite data and in situ measurements was investigated in order to retrieve a relation. The first analyses resulted to a correlation pattern between the sea surface temperature and ¹³⁷Cs activity concentrations. The correlation was evident in both in situ and satellite SST data, whereas correlation is also apparent between the IPar, Par and ¹³⁷Cs measurements. Sea surface salinity measurements showed no conspicuous correlation with ¹³⁷Cs measurements. Except IPar and Par, the rest ocean color parameters showed uncertain results. This study is an ongoing study and additional information is being retrieved in order to improve the relation between satellite environmental monitoring and radionuclide activity monitoring of the marine environment for the tool creation.

eSurge-Venice (www.esurge-venice.eu), a project funded by the Data User Element (DUE) Programme of the European Space Agency (ESA) and completed recently, was aimed to demonstrate the improvement of the storm surge forecasting with the use of earth observation data. It was specifically focused on the Gulf of Venice in the northern Adriatic Sea.

Small basins like the Adriatic Sea (~1000 km by 300 km) are essentially coastal seas, and the orography plays an important role in shaping the winds. Storm surge forecasting in such a basins is based on predicted fields of sea level pressure and sea surface wind, which force the dynamics of the surge. The surge level has to be added to the initial level of the sea surface at the onset of the simulation: this is the initial field of sea level, which also needs to be set.

The accuracy of the storm surge simulations is therefore constrained by those of the atmospheric forcing and of the initial sea level fields: a better knowledge of both would increase the accuracy of the storm surge forecast itself.

On the other hand, scatterometer winds are not yet suitable to be assimilated into the regional Limited Area Atmospheric models because of the poor re-visitation period (1.5 datum/day maximum). Nevertheless they are very important to understand and quantify the discrepancies of Numerical Weather Prediction (NWP) model fields.

As part of the project, dedicated re-analyses of selected storm surge events have demonstrated that scatterometer data can be used to limit the bias between model forecasts and observations, thus reducing the uncertainties concerning the forcing NWP wind fields. Moreover, coastal altimetry has proven to bring a significant improvement of the knowledge of the initial sea level field across the basin. Indeed, measurements of Total Water Level Envelope (TWLE) have been assimilated into a Storm Surge Model (SSM) with a dual 4d-VAR (4d-PSAS) assimilation technique which allows the assimilation of model errors. Due to the lack of a common terrestrial reference frame, the TWLE observations and the surge level simulated by SSMs cannot be compared directly. However, their profiles should have comparable shapes, up to an additive constant, thus permitting the assimilation into SSMs of their differences.

This contribution describes the methodology able to bring satellite observations of radar altimetry and scatterometry into storm surge modelling. Considering the re-analysed storm surge events, the rms error on the estimation of the maximum surge peak (observed – SSM) reduced by 35% using only scatterometer data, by 11 % using only altimetry data, and 40% using both.

The technique of direct assimilation of altimetry data into the storm surge model, although the promising results achieved, needs further refinements, while the strategy of mitigation of the bias between model and scatterometer winds resulted reliable and easy to set up in the operational context.

The International GNSS Service (IGS) Tide Gauge Benchmark Monitoring (TIGA) Working Group (WG) has recently finalized their reprocessing campaign, using all relevant Global Positioning System (GPS) observations from 1995 to 2015. This re-processed dataset will provide high quality estimates of land motions, enabling regional and global high-precision geophysical/geodetic studies. Several of the individual TIGA Analysis Centers (TACs) have completed processing the full history of GPS observations recorded by the IGS global network, as well as, many other GPS stations at or close to tide gauges, which are available from the TIGA data centre at the University of La Rochelle (www.sonel.org). The TAC solutions contain a total of over 700 stations. Following the recent improvements in processing models and strategies, this is the first complete reprocessing attempt by the TIGA WG to provide homogeneous position time series. The TIGA Combination Centre (TCC) at the University of Luxembourg (UL) has computed a first multi-year weekly combined solution using GLOBK software package. This combination allows an evaluation of any effects from the individual TAC contributions and their influences on the combined solution. In this study we will present the first UL TIGA multi-year combination results and discuss these in terms of geocentric sea level changes.

In this poster, the results of using ship detection algorithms on Sentinel-1A (S1A), TanDEM-X (TDX) and TerraSAR-X (TSX) data are discussed. In total, 138 images were acquired over an area of interest surrounding the Pitcairn Islands, consisting of the Pitcairn Island Exclusive Economic Zone (EEZ) surrounded by a 100 nautical mile buffer zone. 41 images were obtained using S1A, while the rest were acquired using TDX and TSX, throughout the period of January - March 2015. The Pitcairn Islands are a remote island chain in the South Pacific ocean, and, in early 2015, the largest continuous marine reserve in the world was established around them. As such, monitoring the area for illegal fishing activity is of huge ecological importance.

A tailor-made algorithm was used to process the data, making use of Sentinel 1 Toolbox and Next ESA SAR Toolbox (NEST) routines, applying the most up-to-date calibration files. Ships were then detected using the Constant False Alarm Rate (CFAR) detection algorithm, with the variable input parameters adjusted for each sensor and polarisation in order to optimise the ship detection rate. The CFAR detection algorithm evaluates the probability of each pixel within a “guard window” being a false alarm, by comparing each pixel’s statistics against those of its surrounding pixels, which are considered to be background noise. Given a user-defined false alarm probability threshold, this is then used by the algorithm to determine which bright pixels are real detections, and which are false alarms. Finally, the algorithm clusters together pixels that have been classified as real detections and, using a user-defined minimum target size, determines whether the pixel clusters are objects (i.e. ships, in this case) or not.

In total, the algorithm detected 6 ships in the 138 images, 2 of which were in TSX/TDX images, and the remaining 4 were seen S1A images (4 in Extra-Wide swath (EW) mode, 1 in Interferometric Wide swath (IW) mode). However, one ship in a TDX image was not detected by the algorithm, and was instead found while the image was being visually cross-examined. Another ship was not detected in a S1A EW image, and was detected after the image was cross-checked with contemporary Automatic Identification System (AIS) signals taken over the same area. The latter might be due to the fact that the calibration of the S1A EW mode had not yet been completed when the data were processed.

We therefore conclude that the method here presented to detect ships in TDX/TSX and S1A images is relatively reliable, and demonstrates that SAR imagery is a useful tool in monitoring fishing activity in remote areas. However, it needs to be further tested for the S1A EW mode, to include use of the most recent calibration database. Further work is also currently being carried out to determine whether there is a correlation between the detected ship signal, and its physical characteristics, mainly its shape and type (e.g. cargo ship, oil tanker etc.). Finally, work is currently underway to adapt this ship detection algorithm for use with other satellites, such as RadarSat-2.

We present an algorithm to retrieve water inherent optical properties (IOPs) from ocean colour satellite data using an optimal estimation procedure.

The retrieval includes three main steps: The atmospheric characterization, the atmospheric correction and the characterization of the water body.

Regarding the atmospheric characterization, the atmospheric state, by means of aerosol optical thickness, fine mode fraction and wind speed, is retrieved from top-of-atmosphere radiances (LTOA) and given principle components of the water leaving reflectance at all available channels. Our aerosol model is based on the maritime aerosol model from Ahmad et al. (2010). The wind blown ocean surface is described by wave facets whose normals are statistically distributed using the Cox and Munk (1950) model. As to the atmospheric correction, the water leaving reflectance is retrieved for this atmospheric state. Finally a set of five IOPs (absorption of chlorophyll, yellow substance and detritus, and scattering of fine and coarse hydrosols) is retrieved from the water leaving reflectance.

The atmospheric and water body characterization is based on an optimal estimation approach, which allows a consistent treatment of uncertainties and prior knowledge. This approach requires a forward model, which is implemented by means of pre-calculated look-up-tables, based on radiative transfer simulations.

The presented method is as well sensor independent as globally applicable, while the utilization of the apriori state and it's covariance allows a unitary regionalization and the instrumental specifics are quantified by the measurement vector and it's covariance.

The causes of Cévenol events (affecting regularly Southern France) seem to be well understood by meteorologists, however models might be improved by integrating EO data, to increase the accuracy of forecasts.

To this purpose, Sea Surface Temperature data, routinely acquired by satellites, may represent a potential useful input.

We have processed time-series of Medspiration L4 SST data produced over the Mediterranean, centered on the date of the Cévenol event and extending the observation time 10 days around the event. We have also considered 10 years of observations to synthesize the expected climatology for the considered dates.

By simple comparison of the daily SST values with the climatology, we estimated SST anomalies in the temporal adjacency to the event. We repeated the test for all the events classified by Meteo France as “Cévenol” for which we had SST data. 

We found a spatial/temporal correlation between Cévenol events and hotspots near the gulf of Lyon and we believe that assimilation of the spatial information deriving from the localisation of such anomalies in a numerical simulation model at high resolution could help enhancing the capability to refine forecasts for location and time of Cévenol events. 

The launch of the Sentinel-3 mission by the end of 2015 will sustain the operational production of SST data offering the modelling community with continuity of observations for at least 7-years (life expectation for one single satellite). 

Monitoring of Water Quality (WQ) in coastal areas is of concern to many human activities for their relevant ecosystem and economic values. The EU issued the Water Framework Directive (WFD) and the Marine Strategy Framework Directive (MSFD) for fostering the correct management and monitoring of coastal waters and nearby open sea.  The Member States implemented them into their respective national legislation by appointing a central institute for coordination and reporting to EU and by assigning the monitoring to local environmental authorities/agencies: Italy is an example of that, where the measurement of the WQ parameters is demanded to regional authorities (ARPA) which perform regular sampling campaigns at sea and coordinated at national level by a central institute (ISPRA). However such practices are expensive and the area covered and the frequency of the campaigns are limited and need to be integrated by other means.
WQ measurements from Earth Observation (EO) is a common practices based on past (ENVISAT), present (AQUA, TERRA, VIIRS) and future (Sentinel-3) missions, but up to know they lack in spatial details (ranging between 300m and 1km) and accuracy in the proximity of the shoreline, which prevent them to be a concrete support for coastal areas monitoring. The Integrated Coastal Water Management for Mediterranean (ICWM for MED) project – an ESA Integrated Application Programme, under ARTES 20, funded project – has the objective to develop an operational service by integrating different satellite assets (EO, satellite communications and GNSS), real time in situ measurements and other local information, able to provide effective support to the WQ monitoring and surveillance of coastal areas.  The project includes a final demonstration by means of a pilot in a coastal area of the central Tyrrhenian sea. The integrated system is able to provide various products, among which spatially- and accuracy- enhanced WQ measurements (chlorophyll, water transparency and turbidity) by fusing EO data (MODIS/Sentinel-3) and in situ measurements. A test case for the use of Sentinel-2 data is included, in order to exploit its spatial and spectral characteristics.
This paper presents the outcomes of the methodology adaptations to the Sentinel-2 bands – including the evaluation of the most relevant ones – for the provision of high resolution water quality maps. Furthermore the evaluation of the expected accuracy and the preliminary results from the implementation of a real test case into the pilot demonstration area are illustrated.

Sea surface temperature (SST) is a geophysical parameter important for both weather forecasting and climate monitoring. Satellite instruments provide SST measurements by using a retrieval method to convert the observed signals to SST. The satellite SST retrieval can be calibrated using in-situ observations, and its quality is validated against in-situ observations. In both cases some predefined collocation criteria in space and time with in-situ observations are needed. Among the available in-situ SST observations, drifting buoys are widely used for the calibration and validation of satellite SST retrievals. This is because they are less biased in comparison to ships, while they provide global coverage in comparison to moored buoys. Given the demand for high resolution SST observations, which can be provided only by satellites on a global scale, there is an on going effort to produce SST retrievals with finer resolution approaching the pixel level of each satellite instrument. The principal aim of this study is to determine the limits of the collocation criteria both in space and time for which the comparison of SST measurements between satellite and in-situ is within a certain threshold of accuracy. A second target is to estimate the random error of the drifting buoys. For these purposes match-ups between drifting buoys are found using the HadIOD1 climate dataset. The match-ups are restricted to the sub-mesoscale, i.e. 1-50 km and 0-6 hours for this study. While the drifting buoy observations in HadIOD1 start in 1981, match-ups between buoy pairs take place mainly from 2002 onwards. This is due to an increase in both the number of deployed buoys on the ocean, and the number of observations reported by them. The SST absolute difference is less than or equal to the level of 0.1 K, a threshold imposed for climate quality satellite SST retrievals, when the match-up window size is 5 km – 90 minutes or less. The window size providing this level of accuracy is stable through time with the exception of 2005, and to a lesser degree of 2004. This is demonstrated after the elimination of outliers with SST absolute differences more than 0.8 K, discarding less than 2% of the total number of buoy pairs that have differences mainly of ±1 K. The random error of the drifting buoys is 0.08 K, at least 50% less than the estimation of previous studies. This value is obtained from the match-up window of 1 km - 10 min, after the elimination of outliers. Finally, the random error is stable at this level for all years except in 2005, indicating a degradation of the quality of the drifting buoys for this year.

The aim of this study is to estimate water quality parameters in the Three Gorges Dam area quantifying turbidity and suspended solid sediment retrieved by several algorithms and different satellite sensors.

The year 2006 is selected as dry year, while 2004 as normal year and 2002 for floods events.

Due to unavailability of ground data to calibrate new algorithms, literature methods which have been previously developed for comparable water quality conditions will be applied and compared, using surface reflectance values from MODIS MOD09GQ (pixel size 250x250 m) and MERIS FR L2 (pixel size 300x300m).

Hence, changes in turbidity and suspended solid sediment concentrations upstream and downstream of the dam will be computed.

To evaluate seasonal and yearly variations the retrieved values will be correlated to precipitation, river discharge and flood events.

The aim of water quality analyses is a first quantitative assessment of how the Three Gorges dam’s flow regulation affects the sediment transport upstream and downstream.

In this work we use a global colocalized ocean surface observations in near nadir Ka and Ku band microwave measurements to comment on their discrepancies and the different physical multi-scale processes involved in both bands. The latter are from The Dual Frequency Precipitation Radar (DPR) onboard the Global Precipitation Measurement (GPM) platform and are colocalized with the wave model Wave Watch III giving access to local wave and wind conditions. Sensitivity of the radar cross section with wind speed (U10) and significant wave height (SWH) are compared and analysed in term of physical processes involved in the scattering phenomenon. We focus on the relationship between Ka and Ku cross sections fall-off with incidence to provide a valuable insight on the roughness structure (slope and curvature) of the ocean surface. Future spatial missions can take benefit of this work for possible inter-validation/comparison in this frequency regime and range of incidence (SWOT, CFOSAT, ...)

Large rivers are key elements of the global hydrologic cycle and important links between terrestrial hydrology with the ocean. The freshwater associated with major river plumes can influence air-sea interaction by affecting the vertical stratification in the surface layer and thus the efficiency of vertical heat transfer. The Mississippi River, the largest river in North America, provides a major contribution of freshwater into the Gulf of Mexico and has implications to ocean circulation in the subtropical North Atlantic. Monitoring of the spatial and temporal variability of the Mississippi plume extension is therefore important to the physics and biophysical interaction at regional and basin scales. In this study, we use data from the ESA SMOS and NASA Aquarius/SAC-D satellites to study the seasonal and interannual variations of sea surface salinity (SSS) in the Gulf of Mexico near the outflow of the Mississippi River. A seasonal cycle is clearly observed in the SSS, with SSS maximum occurring in December-April and SSS minimum in July-August. The averaged magnitude of the seasonal change is about 4 psu. Interannual changes are also observed over the 2010-2014 period, with the lowest SSS in summer 2011 and the highest SSS in summer 2012, and a difference of 3 pss. Therefore, the interannual changes magnitude is comparable to that of the seasonal cycle. Our analysis suggests that the seasonal and interannual variations of SSS near the mouth of the Mississippi River are consistent with variations in the River discharge. Evaporation minus precipitation (E-P) flux plays a minor role.

Recent studies have shown that mesoscale and submesoscale dynamics play an important role in setting frontal structure in the upper layers of the ocean. Small structures at submesoscale have been revealed by high-resolution infra-red tracer sensors however data are usually sporadic­ as sensors are sensitive to clouds. Meanwhile, mesoscale eddies are currently best observed by altimetric satellites and tracers such as the sea surface temperature (SST) are available systematically at low-resolution using microwave sensors (not sensitive to clouds). All these observations are complementary and combining them would provide a refined assessment of the ocean surface and thus, improve our understanding of mesoscale-submesoscale interactions as well as frontogenesis processes. Studies (Dencausse et al., 2013) have shown that Lagrangian advection with altimetric geostrophic velocities can be used to stir large scale ocean tracer fields and reconstitute small scale fronts and eddies. The goal of this work is to apply several methods of Lagrangian advection to a low-resolution SST field and compare the resulting high-resolution tracer with high-resolution SST observation on a cloud-free day. The first one is a simple Lagrangian advection using geostrophic velocities from AVISO. Fronts and eddies are consistent with observations but filaments are unrealistic. The second method simulates diffusion as well as source and sink along the advection process using the nudging of the low-resolution SST from AMSR-E, and enables to smooth unrealistic filaments. Results from both methods are compared with an Infra-Red High Resolution SST image from MODIS in terms of RMS as well as frontal structures.

Winds in the coastal zone have importance for near-shore wind farm planning. Recently the Danish Energy Agency gave new options for placing offshore wind farms much closer to the coastlines than previously. The new tender areas are located from 3 to 8 km from the coast. Ground-based scanning lidar located on land can partly cover this area out to around 15 km. In order to improve wind farm planning for near-shore coastal areas, the project‘Reducing the Uncertainty of Near-shore Energy estimates from meso- and micro-scale wind models’ (RUNE) is established. The measurement campaign starts October 2015 and has 3-month duration at the Danish North Sea coast at around 56.5°N, 8.2°E.

Ocean surface winds derived from Sentinel-1 will be compared to the ground-based scanning lidar observations of winds as well as to winds observed at the coastline, at a floating wind lidar buoy and at a wave buoy. The various observation types have advantages and limitations; one advantage of both the Sentinel-1 and the scanning lidar is that they both observe wind fields covering a large area and so can be combined for studying the spatial variability of winds. Sentinel-1 are being processed near-real-time at DTU Wind Energy (Badger et al. 2016) using GFS winds as input. Wind direction can be checked from the various other observations. Sensitivity to possible deviations in wind directions in the near-shore area will be investigated. Furthermore, oceanic features not related to winds but to e.g. surface current, breaking waves, etc. will be investigated. The plan is to establish high-quality coastal wind speed cases based on Sentinel-1 for quantification of the coastal winds, for verification of wind resource modelling best practices in the coastal zone.

The study is supported by RUNE and New European Wind Atlas projects and satellite data from Copernicus Sentinel-1.

References

Badger et al. 2016 abstract (same issue)

1. Introduction

 

The Chinese and French Space Agencies propose to jointly carry out an innovative mission, CFOSAT (China France Oceanography Satellite project) devoted to the monitoring of the ocean surface and its related science and applications. CFOSAT will embark both a wind and a wave scatterometers, enabling a simultaneous measure of the wind and the wave vectors with a global coverage for the first time. The launch is planned for mid-2018.

The satellite embarks two payloads; both are Ku-band radar scanning around the vertical axis:

-       the wave scatterometer SWIM, a rotating 6-beams radar at small incidence (0 to 10°) [1, 2],

-       the wind scatterometer SCAT, a fan-beam radar at larger incidence angles (30° to 50°) [3].

 

This paper focuses on the processing of the SWIM data for the retrieval of the 2D wave spectrum and, especially the estimation of the speckle spectrum.

 

 

2. Overview of the SWIM data processing

 

See pdf for fig. 1.

Figure 1. Overview of the SWIM L1 and L2 products and the associated processing.

 

The SWIM data are of three kinds [4]: the normalized radar cross-section profiles from 0° to 11°, the SWH and wind speed from the nadir beam and the wave spectrum from the 6°, 8° and 10°. The processing steps are following:

-       inversion of the measured power to normalized radar cross section (inversion of the radar equation) and geolocalization of each radar gate,

-       estimation of averaged normalized radar cross section per step of 0.5° in incidence and 15° in azimuth,

-       retracking of the nadir waveform to estimate the SWH and the wind speed,

-       estimation of the modulation spectrum from the s⁰ fluctuations and correction of the spectrum from speckle and impulse response,

-       computation of the wave spectrum from modulation spectrum through the computation of the modulation transfer function.

There are two major issues: the estimation of the speckle spectrum and the estimation of the modulation transfer function. The paper addresses the issue of the speckle spectrum.

 

See pdf for fig. 2.

Figure 2. Illustrations of the SWIM products for wave spectrum computation.

 

3. Estimation of the speckle spectrum

The wave spectrum is estimated from the density spectrum of the fluctuation of the backscattering coefficient. This density spectrum is equal to [1, 2, 5]:

 P_ds₀(k,Φ)=δ(k)+R(k)P_m(k,,Φ)+1/(N_impL_dis)P_sp(k)+1/(N_impL_dis)P_b(k)

with ds₀ the backscattering coefficient fluctuations (due to the long slopes), R the spectrum of the radar impulse response, P_m the modulation spectrum (linearly linked to the  wave spectrum), P_sp the spectrum speckle, P_b the thermal noise spectrum, N_imp the number of pulses averaged in time, L_dis the number of averaged range gates. The aim is to estimate P_m.

The speckle spectrum depends on the real number of independent pulses [5]. The dependency of the spectrum to the sea state conditions is still under investigation. Three main methods are used:

-       analytical methods,

-       estimation from the noise floor,

-       use of cross-spectrum computation.

The three methods will be detailed and discussed.

 

4. Results on simulated data and airborne data

 

The three methods have been tested on simulated data and on real airborne data. The results will be shown and discussed.

See pdf for fig. 3.

Figure 3. Illustrations of speckle estimation with different methods.

 

 5. References

[1] Enjolras V., L. Rey, T. Amiot, C. Tison, P. Castillan, SWIM, a state of the art multi-incidence beams Ku-band waves scatterometer to go beyond current radar systems, in IGARSS’09, July 2009

[2] Hauser, D., Soussi, E., Thouvenot, E., and Rey, L.,“SWIMSAT: a real aperture radar to measure directional spectra of ocean waves from space  main characteristics and performance simulation”. Jour. Atmos. Oceanic Tech., 18, 2001

[3] Zhu J., X. Dong, W. Lin, X. Xu, Calibration and estimation of attitude errors for a rotating fan-beam scatterometer using calibration ground stations, IEEE JSTARS, PP(9), 2014

[4] C. Tison, D. Hauser, L. Delaye, T. Koleck, N. Lamquin, M. Planells, F. Gouillon, P. Castillan, Processing of the CFOSAT-SWIM data: algorithm prototyping and simulations, IGARSS’15, 2015

[5] D. Hauser, G. Caudal, R. Valentin, C. Legac, N. Pauwels, C. Tison, A study of speckle properties over the ocean surface from the airborne radar Kuros, ENVIREM, Gif-sur-Yvette (France), 9-10 Juin 2015

 

Conventional spaceborne altimeters determine the sea surface height with an accuracy of a few centimeters. Although satellite altimetry may be regarded as a mature technology, altimeter observations collected over coastal regions suffer from numerous effects which degrade their quality. For example, land and bright targets contaminate the radar footprint, and the range and geophysical corrections (wet troposphere, tides, and high-frequency atmospheric and ocean signals) are notoriously difficult to model in the coastal zone. The Norwegian coast adds further complications, due to the many islands, mountains, and deep, narrow fjords.

As the first of its kind, the European Space Agency's CryoSat-2 satellite carries a Synthetic aperture Interferometric Radar ALtimeter (SIRAL) which can operate in Synthetic Aperture Radar (SAR), SAR Interferometric (SARIn), as well as conventional Low Rate (LR) modes. When operating in SARIn mode, the altimeter measures the phase difference of the backscattered signal at two antennas, from which the position of any backscattered point may be derived. Thus, the SARIn mode may help discriminating and mitigating land contamination signals from off-nadir land targets (e.g., steep cliffs) over coastal regions.

Conventional altimetry (Envisat, Jason-2) and specific coastal products (CTOH, PISTACH) have recently been tested along the Norwegian coast, where it was concluded that the coastal products did not offer an improvement over the conventional products they are based on. We therefore investigate the potential for CryoSat-2 data to provide improved ocean measurements in the Norwegian coastal zone. In particular, we make use of CryoSat-2's SAR and SARIn modes and determine coastal sea surface heights in specific coastal regions, and compare results with independent sea-level observations.

Ship monitoring is a key service for most of the authorities with competences on sea surveillance. Especially important becomes to the European Member States with coastline facing the Mediterranean Sea or the Atlantic Ocean where illicit activities, such as border-crossing, drug dealing or terrorist attacks are taking place. Classic approaches to tackle the problem by means of in-situ surveillance assets (like visual inspection by plane or ship, or ground-based sensors) have shown proven limitations with respect to the associated budget and the area that they can monitor at the same time. Along the last two decades, alternatives to overcome such drawback have been made available in terms of Remote Sensing (RS) sensors boarded on satellite platforms (Earth Observation –EO-). Synthetic Aperture Radars (SAR) and Optic sensors have been the most proliferated technology to support service provision in the field of Maritime Surveillance Awareness (MSA). They have shown proven performance for ship detection and tracking (in support of cooperative polling) in wide areas under almost any observation conditions. Different research studies (most of them cofinanced by public institutions) have shown that EO-based ship monitoring is a perfect complement to classic surveillance methods and outcomes can be used in operational and tactical missions. Example would be the efficient delineation of the paths that in-situ assets should follow to maximize the chances to locate targets of interest.

In this context, GMV have started research activities in the field of MSA ten years ago within the framework of different research projects, like MARISS, LIMES, DECLIMS, NEREIDS or SAGRES. The gathered experience has permitted developing a Ship Monitoring System (SIMONS) capable to cover most of key user requirements. SIMONS has been developed as a standalone application that run in background within a processing layer that can be integrated with any external cartographic visor. Modular design has been the design backbone that provides flexibility, scalability and processing efficiency. The main features of SIMONS are:

 

EO image post-processing to detect and categorize ships at sea. Main concerns are putted in the detection of small and non-metallic ships, and the provision of enough parameters for their classification.

EO image post-processing to delineate the coast by uniquely using the information within the image

Advanced processing of cooperative means with alarm detection, route propagation and constraint-based track delineation (coastline, bathymetry, harbour…)

Integration of the information derived by in-situ or local devices, such as coastal radars, ship radars or underwater sensors

Integration of supporting information derived from statistics and external reports that can shed light on the best way to exploit the available assets. Strategic analysis is benefitted to decision making

Standardization of the dissemination mechanisms and fully integration into user systems. SIMONS publish the output reports via web services (WFS, WMS, CSW) that can fed open-source or any user-customized cartographic visor. SIMONS can interact with the EMSA system ImDATe as the same formats are adopted

Capability to success in load and performance tests that assure that system performance is not dropped down when the load of the system increases

 

In the current paper, the latest operational campaigns conducted with SIMONS will be presented and analysed. Among others, a key milestone has been the detection of a small boat with irregular immigrants on board that was under a Search and Rescue operation within the Alboran Sea. The EO-based detection report served authorities to quickly locate the target within an area of 50,000 km2, which otherwise would need more time and efforts to be located. Other experiments that would be topic of discussion will be the test exercise conducted in the Tokyo Bay where accurate ground-truth is available. The results will show progress in the following topics:

Target detection of small, non-metallic and fast boats

Estimation of macro-scale features and target classification by only using EO data

Advance EO processing to diminish image artefacts and distortions

Discrimination of any other target different than ships within the sea

Target tracking by fusing cooperative and non-cooperative (EO) data

Constraint-based track reconstruction

Detection of anomalous behaviour to support decision making

Environmental assessment to locate areas where the probability that certain activities of interest occur is higher

Monitoring of specific ports and shore areas

Complement the information provided by ship and airborne patrols

Sea level changes in the Eastern Mediterranean Sea are examined through the use of altimetric data (SSALTO/DUACS gridded DT MSLA and along-track DT SLA (Mediterranean Sea)), as well as data from tide-gauge stations in our research area.

Annual Sea Level Anomalies for the period between 1993 and 2014 were obtained by averaging daily values of SSALTO/DUACS gridded Delayed-Time Map of Sea Level Anomalies for the Mediterranean Sea. The results showed a small increase of Sea Level Anomaly (SLA) over the period under study. SLA in the Eastern Mediterranean was minimum in 2007, while in 2010 was maximum.

Annual Sea Level Anomalies from 1992 to 2014 were also computed at locations near tide-gauge stations in the Eastern Mediterranean Sea, using along-track DT SLA from 14 satellite missions. In particular, averages of SLA of the nearest points to the tide-gauge stations, as well as averages of SLA of all the points, in a distance of 30 km and 50 km respectively from the stations were used in order to estimate annual SLA.

Annual Mean Sea Level data from the Permanent Service for Mean Sea Level (PSMSL) and Oceanography Center of Cyprus were used in this study. For a fair comparison with the altimetric data, Dynamic Atmospheric Correction was applied to tide-gauge data from 17 stations in our research area. The results showed that sea level variations from altimetry and tide-gauges are correlated and of a similar magnitude.

In order to improve our knowledge of primary production within intertidal mudflats, satellite remote-sensing (RS) was used to provide a large-scale, synoptic view of their main primary producers: benthic microalgae and cyanobacteria – commonly referred as microphytobenthos (MPB) – which form transient biofilms at the surface of the sediments during low tide. Although MPB primary role in the functioning of coastal ecosystems is increasingly studied, to date RS time-series have not been analyzed to their full potential.

In this study, high resolution remotely-sensed data were analyzed in order to characterize MPB seasonal variability as well as its spatial interactions with local benthic communities. For this purpose, a site within a macrotidal bay (Bourgneuf Bay) located along the French Atlantic coast was selected. A focus was placed upon the analysis of a BACI (Before After Control Impact) experiment launched in July 2014, which consisted in the removal of wild oysters from a reef. The main objectives of this work were: 1) to characterize MPB seasonal variability in order to identify local seasonal trends, 2) to analyze time-series of MPB spatial distribution around oysters reefs, and 3) to analyze MPB spatial response to the ecological experiment.

Time-series datasets were built using multi-sensor data coming from Landsat and SPOT satellites (including SPOT-4 and SPOT-5 Take-5 data). Upcoming Sentinel-2 (and later Sentinel-3) data, which will be made available thanks to the Sentinel-Copernicus programs, are to be included in order to complete this time-series. Two atmospheric corrections were evaluated: FLAASH and MACCS. MPB was estimated using a proxy of the chlorophyll a, the Normalized Difference Vegetation Index (NDVI), based on the RS Red and IR spectral bands. Hyperspectral field measurements were also conducted in order to ground-truth the remotely-sensed data. Algorithms enabling the homogenization of signals coming from multiple sensors were developed and consistently applied to the data in order to limit the background noise caused by differences in instrument features

The first results of MPB seasonal trend analysis revealed higher NDVI during spring and fall. In the context of the ecological experiment, spatial responses of MPB to the removal of oysters highlighted their impact on MPB spatial distribution, tending toward the hypothesis of a top-down control between MPB and oyster communities. Further analyses of upcoming satellite data will be performed in order to monitor the patterns of recolonization and development of MPB biofilms in response to the ecological experiment.

This work was also part of a broader issue about the analysis of spatial and temporal dynamics of MPB using multi-sensor RS data, which are currently studied through a PhD supported by the CNES. With environmental issues such as coastal erosion or climate change, MPB role in coastal ecosystems is increasingly brought to light. In this context, the use of existing and future data of Sentinel-2 and -3 will enable the building of high spatial and temporal resolution time-series for the purpose of the observation of coastal zones.

Time series of regional sea level budgets provide a better understanding of the causes for sea level variations as well as a better understanding of the errors of the measurement systems involved. This study therefore aims at closing the budgets in the Northern Atlantic at sub-basin scale by using data from the Jason satellites, GRACE and the ARGO floats. Time series for altimetry are created using along-track data from the RADS database. A new averaging method is applied to altimetry data to overcome problems with conventional gridding (a large ground-track seperation at the equator) and the latitude weighting (underweighting at high latitudes). Additionally, geographical dependencies of the intermission biases are considered. Using the full covariance matrices provided by the Center for Space Research (U. Texas), GRACE solutions are statistically filtered to minimize hydrological leakage and reduce striping. The ARGO data are statistically interpolated to a grid with full variance-covariance matrices. For all three systems, the errors are propagated to provide time series of monthly averages over the considered region including formal error bars. We will show that our approach allows us reach closure of the sea level budgets.

In coastal areas the knowledge of wave period is essential for bathymetry estimation. Typically it is obtained using wave buoys and/or wave models. While wave buoys provide near-continuous data, these measurements may significantly differ from one point to another. Wave models have been used to propagate wave measurements throughout the entire zone. However such models rely on local data such as bathymetry or tidal component. We propose a method to estimate the wave period field in coastal areas entirely from satellite images, based on linear wave theory and wave tracing.

Wave tracing consists of following the path of a wave with a known wavelength from the open sea to the coast. From a given patch of the image, the main wavelength and its direction are calculated and used to locate the next patch until the coast is reached. The entire image is divided into tiles, each of them potentially used to spark the first step of wave tracing. The use of linear wave theory implies the consideration of two water domains: 1) the deep sea domain defined by a bathymetry greater than half the dominant wavelength and 2) the intermediate and shallow water domain. In the deep water domain, linear wave theory gives an immediate estimation of the wave period regarding the wavelength and the gravitational acceleration. In the intermediate and shallow water domain the wave period depends also on the bathymetry. For this reason we use the derivative of the linear wave theory from deep sea domain to the intermediate and shallow water domain in order to estimate the wave period [1].

We verify our method on a Hawaiian Islands site for which a large amount of data is available. The in-situ measurements for wave period comparisons were provided by the National Data Buoy Center network. Moreover, the Hawaiian Islands dispose of an online bathymetry chart. Using Landsat 8 archive and Sentinel images, the first results of wave period estimations show good agreement with buoy measurements both for wave period and wave direction. Using the wavelength and the wave period obtained, we perform bathymetry estimation along wave traces. Comparison with in-situ bathymetry reveals promising results.

In summary, this work founded by the ESA Living Planet Fellowship shows that satellite imagery may be used in an operational way to extract bathymetry information. The proposed approach is not limited by the use of a particular remote sensing instrument. It may theoretically be applied to any kind of optical and SAR images provided that the waves are correctly captured.

[1] Dalrymple, R. A., Kennedy, A. B., Kirby, J. T. et Chen, Q.(1998). Determining depth from remotely-sensed images. Coastal Engineering Proceedings,1(26).

The Infrared Sea surface temperature Autonomous Radiometer (ISAR) was developed to provide reference data for the validation of satellite SST_skin temperature data products, particularly the Advanced Along Track Scanning Radiometer (AATSR). Since March 2004 ISAR instruments have been deployed nearly continuously on ferries crossing the English Channel and the Bay of Biscay, between Portsmouth (UK) and Bilbao/Santander (Spain). The resulting twelve years of ISAR data, including an individual uncertainty estimate for each SST record, are calibrated with traceability to national standards (NIST and NPL, FRM4STS). They provide a unique independent in situ reference dataset against which to validate satellite derived products. We present results of the AATSR validation, and show the use of ISAR fiducial reference measurements as a common traceable validation data source for both AATSR and SLSTR. ISAR data were also used to review performance of the OSTIA SST analysis before and after the demise of ENVISAT when AATSR inputs ceased This demonstrates use of the ISAR reference data set for validating the SST climatologies that will bridge the data gap between AATSR and SLTSR.

This paper proposes a novel multilevel approach to ship classification in Sentinel-1 SAR images based on artificial neural networks and a robust feature extraction and selection scheme. After performing some standard SAR image pre-processing and CFAR-based ship detection algorithmic procedures, a proper feature extraction methodology is followed. A number of diverse types of features, such as scale, shape, intensity and textural ones, are extracted in order to form the utilized feature pool. A two-stage hierarchical feature selection algorithm is utilized next in order to be able to discriminate effectively civilian vessels into four distinct types: cargos, passengers, tankers and small ships. In our analysis, scale and shape features are utilized in order to discriminate smaller types of ships present in the available SAR data, or shape specific ships. Then, the most informative texture and intensity features are incorporated in order to be able to better distinguish the civilian types with high accuracy. A feature selection procedure that utilizes heuristic measures based on features’ statistical characteristics, followed by an exhaustive research with feature sets formed by the most qualified features is carried out, in order to discriminate the most appropriate combination of features for the final classification. The utilized sets of features are finally fed into a two-stage artificial neural network, which effectively classifies the different types of ship targets.

In our analysis, a set of 20 Sentinel-1 SAR data with 10m resolution, coming from the Xios Island in the Aegean Sea of Eastern Greece were used to analyse the detailed characteristics of these types of ships. A total of 350 ships with available AIS data were used in the classification process. The experimental results show that the proposed scheme is able to classify ships at an average precision exceeding 85%. Given the fact that the spatial resolution of Sentinel-1 data is 10m, the specific classification accuracy are very promising and can provide reliable services to near real-time maritime monitoring and sea border surveillance applications. Further investigation of additional features and proper feature selection is currently in progress.

Mineral-rich suspended matter (MSM) is a main optical component of many inland waters, especially those characterized by high terrestrial influence. High MSM concentrations in the areas of river intrusion, land erosion and bottom resuspension cause the reflectance values to shoot up, impairing the water’s light regime with possible ecological consequences.

During field studies in lakes located in catchments dominated by glacial erosion in the Swiss Alps, we encountered MSM concentrations at the surface water layer of up to ~200 mg L^-1. We found low absorptive, highly scattering glacial flour that caused high reflectance values throughout the visible and near infrared domain.

This type of water is found less commonly treated by bio-optical modeling. The high SM content might oppose challenges in terms of non-linear effects. The objective of this study is to compare two current simulation tools for higher MSM concentrations within the realistic context of the catchments of glacial erosion. Forward simulations based on measured inherent optical properties are run with WASI and Hydrolight (Seqouia Scientific, Inc.). WASI is a free-of-charge software tool to simulate and invert light spectra based on Albert and Mobley (2003); the analytical solutions implemented in WASI had been established against reference spectra generated by Hydrolight which solves the equation of radiative transfer numerically. The reference spectra considered SM concentrations up to 50 mg L^-1. With this, we test the suitability of WASI for higher SM concentrations. The closeness of agreement between WASI and Hydrolight outputs indicates the limitations of the analytical solution and points out any necessities for future adaptations to (extremely) turbid environments.

To get the spatial picture a Landsat 8 scene was analyzed. Atmospherically corrected data matched well with our in-situ reflectance measurement (Cal/Val). We conclude the study by a constituent retrieval based on satellite data in WASI.

We present results from ESA’s Climate Change Initiative (CCI) for sea surface temperature (SST) on progress towards including passive microwave SSTs in the climate data record. Measurements from a passive microwave radiometer (AMSR2) are incorporated into an optimal estimation framework similar to that currently employed for infrared instruments. The scheme has been augmented to make it robust to systematic differences between simulated and actual radiances with the development of a “bias-aware” retrieval method.

Microwave sensors offer the prospect of more complete coverage of SST than that derived from infrared wavelengths by virtue of being able to make measurements through cloud. The availability of additional observations from persistently cloudy regions potentially reduces uncertainty in the SST climate record. However, microwave SSTs correspond to a different radiometric skin depth having a different mean SST, are of poorer spatial resolution and are less accurate in general than those derived from infrared wavelengths. In these regards, their introduction into the climate record for SST could adversely affect uncertainty and long-term stability of observation. Thus, to be of use in the climate record, microwave SST retrievals must achieve usefully low biases (<0.1 K) once skin effects are accounted for and good observational stability (<0.05 K / decade).

An optimal estimation retrieval scheme requires a forward model that can simulate the radiances that would result from a given initial estimate of the geophysical conditions in the ocean and atmosphere. For this application, we use the RTTOV radiative transfer model with a priori information on the conditions taken from ERA-interim analyses. Using a database of matches between in situ buoymeasurements and satellite overpasses with a standard optimal estimation scheme, significant biases are apparent between the modelled radiances and those observed by the instrument. These show dependence both on the observed radiance and the wind speed. The radiance dependence is thought to be dominated by calibration effects. The wind-speed dependence results from limited accuracy in estimating surface emissivity, particularly for the low-frequency channels that are also most sensitive to SST.

We have augmented “classic” optimal estimation to be bias-aware, such that bias corrections are retrieved alongside the physical parameters. This approach significantly improves the retrieval accuracy over standard optimal estimation, greatly reducing the instrument calibration and modelling sources of bias.  Applying the scheme to a database of matches to independent drifting buoys gives results approaching the requirements in terms of low bias, with standard deviations comparable to results from previous investigators.

Freshwater water bodies are usually affected by human intervention and activities worldwide, thus their quality and composition get modified. This renders freshwater bodies less suitable for uses that they would have been suitable for in their natural unmodified state. In Europe, in order to alleviate the degradation of surface and ground waters, the Water Framework Directive (WFD) 2000/60/E.C. has been established. One of the main purposes of WFD is to ensure good ecological and chemical status of all surface waters in the European Union countries by 2015.

Towards the definition of the ecological quality, biological, physico-chemical and hydromorphological quality elements (QE) have to be assessed and compared with reference conditions. In many cases, water quality data and especially the historical one either they do not exist or they are very limited.  Databases from in situ measurements often refer to sporadically collected data or to poorly documented, non-consistent information. Nevertheless, a substantial part of this “missing” information has been recorded in the historical archives of satellite imagery, but has never been retrieved. Therefore, satellite remote sensing can be used to map and monitor QE, with the aim to reconstruct the historical evolution of water quality indicators, to reduce cost and time and to face problems of accessibility to the field. To this end, Landsat imagery has been used to estimate QE of several lakes, but typically for relatively short observation periods and/or using a limited number of images.

Lake Koronia is part of Mygdonia basin (Central Macedonia, N. Greece). It is one of the most important Ramsar wetlands of Greece and it is protected as a Natura 2000 site. However, it faces serious environmental issues and water management problems, such as decreasing water levels, deterioration of water quality, water salinization and almost extinction of the ecosystem. During some time periods through the year it is almost dry, while its water depends mainly on precipitation. As a result, the amount, but also the spatial and temporal resolution of in situ data that have been collected during the past decades is limited.

The purpose of this study is to create a dense historical water quality profile of Lake Koronia, following a remote sensing approach. To this end an effort of monitoring the QE, such as total suspended matter, temperature, turbidity, phosphorus and nitrogen content, conductivity and water depth, is carried out by using multispectral satellite images of a time period of 30 years, namely from 1984 to 2015.  The data are images of Landsat 4-5/ΤΜ (Thematic Mapper), Landsat-7/ETM (Enhanced Thematic Mapper), Landsat-8/OLI (Operational Land Imager) and Landsat-8/TIRS (Thermal Infrared Sensor).

Apart from mapping the temporal and spatial water quality variability of Lake Koronia for the past three decades, the results are expected to contribute to: (a) the definition of the optimal image processing routines for the QE estimation and external calibration procedures based on multispectral satellite images and in situ measurements, (b) the assessment of the correlation between QE and Landsat bands and (c) the establishment of procedures that shall allow the compatibility of past satellite information with water quality information derived from future Sentinel-2 data.

Dynamic topography is one of the youngest branches in Oceanographic research and although the subject has been tackled for a long period, research has been scarce compared to other topics. The latest years have brought a shift of view and the scientific community has engaged in a more active approach, mainly in the context of the acute global climate change and the constant need to a have a better understanding of our oceans, the biggest and yet not fully explored systems on our planet.

This research focuses on the North-Western compartment of the Black Sea and aims to identify the changes in dynamic topography over three temporal intervals represented by 2003, a year with hydrological drought, 2005-2006 and 2008-2010, portraying wet periods. The context is to analyze and quantify the influence of the Danube River on the variation of the sea surface topography.

The chosen area represents the discharge zone of the Danube River and the constant intake of terrigenous material is be one of the main causes for changes in thermohaline circulation with echo in the sea surface height fluctuation. Also, the river built a very well developed deltaic system, which faces constant geomorphological transformations, altering the paths of coastal currents and on a bigger scale, having a large impact on the whole marine circulation. On top of that, the area is a crossroad for wind masses, with a very active circulation coming from different diverse neighboring regions, which makes it susceptible to a diversity in weather conditions. Variations on different hydrological parameters such as temperature, correlated with a steady sea level rise in the latest decades, have led to the idea of investigating the triggering factor for all these differences, the impact of climate change on a smaller scale.

The methods used in the research focus on the analysis of the altimetry data from various satellite missions such as Jason 1,2 and Topex/Poseidon integrated with Cryosat-2, ENVISAT and Landsat for additional analysis. The data is acquired using the Copernicus data portal, the Aviso portal and the Earth Explorer portal and processed with BRAT (Basic Radar Altimetry Toolbox) and the STEP application. The time intervals were chosen for a better comparison of the results, as they portray extreme variations on weather conditions and therefore noticeable differences in sea level height. Other local meteorological and hydrological data was used for creating the link between the Danube system and the Black Sea and validating the outcome.

The results create a clear connection between the Danube river hydrological regime in every considered time period and the differences in the dynamic topography of the area, strengthening the synergy between the two systems and generating a new perspective on how the marine circulation in the Black Sea is affected by these variations caused by the climate change.

The investigation is devoted to the study of the variability of the dynamics of the Caspian Seabased on TOPEX /Poseidon and Jason-1/2 satellites altimetry data. The technique of calculating the synoptic dynamic topography (DT) as a superposition of the mean sea climate DT calculated by thermohydrodynamic models of Hydrometeorological Center RF and field of sea level anomalies calculated by satellite altimetry data is considered. An analysis of the geostrophic currents vorticity confirmed the predominance of cyclonic circulation in the Northern and Middle Caspian and anticyclonic in the Southern. After 2008, the swirl in almost all parts of the sea rose, indicating a change in the regime of atmospheric circulation over the water area of the Caspian Sea. After 2008, the vorticity in all parts of the sea increased, indicating a regime change in atmospheric circulation over the water area of the Caspian Sea. Analysis of variation of average velocity and vorticity shows that the average velocity is inversely proportional to the vorticity. Since 1993 to 2007. vorticity rose at a rate of -0,17±0,02•10^-7 per year, , and average velocity has increased at rate of +0,11±0,06 cm/year. After 2008 the situation has changed to the opposite. The vorticity has increased at a rate +0,75±0,12•10^-7 per year, average velocity rose at rate of -0,47±0,19 cm/year.

This study was supported by a series of grants of the Russian Science Foundation (No 14-17-00555).

The  White Sea are the seas of the Arctic Ocean. Today complicated hydrodynamic, tidal, ice, and meteorological regimes of these seas may be investigated on the basis of remote sensing data, specifically of satellite altimetry data. Results of calibration and validation of satellite altimetry measurements (sea surface height and sea surface wind speed) and comparison with regional tidal model show that this type of data may be successfully used in scientific research and in monitoring of the environment. Complex analysis of the tidal regime of the White Sea and comparison between global and regional tidal models show advantages of regional tidal model for use in tidal correction of satellite altimetry data. Examples of using the sea level data in studying long-term variability of the Barents and White Seas are presented. Interannual variability of sea ice edge position is estimated on the basis of altimetry data.

This study was supported by a series of grants of the Russian Foundation for Basic Research (No 16-05-00944).

Penetration of sunlight into the sea and its interaction with sea water, dissolved as well as particulate suspended materials is an important physical phenomenon in the ocean that governs the light availability and associated biological processes like primary production or plankton distribution in the water column.

 

We examine here the in-situ optical hyperspectral measurements of downwelling irradiance (Ed), photosynthetically active radiation (PAR), backscattering coefficient (bb) and chlorophyll fluorescence in the Sognefjord and the Trondheimsfjord in Norway to understand the vertical variations in the magnitude and evaluate the spectral composition of the underwater light field at sampling stations along the two fjords. The in-situ hyperspectral measurements are fundamental in understanding the dynamics that drive these optically complex coastal zone systems for successful retrieval of ocean color bio-geo-physical products through satellite remote sensing that have significant applications in monitoring water quality, ecosystem  assessment, detection of harmful algal blooms and phytoplankton community structure.

 

Investigation revealed a decrease in the depth of occurrence of 1% PAR with increased turbidity as deduced from values of bb. Profiles of bb matched strongly with those of chlorophyll fluorescence at stations upstream. One-to-one correspondence in the depth of occurrence of 1% PAR and maximum chlorophyll fluorescence indicated significant contribution of phytoplankton to the attenuation of light in the fjords. Spectral analysis of 1% light penetration depth at stations outside and end of the fjord exhibited a shift in the relative penetration of light with green wavelengths (570nm) penetrating deeper at the end of the fjord than at outer stations (blue-green). The shift is attributed to greater particle load in suspension in the fjord from terrestrial run-off. These shifts in the dominant wavelengths could lead to implications for phytoplankton community distribution and utilized for laboratory incubation experiments mimicking actual light field conditions.

 

An attempt is made to test algorithms and derive 1% PAR depth from measurements of above water remote sensing reflectance (Rrs) performed during the cruise in concurrent with in-situ profiles for subsequent retrieval of the same from ocean color satellite sensors such as the Sentinel-3 OLCI and ENVISAT MERIS.

Planktonic algae - phytoplankton, causing algal blooms and affect its quality, especially near the coast where phytoplankton is booming. Coastal area is considered as a special dynamic boundary ecotone structure within the concept of ecotone "water-land". Strong anthropogenic impact on urban and agricultural landscapes and remote coastal waters with a transfer of substances and materials is carried out through the buffer zone - Ecotone, which consists of five blocks. First block is a water block - "Aquatic", where there is an active transformation of land substances, recreational, economic use of the territory and also four land blocks. Basic research has been conducted on Lake Ladoga (N 60.8, E 31.5) and also on Chudskoe (Peipsi) lake by a sharing data of ship and satellite measurements (N 58.4, E 27.6) in August-September 2014.

Earth remote sensing data (RSD) are used to distinguish these blocks and implement a full-scale real-time monitoring of their condition, as well as the distribution of phytoplankton in the waters.

The purpose of this project was to create the operational monitoring of phytoplankton "fields" in Lake Ladoga on the basis of satellite imagery.

Following problems were solved.

1) Determination of the optimal data types for operational detection of phytoplankton (free MODIS data - refresh rate at least 1 image/day. LANDSAT-8 – for further test of the algorithm phytoplankton identify).

2) Determination of the optimal data and places of ground truth measurements (i.e. synchronized with satellite imagery), taking into account the timing of the satellites passage over the place shooting (for experimental part and development of interaction between the monitoring groups).

3) Processing of archival field data 2005 - 2012: field data and them maximum corresponding satellite images (development of an algorithm to identify phytoplankton, its spatial distribution and quantification of the biomass).

4) Training of the operational cooperation scheme for ground truth and satellite groups of monitoring for the transfer of information on: the distribution and biomass of phytoplankton and fronts of dense clouds.

5) Feedback - getting operational monitoring ground truth data: coordinates, date and time of measurements, a qualitative assessment of phytoplankton biomass, a representative sample (i.e. uniform distribution in the range about of 500 m from the place of sampling).

 

Low-resolution data from AQUA and TERRA satellites - multispectral instrument MODIS, were processed in accordance with the approach to color scanner data processing that is based on assessments of the relationship λ1/λ2 signals - "color index" for wavelengths λ1 ≈ 440 ... 490 nm and λ2 ≈ 520 ... 560 nm. This is due to the fact that the first range corresponds to a maximum absorption coefficient of chlorophyll-A, the second - its minimum. Also phytoplankton chlorophyll-A absorbs radiation in the red (about 675 nm) region of the spectrum.

Were selected relationship B10/B12 (and close B9/B11) channel radiometer MODIS, corresponding to the spectral ranges B10 = 483 - 493 nm and B12 = 546 - 556 nm. For LANDSAT-8 data were used 2 spectral channels: 2nd (450 - 515 nm) and 3rd (525 - 600 nm). The color index is calculated as the ratio B2/B3 of the spectral brightness.

Were obtained concurrences of "phytoplankton fields" outlines for processed satellite data different spatial resolution (1000 m/pixel for AQUA/TERRA and 30 m/pixel LANDSAT-8). When generalization (coarsening) images LANDSAT-8 to 1000 m/pixel matching of these outlines increases significantly.

Following the thematic processing of all results and data, including ship measurements were exported into an integrated GIS project "Plankton" for further combined analysis of ground truth and satellite data.

Comparison results of the quantification method with ground truth data showed that requires consideration of several points.

1. Satellite image makes it possible to identify fields of phytoplankton distribution at the time of the shooting only.

2. Coarsening of satellite images can lead to a spatial shift in pixels for "phytoplankton fields".

3. You should specifically consider and handle various anomalies (e.g., bandwidth inversion) in the image because of hardware or other failures.

 Results of experimental studies on the Ladoga and Chudskoe lakes were ranked in the table in ascending concentrations of chlorophyll-A, in the adjacent column - color index data. The general course of these values is in good agreement.

The correlation coefficient for these two data sets is very large - about 1 (r = 0.965608 and r = 0.949613 accordingly) - high accuracy quantitative evaluation of concentration of phytoplankton, in good weather conditions for satellite imagery. Interpretation of the correlation coefficient r: 0.7 < r < = 0.9 is a strong correlation, 0.9

 Conclusion

The results of available data processing show a high correlation of satellite data of high and low resolution, which proves the ability to use it in the daily monitoring of phytoplankton. Development of proposals for the organization of operational data monitoring shows a fairly high level of accuracy quantify with ground truth (ship) and satellite data (the concentration and spatial distribution of phytoplankton). Thus, the proposals being developed by the operational system of remote monitoring can significantly reduce the need to contact (ship) measurements within acceptable weather conditions for satellite monitoring.

Results quantification and analyze the spatial distribution, allow largely replaced traditional ship measurements and, simultaneously, to increase the amount of the definitions these water environmental parameters.

New level of quality can provide a snapshot of higher relevance and high spatial resolution, such as satellite imagery of the European Space Agency (ESA).

The presence of more than 10% cloud cover severely limits the possibility of using satellite images to determine planktonic algae from satellite data. However, it is shown that operational monitoring can be use and for such images by determining the cloud and mask processing sections of images with a low density of the clouds.

Marine aquaculture is progressively becoming a key component of a sustainable food supply, and its role has been recognized by the European Blue Growth strategy, which aims at increasing the exploitation of marine resources in compliance with EU Directives designed for achieving or maintaining a Good Environmental Status, such as the Water Framework Directive and the Marine Strategy Framework Directive. To this regard, aquaculture activities need to be optimized, in order to maximize the profitability of the industry taking into account both the constraints set by the environmental legislation and by the competition for maritime space with other sectors.  This harmonization could be achieved within the framework of the Maritime Spatial Planning Directive, which requires novel tools for its cost-effective implementation. The integration of Earth Observation and simulation models represent, in our point of view, a very promising approach to the development of such tools. In fact, aquafarming requires good environmental conditions (e.g. temperature, transparency, food availability for bivalves) for efficient production and shall be set out of risky areas (storms, waves, eutrophication, harmful algal blooms..).. EO provides a synoptic view of a large panel of required environmental data, which could then be used as input for a suite of simulation models and enable one, on one side, to predict the biomass yield and, on the other, to assess the environmental impact of aquaculture activities.    

The SMART project (Sustainable Management of Aquaculture through Remote sensing Technology), supported by ESA DUE-Innovator program, make use of EO for three aspects of aquaculture i) the monitoring of natural marine conditions for operational aquaculture and the optimization of locations for farming, ii) the mapping of potential in terms of carrying capacity for shellfish and iii) the impact induced by the extensive farming (e.g. shrimps) in terms of release of organic material. We will present the status of the project as well as the results concerning the first two aspects, namely the (future) use of Sentinel-3 (today of GHRSST and GlobColour) derived information (SST, Chl-a, SPM) as inputs for growth model of shellfish (mussels). The results of the validation with in situ-truth collected along the western Adriatic coastline were very encouraging and, on this basis, the SMART approach could be transferred to other areas, in order to map optimal sites for farming. Lastly, an innovative coastal POC (Particulate Organic Carbon) derived from Sentinel-3 OLCI-like measurements built with MERIS data has been developed, and validated and it will be used for further improving the modeling of carrying capacity  of shellfish culture. In addition, Sentinel-2 products are also used for land mapping and shallow water classification to help the optimal farm zoning. The project will expand results and analysis on Algerian sites (mussels) and Vietnamese sites (shrimps). Results will be presented at the Symposium.

Continental and oceanic processes converge along the coast to create landscapes typical of rapid change (Keller 1996). The Adriatic Coast of Albania (about 230 km from the Buna River mouth to Vlora bay) is distinguished for its strong dynamics expressed by the large accumulation and seaward advancement of the coastline in several sectors and by the strong and rapid erosion and advancement of the sea in the mainland in several other sectors. These opposite processes lead to large and rapid changes of the coastline configuration (Bedini 2007).

Change detection is the process of identifying differences in the state of an object or phenomenon by observing it at different times (Singh 1989). Timely and accurate change detection of Earth's surface features provides the foundation for better understanding relationships and interactions between human and natural phenomena to better manage and use resources (Lu et al. 2003). In general, change detection involves the application of multi-temporal datasets to quantitatively analyze the temporal effects of the phenomenon (Lu el al. 2003). Remote sensing data are an important source of information for change detection studies.

This poster presents the analysis of multi-temporal Landsat imagery 1984-2015 to detect changes that have occurred during this time interval in the Adriatic coastline of Albania. Despite the short time interval, the presence of a very dynamic coastline has led to significant changes. For the analysis of the multi-temporal remote sensing data GIS and visualization techniques were used. The results indicate the coastal sectors that have been subjected to erosion and the coastal sectors in which accumulation has taken place. The study shows the applicability of historical Landsat data to coastline change detection, despite the limitations of the moderate spatial resolution of the data.

 

References

Bedini E. (2007). Use of GIS and Remote Sensing data to detect change along the coastline segment between Shkumbini and Semani rivers, central Albania. Bulletin of the Geological Society of Greece vol. XXXX, pp. 1916-1924.

Keller, E. (1996). Environmental Geology, Seventh Edition. Prentice Hall. 560 pp.

Lu, D., Mausel, P., Brondizios, E., and Moran, E. (2003). Change detection techniques. International Journal of Remote Sensing, 25, 2365-2407.

Singh, A. (1989). Digital change detection techniques using remotely sensed data, International Journal of Remote Sensing, 10, 989-1003.

Humic lakes are abundant in the temperate and cold regions of the boreal zone (especially in Finland and Sweden). Humic lakes typically have fairly low TSM and Chl-a concentrations, and due to the high absorption by Coloured Dissolved Organic Matter (CDOM), the lakes can have low water leaving reflectances. The currently available EO water quality processors cannot properly interpret those and, hence, alternatives are needed.

 

The Sensor-Independent Ocean Colour Processor (SIOCS) is an earth observation data processor for converting water leaving reflectances into Inherent Optical Properties (IOPs), which can then be converted to concentrations using local Specific Inherent Optical Properties (SIOPs). The processor was developed by Brockmann Consult and Free University of Berlin (FUB) and utilizes look up tables (LUTs) in the inversion. The original database for the LUTs was developed using data simulated with the Matrix Operator Model (MOMO), which was parameterized with a general SIOP model. One of the advantages of SIOCS is that it allows the use of other LUTs in the processing. Thus, it is possible to generate lake type specific LUTs in order to improve the inversion. In FP7 project Global Lakes Sentinel Services (GLaSS) LUTs (for MERIS, S2 MSI and S3 OLCI) were generated for Finnish lakes using HydroLight simulations and Finnish SIOPs. These LUTs include cases where the absorption by CDOM is very high (up to 60 m^-1 at 400 nm).

 

The accuracy of the inversion was tested with simulated and measured water leaving reflectances. The simulations were done with Hydrolight by using Chl-a, TSM (converted from turbidity) and aCDOM data from 5553 lake stations as input data. The inversion result with simulated data (S3 bands) was very good (R² values: aCDOM = 0.98, Chl-a = 0.93, TSM = 0.99) and clearly better than with band ratio algorithms (R² values: aCDOM = 0.89, Chl-a = 0.70, TSM = 0.97). With measured (with an ASD spectrometer) water leaving reflectances (N = 15) the R² values are lower (aCDOM = 0.98, Chl-a = 0.47, TSM = 0.65), but still better than with band ratio algorithms (aCDOM = 0.90, Chl-a = 0.27, TSM = 0.64). Besides the inversion of water leaving reflectances, a high quality atmospheric correction is crucial for retrieving good results from satellite image data. For humic lakes, an adequate AC is still missing.

Along-track altimetric measurements of Sea Surface Heights (SSH) are very well suited to quantify across-track currents, but the spatial resolution of the derived 2D velocities from conventional altimeters is restricted to scales above 100-150 km with the added limitation that some regions have limited number of available altimeters which lead to errors in the location of currents. Infrared measurements of Sea Surface Temperature (SST) are well suited to locate flow patterns, but the extraction of quantitative estimations of ocean currents from SSH data is complex. To overcome these constrains we have developed a methodology to exploit the synergy between SST and SSH measurements to provide enhanced 2D surface currents. Our approach combines the usage of advanced signal processing techniques, such as wavelet analysis, with dynamical approaches like the Surface Quasi-Geostrophic (SQG) equations. This approach allows for the derivation of the velocity field associated to structures smaller than 30 km, not accesible through the standard SSH maps. Here we will show some examples of the capabilities of our technique for the better characterization of small scale dynamics by comparing the performance of existing satellite information with in situ experiments. Moreover, we will compare the results achieved from conventional altimetry versus those derived from SAR altimetry as offered by the CryoSat-2 mission and the forthcoming Sentinel-3A.

Among the major marine pollutions, oil release into the sea can causes serious biological and economic impacts. Damages caused by oil spill may require decades for the coastal areas to recover. Oil pollution may be intentional, as for instance wastewater discharges, and accidental, as the major oil spill disasters such as Deepwater Horizon oilspill or The Torrey Canyon oil spill. Indonesia is particularly vulnerable to oil spill since national waters, which account from 80% of its territory, involve a very active maritime traffic and a large number of oil platforms onshore and offshore. The numbers of incidents involving oil tankers, offshore platforms or oil pipelines raise increasing concerns among stakeholders. And, the development of accurate and reliable tools for the detection and forecasting of  such oilspills has become a major issue. Within the framework of INDESO project (Infrastructure Development Space Oceanography), we address this issue and aim at developing an operational SAR-based system for oilspill monitoring in Indonesian waters from space. The SAR-based (synthetic aperture radar) monitoring of oil spill is effective due to its all weather and day/night capabilities. In this work, we focus on the backtracking of an oilspill detected from SAR observations. As a case-study, we consider one large oil spill event that happened in Indonesian waters in 2009, referred to as the Montara oilspill. On 21 August 2009, the Montara Wellhead Platform had an uncontrolled release of hydrocarbons from one of the platform wells. The estimate provided by the operator, PTTEP Australasia (Ashmore Cartier) Proprietary Limited (PTTEP AA) was that 400 barrels (or approximately 64 tonnes) of crude oil were being lost per day. The uncontrolled release continued until 3 November 2009 and response operations continued until 3 December 2009. Montara is located in Australian waters but still very close to Indonesian waters. In this work, we develop a Langragian analysis and associated numerical inversion tools with a view to further analyzing the oil spread due to the Montara Wellhead Platform. Our model relies on a 2D Lagrangian transport model developed by CLS (Collecte Localisation Satellite). Our model involves four main parameters : the weights of wind-related and current-related advection, the origin and the duration of the oil leakage. Given SAR oilspill detections, we propose a numerical inversion of the parameters of the Lagrangian model, so that the simulated drift match the SAR observations of the oil spill. We demonstrate the relevance of the proposed model and numerical scheme for the Montara oilspill and further discuss their operational interest for the space-based oilspill backtracking and forecasting.

Climate change is known to impact on glacier mass balances, leading to a net loss of ice mainly caused by higher temperatures and poorer precipitation trends. The evidence for this is particularly strong in high-altitude areas, such as the centralHimalaya. Increasing melting rates induced by this phenomenon can have various effects among which the increase of both the number and size of glacial lakes or the creation of potentially dangerous lakes with risk of outburst floods.

This work aims to characterize water reflectance of glacial lakes in the visible near-infrared, in the Mount Everest region (Nepal) through in-situ radiometric data and satellite acquisitions. The shape and the intensity of water reflectance are here used as a proxy of water color.

A field campaign performed in October 2014 with a WISP-3 spectroradiometer allowed the water reflectance in five Himalayan lakes to be measured. Overall, three groups of spectra with increasing degrees of brightness were clearly distinguished with consequences on color: i) dark blue waters in highly transparent lakes with Secchi disk (SD) depth of about 8-9 m; ii) turquoise waters with SD of about 1 m and a reflectance peak in the green; iii) grey waters with SD of few centimeters and water reflectance typical of turbid lakes with a significant signal beyond 700 nm. In particular, highest reflectance was measured in the lake directly fed by the Imja Glacier, which ice-tongue is protruding into the lake waters. It is a typical supra-glacial lake gradually widening with years due to the melting and retreat of the glaciers, representing now a hazard for the local populations because it is at risk of outburst.

Two satellite images, one from GeoEye-1 and one from Landsat-8 Operational Land Imager (OLI) acquired in the same period of fieldworks were used to up-scale water reflectance observations. The vector version of the 6SV code (Second Simulation of a Satellite Signal in the Solar Spectrum) was applied for removing atmospheric effects. The code was run with a continental aerosol model and measured values of O3. AERONET provided water vapor values and aerosol optical depth. The 6SV-derived reflectance over water bodies showed a serious contamination due to adjacency effects that were corrected for based on match-ups of GeoEye-1 with WISP-3 spectra.

Satellite data were then used to investigate variability of water color at regional (with OLI) and finer (with Geo-Eye) scales. The supervised classification technique ‘Prediction tool’ available within the Global Lakes Sentinel Services (GLaSS) project was applied to OLI image. A classification of 119 different lakes according to the three in-situ defined water color classes was produced. The majority of lakes were classified with predominantly grey waters. In particular, grey lakes seem to be characterized by larger surfaces and an elongated shape, which extends the feeding glacier tongue likely to be its continuation. The 2 m spatial resolution data from GeoEye-1 allow water color variability within a single lake to be assessed.

This study was part of the global use cases of the project GLaSS. It shows the potentiality of satellite imagery for understanding one of the effects of climate change in poorly accessible areas, such as the Himalayas. Moreover, as some of these lakes are risky for outburst flood, further regional scale studies with Landsat-8 and Sentinel-2 data might be useful for monitoring the level of attention on these moraine dammed lakes.

The new generation of sensors onboard the Landsat-8 and Sentinel-2 spacecrafts supply data with improved spatial, spectral and temporal resolution increasing the capabilities of earth observation for water quality analyses and monitoring in inland and coastal zones. In particular, by combining the data from the Multi Spectral Instrument (MSI) on Sentinel-2 and the Operational Land Imager (OLI) on Landsat-8, a weekly based observation of small/medium size targets will be possible. Nonetheless, suitable algorithms to retrieve water quality from optical data are needed and a first issue deals with the correction of atmospheric effects. According to Vanhellemont & Ruddick (2015), the “monitoring of water quality by satellite ocean colour data requires high quality atmospheric correction and especially the accurate quantification of the aerosol contribution to the top of atmosphere radiance”. Moreover, the influence of the surrounding pixels in terms of adjacency effects also plays a crucial role in the atmospheric correction for inland and nearshore coastal waters [Sterckx et al., 2015].

In this work, we present the correction of atmospheric and adjacent effects of OLI imagery based on the Second Simulation of a Satellite Signal in the Solar Spectrum (6SV) radiative transfer code. The algorithm follows the procedure explained in Bassani et al., 2015. Shortly, the 6SV code, that also allows to take into account for the multiple scattering, is used to simulate the crucial contributions for atmospheric correction over water, including the specular reflection, foam contribution and sun glint (Kotchenova et al., 2006). To validate water-leaving reflectance and assess its accuracy, we used atmospheric and above-water data acquired at AERONET-OC sites (Zibordi et al., 2009). In particular, the performances of the atmospheric correction is evaluated by comparing the results obtained using the 6SV default aerosol models and the aerosol size distribution and refractive index available from the AERONET sites; the assessment of water-leaving reflectance with or without the correction of adjacency effect is performed evaluating the OLI-derived apparent surface contrast respect to above-water AERONET-OC measurements.

Among the AERONET-OC sites, three study areas are considered, including two coastal areas and a lake. The two coastal sites are the Acqua Alta Oceanographic Tower (AAOT) in the northern Adriatic Sea (Italy) and the Lucinda Jetty Coastal Observatory (LJCO) in the Great Barrier Reef (Australia). The third AERONET-OC site is the Lake Varnen (Sweden). These sites cover a significant range of both atmospheric (from boreal to tropics) and water quality conditions. Additional AERONET sites, performing only atmospheric measurements, (i.e., Garda and Maggiore lakes in Italy) are selected for further tests on OLI imagery, synchronous to fieldwork activities. Few examples with Sentinel-2 imagery are also showed for a selection of target areas.

Acknowledgment. The authors thank Giuseppe Zibordi and Thomas Schroeder, AERONET PIs, for their efforts in establishing and maintaining the AAOT and Lucinda stations.

Biblio.

Vanhellemont and Ruddick, Remote Sensing of Environment, 2015

Sterckx et al., Remote Sensing of Environment, 2015

Kotchenova et al., Applied Optics, 2006

Bassani et al., Atmospheric Measurements Techniques, 2015

Zibordi et al., Journal of Atmospheric and Oceanic Technology, 2009

Owing to their ability of releasing inorganic carbon in the ambient aquatic habitat and emitting, along with CO2, dimethylsulfide (DMS) into the atmosphere, coccolithophores are subsumed under the category of ecologically and climatically important players. Their blooms occur in a great variety of provinces across the World’s Oceans, and frequently extend over vast areas. Thus, spaceborne estimations of the above impacts produced by coccolithophores are veritably choiceless means in this regard. To achieve this goal, a sequence of dedicated techniques is required to precisely restore the desired information.

This presentation is focused on the description of our original the ocean colour image processing to the effect of a precise quantification of coccolithophore E. huxleyi bloom areas.

The initial OC CCI product employed in this study is constituted by concatenated and averaged over 8 days SeaWiFS, MERIS, and MODIS-Aqua ocean colour images that were collected during the time period 1998-2013.

In the first stage of satellite data processing, all images were reprojected to a polar projection in order to obtain a more detailed view of the target seas. Spectral remote sensing reflectance, Rrs(l) is the basic radiometric parameter that was used throughout all further processing procedures and ensuing analyses. Reprojection was then followed by the innovative technique of gap filling necessitated by the need of cloud masking reduction. This was done by means of averaging of both Rrs(l) values in pixels surrounding a selected central one (i.e. each of the masked ones) and the respective Rrs(l) values recorded in the immediately previous and subsequent images. This stage of processing was then followed by converting the images into RGB images using three spectral channels and applying respective weighting coefficients confidently established. RGB images allowed to tentatively visualize both the coccolithophore bloom areas (they appear as turquoise areas) and their gradual displacements along the west European coast from the North Atlantic Seas to the Arctic Seas. A rigorous interannual periodicity in the timing of blooms onset in each of the studied seas was revealed.

Quantification of bloom areas was performed using our original algorithm for forming binary masks.  The algorithm is based on the spectral specific features of Rrs(l) inherent in the optical properties of coccolithophore cells. The masks thus formed permitted to quantify the bloom areas with high precision, but also establish the dates of the bloom areas when they are largest possible.

The application of the above processing procedures resulted in generating temporal series of dynamics in bloom areas location and extent for each of the seas studied.

It is noteworthy that since 2001 to present there was a drastic drop of bloom incidence and extent in the Bering Sea. Interestingly, this drop was preceded by anomalously intense and extensive coccolithophore blooms in the period 1998-2001. This remarkable change in the interannual bloom sequence pattern obviously points to some significant alterations in the ensemble of bloom forcing factors/mechanisms that occurred in the early 2000s. Unlike the Bering Sea, the pattern of the interannual coccolithophore bloom occurrence in the Barents Sea is much more regular and has a quasi-sinusoidal form. The pattern of interannual incidence of coccolithophore blooms in the North, Norwegian and Greenland Seas is also regular, although the bloom intensity there is invariably much lower than it is in the Barents Sea.

This study is the first step to our ongoing work aiming at obtaining multi-year quantitative data on the inorganic carbon generation and gas emission rates within coccolithophore blooms in the aforementioned seas.

The protection of ports in coastal areas exposed to the action of the sea is carried by sea hydraulic structures (breakwaters) which are exposed to the influence of sea waves. The monitoring of the surface plates movement, constituting the protective mantle and the superstructure (concrete structure) is made by visual comparison of images obtained during follow-up visits. No quantification of these shifts is yet performed. In the design phase of a port protection structure, models are built to study the effect of waves on the protection blocks. During the testing phase waves are generated and periodically are measured the effects of waves in the model.

Presently, photogrammetric monitoring of the model ports is being held, including the use of orthophotos and point clouds to measure displacements of protection blocks. With these methods there were obtained results of the change in overall topography of the model structure. However the real 3D motion of each block is indeterminate due to the absence of data on the geometry of the blocks. The extraction of these geometric data from the images, along with the spatial data of the point cloud, will deepen knowledge of the intended movement.

This paper presents an alternative monitoring solution based on image segmentation in order to obtain displacements of the model protection blocks so to monitor its development over time. Time sequence orthogonal images were obtained during the action of blocks movement. In this study, blocks are identified in the image and shifts are quantified. Validation is held by crossing image segmentation results with point clouds data obtained in laboratory by photogrammetric methods at the same moment images were taken. The present study intends to show the positive contribution of image processing techniques for the laboratory monitoring of breakwater models.

It is widely recognized that fully-polarimetric (FP) SAR measurements offer an unprecedented amount of scattering information on the imaged scene that can be successfully exploited for observing sea oil slicks. Nevertheless, in the ocean domain, FP SAR-based approaches are strongly limited by the reduced area coverage with respect to single-/dual-polarimetric SAR architectures. Hence, new coherent dual-polarimetric SAR architectures, termed as compact-polarimetric (CP) have been developed and investigated for a broad range of applications. It was shown that they often allow obtaining performance that tends to the FP SAR architecture one, while offering a doubled swath. Two different CP modes, the Hybrid-Polarity (HP) and the slant linear (π/4), are considered. The former consists of transmitting a circularly polarized wave while receiving in a traditional orthogonal h-v basis, while the latter is characterized by a slant linearly-polarized transmitted wave and a h-v receiving basis. Independently of the considered SAR architecture, their capability to detect sea oil slicks and distinguish them from weak-damping look-alikes is physically based on the slick-free sea surface Bragg scattering model and on the departure from the Bragg behavior that is in place over the surfactants. In addition, such departure is proportional to the physical properties of the surfactant, thus allowing to map its damping status by using a wide set of continuous estimators (polarimetric features).  In this study, an eigenvalues model is developed in order to analyze the different performance of CP and FP SAR configurations for sea oil slick observation. In addition, the subtle different behavior of CP modes is investigated.To develop the eigenvalues model second-order descriptors for polarimetric scattering are needed. The 3x3 scattering coherency matrix T and the 2x2 wave coherency matrix C, both Hermitian and positive semi-defined, are considered for FP and CP SAR architectures, respectively. Assuming reflection symmetry, which applies over both slick-free and slick-cover sea surface, T and C can be expressed as proper combinations of scattering amplitudes in order to derive the analytical expression of the eigenvalues. Hence, considering the slick-free sea surface X-Bragg model, the coherency matrices for both FP and CP SAR architectures can be simulated in order to compare the theoretical eigenvalues model to the one obtained through actual C- and L-band FP SAR measurements (also used to emulate CP SAR data). Then, the results of the developed eigenvalues model are used to interpret the different FP-CP behavior of polarimetric features such as entropy and degree of polarization. In fact, they can be directly expressed in terms of eigenvalues, and thus a physical mapping between FP and CP SAR architectures can be performed analyzing the underlying differences that lead to a slightly different behavior of the various CP modes. Hence, to validate the robustness of the proposed model, a wide set of wind speed conditions and angles of incidence are considered to evaluate their effects on the eigenvalues behavior.

Ongoing climate change in subarctic and arctic seas is caused by a wide spectrum of   driving factors, among which the fluxes of CO2 in the atmosphere-ocean system are thought to be of paramount importance.  Meanwhile, there are only tentative quantifications of annual production of inorganic carbon in the World’s Oceans. Coccolithophores, especially E. huxleyi microalgae, are the main producers of inorganic carbon in marine biosystems. 16-year (1998-2013) trends in both the incidences of E. huxleyi outbreaks and bloom extent dynamics in the North, Norwegian, Greenland, Barents, and Bering Seas has been obtained from satellite observations.  OC CCI database encompassing the concatenated observations from SeaWiFS, MERIS and MODIS-Aqua was employed. Original techniques of filling of cloud-related gaps, spectral-based identification and masking bloom areas as well as quantification of bloom areas were developed and applied.

Analysis of the revealed seasonal and interannual variations in E. huxleyi bloom timing and spatial extent has shown that, first of all, there is a striking difference in respective patterns inherent in the Bering and studied North Atlantic –Arctic Seas.

Unlike the North Atlantic and Arctic Seas, the Bering Sea exhibited a very irregular pattern of E. huxleyi blooms: between 1998 and 2001 (with exception of 1999), there were extensive blooms with maximum areas of up to (2-2.5)×10⁵ km² that were further followed by the long period (up present) of very low, nearly completely damped down activity of this sort. It is also noteworthy that E. huxleyi blooms occurred there not only in autumn but also in early spring and even summer.  

Contrarily, in the Barents Sea for instance, the blooms occur annually throughout the entire observation period and invariably in August – September with the areal extent varying in the range ~(1-3.5) 10⁵ km². The Barents Sea proves to be the main arena of E. huxleyi extensive blooms along the route of their propagation from south to north: this phenomenon is far less pronounced in the studied North Atlantic Seas, and further up, in the Greenland Sea where the maximum bloom areal extent is about one order of magnitude less than that in the Barents Sea.   With its propagation from the North Sea to the Barents Sea, the timing of blooming onset sequentially shifts from June-July to July August, and finally to August-September. For this group of seas, no stable trend was established across the observation time period.

The observed striking difference in the patterns of the time series of E. huxleyi bloom incidence and areal extent in the Bering Sea on the one hand and North-Atlantic and Arctic Seas on the other hand explicitly points to definitely different mechanisms controlling the phenomenon in these two marine environments. 

Presently this study is further unfolding with the goal of quantitatively assess the production of inorganic carbon within the revealed E. huxleyi blooms as well as the associated  emissions of DMS and CO2 into the atmosphere. 

The level-4 daily Optimal-Interpolation product is a combination of a water typed merge of chl-a products and an optimal interpolation based on the kriging method with regional anisotropic models [1]. The Level 4 product basically provides a global continuous (cloud free) estimation of the surface chl-a concentration at 4 km resolution over the world and 1 km resolution over the Europe.

 

To provide to end users the best quality products, most popular algorithms have been assessed using an in-situ database [2] on three defined water types, each associated with a predefined radiometric shape (oligotrophic waters, chl-a dominated waters and coastal turbid waters). One specific algorithm has been retained for each water type depending on its validation results [2]. The merging between the algorithms is obtained as a weighted sum depending on the membership probability to the considered water type. The anisotropic covariance models allow both seeking for observations along the principal covariance axes, typically coastal gradients or fronts, and giving more important weights to the observations close to the main covariance axes.

 

The Level 4 product avoids end users to consider typical lack of data as observed during cloudy conditions and the historical multiplicity of available algorithms such as involved by case 1 (oligotrophic) and case 2 (turbid) water issues in ocean colour.

 A total product uncertainty, i.e. a combination of the interpolation and the estimation error, is provided for each pixel.

 These products are freely distributed in the frame of the Copernicus - Marine environment monitoring service.

 [1] Updates of B. Saulquin, F. Gohin and R. Garrello. Regional objective analysis for merging high resolution MERIS, MODIS/Aqua and SeaWiFS Chlorophyll-a data from 1998 to 2008 on the European Atlantic Shelf, 2010. IEEE Transactions on Geoscience and Remote Sensing, 99, 1-12

 [2] http://catalogue.myocean.eu.org/static/resources/myocean/quid/MYO2-QUID-OC-GLO-ATL-L4-ACRI-V1.4.pdf

 

Indexes: 1) Interpolation, 2) Ocean color, 3) Optimal merge of algorithms.

River waters provide the major link between the catchment and the coastal waters. Extensive monitoring of coastal waters is needed to address the requirements of national and international legislation (e.g. EU directives) and to reliably determine the impact of riverine nutrient fluxes on the coastal water quality. We present methods developed for these purposes using MERIS instrument time series, but also directly applicable for the forthcoming Sentinel 3/OLCI instrument data. We examined the relationship between catchment characteristics (based on GIS data), modelled riverine nutrient fluxes, and Earth observations (EO) of turbidity in coastal waters. For defining the influence of land cover to the water quality on estuaries, the combined use of all available data sources, such as automated discharge information, models, EO and GIS data, is important.

 

The study area is located in the northern Baltic Sea. Data on river discharges was obtained from 23 catchments along the Finnish coast equipped with measurement stations. Turbidity was calculated from images by the MERIS instrument (years 2003-2011, using C2R processor from the BEAM software). EO dataset covers the coastal waters of Finland. To describe the catchment characteristics, statistics of the available land use data of were calculated. The utilized GIS land use data consisted of for instance Corine Land Cover 2006 (CLC2006), shoreline, top soil and field plot register data. Altogether, the catchments (n = 94) varied in their morphometric and land-use characteristics and covered 65% of Finnish territory and 72% of the Finnish catchment area of the Baltic Sea. In addition, the available information from nutrients was included in catchment characteristics. The riverine fluxes of nitrogen (N) and phosphorus (P) [kg/m²/a] were modelled with the so called 'concentration-discharge-model' using quality checked water quality measurements from period 2000–2011. With the purpose of examining the effects of different kinds of catchments on the turbidity values of the receiving sea water, the catchments were divided into classes according to their main land use characteristics.

A multivariate regression was utilized to study the relationship between land use, nutrient fluxes and turbidity on the coastal waters. The outcome was that the statistics calculated from the land use data, such as the percentages of agricultural fields, lakes, rivers and soil complemented with the modelled nutrients showed clear relationship (r² = 0.72) with the EO turbidity values observed at the estuaries.  

We also studied the correlations between the observed automated river discharge data and coastal EO observations of turbidity. The correlations between the two data sources were dependent of the volume of discharge and the size and the characteristics of the catchments.

To identify the influence of river outflow on the coastal water quality, cross-sections starting from the outlet of the catchment and extending to open sea were generated using EO turbidity observations. The cross-sections proved to be useful tools for defining the spreading and impact of riverine nutrient fluxes. Finally, maps describing the impact area of incoming river water to the coastal waters were formed . These impact maps were calculated using the annual maximum of observed EO turbidity values, separated to four discrete classes. The developed methods, impact maps, transect and correspondences between discharge and coastal turbidity values are currently utilized in the monitoring of coastal waters nearby river sites.

Lakes represent important ecosystems which offer a variety of ecosystem services. Several stressors such as climate change, eutrophication, morphological alterations and contaminations, however, threaten their ecological functions. To address these threats various national (e.g. US Clean Water Act) and international (e.g. European Water Framework Directive) legislations have been passed which include regular monitoring of water bodies. Indicators which can be obtained by remote sensing include chlorophyll-a (Chl-a), suspended particulate matter (SPM), coloured dissolved organic matter (CDOM) and benthos coverage. Sentinel-2 offers unprecedented opportunities to investigate and monitor indicators of lake ecology at spatial and temporal scales which are not covered by traditional point based in situ samplings. Sentinel-2 is expected to have a high potential for deriving information on optically active water constituents and benthos. Within in the project LAKESAT, we tested Sentinel-2’s practical capability to deduce concentration of Chl-a, SPM and absorption by CDOM (a_CDOM) in optically deep water. Bathymetry, information on benthos, SPM and a_CDOM were the parameters of interest in optically shallow water. We used Sentinel-2 sample data (L1C) acquired on 13^th August 2015 covering the southern part of Lake Starnberg (11°19’14’’ E, 47°49’34’’ N, Germany), an oligotrophic, deep peri-alpine lake. L1C data were processed with the Sen2Cor processor to obtain ground reflectance and 20x20 m spatial resolution. At seven points radiometric measurements were taken concurrently to satellite overpass (-1 and +2 hours). In situ measured irradiance reflectance spectra were compared with Sen2Cor spectra to check the suitability of Sentinel-2 L2A-product for investigating water bodies. The present version of Sen2Cor performs the correction with fixed visibility of 23 km, which leads to an overcorrection for the day under investigation. Sen2Cor irradiance reflectance spectra were converted into remote sensing reflectance spectra. To analyse the remote sensing reflectance spectra for water constituents, benthos and bathymetry we applied the freely available, bio-optical modelling software WASI-2D. WASI-2D inversely models the remote sensing reflectance according to radiative transfer in water using constant and variable parameters. The variable parameters are iterated within a pre-defined range until either the closest match between the modelled and the measured Sentinel-2A spectrum is found or a maximum number of model iterations is achieved. We resampled the WASI spectral data base to the Sentinel-2A spectral response curves. Water samples and lake specific bottom reflectance spectra served to adapt WASI-2D to the study area. Inverse modelling of the Sentinel-2A remote sensing reflectance spectra provided reasonable concentrations ranges of SPM and a_CDOM in optically deep water. Owing to low Chl-a concentration (~1 µgl^-1) and little spectral information in blue and green wavelengths impeded separating of Chl-a. Apart from SPM and a_CDOM, bathymetry and fractional coverage of bottom substrates, i.e. sandy sediment and predominant macrophytes (Chara spp.), were successfully derived in optically shallow water. Compared to deep water, SPM was more variable and a_CDOM  was higher in the shallow water. Sandy sediments predominated in the shallow water, macrophytes occurred in patches. Bathymetry values reasonably ranged between 0.15 m and 12.77 m. Comparisons with echo sounding measurements showed significant correlations (r²= 0.92) with an underestimation in the very shallow water (up to 1 m) and close to the deeper water (> 3.5 m) and a tendency of overestimation in the range from 1.5 to 2.5 m. As soon as all Sentinel-2A data are available we will transfer the model set up to the entire lake. The results indicate that Sentinel-2A is suitable for observing indicators of lake ecology and encourage considering these satellite data for lake monitoring.

The Constant False Alarm Rate (CFAR) has been widely used in the automatic detection of targets (basically associated to ships) applied to Synthetic Aperture Radar (SAR) imagery. The success in the detection of ships depends on two aspects: (i) The spatial resolution of the SAR image used, which is related to the processing level; and (ii) The size of the ship, as the capability of the CFAR methodology for positive detections increases when the size of the target (ship) is bigger. Some other factors might also compromise this capability, such as the speckle associated to the SAR image, the environmental conditions when the image was acquired or the course of the ships respect to the orientation of the image. In this work we present the comparison of the results (in terms of small ships detection) obtained with the CFAR method, using RADARSAR-2, ALOS and ENVISAT SAR SLC images. We used the historic positions of ships from the Automatic Identification System (AIS) in order to optimize the parameters used in CFAR. The ships detection obtained with CFAR are in good agreement with the data retrieved from the AIS in terms of ship length and position.

Accurate knowledge of spatial and temporal ocean surface currents at high resolution is essential for a variety of applications.The altimeter observing system, by providing global and repetitive measurements of the Sea Surface Height (SSH), has been by far the most exploited system to estimate and monitor ocean surface currents in the past 20 years. However itdoes not allow observing currents departing from the geostrophic equilibrium, nor is capable to resolve the shortest spatial scales of the currents. In order to go beyond these limits, new sensors and new methodologies must be explored. In this study, we investigate howthe higher spatial and temporal resolution information from Sea Surface Temperature (SST) images can improve the altimeter derived currents by adapting a method first proposed by Piterbarg et al (2009). It consists in inverting the SST evolution equation for the velocity by prescribing the source and sink terms and by using some background information (here the altimeter derived geostrophic currents) in order to remove the uncertainty of the along-gradient velocity.The method feasibility is first tested using an Observing System Simulation Experiment (OSSE) based on model outputs from the Mercator-Ocean system. Up to 30% of improvement is obtained globally, for both component of the velocity, with maximum improvement in the Equatorial band. The method is then applied on real SST data. The use of both microwave and infrared measurements is investigated. The advantages and limitations of these data for applying the proposed methodology are discussed.

Sea Surface Temperature (SST) and ocean-surface winds have been identified as essential variables by the Global Climate Observing system (GCO). Satellite observations have aided the understanding of air-sea interactions and the important role these two parameters hold in climate related studies, atmospheric and oceanic modelling, bio-chemical processes and oceanic CO₂ studies.

The diurnal variability of SST, driven by the coincident occurrence of low enough wind and solar heating, is currently not properly understood. Atmospheric, oceanic and climate models are currently not adequately resolving the daily SST variability, resulting in biases of the total heat budget estimates and therefore, demised model accuracies.

The ESA STSE funded project SSTDV:R.EX.-IM.A.M. aimed at characterising the regional extend of diurnal SST signals and their impact in atmospheric modelling. This study will briefly present the final project findings regarding the analysis of hourly SEVIRI SSTs from SEVIRI over the Atlantic Ocean and the European Seas, revealing the regional extend of diurnal warming. As satellite SSTs are representative of the upper centimetre of the water column, they do not provide information of the vertical extend of diurnal signals. Drifting buoys provide measurements close to the surface but are not always available. Moored buoys are generally not able to resolve the daily SST signal, which strongly weakens with depth within the upper water column. For such reasons, the General Ocean Turbulence Model (GOTM) was used to resolve the vertical temperature structure of the upper water column and provide the link between surface temperatures and the ones observed at some depth. The model proved able to reproduce signals observed from satellite and in situ instruments, thus can be a candidate model for operational analysis of the daily SST variability. Such an analysis can be useful for the already operational L4 SST analysis products. Moreover, the project aimed at characterizing how the diurnal SST signals impact atmospheric modelling. Hourly SST fields, were used to initialize the high resolution Weather Research & Forecasting (WRF) model. The perturbations in the atmospheric model, associated with the daily SST cycle were assessed through comparisons of the modelled 10-m wind fields against the ESA's ENVISAT ASAR 10-m winds and in situ measurements at various atmospheric levels, from meteorological masts located offshore. The project resulted in expanding the scientific background for understanding the spatial and temporal variability of key climate variables and their representativity in atmospheric and oceanic models. 

Under the background of the global changes, previous studies show that large scale variations of the sea level rise have great impacts on the regional seas/oceans. The merged satellite altimeter sea level data products play a very important role in these studies. However, up till now, there is not a significant and convenient system of indices based on the altimeter data which can fully provide a reasonable and physical description of the large scale features of the global oceans' sea level changes.

 

To quantitatively show the impacts of the large scale sea level change features of global major oceans on the regional oceans, the relationship between SCS (South China Sea) sea level variations and the Pacific/Indian Oceans is investigated in this study as a case study, by using AVISO all satellite altimetry merged SLA (Sea Level Anomalies) products in which the ERS and Envisat products are included. The MLR (Multiple Linear Regression) model is mainly used to retrieve the SCS regional mean sea level anomaly time series from the mean sea level anomaly time series of the North Pacific Ocean, the Indian Ocean and the zonal volume centroid anomaly time series of the North Pacific Ocean and the Indian Ocean. The zonal volume centroid index is recently proposed by Fang and Zhang (2015), which can help capture some important basin scale features of the sea level changes. All these time series are smoothed by a five-month-running-mean filter to remove the monthly and seasonal signals. In physics, the volume centroid changes reflect the changes of the spatial distributions of the upper layer oceans volumes for a specific ocean area.

 

Primary results of this study reveal that the zonal volume centroid anomalies of the North Pacific Ocean and the Indian Oceans significantly enhance the correlation between the mean sea level anomaly of SCS and the sea level variation features of the adjacent major oceans. The simultaneous correlation coefficient between the original sea level anomaly time series of SCS and its MLR results is higher than 0.8 and if the time leads are considered, the correlation is higher than 0.85. The regional scale SCS sea level changes are well reconstructed by simple linear combinations of the large scale sea level indices, namely, the mean sea level and the zonal volume centroid of the Pacific and Indian Oceans.

 

The study suggests that:

1) the volume centroid, a recently established new index, can deepen the studies of the regional sea level connections to the sea level changes of the large scale oceans. For example, this study shows that the variation of SCS sea level is not only related to global mean sea level rise, but also to the non-uniformity of the sea level trends.

 

2) to make full and better use of the merged satellite altimeter products, a reasonable and physical sea level index system at global/basin scales should be established as soon as possible to benefit sea level related studies. These indices should be able to provide as complete features at basin scales as possible for the global major oceans.

 

3) future validation missions for ESA's observing systems should also include the "large scale index" validation, because these large scale indices, to some extent, have more direct impacts on this planet's activities.

The GLaSS project develops a prototype infrastructure to ingest and process large amounts of Sentinel-2 and Sentinel-3 data for lakes and reservoirs. To demonstrate the value of satellite observations for the management of aquatic ecosystems, global case studies are performed addressing different types of lakes with their respective problems and management questions. The aim of this study was to analyse trends in the trophic level evolution in clear deep lakes world wide, which, being characterised by good quality state, are important socio-economic resources for their regions. The selected lakes are situated in Europe (Garda, Maggiore, Constance and Vättern), North America (Michigan) and Africa (Malawi and Tanganyika), and cover a number of eco-regions (continental, perialpine, boreal, rift valley) distributed globally.

To evaluate trophic level tendency we mainly focused on chlorophyll-a concentrations (chl-a), which is a direct proxy of trophic status. The chl-a concentrations were obtained from 5216 cloud-free MERIS imagery from 2002 to 2012 based on the following validated algorithms: C2R (Case-2-Regional) for Garda and Maggiore; MIP (Modular Inversion and Processing System) for Lake Constance; FUB (Free University Berlin) for Vättern and Michigan; CC (CoastColour) for Lake Michigan and CC and WISP-3 for Malawi and Tanganyika.

The ‘ROIStats tool’ available within the GLaSS project was used to extract chl-a in a number of region of interests located in pelagic waters as well as some few other stations depending on lakes morphology. For producing the time-series trend, these extracted data were analysed with the Seasonal Kendall test. For each year or season, an analysis on phenology was also conducted, identifying the initial date of chl-a increase and the events of phytoplankton growth.

In general, the results show almost stable conditions, with a slight increase of trophic status for Maggiore, Constance and the Green Bay of Lake Michigan, a slight decrease in trophic status for Garda and Tanganyika and absolutely stable conditions for Vättern and Malawi. The phenology analysis identified a few years with higher values of chl-a: 2005 for Garda and Constance, 2011 for Maggiore, 2007 for Vättern, 2003 and 2008 for Malawi and 2004 for Tanganyika. No specific year was identified for Lake Michigan. The slight changes observed might be due to a combination of meteo-climatic conditions (e.g., windy and cool winter that facilitates the water column circulation) and anthropic impacts (e.g., under-dimensioned water treatment plants).

The results presented in this study showed the great capability of MERIS to perform trend tests analysis on trophic status with focus on chl-a concentration. Being the chl-a concentration also a key parameter in water quality monitoring plans, this study also supports the managing practices implemented worldwide for using the water of the lakes. Since four of the seven lakes investigated in this study are transboundary, this work might be used to also emphasize the advantage of integrating satellite technologies into synoptic and self-consistent monitoring plans of lakes at catchment scale.

Nonetheless, as 10-years long time-series might be too short for understating trophic level trends and their relations with climatic/anthropogenic processes we look forwards to the new missions. In particular, we expect that the results presented here will benefit from the MERIS successor, called OLCI onboard of Sentinel-3 that will ensure data continuity collection as needed by time-series analysis. In addition, the already launched Sentinel-2 provides further information on water quality in lakes, enhancing both the temporal and spatial resolution.

The two-dimensional variational ambiguity removal scheme (2DVAR) provides a spatial analysis of the sampled ocean vector winds to resolve the local Advanced Scatterometers (ASCATs) dual wind vector ambiguity. Like other variational meteorological data assimilation systems in Numerical Weather Prediction (NWP), it combines prior NWP information (background) with observations, in this case from ASCAT. Although 2DVAR is in general effective, it may select the wrong ambiguity under certain conditions, e.g., when the background mislocates frontal (convergence) areas or low-pressure centers, or when it misses convective systems. The background in 2DVAR consists of forecasts from theEuropean Centre for Medium-range Weather Forecasts(ECMWF). The relative influence of the background and the ASCAT wind fields in the resulting 2DVAR analysis field can be controlled by adjusting the background error spatial correlation structure, and the background and/or observation error variances. In this paper an adaptive 2DVAR approach is proposed to improve ASCAT ambiguity removal, using background error spatial correlations estimated from the autocorrelation of observed scatterometer wind components minus ECMWF forecasts (i.e., numerical structure function, NSF), and using background and observation errors estimated from triple collocation (TC) analysis on collocated buoy, ASCAT, and ECMWF data. The triple collocations are segregated into several categories according to the ASCAT-derived parameters that have proven to be effective in detecting the correct position of frontlines and low-pressure centers. Verification using a typical cyclone case and collocated ASCAT and buoy winds shows that the 2DVAR analysis as well as the ASCAT ambiguity removal is improved significantly by putting more weight on the ASCAT observations using empirically determined spatial background error structure functions and situation-dependent background/observation error variances.

Note that in 2DVAR the analysis’ objective is to fit all scales present in the ambiguous scatterometer winds, while in Numerical Weather Prediction the degrees of freedom in the forecasting model should be initialized without creating small-scale noise.Although the NSF proves to be effective for ASCAT AR, its long tails will also impact NWP parameters outside the swath. Further studies are therefore required to verify whether NSF is also beneficial to thehigher dimensional variational (e.g., 4D-var) data assimilation schemes.

Atmospheric correction is a critical step and can be a limiting factor in the extraction of aquatic ecosystem parameters from remote sensing data of coastal and lake waters. Atmospheric correction models commonly in use for open ocean water and land surfaces can lead to large errors when applied to hyperspectral images taken from satellite or aircraft. In the case of open ocean models, the main problems arise from neglecting the adjacency effect, which originates from multiple scattering of upwelling radiance from the surrounding land and can have significant influence in the range of several kilometres. Another relevant source of errors originates from the method to determine aerosol parameters, which is based on the assumption of negligible reflectance of water in the infrared, which is no longer valid in turbid and shallow waters. In case of land models, the specular reflections at the water surface (sun glint, sky glint) usually cannot be taken into consideration, and the common dark dense vegetation method to determine aerosols is not applicable for the water areas, but requires the nearby presence of forest areas in the image.

To better understand the challenges for developing an atmosphere correction model suitable for lakes, we compare Sentinel-2A and airborne hyperspectral data (HySpex) of two Bavarian lakes (Klostersee, Lake Starnberg) with in-situ measurements performed with RAMSES and Ibsen spectrometer systems and a Microtops sun photometer. The modified atmospheric correction model for coastal and lake waters, which will be developed, is supposed to find primarily application in remote sensing data from HySpex, EnMAP and Sentinel 2 sensors, but will be adaptable to any kind of remote sensing data.

With the current lack of a suitable model for coastal and lake water, atmosphere parameters are derived from the HySpex and Sentinel 2A data using ATCOR and Sen2Cor, respectively, and these models are also used for atmospheric correction and conversion of the image data to units of water leaving remote sensing reflectance. The derived atmosphere parameters are the Ångström exponent of aerosol, aerosol optical thickness, ozone and water content. These are compared with the corresponding parameters derived from the sun photometer measurements. Furthermore, spectra of the downwelling irradiance calculated with these atmosphere parameters are compared to direct field measurements with the above mentioned spectrometer systems, and similarily the remote sensing reflectance spectra obtained for Sentinel-2A and HySpex are compared with the field measurements. The simulation of downwelling irradiance spectra at the water surface during the time of the HySpex and Sentinel 2A overflight is performed using MODTRAN 5. Finally, the influence of uncertainties of the atmosphere parameters on atmospheric correction will be discussed for the HySpex and Sentinel 2A data.

Algeria coastal zone is one of the most vulnerable in the southern shore of the Mediterranean Sea. It is facing natural and anthropogenic pressure and hazards. The Algerian margin is seismically very active with tsunami risks, it also knows erosion/accretion phenomena, drought and salt-water intrusion, etc., but the most noticeable effect is the urban sprawl and its consequences on the environment (pollution, loss of biodiversity and economic value, etc.).

The aim of this work is to produce coastal vulnerability index (CVI) maps to erosion and flooding at different spatial scales. This index, developed by Gornitz & White (1992), is achieved by integrating in a GIS, different factors contributing to the vulnerability of this coastal zone. The lack of data, particularly time series over naturel process and even socio-economic data is an impediment for decision making. Many relevant parameters were derived from remote sensing, combined with other data, they are analyzed with a multicriteria method in three sub-indexes; coastal physical characteristics, coastal forcing and socioeconomic factors, which combined gives the CVI. This index turned out to be a relevant tool for coastal planning and decision-making. It identifies the vulnerability of some physiographic units and give at local, regional and the national level a broad image of coastal zone sensitivity to erosion and flooding. The results of this contribution will be proposed for improving the “IMCA” (integrate management of the coastal area) policies in the country.

Derivation of high-resolution (HR) spatial distribution of data is a fundamental problem in Earth observation, either in the case of physical variables or when the data is obtained as the output of a classification process. We show that the problem can be resolved through information fusion at different scales.  In the case of physical intensive variables of fully developed turbulence, as encountered in satellite Oceanography and Ocean/Climate interactions,  we show that the derivation of HR data can be solved using a new method based on an approximation of the energy cascade, expressed in a microcanonical formulation, and associated to turbulent signal provided by HR satellite data. The generality of the approach offers the opportunity to infer different oceanic signals from Low Resolution (LR) to HR. The basic idea is to use optimal cascading to decrease the spatial resolution of the HR signal, then use the signal available at LR, transmit that information along the scales back to higher spatial resolution using the cascade to obtain a new HR signal. The process has been successfully used to obtain oceanic currents [1, 2], oceanic partial pressure of CO2 [3], and more recently sea surface salinity. Our methodology could be extended to many Essential Climate Variables both in the ocean and atmosphere critical for characterizing Earth’s climate and its changes.

The western part of the Black Sea is characterized by heavy ship traffic with the main routes from the Bosphorus Strait to ports of Bulgaria, Romania and Ukraine. The level of anthropogenic pollution in the region is very high, notably that of deliberate and accidental spills of oil and oil products from ships.  Another major pollution source is the outflow of the Danube River, the largest of the rivers running into the sea.

We discuss the influence meso- and submesoscale hydrodynamic processes, such as eddies and internal waves, have on the transport of pollutants in the surface layer of the sea in this region. The research is based on satellite remote sensing data obtained by ERS-2 SAR, Envisat ASAR, Sentinel-1 radar instruments and also visible and infrared sensors of the Landsat series.

In SAR images, ship spills are often distinctly manifested as slick bands of low backscatter resembling ship wakes. Sometimes, they are subjected to significant shift and deformation under the impact of wind and currents. Comparing the true route of a vessel with its wake-like spill we can, on the one hand, obtain more information about current components and, on the other hand, investigate the impact of various hydrological processes on ship spill patterns.

Visible satellite data due to drastic differences in optical properties of sea and turbid river waters allowed determining the area of Danube waters propagation and frontal zones associated with the river plume. Moving and oscillating powerful non-stationary hydrologic fronts induce multiple vortical structures and internal waves. Their spatial distribution and seasonal variability was investigated.

Spreading forecasts for pollution revealed in satellite images were made based on FOTS model developed by Marine Hydrophysical Institute in Sevastopol. FOTS assimilates data available  from various internet archives of satellite and meteorological information to calculate full surface currents in a particular region. Fields of full surface currents were used to derive trajectories of floating objects, for instance, oil slicks.

The work was accomplished with partial financial support by the Russian Science Foundation, grant # 14-17-00555. The Envisat ASAR and MERIS data were obtained in the framework of European Space Agency projects.

The Southern Ocean plays a key role in the global climate system. However, in-situ monitoring of the changing dynamics within this inhospitable and inaccessible region is challenging. Thus satellite observations are a vital tool in understanding the response of the Southern Ocean’s circulation to a warming climate. Here we assess a range of different data sources and methods for measuring the surface circulation of the Southern Ocean via the geodetic approach. This involves calculating the dynamic ocean topography (DOT) by subtracting a geoid from sea surface height (SSH) measured by satellite altimetry. We compare the effect of using satellite-only geoid products (e.g. from GOCE) in this calculation against the use of geoid products derived from a combination of satellite and surface measurements. We also investigate the difference between the DOT calculated as the sum of a time-averaged mean sea surface (MSS) and sea level anomalies (SLA), and the DOT determined directly from the along-track SSH measurements. The impact of various MSS products is also examined. Finally, we consider a range of filtering strategies for the DOT, based on formal and informal errors. In-situ observations and ocean modelling data are used as a benchmark by which to assess the various DOT estimates. The use of the geodetic approach for accurate calculation of the DOT provides us with an extremely valuable tool to study the surface circulation of the oceans. Understanding the advantages and disadvantages of different data sources and methods used to determine the DOT is an essential step in improving our knowledge of Southern Ocean dynamics.

Fifteen gas and oil deposit fields were discoveredon the Northeastern shelf of the Sea of Okhotsk. Eight of them are large-scale deposits and two are unique ones. Currently, oil production is carried out on four platforms: Berkut, Molikpaq (Piltun-Astokhskoye-A - PA-A), Piltun-Astokhskoye-B (PA-B) and Lunskoye-A (Lun-A) located at the distance of 12 – 25 km from the coast where water depth is 32 – 48 m (Figure 1a). Through the pipeline system oil and gas are supplied to the De-Kastri Oil Terminal in the Tatar Strait of the Japan Sea and to Prigorodnoye Production Complex located alongside Aniva Bay, Okhotsk Sea. Ice-class tankers are used for the transportation of oil. The most difficult oil operation and transportation conditions in the Okhotsk and Japan Seas are associated with storm winds and high waves which are observed during a deep cyclone passing. In the cold season the moving pack ice has a pronounced effect on oil platforms and tankers. The quantitative information about the sea state, hazard weather phenomena and ice conditions independently of sun illumination and cloudiness can be obtained by satellite microwave active (radar) and passive (radiometer) remote sensing sensors. Images acquired by Sentinel-1A SAR-C and ALOS-2 PALSAR-2 allow to map the pack ice structure and surface roughness, locations and dimensions of hummock ridges, polynyas and ice openings, detect oil platforms, icebreakers, ships and ice wakes resulted from the interaction of ice with obstacles, classify ice types, distinguish individual ice fields, oil spills, etc. The Landsat-8 visible and infrared images having the same spatial resolution as satellite SARs are used for study ice structure, types and snow cover as well as for SST retrieval under cloudless conditions. The signatures associated with oil production, processing and transportation as well as with the changing environmental conditions were revealed on SAR-C with VV-polarization, PALSAR-2 with HH and HV polarization and on Landsat - 8 visible and IR images of Sakhalin shelf acquired in January - May 2015. Examples of SARs and Landsat-8 visible images obtained in February and in March when all oil platforms were within pack ice are shown in Figures 1, 2 and 3. The signatures were interpreted by joint analysis of satellite and relevant data including air temperature, wind speed and direction. This work was partly supported by the FEB RAS Research projects15-I-1-009_о and 15-I-1-038. Authors thank the European Space Agency for Sentinel-1A SAR-C images, U.S. Geological Survey for the Landsat-8 images and Japan Aerospace Exploration Agency (JAXA) for ALOS-2 PALSAR-2 images. PALSAR-2 data include material copyright by METI and JAXA.

The Yellow River (in Chinese, Huang He) is the most sediment-filled river and the sixth-longest one in the world. It flows from the Bayan Har Mountains empting with 8,000 square kilometers large delta mouth into the Bohai Sea near the city of Dongying. The Yellow River carries on to the sea a very large amount of silt (1.6 billion tons annually) while descending from the Loess Plateau. However, since 1972 it often runs dry before reaching the sea due to increased agricultural irrigation. This phenomenon led to a slightly shrinking since 1996 through erosion that is still building and rebuilding the river delta.

The Yellow River is of paramount importance for safe navigation, local economy and environment due to the presence of floods, farms, aquacultures and pollution. Devastating floods and course changes often happens due to the continual elevation of the river bed (sometimes produced by man-made erosion upstream). There have been 11 main floods in the past century, each causing tremendous loss of life and property. Main causes are the large amount of fine-grained loess carried by the river from the Loess Plateau, and the collapse of upstream ice dams in Inner Mongolia with a consequent sudden release of large amount of impounded water. Aquacultures are also present along the riversides: the fish ponds system of Wangcun Town (10 square kilometers size) is the largest center in North China. In addition, for saving excess water and for flood control and electricity generation, several dams and hydroelectric power stations have been built along the Yellow River. Official reports assessed that severe pollution has made one-third of China's Yellow River as unfit for drinking, aquaculture, industrial and agriculture use, mostly due to factory discharges (70%) and sewage from fast-expanding cities (23%).

In this study, actual L- and C-band polarimetric SAR data are used to investigate the scattering properties of the Yellow River delta area. The SAR dataset consists of fully-polarimetric Radarsat-2 and ALOS PalSAR-2 SAR acquisitions that are partially overlapped. The wide area of the Yellow River mouth they cover is a challenging area characterized by different kinds of soil: forest, canopy, mud, intertidal zones, sandy terrains, swamps, reservoirs, ponds, etc. The basic idea of this study is to exploit the large amount of physical information provided by polarimetric SAR data to classify different kinds of land-cover according to their scattering properties and to analyze their backscattering coefficient varying the transmitted polarization. In fact, each kind of terrain is characterized by a random combination of elementary scattering mechanisms: single-bounce from a flat or slightly rough surface (sea, sand, mud), double-bounce due to ground-trunks interaction and random scattering typical of forest. Hence, this information can be inferred from fully-polarimetric SAR measurements by using proper modeling tools to interpret a large set of polarimetric features, namely the polarimetric entropy, mean scattering angle, amount of unpolarized backscattered energy, etc. In addition, to improve the Yellow River delta classification results, once the scattering-based classification is performed, single-polarimetric intensity information can be taken in account by using conventional backscattering models. In fact, such electromagnetic models allows predicting the backscattering coefficient for different types of surfaces and terrain mixtures depending, among other things, on their dielectric properties. This results in the possibility to distinguish between land-cover characterized by the same scattering behavior while providing a different backscattering coefficient. Furthermore, the key role played by the incident wavelength into the scattering process is also investigated and discussed. Hence, the classification results are compared with ground truth data collected over the Yellow River delta during an in-situ ship-based campaign during which the characteristics of the soils have been recorded.

Oceanic shear front and convergent front are two types of oceanic front. Shear front and convergent front have different generating mechanisms resulting in different flow field which leads to the different SAR image features. In this paper, we derive the relationship of radar look direction angle with the hydrodynamic modulation, as well as it with the velocity bunching modulation. The experimental results of SAR images of shear front and convergent front with different angles of radar look direction are presented. The analysis of the simulation results indicates that shear fronts appear as bright or dark line under different conditions while convergent fronts appear as bright line.

Bathymetry data is considered as an essential data for hydrodynamic modelling in coastal zone studies since it affects quality and reliability of a model result. The Ythan estuary, located in Aberdeen, Scotland is a test site for the hydrodynamic modelling and it has no complete bathymetry data available. There is only an availability of a terrestrial LiDAR DEM, which has some areas without data and uncertainty of elevation values along the middle of the estuary channel. Collecting a new bathymetry data for the entire estuary by using an echo sounder mounted on a boat is costly and time-consuming task. Thus, making use of the existing LiDAR DEM to create a new bathymetry data for the modelling is a challenge. This study aims to generate a new bathymetry data from an integration of the terrestrial LiDAR DEM and elevation data collected by a single beam echo sounder. The echo sounder is used to collect elevation values in the areas without data and in the middle of the estuary channel. The elevation values derived from the echo sounder are interpolated to obtain prediction elevation values in the target areas. The interpolation method used is the ordinary kriging of which the cross validation result gives the smallest root mean square error of 0.19. Then, the predicted elevation values are merged to the LiDAR DEM to obtain the new bathymetry data. The elevation values derived from the echo sounder are not only used to create prediction elevation values for the target areas, but they are also used to validate the elevation values of the LiDAR DEM. The echo sounder data show a strong positive relationship with the LIDAR DEM data with R² of 0.8. The strong relationship is found at bed elevation of 0.1m onwards above MSL, which is under water depth between 0.33 – 1.18 meters. In contrast, there is no relationship found between those two data at bed elevation of 0.1m below MSL, which is under water depth deeper than 1.18 meters. The offset of the elevation values of those two data set also show that the LiDAR DEM and the echo sounder data are generally greater in the middle of the estuary channel where water depth is deeper. Whilst the smaller offset values are found in shallower areas closer to the river bank. These results suggest that bed elevation of 0.1m downwards below MSL, which is under water depth deeper than 1.18 meters, could affect quality of return signals of the terrestrial LiDAR. Therefore, elevation values of LiDAR DEM, which are higher than 0.1m above MSL and are not under water depth deeper than 1.18 meters, are reliable to use to create a new bathymetry data. The new bathymetry derived from the study clearly shows that it is possible to use the terrestrial LiDAR DEM coupling with the single beam echo sounder to generate the new bathymetry data for hydrodynamic modelling, which could reduce cost and time occurring during fieldwork for coastal zone studies of which study areas is in shallow area.

Eutrophic shallow lakes represent a group of lakes situated in heavily populated areas where the presence of cyanobacterial blooms periodically prevent activities in or on a lake leading to socio-economic problems. A socio-economic analysis within the GLaSS (Global Lakes Sentinel Services) project showed that these lakes are often used for drinking water production. Therefore, it is important that earth observation (EO) methods could provide adequate monitoring of these environments. GLaSS prepared showcases on the use of EO data for different lake types. This study on eutrophic lakes comprises eight representative European water bodies that were investigated by using satellite data. Different processing schemes (FLH – Fluorescence Line Height, MCI – Maximum Chlorophyll Index, MPH – Maximum Peak Height, CoastColour) were applied to MERIS Level 1 products. The best Chl-a estimates were received from FLH products, which indicated especially good results for Estonian lake Peipsi (R² = 0.74) and Italian lake Trasimeno (R² = 0.90). The MCI algorithm worked also rather well but MPH Chl-a results showed in some cases notably weaker concurrencies with in situ data. FLH and MCI algorithms were also more successful in following the seasonal trends and year-to-year variation of Chl-a values. For German lake Müggelsee the CoastColour products supported the temporal differences very well.  In order to characterise the presence of cyanobacteria specific MPH products and quality flags were used and found to be in good agreement with in situ measurements from lake Trasimeno and Müggelsee. Two small Dutch lakes (Paterswoldsemeer and Westeinderplassen) were analysed with Landsat 8 data. A special band ratio algorithm was sensitive to Chl-a seasonal trend for lake Paterswoldsemeer but not for the clearer lake Westeinderplassen. It is expected that Sentinel-2 and Sentinel-3 instruments will provide the adequate data for further investigations of such inland waterbodies. 

The research focuses on the relationship between the phytoplakton absorption coefficient (a_ph(442)) and chlorophyll-a (Chl-a) concentration in Estonian large lakes. Physical, chemical and biological parameters of lakes Peipsi and Võrtsjärv have been investigated regularily and thoroughly for decades. Cyanobacteria dominate in phytoplankton community in both lakes during summer and autumn, but surface blooms are common in lake Peipsi (Gloeotrichia echinulata, Aphanizomenon ssp., Microcystis ssp.). Phytoplankton is one of the most examined and monitored parameter that determines optical properties of the waterbody. It is also an indicator of ecological state of natural waters and therefore, it is important to provide a regular monitoring of this parameter.

In addition to field measurements the satellite products enable us to supplement and enrich our knowledge about the water quality seasonally and spatially. Nowadays implemented technology of satellite remote sensing has become a useful tool to monitor not only open sea areas but also optically complex inland water bodies. During its ten year mission MERIS (MEdium Resolution Imaging Spectrometer) onboard of  ENVISAT (ENVIronmental SATellite) has offered valuable information about water quality parameters of coastal waters and large lakes. MERIS products and algorithms have passed several improvements in order to minimize atmospheric scattering and therefore evaluate the water-leaving radiance reflectance more successfully (ESA, Evolution of the MERIS IPF, 2012).

Results of the study indicated that power law (C_Chl-a = 29.5*a_ph(442)^0.75) describes the positive relationship between a_ph(442) and C_Chl-a most successfully (R² = 0.76, N = 350). MERIS Case II water algorithm derives C_Chl-a values from absorption coefficients (442 nm) of phytoplankton (a_ph(442)) using empirical formulas (Doerffer & Schiller, 2007). The evolved empirical algorithm based on data from 2010 to 2013 was also applied to MERIS satellite images (2010-2011) in order to investigate whether the reliability and accuracy of determining the concentration of Chl-a by lake-specific conversion factors will increase. MERIS 2^nd and 3^rd reprocessed water quality products were investigated, both corrected against adjacency effect caused by land pixels. Results showed that these products overestimated a_ph(442) and therefore C_Chl-a values. Moreover, both reprocessed products failed with determining relative proportions of absorption of phytoplankton and other optically active substances in total absorption and therefore, applied empirical relationships did not improve C_Chl-a retrieval compared with MERIS standard algorithm. All in all, investigated products, especially MERIS 2^nd reprocessed products followed the seasonal and spatial variability of a_ph(442) and C_Chl-a values rather well.

Doerffer, R., Schiller, H. (2007). The MERIS Case 2 algorithm. International Journal of Remote Sensing, Vol 28, No 3–4, pp. 517–535 pp. DOI: 10.1080/01431160600821127.

ESA, Evolution of the MERIS IPF (2012). https://earth.esa.int/documents/700255/707220/MERIS_IPF_evolution.pdf/c89ebffd-c453-437b-b3f9-1afbc2a07e5b. 

Altimetry-derived maps of Lyapunov exponents (LEs) provide proxies of (sub-)mesoscale transport fronts. They are being increasingly used in physical, biogeochemical, and ecological applications, ranging from real-time support to field studies to co-localisation of animal tracking with Lagrangian Coherent Structures. Their calculation however is more complex than standard Eulerian diagnostics, because it requires a Lagrangian algorithm which integrates the velocity field. During the past 20 years, in parallel with the altimeter measurement Level2 (a.k.a [O/I]GDR) to Level3 and Level4 (along-track cross-calibrated SLA, and multiple sensor merged maps) processing, different applications and derivated Level4+ products were developed by AVISO+. In order to better serve the users need, and in collaboration with different laboratories (LOCEAN and CTOH), the LEs and vectors are computed over the 21-year altimeter period and over the global ocean within the SSALTO/DUACS project. This product provides the position, and intensity, and orientation of fronts induced by the mesoscale eddies and underlining part of sub-mesoscale activity. We present here the Lyapunov products that are available on AVISO+ since early 2015.

The ocean surface is a complex natural environment including many parameters such as salinity, temperature, surface wind, waves with various scales, currents, etc. The observation of sea state and the analyses of oceanic parameters allow obtaining/providing the information in various applications: meteorology, collision, rescue, oil slick observation, ship detection…Among the radar systems used to observe and/or characterize the ocean surface, SAR satellites are preferred due to many advantages: stable operations in most meteorological conditions, revisit period, high resolution, etc. In order to describe the relation of radar backscattering and sea state, we can find in the literature many electromagnetic (EM) models such as Small Perturbation Method (SPM), Two-Scale Model (TSM), Small Slope Approximation (SSA), Method of Moment (MoM), and Forward–Backward Method (FBM). The TSM and SSA are asymptotic methods while the MoM and FBM are exact ones. They are used to model the sea state in the general conditions with different wave scales. Meanwhile, the SPM describe the relation of ocean surface roughness and radar scattering in the condition of slightly rough sea state. Based on the inversion of these models, we can retrieve the oceanic parameters from SAR images.

Among the indicated oceanic parameters, sea surface wind plays a crucial role for the studies of the other parameters such as waves, currents, marine meteorology, and the coupling of oceanic and atmospheric systems. Furthermore, strong wind could affect significantly the studies of oil slick observation and ship detection. The extractions of wind field could be carried out on a variety of SAR images in L-band (JERS-1), in C-band (ERS-1/2, Envisat, Radarsat-1/2), and in X-band (TerraSAR-X, Cosmo-SkyMed). Certainly, there are a lot of studies which estimate wind speed and wind direction from SAR images. Regarding wind direction extraction, this is a crucial parameter to estimate accurately wind speed, especially for strong wind. It can be obtained by in situ measurements or numerical meteorological models. However, if wind streaks on SAR images are visible enough, the most well-known way is the extraction of wind direction from SAR images by three principal methods: Fast Fourier Transform (FFT), Local Gradient (LG), and Continuous Wavelet Transform (CWT). Among them, the LG method has been widely used because it can give high-resolution wind directions and good agreement with in situ measurements [1]. Concerning wind speed retrieval, the scatterometry-based approaches [2] have been widely used in a lot of studies because they are quite simple to apply and give good agreement with buoy measurements. Besides, we can find some alternative methods such as azimuth cut-off, neural network, and dual-polarization. In the scatterometry-based approaches, the dependency of radar backscattering with respect to wind speed and the geometry of observations (i.e. incident angle, relative wind direction) is described by the geophysical model function (GMF). Depending on radar frequency and co-polarization (VV- or HH-pol), several empirical GMFs (e.g. CMODx, XMODx) are constructed and validated by a series of satellite scatterometer missions. This probably leads the scatterometry-based approaches to be less general and less flexible. Thus, recently there have been many studies which use the EM models (composite model TSM, SSA-1, 2) to retrieve wind speed from SAR images. Compared to the TSM and SSA-1, 2, the SPM [3] is preferred in this paper due to its simplicity and rapidity. Additionally, it gives good agreement with the other models in the wind velocity of 3–10 m/s and for moderate incident angles (30°–45°).

The SAR images used in this paper are acquired by the Sentinel-1 satellite along the Iroise Coast (France). Compare to the other coastal places in France, Iroise Coast (Fig. 1) is an ideal place to study the estimations of wind speed because most wind velocities (weak, moderate and strong) are present and change according to seasons. The Sentinel-1 was launched in April 2014 by the European Space Agency (ESA) with the aim of providing an independent operational capability for continuous radar mapping of the Earth. It operates in C-band with dual-polarization for all modes (VV+VH or HH+HV). Compared with the other C-band satellites (i.e. Envisat, ERS-2), the Sentinel-1 images have a higher spatial resolution. They are acquired in three principal modes: Stripmap (SM) with 80 km swath and 5 m x 5 m resolution, Interferometric Wide Swath (IWS) with 240 km swath and 5 m x 20 m resolution, and Extra Wide Swath (EWS) with 400 km swath and 20 m × 40 m resolution.

The wind speed estimated by the SPM (EM model) and CMOD.5 [4] (empirical model) is compared in Fig. 2 for a sample of VV-pol images and in Fig. 3 for two samples of HH-pol images. From the obtained results, we can conclude here some important points. For VV-pol images, the SPM has good agreement in wind speed estimations with the CMOD.5 and in situ measurement for low and moderate wind, and for incident angles above 40° or above 30° with the tolerance of 2–3 m/s. To estimate more accurately strong wind and for incident angles below 30°, the other EM models such as TSM, SSA should be studied. Regarding HH-pol images, the SPM-based wind speed is significantly overestimated despite relative wind direction and radar incident angle. On one hand, the reason is due to the increase of the difference between VV- and HH-pol radar backscattering with incident angle when using the SPM. On the other hand, the wind speed overestimations for HH-pol images concern the sea surface roughness spectrum used in this paper (Elfouhaily model [5]). Hence, the other ocean wave roughness spectra are under study in the perspectives of this work.

Abstract

The NOrwegian Radar oil Spill Experiment 2015 (NORSE2015) took place at the Frigg field in the North Sea in June 2015. The campaign was realized as a collaboration between Norwegian Clean Seas Association for Operating Companies (NOFO), UiT The Arctic University of Norway and Jet Propulsion Laboratory (JPL)/National Aeronautics and Space Administration (NASA). The objective of this contribution is to give an overview of i) NORSE2015 and ii) the first results from ongoing analysis of the remote sensing data collected at the Frigg field.

 

During NORSE2015, four experimental oil spills were released on the ocean surface. Three of the experimental oil discharges were of mineral oil types with different oil-to-water fractions. The fourth spill was a plant oil type, released to simulate a biogenic film.

While carefully monitoring the environment, the slicks were deliberately left untouched on the water avoiding any inference from mechanical or chemical recovery within their lifetime. This allowed for several hours of data collection by airborne and space-borne synthetic aperture radar (SAR) during the development and weathering of the slicks.

 

The NASA L-band Uninhabited Aerial Vehicle Synthetic Aperture Radar (UAVSAR) carried by a GulfStream-III participated in the experiment and collected 22 quad-polarimetric SAR (QP SAR) images of the four slicks within an eight hour period. Supplementing this time series, QP SAR imagery from Radarsat-2 (C-band), dual-polarization TerraSAR-X data (X-band), and hybrid-polarity SAR data from RISAT-1 (C-band) were collected.

 

Scientists from the Norwegian Meteorological Institute released iSphere drifters (which experience windage), self-locating datum marker buoys (completely immersed, no windage) and radiosondes synchronized with the release of the oil. In addition, observations and photography of the slicks taken from the release vessel, optical/infra-red imagery, and sampling of the oils were also recorded.

 

The overall aim of NORSE2015 is to increase the understanding of the interaction between the polarization dependent electromagnetic radar signal and its interaction with the emulsions of oil and seawater of varying dielectric properties. With the unique data set available, the NORSE2015 team has launched a multitude of studies on the SAR measurements. The current research includes, but is not limited to, drift and development analysis of the slicks, comparison across the SAR sensors operating at different frequencies (X-, C-, and L-band), and investigations of hybrid-polarity versus QP radar measurements. The preliminary results are to be discussed at the symposium.

 

Acknowledgement

This work was carried out in part at the Jet Propulsion Laboratory, California Institute of Technology, under a contract with NASA. NORSE2015 was partly financed by CIRFA – Centre for integrated remote sensing and forecasting for arctic operations (RCN grant no. 237906).

Coastal areas are of high ecological and economic value however they are, today, subject to intense human-induced environmental pressures. An effective monitoring system is therefore vital for the operational management and safeguarding of the coastal areas. Satellite remote sensing has the potential to monitor the coastal area and its dynamics in a synoptic and cost-effective way. However, due to frequent cloud coverage, combined with the high dynamic nature of near shore coastal waters, one single sensor such as for example MODIS has shown to provide only limited information on the Total Suspended Matter (TSM) dynamics. To monitor coastal areas a 250 m spatial resolution is often put forward (e.g. MODIS 250 m channels, MERIS, Sentinel-3).  However this spatial resolution might be inadequate to monitor small scale feaures in near shore areas such as ports and estuaries i.e. areas which are facing intensified anthropogenic pressures from maintenance of capital dredging activities, large scale construction works etc.   The remote sensing community has therefore  already been looking to some extent into the usage of non-OC sensors to fill the gaps, both in time of temporal and spatial coverage.

Although designed as a land mission only- the PROBA-V instrument, providing a daily coverage at 300 m and a 5-daily coverage at 100m resolution, opens opportunities for retrieval of coastal products.  The high revisit time, coastal coverage and the good image quality, provides opportunities to expand its current use from the typical land applications to coastal water applications.

At the conference we will present the first results of a study which has been recently set-up to analyze the feasibility of the use of PROBA-V data for coastal applications, more specifically for the monitoring of TSM. The study will take into account the possible limitations imposed by the design of the PROBA-V instrument as a land mission  such as the broad spectral bands and reduced radiometric sensitivity at low radiances (i.e. lower SNR compared to typical OC sensors for typical water scenes) of the instrument. This activity will also look into the selection of both an optimal atmospheric correction scheme and TSM retrieval algorithm.  

The United Nations Sustainable Development Goals (SDG) provide for a new policy framework for the management of freshwater resources, with SDG 6 aiming to ensure availability and sustainable management of water and sanitation for all. Global availability of timely, scientifically credible and policy-relevant environmental monitoring data is crucial for accomplishing this goal. The Global Expanded Monitoring Initiative (GEMI) was launched to meet this requirement. It builds on existing initiatives, in particular the GEMStat water quality database by the United Nations Environmental Programme for ambient waters. GEMStat is managed by the UNEP GEMS/Water Data Centre at the German Federal Institute of Hydrology. It currently consists of in situ monitoring data provided through a network of National Focal Points (NFPs). Despite decade-long efforts, representativeness and continuity of the data records in GEMStat remain insufficient due to financial and administrative constraints. The GEMS/Water Data Centre is trying to improve this situation by promoting the use of advanced and more efficient monitoring, data management and reporting tools.

The ESA DUE Innovator III project SPONGE (Spaceborne Observations to Nourish GEMStat) enables a collaboration of remote sensing specialists with GEMS/Water and NFPs in six selected countries, namely Brazil, Finland, Ghana, Guatemala, Japan and Tanzania. Country-specific requirements are identified concerning environmental issues and corresponding monitoring needs, including economical limitations for in situ measurements. Each NFP proposed up to 7 test sites in lakes and rivers. In situ sampling in these sites is carried out or commissioned by national environmental agencies at monthly or lower frequency. Site-specific acquisition calendars for Sentinel-2 and Landsat-8 were provided to the agencies in order to maximise the number of comparable reference measurements acquired throughout 2016. On this basis, the project is demonstrating during a one-year pilot service how ESA’s new Sentinel-2 satellite and Landsat-8 can contribute spatio-temporally comprehensive ambient water quality information to supplement limited in situ monitoring activities, and whether this information could be used as a sole proxy for the large number of unattended water bodies, especially in developing countries and at high Northern latitudes.

Sentinel-2 and Landsat-8 enable the direct retrieval of optical water quality parameters at high spatial resolution, and with an average temporal resolution of 10 days across all sites using both satellites. With the in situ data provided by the NFPs to GEMStat as reference, several algorithms are assessed in order to identify the most adequate procedures for each test site. Candidate algorithms include an adapted Case-2-Regional neural network, spectral optimization algorithms and linear matrix inversion. Products are provided to GEMStat for access by the individual NFPs. Product design, format and transfer are accordingly specified to meet the needs of both user groups. The number of parameters that are feasible with the given sensors and methods are limited to suspended matter, transparency, turbidity, and with reservations chlorophyll-a and coloured dissolved organic matter. Apart from traditional in situ sampling and high resolution remote sensing products, SPONGE will also assess the potential of automated monitoring stations and citizen monitoring as well as complementary Earth observation techniques.

Based on the pilot service, the project will define guidelines on the optimal combination of remote sensing (few parameters, relative accuracy, high spatio-temporal resolution, inexpensive) and in situ measurements (all parameters, absolute accuracy, low spatio-temporal resolution, expensive) in view of global water quality information needs and SDG 6.

Shell farming and oyster farming in particular is an important economic activity in France with several growing areas located along the Atlantic coast. Because of various sources of ecological and economic pressures, the oyster production in those areas underwent in the past decades a constant evolution of the areas occupied by cultivated oysters, some production sites being abandoned and others re-cultivated. In abandoned fields, wild oysters tend to grow rapidly and become an increasing source of competition for nutrients to the detriment of cultivated oysters. It becomes therefore necessary to develop new approaches designed to provide local authorities with the right tools to monitor this evolution and to take necessary actions.

Some recent studies demonstrate the interest of the international community to detect and map oyster habitat in tidal flats using high resolution satellite data. In these studies, the detection of oyster reefs is mainly achieved by exploiting the radar backscattering properties specific to these land covers in SAR remote sensing data (C- and X-band). Unfortunately, the spatial resolution of these data does not always enable a fine detection of oyster beds contours and prevents the distinction between abandoned and cultivated fields. 

In previous studies, we proposed two distinct processing chains to detect oyster fields in very high resolution optical data through texture analysis. The first method is an unsupervised segmentation approach relying on the extraction of Haralick texture descriptors to detect oyster beds contours. The second one is a supervised classification approach based on the use of multivariate probabilistic models to represent the distribution of wavelet coefficients obtained by an orthogonal wavelet decomposition of the image. This second approach enables to separate oyster beds in two classes, i.e. cultivated oyster fields and abandoned fields.

In this paper, we propose to combine the results obtained with both abovementioned approaches with a very high resolution panchromatic Pléiades data in order to enhance the mapping accuracy of cultivated and abandoned oyster fields. Both approaches have indeed complementary advantages, the first bringing fine contours of oyster beds and the second producing coarser contours but with an efficient distinction between cultivated and abandoned fields. This fused result is then further post-processed to derive two by-products: a map of abandoned fields and a map of cultivated oyster tables including table lengths. This second by-product can be useful to local land managers as table lengths are directly related to the oyster production. In the full version of this paper, the accuracy of the fused result will also be compared with accuracies of both previous approaches by including ground-truth data in the analysis.

The project GLaSS prepares for the Sentinel satellite series by setting up a core system processing facility, preparing and testing algorithms, and providing the larger audience with examples based on global lakes use cases of lakes of very different type.

Shallow lakes are often turbid, because of local re-suspension caused by wind-waves or turbid river discharge. In waters that are rich in sediments the ecology can be affected: reduction of water transparency inhibits primary production due to the lack of light in the water column. This also has a detrimental impact on organisms higher up in the food web. Lake management or restoration projects usually aim to bring these lakes from turbid to clear water state.

We chose two lakes to study in more detail. The first is Lake Markermeer (Netherlands), where the soft bottom leads to frequent resuspension events. It is part of a restoration project (Marker Wadden) in which several islands will be built to reduce fetch of the wind waves and thereby resuspension, improve the underwater light field and increase the biodiversity, also by introducing gently sloping shorelines.

After atmospheric correction, turbidity was derived from Landsat 8 data and compared with in situ turbidity measurements to tune the algorithm. Based on this algorithm high resolution maps of the lake can be created. With Sentinel 2 the development of the Marker Wadden project will be followed. This can be compared against long-term time series from MERIS.

The second is Lake Böyük Şor (Azerbaijan), which is under high environmental pressure, which includes oil pollution. A restoration project is ongoing here to improve the water quality. For Lake Böyük Şor oil potential and turbidity maps were made based on TOA data, and scaling of band ratio results to the same level of in situ data. Especially for the oil maps, the evolving efforts and results of the restoration project could be followed.

The GLaSS project prepares for the Sentinel satellite series by setting up a core system processing facility, developing and testing algorithms, and providing users/stakeholders with examples based on global use cases of lakes with very different water types. Based on the examples, GLaSS has produced training material targeting graduated students, professional ecologists, lake specialists and others who would like  to learn how Earth Observation, especially data from Sentinel-2 and Sentinel-3 can be used to derive and analyze water quality parameters.

Students can learn which satellite data is appropriate for their purposes and how to access, download and open the images. They can learn stepwise how to apply atmospheric correction, validate the reflectances, process the images in order to retrieve water quality parameters such as chlorophyll-a, total suspended matter, turbidity and coloured dissolved organic matter (CDOM).  They will have access to the GLaSS tools developed in BEAM (and soon also the SNAP software), for example to the lake classification and extraction of parameters from a region of interest (ROI) with calculation of basic spatio-temporal statistics.

The examples cover lakes from all over the world that are subject to a large range of socio-economic situations and which present very distinct optical properties, from highly scattering to mainly absorbing waters, from highly polluted or eutrophic waters to deep clear and blue lakes, and from pure glacial lakes to mine tailing ponds.

The GLaSS training material is made available via the projects website via the ESA´s LeanEO! facility and the GEOs GEOSS portal. It can therefore easily be used for self study, or downloaded by universities or other entities to be used in courses such as aquatic ecology, remote sensing, environmental sciences and technology. An overview of the training material and its potential usage will be presented at the conference.

The increase in maritime traffic, and particularly the transport of chemical products, necessitates  enhancement of methods of prevention and intervention towards the spills of chemicals in the environment. Maritime pollution by chemical products occurs at much lower frequency than spills of oil, however the consequences of a chemical spill can be more wide-reaching than those of oil. While detection and characterization of hydrocarbons have been the subject of numerous studies, detection of other chemical products at sea using remote sensing has been little studied and is still an open subject of research. To address this knowledge gap, an experiment was conducted in May 2015 over the Mediterranean Sea during which controlled releases of hazardous and noxious substances were imaged by remote sensing systems. The aim of these experiments and subsequent analysis is to establish a procedure for collecting evidence of illegal maritime pollution by noxious liquid substances using airborne sensors.

In this paper we discuss the experimental procedure, which was carried out in collaboration with the French Navy and Customs, and report the main results from the airborne radar imaging campaign. We develop an accurate method for using multifrequency radar sensors to detect and quantify impact of chemical products at sea. We conclude by demonstrating the capability of radar imagery to distinguish two different substances within the same spill.

The aquaculture industry in Denmark is ready to significantly expand in the coming years. Currently, the Danish environmental protection agency is examining applications for a dozen new marine farms. To align requirements of fish farmers with the concerns of the tourist sector and nature protection organisations, impacts have to be thoroughly assessed and sites carefully selected.

The process of selecting a site is complex and many factors are feeding in to it; yet, it is crucial to find the right combination of physical conditions that best support cost-efficient production, while at the same time ensuring a low environmental footprint. Information from Earth Observation (EO) can be highly valuable in this process, because it can not only provide data on the current state of potential sites but also historical information.

In this project we focused on two variables which can be derived from EO: chlorophyll-a (chl-a) and sea surface temperature (SST). Chl-a serves as an indicator for water quality and SST is crucial for the fish’s health and feed management. We used the ENVISAT Meris time series from 2002-2012 to derive Chl-a, and SST was taken from the Group for High Resolution Sea Surface Temperature (GHRSST) for the same time period. In addition, both variables were also modelled with ECO Lab, an ecological modelling tool.

Then we analysed and statistically compared the EO with the modelled data and evaluated both products also with in-situ data. In a geospatial analysis we were looking for reoccurring patterns to spot if certain areas in the Danish waters are better suited for aquaculture than others, based on the 10-year record and we investigated if and how many times, certain chl-a and SST values which are critical for the fish’s survival were exceed. In addition, we also analysed data in proximity to existing fish farms to see if we can detect a difference in chl-a as compared to the surrounding waters and hence monitor impacts from the farms. Finally, the results were compiled in a suitability map showing areas that are better suited for farming activities than others based on the analysis of chl-a and SST.

The project presented here was carried out under the framework of AQUA-USERS (http://aqua-users.eu), an EU-funded project with the objective of providing the aquaculture industry with user-relevant and timely information based on satellite data and innovative optical in-situ measurements.

Cyclone is defined as the circular motion of fluid in a closed area rotating in the same direction of an earth which is characterized by strong winds counter clockwise in the northern hemisphere and vice versa in the southern hemisphere. However, the main reason for tropical cyclone is the sea surface temperature, and low pressure and high precipitation rate in nearby areas. Depending upon the wind speed cyclone can be categorized into 5 major categories I.e. category-1, 2, 3, 4 and 5. Cyclone genesis, intensity, track and landfall have to be predicted with accuracy throughout the life cycle for a better disaster management in planning the mitigation efforts to save loss of human life and property. Upon the integration of in-situ measurements and satellite observations, the study of cyclones can be done. On the contrary, in-situ measurements pose the greatest limitation on studying the temporal and spatial coverage of cyclone. So therefore, satellite observations have been a key component of cyclone monitoring and research. Optical sensors on the geostationary satellites have been the only source of information for cyclone studies for a long period of time. It is well known that visible/infrared sensors cannot provide all the information required for monitoring cyclones. The development of advanced sensor technology and the availability of a variety of satellites orbiting the earth, it has been possible to use multi-spectral, multi-sensor combination to improve the understanding of physics of the tropical cyclones for a better and accurate forecasting. Thus a combination of visible and infrared data from the geostationary satellites and active and passive microwave data from atmospheric sounders, scatterometers and altimeters has improved the cyclone prediction capability skills. The main purpose of this study is to temporally study the effects of Cyclone Nanauk an Indus Delta (Wetland) and to know the relationship established among SST, SSH, wind vectors and wind speed. For this remote sensing plays a vital role in cyclone monitoring and assessing the damage ahead of cyclone. Study area was Indus Delta forms where River Indus is entering into the Arabian Sea, Pakistan. The total area of Indus Delta is about 41,440 km² (16,000 square miles) and is approximately 210 km across where it meets the sea. The vital ecological importance of this area is its being the largest arid mangrove forests in the world, as well as homes for many birds, fish and the Indus Dolphin. Information on the location, intensity and track of the cyclones were first confirmed from Pakistan Meteorological department. Then desktop work was started with the acquisition of Landsat 8 image then pre-processing was applied, that includes stacking of bands, digitizing Indus Deltaic region and latterly sub setting of this area. Now spectral indices were applied to enhance water and vegetation. Normalized Difference Vegetative Index (NDVI) and Modified Normalized Difference Water Index (MNDWI) were calculated. Unsupervised Classification of Images was performed on land covers on the basis of their inundated and non-inundated area extent. Inundated areas were all calculated in square kilometers (km2). This work needs to be extended with the data of sea surface temperature, sea surface height, wind vectors and wind speed so that the reason of tropical cyclone Nanauk can be studied and it will play a critical role in disaster management practices. With sea surface temperature, wind vectors and wind speed, cyclones emergence time can be ascertained and sea surface height can be useful to give wave height which will assist in early warning systems of cyclones by measuring and comparing it with the normal wave height. This is how this research can play its vital role in predicting the emergence of cyclones and other sea level anomalies.  Therefore, along with other remote sensing and in-situ data, prediction and the severity of cyclones in terms of intensity of given storms can be forecasted. Or in other words, with this altimetry tracking, monitoring and measurement of the ocean surface including waves, winds can help in providing near real-time estimates of wave height and storm surge at landfall. Due to the non-availability of data , damage assessment in terms of inundation of sea water in Indus Delta was done prior to analyze  SST, SSH, wind speed and wind directions. 

A consensus exists that most applications of ocean surface current information (just like other oceanic and atmospheric state estimates) require useful measures of confidence.  One challenging aspect is to provide measures of confidence that are as local as possible, whether in space or in time. Regional characterizations of a combined geostrophic and Ekman surface current estimate at 15-m depth, as given by ESA's GlobCurrent analyses (2002-2014), are provided at a spatial resolution defined by about 15 regions covering the global oceans. The proposed regions are based on descriptions of well defined climatological surface currents and follow the contours of the project's mean dynamic topography.  Comparisons include relative differences in Lagrangian separation between actual drogued drifting buoy trajectories and trajectories following the combined current.  Hypotheses to explain some of the skill variations found within and between regions are further explored.

For ecologically sensitive regions of the global ocean like the Baltic Sea, wind induced upwelling is one of the main factors affecting the circulation and the ecosystem of the region. With the help of satellite remote sensing which allows one to obtain spatially detailed, regularly and continuously repeated datasets, more detailed coastal upwelling studies for the Baltic Sea and its coastal lagoons became available.

In this study, the applicability of Terra/Aqua MODIS Infrared imagery and Landsat Thermal Infrared scenes for documentation and calculation of the SST changes during the coastal upwelling in the SE Baltic Sea and Curonian Lagoon is shown. Upwelling events that took place during warm period of 2000 – 2014 are analysed and standard oceanographic characteristics describing frontal zones are calculated. The maximum observed SST gradients across the front were up to 1.2 °C/km, temperature drop up to 15 °C, with total upwelling-affected area being up to 18000 km². It is also recorded that the horizontal scale of the upwelling can be from 100 to 400 km alongshore, up to 70 km cross-shore. The duration of the upwelling in this part of the sea may reach up to several weeks. The areas of the most frequent upwelling development are indicated.

We also show that intensive coastal upwelling events strongly influence not only the SE Baltic coastal zone, but also impact the hydrological regime of the northern part of the Curonian Lagoon. Under certain wind conditions, coastal upwelling along the SE Baltic Sea coast, depending on its scale and intensity, may lead to an intensive intrusion of cold and salty marine waters to the Curonian Lagoon resulting in hydrodynamic changes and pronounced density drop extending up to 30-40 km down the lagoon. Availability of satellite data helps to understand the hydro-dynamical processes taking place during the intrusion of the Baltic Sea waters to the Curonian Lagoon.

This work is supported by "Lithuanian Maritime Sectors' Technologies and Environmental Research Development" project Nr. VP1-3.1-ŠMM-08-K-01-019 funded by the European Social Fund Agency.

Satellite altimetry missions now provide a more than 20 years record of continuous measurements of sea level along the reference ground track of TOPEX/Poseidon. These measurements are used by different groups to build the mean sea level rise record, which is an essential climate change indicator. Estimating a realistic uncertainty on the sea level rise rate deduced from satellite is of crucial importance for climate studies such as sea level budget closure.

Ablain et al., 2015 estimated the GMSL trend uncertainty 0.5 mm/yr (90% confidence interval) by a careful study of the differences between altimeter standards. In this study we use two approaches to improve this kind of studies:

A probabilistic ensemble approach where a large number of members are used to derive an uncertainty envelope for the MSL,

A generalized least squares method using modeled error covariance to estimate a map of realistic uncertainties on regional sea level trends.

After ENVISAT, and now Sentinel-1, co-polarization SAR imagettes have been widely utilized to extract ocean wave spectral information, with adjusted modulation transfer function (MTF) for both HH and VV polarization. So-called velocity bunching contribution shall affect both VV and HH SAR image contrasts, these data can serve to better investigate the distinct tilt and hydrodynamic MTFs. However, due to rare observations of swell wave in cross-polarized channels, the MTFs in HV and VH for inversion algorithm are still unknown. Thanks to Radarsat-2 cross-polarized capabilities, this analysis can already be extended to analyze simultaneous co- and cross-polarized acquisitions. This opens new opportunities and helps to explore SAR capabilities to image large-scale waves using cross-polarized datasets. The inversion scheme included in ASAR as well as Sentinel-1 wave mode is based on estimation of cross-image spectra, which is able to resolve the 180 direction ambiguity without a priori information. By analogy, this effective algorithm is applied to Radarsat-2 SLC Quad-polarization products to examine its behavior in cross-polarizations. Preliminary results are encourging since the imaginary part of cross-image spectral is clearly visible in both VH and HV channels, demonstrating the applicability of cross-polarizations in retrieving ocean waves. Further efforts would be devoted to estimating the MTFs for HV and VH polarizations and investigating the influence factors on wave spectra inversed from quad-polarized data. Results will be presented and be discussed to help improve inversion algorithms for both dual and cross-polarized SAR data.

Mine tailing ponds, filled with hazardous waste, can lead to local environmental disasters. Every now and then it happens that a tailing pond construction breaks and the toxins are released in a river. Not everywhere in the world the information on the exact location of these ponds is well organised and mines are found in the most remote locations, which are often also nature areas.

Therefore, location of mine tailing ponds with Earth Observation is of great interest. The project Global Lakes Sentinel Services (GLaSS) created a method for this, based on the distinct colour of these ponds. In preparation of Sentinel-2 data, Landsat-8 is used as a proxy to develop an algorithm. Three different study areas are selected: Azerbaijan, Finland and Mongolia. For these areas all available images for the year 2014 were acquired from the U.S. Geological Survey as  level 1 products. The images were converted to top-of-atmosphere (TOA) reflectance. From the available flags provided with the images only the flag for high probability of snow could be used. The other flags were either not discriminative enough of tended to flag out mine tailing ponds because of their distinct spectral shape. To select only water bodies, several land masks were applied to the TOA reflectance images. One was selected, that did not flag out the mine tailing ponds. In evaluating the images it was found that images with snow were not useable, because, the Landsat-8 snow flag was also flagging out parts of mine tailing ponds. Consequently the dataset was reduced to only the months with low snow cover potential, the months July through September for the Northern Hemisphere. After applying all these masks and restrictions, the remaining data mostly comprised of mine tailing ponds and cloud shadow. By evaluating the spectra of the remaining data, several band ratios were applied to reduce the dataset to primarily mine tailing ponds. This ultimate selection was used to trace the original water bodies they are a part of, using a binary propagation algorithm.

The final results show mine tailing ponds across all three study areas. There were still some false positives and false negatives, but the overview provides clear insights in mining tailing ponds locations. With the availability of Sentinel-2 data it is expected that it will be possible to elaborate and expand this method to increase accuracy and reliability.

The normalized radar backscatter (NRCS) of SAR signals strongly depends on the radar incidence angle (Figure 1A). In cross polarization (HV and VH), the NRCS signal-to-noise-ratio (S/N) is much lower than in HH and VV (Figure 1B). Exact thermal noise antenna pattern is presently absent or incorrect in the Sentinel 1A data product. Existing software tools seems to lack the ability to  correct for the thermal noise.

We present the Nansat software based solution to the thermal noise correction and incidence angle normalization. Based on analysis of many SAR scenes, mean dependences of the NRSC (HH polarization, Figure 2A) and of the thermal noise on the incidence angle (HV, Figure 2B) are computed. This procedure is automated and provided as an open source tool at github.com. The results (Figure 3A and 3B) are promising and, as demonstrated, the method is used in operational sea ice classification at the Nansen Center.

In inland waters, bad water quality due to cyanobacteria blooms is an increasing problem. Cyanobacteria can form thick floating mats called ‘scums’, which produce bad odour and also toxins, that may lead to dead fish, water fowl, pets and cattle. The bad smell is of public concern at locations where houses are close to the water, while scums and toxins lead to closure of bathing locations and therefore to high costs for the tourist industry in the area.

The CyMonS-service offers water authorities high-frequency monitoring with a spatial overview at low costs, combined with a daily forecast for cyanobacteria scums in surface water bodies. Consequently, decisions by the water authorities can be taken on a daily basis, including the possibility to re-open a bathing water location during the season which normally stays closed until the next inspection (usually one or two weeks later). Optical field instruments such as the WISP-3s (Water Insight SPectrometers) and EcoSpot are used in combination with Earth Observations (EO) to provide a frequent and synoptic overview of the current situation. The algal forecasting model EWACS (Early Warning Against Cyanobacteria Scums) uses the WISP-3, EcoSpot and EO data to produce scum forecasts. These warnings allow water authorities to take appropriate measures on time, such as closing of inlets, aeration of the water column and warning the public. The long term dataset based on EO and field observations is used to find the sources of the scums and can serve for Water Framework Directive Reporting.

In 2014 the first positive feedback was obtained from local water authorities in the Netherlands. Together with these users the service is being improved, e.g. by automatic processing of Sentinel-2 data and merging the results into the users’ data management and decision tool systems. The results will underline the importance of Sentinel-2 for the monitoring small inland water bodies with a high accuracy and frequency. 

Since existing wind measurements in near-shore and offshore areas are both sparse and scarce, simulations from state-of-the-art mesoscale models are being used for wind resource predictions. In coastal and near-shore areas, such models are rather inaccurate and uncertain, primarily due to their numerical approximations, which do not resolve the large changes in local topographic features and atmospheric stability well. The RUNE project is designed to investigate cost-effective measurement solutions for improving wind resource modelling of coastal areas, since the accuracy of modelled wind resource predictions can be significantly improved by using local wind measurements to ‘calibrate’ the models.

Within the framework of the RUNE project, the wind over a coastal area is measured by land-based lidar systems, an offshore lidar buoy and satellite radar remote sensing (SAR and scatterometers). In addition, simulations using the Weather Research & Forecasting (WRF) mesoscale model are conducted. The purpose of the analysis is to evaluate the uncertainty of the modelled wind in the coastal zone and further improve it. To achieve this, WRF is run using a high-resolution satellite SST reanalysis product from the Danish Meteorological Institute (DMI), specifically developed for the North Sea and Baltic Sea region. Moreover, based on recent findings of the ESA STSE project SSTDV:REX-IMAM on the diurnal variability of SST and its implications in atmospheric modelling, WRF is also forced with optimally interpolated hourly SST from a geostationary sensor for selected cases. To improve the representation of the coastline, the elevation, topography and land use, the CORINE land cover database and the SRTM elevation database are implemented in the description of the WRF domains; with a spatial resolution of 100m to 250m, the CORINE land cover information represent a more accurate classification of land uses for the entire domain. 

SST, land cover, and elevation information from Earth Observation platforms are unique due to their extended spatial coverage and resolution, such that they can be implemented in the mesoscale model to better represent the actual conditions in the study area. Such improvements are expected to strengthen the model's ability to represent land-sea and air-sea interactions, the atmospheric stability and the local topographic features that partly affect the coastal zone.

A combination of the GOCE-based geoid models and mean sea surface models provided by satellite altimetry allows modelling of the satellite-only mean dynamic topography (MDT). Such MDT models are significantly affected by a stripping noise due to omission errors of the spherical harmonics approach. Appropriate filtering of this kind of noise is crucial in obtaining more precise satellite-only MDT models. In our study we use the nonlinear diffusion filtering based on a numerical solution to the nonlinear diffusion equation on closed surfaces (e.g. on a sphere, ellipsoid or the discretized Earth’s surface), namely the regularized surface Perona-Malik model. A key idea is that the diffusivity coefficient depends on an edge detector. It allows effectively reduce the noise while preserve important gradients in filtered data. Numerical experiments present nonlinear filtering of the satellite-only MDT obtained as a combination of the DTU13 mean sea surface model and the GO_CONS_GCF_2_DIR_R5 geopotential model. They emphasize an adaptive smoothing effect as a principal advantage of the nonlinear diffusion filtering. Consequently, velocities of the ocean geostrophic surface currents derived from such filtered satellite-only MDT models contain stronger signal.

The free availability of Sentinel-1 image data and their frequent revisit capabilities has prompted us to analyse their suitability for the classification and monitoring of coastal areas. Within the framework of a coastal Thematic Exploitation Platform (TEP) development study under the auspices of ESA, we experimented with the information content and quality of Sentinel-1 images for coastal applications. In particular, we were interested in the analysis of urban and agricultural areas as these two categories represent typical examples of near-coastal land cover apart from non-accessible and uninhabited remote areas. Evidence of human and economic activities is often concentrated in coastal areas and their classification together with the monitoring of any changes is a primary task in remote sensing.

Among the Sentinel-1 image data acquired over land, the level-1 Ground Range Detected (GRD) products comprise High Resolution (HR) detected data taken routinely in Interferometric Wide swath (IW) mode. These data are available (prior to geocoding) with a pixel size of 10×10 meters and correspond to about 5 looks. They should have a nearly uniform Signal-to-Noise Ratio (SNR) and also a stable Distributed Target Ambiguity Ratio (DTAR). Thus, automated and rapid image analysis should be possible without repeated and detailed quality checks of each image. ESA also provides a toolbox for further image processing and analysis steps that be exploited for quality assurance.

Over Europe, ESA defined an operational sub-mode providing a standard combination of VV and VH data thus permitting a continuous analysis of the surface scattering characteristics over the European countries. We applied the ratio of the VV and VH polarization channel data to investigate their suitability for automated land cover analysis in form of semantically labelled image patches and we studied the impact of terrain slope, incidence angle and noise on the attainable results. The effect of random noise reduction can be seen clearly for extended targets such as agricultural fields.

We also looked into the potential of analysing image time series, where we were interested in local changes appearing in sequences of geographically superimposed and co-registered images ideally taken from identical viewpoints.  Here, the analysis of pixel stacks yields pixel change maps, i.e., a systematic mapping of the locations where changes occurred. This technique works well for low noise data, while strong noise may be confounded with real physical phenomena. A potential remedy is the exclusion of intermittent changes that do not persist in more than one image.

Our purpose was to obtain fast classification results. Therefore, we set out for “cascaded learning”, a layered technique that minimizes computing times during classification by applying a multi-stage classifier, where lower level stages will be skipped when a conclusive decision can already been made on a higher level. This pruning of lower level computations can, depending on the homogeneity of image content, lead to considerable gains in run-time during the automated classification of big images.

This becomes especially apparent when we compare urban and agricultural scenes. The given pixel size of 10 meters leads to locally extended homogeneous areas in the case of big agricultural parcels, while urban scenes containing irregular architectural structures of historic settlements do often not allow the identification of small-sized objects, however, as the architectural styles differ among the European nations, the learning of categories has to cover various cases. We will demonstrate what can be attained in the case of Sentinel-1 data for rapid land cover classification and change detection.

The ability to understand the close links between lacustrine environments and the human-altered landscapes in which they are embedded in depends strongly on the spatiotemporal availability and quality of key bio-optical parameters such as transparency, Chlorophyll A concentration (ChlA), Colored Dissolved Organic Matter (CDOM) absorption coefficient, Inorganic Suspended Matter (ISM) concentration. However, the current European monitoring based on the data collected from field observations do not characterize finely and reliably the spatiotemporal dynamics of these parameters. As an example, for the application of the Water Framework Directive in France, on the one hand only 380 lakes/reservoirs are monitored among the 480 involved water bodies, and on the other the collected measurements remains very punctual and discontinuous in time (four dates per year every 6 years ).

Remote sensing technique is often portrayed as a powerful tool to overcome this difficulty, as it provides synoptic and accurate spatio-temporal information on the desired bio-optical parameters. However, the different “ocean color” sensors that are able to monitor these parameters, such as SeaWiFS (1km, NASA), MODIS (250 m, NASA), MERIS (300 m, ESA), have a spatial resolution which is often too low to correctly study the majority of lake/reservoir ecosystems. Typically, only 2% of 480 French lakes/reservoirs are overlapped by at minimum 5 pixels of 250-m spatial resolution. The successful launch of Sentinel-2 satellite that carries the Multi-Spectral Instrument (MSI) sensor allows interesting research perspectives to monitor bio-optical properties of these lakes/reservoir due to (i) its high spatial resolution (typically between 10 m and 60 m); (ii) its suitable signal to noise ratios, and (iii) its good spectral performances ( 13 bands).

The presentation aims at describing an ongoing project so-called TELQUEL. The main objectives of this project are to provide bio-optical algorithms for MSI/Sentinel-2 data to retrieve the water transparency, and ChlA, CDOM and ISM concentrations. The first step consists in developing a specific atmospheric correction algorithm over lake areas that takes into account the altitude effects and the adjacency effects induced by the land surface surrounding the lakes. The second step of the project consists of establishing bio-optical relationships between MSI/Sentinel-2 reflectances and the concentrations of biogeochemical parameters. A final step is dedicated to use the retrieved bio-optical properties and suspended matter concentrations for enhancing the lake water quality evaluation and simulation. We will focus our presentation on the methodological approach to correct for the atmospheric effects for lake/reservoir targets (dark target). Preliminary results on the bio-optical relationships obtained over the test-site of the project, namely the Naussac reservoir (10 km²) (close to Le Puy-en-Velay, France) will be presented.

Cryosat-2 offered the first ever possibility to perform coastal altimetric studies using SAR-Interferometry as well as SAR altimetry in preparation for the Sentinel-3 mission.  With this technological leap forward it is now able to observe sea level in very small water bodies and also to provide coastal sea level very close to the shore.  

We perform an investigation into the retrieval of sea surface height in the North Sea and around Denmark and performing the first inter-comparison between SAR altimetry from Sentinel-3 and Cryosat-2. The availability of a radiometer on-board Sentinel-3 used for i.e. the wet troposphere corrections is also tested and compared with similar wet troposphere correction for Cryosat-2 based on a model (ESMWF)

The use of the Cryosat-2 and Sentinel-3 is furthermore investigated for possible assimilation into sea level forecasting along the coasts of Denmark. This is a part of the EU sponsored project LOTUS in which the possibility of new Sentinel-3 downstream services are outlined . The advantage of the SAR data compared with conventional altimetry in the coastal zone, is the fact that the  increased spatial resolution of Sentinel-3 and Cryosat-2 SAR provide valuable sea level observations within the narrow Straits around Denmark which are crucial to constrain the waterflow in and out of the Baltic Sea. These narrow straits were not well monitored by i.e. ENVISAT due to land contamination within the altimeter footprint.

The European Space Agency (ESA) Soil Moisture and Ocean Salinity (SMOS) satellite has provided sea surface salinity (SSS) data for over six years. It has been shown that the operational ESA Level 2 SSS data have significant spatially and temporally varying biases between measurements from ascending passes (SSS_A; SMOS moving south to north) and descending passes (SSS_D; SMOS moving southwards). Therefore, without calibrating SSS to some reference salinity it is difficult to use these data to study long-term, basin-wide processes.

The data used for this study were from two SMOS SSS climatologies one based on SSS_A and the other from SSS_D. These climatologies have been used to calculate salinity anomalies, which are shown to have significantly reduced the spatio-temporal biases. This presentation will show how these SSS anomalies can be used for oceanographic studies, including the detection of planetary waves/large eddies in the South Indian Ocean.

Planetary, or Rossby, waves (RW) are important in oceanography (e.g. in maintaining western boundary currents). Prior to the availability of satellite measurements of sea surface height (SSH), only an indirect confirmation of the existence of RW had been possible from sparse in situ data. Signals attributed to RW have now been fully observed and characterised in satellite-derived fields of SSH, sea surface temperature (SST) and chlorophyll-a. At mesoscale wavelengths (<300 km) these signals seem to be dominated by non-linear eddies and at large length scales eddies and RW have the similar propagation speeds.

Our analysis shows geophysical signals in the South Indian Ocean that are consistent among SSS anomalies, SST and SSH, and with previously published results, in detecting large-scale non-linear eddies/RW in the region.

Strong seasonality, driven by the semi-annually reversing monsoon winds and associated surface currents, distinguishes equatorial Indian Ocean (EIO) from other equatorial regions of the global ocean. Intra-seasonal variability is the next dominant variability influencing the dynamics of EIO. Though the intra-seasonal variability is comparatively a well researched and reasonably well observed phenomena in the atmosphere, in the ocean it is still in its infant stage. In the present paper we document the bio-physical coupling at the central EIO mediated by the atmospheric Mixed Rossby Gravity Waves (MRGW) using a high-frequency co-located in-situ physical and biogeochemical time-series observation at 2 locations. The 1^st time-series was at 80.5^oE and equator during 25^th September to 3^rd October, while the 2^nd was at 83^oE and equator during 4^th to12^th October 2011. The characteristic feature of the vertical temperature distribution was a gradual cooling and deepening of the warm upper ocean during 1^st time-series followed by a warming but continued deepening of the upper ocean during 2^nd time-series. Concomitant with temperature, the vertical salinity structure showed a gradual increase in salinity with the subsurface high salinity core located in the thermocline showing high frequency oscillation. The subsurface chlorophyll maximum (SCM), which is a characteristic of the EIO, showed gradual deepening trend overriding high frequency oscillation. The Chi-squared analysis of the above data showed that the dominant variability was in the 4 to5-day periodicity. A similar periodicity was also seen in the atmospheric parameters lending support to the notion that this high frequency upper ocean variability was mediated by the atmospheric forcing in the similar frequency band. These oceanic MRG waves are important in entraining essential nutrients to the oligotrophic upper ocean in the EIO thereby triggering episodic high chlorophyll biomass in the euphotic zone, a phenomenon hitherto unexplored and documented.

One of the important water quality parameters related to human health is phytoplankton (cyanobacteria and algae) water blooms. Due to intensive landscape use we observe massive water blooms occurrence in many freshwater lakes, slowly running rivers and shallow coastal water, which is mostly due to anthropogenic eutrophication. Monitoring of such outbreaks is becoming a subject of many national and international programs.

Standard methods of phytoplankton quantification are based on the laboratory microscopic analyses and complementary spectroscopic assessment of chlorophyll concentration, and ultimately on in-situ measurements of induced chlorophyll fluorescence. Remote sensing may offer more effective and spatially precise methods for monitoring of cyanobacterial occurrence.  

We report on possible application of airborne imaging spectroscopy to map spatial distribution and concentration of cyanobacteria in inland waters. Brno water reservoir (Czech Republic) was selected as a model water body known for high seasonal concentrations of cyanobacteria. An airborne hyperspectral image (scanner AISA Eagle, Specim Ltd.) of Brno water reservoir was acquired in September 2013. Complementary, in-situ spectral measurements of the water surface were carried out with a FieldSpec – 4 (ASD Inc., USA) and water sampling with YSI probe (YSI, Inc. USA). We have done both categories of measurements from a boat during airborne data acquisition.

We applied two selected spectral indices (R₇₀₀/R₆₇₅ and R₇₀₀/R₆₀₀) which are sensitive to chlorophyll (Chl) and phycocyanin (Pc) concentration, respectively. The model was calibrated on field spectral measurements and consequently applied on the airborne hyperspectral image.

The results document: 1/Statistically significant correlations between spectral indices and concentrations of chlorophyll and phycocyanin; 2/Chl and Pc concentrations estimated from the airborne hyperspectral data display similar spatial pattern because of chlorophyll presence in both organisms (algae and cyanobacteria) and the accuracy of the Pc model was lower than that of Chl model.

This work has shown that airborne imaging spectrometry in combination with spectral indices can be used for effective assessment of cyanobacterial occurrence in inland waters. Nevertheless, other methods, such as radiative transfer modeling, can provide perhaps more reliable estimates of Chl and Pc concentrations because they can also account for other optically active water constituents.

Satellite ocean colour remote sensing provides a valuable source of information on the status of marine ecosystems. From 2002 till 2012 the MEdium Resolution Imaging Spectrometer (MERIS) ocean colour sensor onboard the ENVISAT satellite of the European Space Agency measured radiances emerging from the sea to quantify optically significant constituents; chlorophyll a (Chla, obtained through MERIS standard Algal Pigment Index 1 algorithm), total suspended matter (TSM) and yellow substances (YS). This data provides information on phytoplankton biomass and transparency through Chla and TSM, respectively, at large temporal and spatial scales, allowing an understanding of ecological dynamics. The concentrations of Chla and TSM also contributes to assessments for the “good ecological status” of the European Union (EU) Water Framework Directive 2000/60/EC, and for the “good environmental status” of the EU Marine Strategy Framework Directive, 2008/56/EC.

The present study evaluated the temporal variability of the MERIS water constituents off Sagres, in southwest Iberian Peninsula. This is one of the MERIS validation study sites, where in situ water samples and radiometric measurements were taken from 2008 till 2012, to assess the uncertainties related with the MERIS ocean colour products. To study the time series for Chla, TSM, and YS, a Seasonal-Trend decomposition procedure based on Loess (STL) was applied to decompose the trend (T_t), seasonal (S_t), and irregular (I_t) components using non-parametric regression. The STL decomposition was chosen over other decomposition methods, because it has several advantages, including changing the seasonal component over time and specifying a decomposition that is robust to outliers.

The authors have developed an algorithm that selects the best STL model for each combination of the seasonal and trend smoothing parameters, supported by the idea that the best data fitting will lead to a model that best describes the stochastic behaviour of a time series; i.e. the one that best captures the dynamics of the time series. The selection is based on the Root Mean Square Error accuracy measure, and the new stl.fit() procedure in R software (version 3.2.1). The daily satellite products, extracted from MERIS Level 2 reduced resolution images, were aggregated into monthly means, and the missing observations were estimated by linear interpolation.

Overall, the study characterizes the seasonal and trend patterns of the satellite-derived data, and then relates these patterns to environmental changes and possible causes. These components are useful for a better understanding of the temporal variability, and also explain the influence of each of the MERIS constituent products.

To assess the potential of a wave energy site, the wave climate must be established from historical data over a period of several years. With their 20-year long time-series, satellite altimeter measurements have increasingly become one of the most important sources of information on the sea state. The record of the wave conditions over a representative period builds up a picture of the wave climate, such as its variation with the seasons and from year to year.  The same analysis can be done for the tidal energy sites.

Wave climate data combined with numerical model outputs give the electrical power response over time and from that, its average level and its variability can be derived. These in turn will help determine the revenue stream from the sale of that electricity for a wave farm at that location, and therefore its commercial attractiveness. As more empirical data becomes available from operational machines, the power capture can be defined directly from measurements, although a role will remain for modelling tools in adapting this to improved energy converter designs and upgrades to control algorithms.

For marine energy resource assessment, TechWorks Marine Ltd. is assimilating altimetry data into the numerical wave model: WAVEWATCH III. In the light of RESGrow (collaborative project funded by ESA and managed by TechWorks Marine), two types of products are being developed for ocean energy resource assessments and marine operational services. One is a Marine Climatology Information(MCI) product for design and operational planning and feasibility studies whereas the other is a Nowcast Operational Service (NOS) for feasibility studies based on a near-real time basis (shipping and offshore operations).

Models are key tools for understanding the dynamics of the carbon cycle in the ocean. Satellite ocean colour imagery provides data at suitable temporal and spatial resolutions for model validation. However, there may be a disconnect between the variables used within a model, and those derived from satellite data.  The POCO project aims to evaluate satellite products for estimating pools of carbon in the ocean, with a focus on their suitability for use in model validation studies.Phytoplankton carbon was identified as a key variable of interest arising from user consultation conducted through POCO. Several candidate algorithms were selected and applied to satellite data from the ocean colour climate change initiative (OC-CCI). Presented here are the results of algorithm intercomparison and validation against a global database of phytoplankton carbon measurements. 

A pilot satellite-based investigation of modulations exerted upon mixed-layer phytoplankton fields by cyclones was performed for the first time across a selected part of the Arctic Ocean, the Barents Sea (BS). Resorting to a synergistic approach, cyclones were first identified from NCEP/NCАR data for the summer period during 2003–2013, and their propagation throughout the BS was further surveyed. The above-water wind force was retrieved from QuikSCAT data. These data were further accompanied by
ocean colour data from SeaWiFS and MODIS to examine the spatial and temporal distributions of surficial phytoplankton chlorophyll concentration (chl) dynamics along the trajectory of the cyclone’s footprint across the sea. Sea surface temperature was retrieved from MODIS data. The specific trajectory of cyclone passage across the BS area, depression depth, and wind speed proved to be conjointly the main factors determining the sign, amplitude, and duration of modulations of phytoplankton chl. The spaceborne data obtained over more than a decade indicate that, on balance, the cyclone passage led to increase in chl within the cyclone footprint area. On average, this increase did not exceed 1–2 μg l –1 , which is nevertheless appreciable given that the mean chl within the cyclone footprint rarely exceeded 1 μg l –1 . However, chl enhancement within the footprint area lasted only within the range of a few days to a fortnight, with the footprint area generally accounting for about 14% of the BS area. During the vegetation season (April–August, rarely till mid-September), the number of cyclones prone to optical and infrared remote sensing was about 2–3. In light of the above, arguably the cyclones studied are hardly capable of boosting annual primary productivity in the BS. Moreover, it can be conjectured that the same conclusion can be drawn with respect to the pelagic Arctic tracts that are generally less productive and more extensively cloud-covered than the BS. However, this supposition requires further studies in order to advance our understanding of the actual role of cyclones in modulation of Arctic Ocean productivity and ecosystem functioning.

To assess reliably and comprehensively eutrophication status of marine waters compel to develop and implement joint use of data from several monitoring platforms, monitored by different methods. Currently, the HELCOM eutrophication assessment is done based on conventional monitoring station data provided by the contracting parties i.e. countries surrounding the Baltic Sea.  The HELCOM EUTRO-OPER project aims to increase the confidence of eutrophication assessment by including Earth Observations (EO) and automated ferrybox observations to the chlorophyll-a indicator. The inclusion of EO data and Alg@line flow-through data complements the monitoring station measurements in the assessment of the state of the Baltic Sea, particularly in areas and seasons out of reach of traditional methods.  

 

The EO data in this study was MERIS (MEdium Resolution Imaging Spectrometer) chlorophyll-a data processed using BEAM plug-in processor FUB. The inclusion of MERIS observations to the chlorophyll-a indicator serves as an example on the use of Sentinel 3a/OLCI data in possible future assessments. The convetional monitoring station data orginates from HELCOM contracting parties and is hosted at the ICES data center (HELCOM COMBINE database). Alg@line  provides real-time information on the water quality with high-frequency automated sampling onboard several merchant ships on the Baltic Sea.

 

During the project, the accuracy and representativeness of chlorophyll-a produced through EO, Alg@line and conventional water quality monitoring were evaluated in different parts of the Baltic Sea. The results provide thoughout analysis on EO data accuracy in different parts of the Baltic Sea. The comparisons were carried out by analyzing areal distributions, time series and histograms. We also tested the ability of different measures of central tendency of EO data and compared it to the conventional monitoring data to describe eutrophication status on HELCOM assessment areas. We were able to yield spatially and temporally reliable results for the status assessments.

Aquaculture industry is increasingly contributing to the production of protein for human consumption [1]. However, in Portugal the growth of aquaculture has been limited  and there has been substantial  effort by the government in recent years  to promote aquaculture though the allocation of EU money (PROMAR 2007-2013)[2] to develop offshore concessions  for aquaculture development; most of these have been directed to the production of bivalves [3].  As part of  the licensing obligations for new  and existing aquaculture concessions, there has to be a  monitoring plan for  water quality parameters  including regular measurements of  sea surface temperature (SST), salinity (CTD), pH, nutrients, chlorophyll (Chl-a), Secchi depth (SD)  and oxygen (O₂) , as well as  yearly estimates of  the granulometry and organic matter of  the surface  sediments from the seabed  [3].

For this purpose, a weekly sampling campaign was established during 2014/2015 at the offshore mussel aquaculture, Finisterra S.A. located in the SW Portugal. SST, CTD profiles, SD and O₂  were measured at each campaign. Additionally water samples were collected to determine Chl-a, total suspended matter (TSM), nutrients and samples fixed for phytoplankton identification. Although this  in situ sampling regime is a legal requirement, it  is time consuming, costly, localised and intermittent. However, the use of remote data and near real time data sources would markedly improve the environmental information available for  effective decision making and management decisions. A good example of the applicability of remote sensing techniques in the SW Portugal to support aquaculture management is shown [4].  For the same area, the suitability of using remote sensing was tested for  the implementation of the EU directive MSFD applied to the Descriptor 5 Eutrophication [5].

The EU-funded project AQUA-USERS focuses on providing the aquaculture industry with user-relevant and timely information based on the most up to date satellite data and innovative optical in-situ measurements (http://aqua-users.eu). The WISP-3 is a handheld radiometer for monitoring water quality that performs measurements of sky upwelling radiance  (L_sky), upwelling radiance (L_u), spectral downwelling irradiance ( E_d) and can retrieve instant values of Chl-a, TSM also the  spectral diffuse attenuation coefficient (Kd) [6] and  other parameters dependent on the algorithms applied in the instrument.  The WISP-3 is  being used at the Finisterra mussel farm, at the same time as the collection of  in situ water quality parameters, allowing the verification/optimization of the algorithms used to better adjust to the water properties (Case 1). At the farm site, a buoy is deployed, carrying a set of instruments that include  sensors for optical irradiance, temperature, salinity, pH, O₂,  fluorometry, transmissometry and  wave height. The data is sent remotely every hour to a web site by file transfer protocol.

The objective of this contribution is to continue to relate  the data series on  in situ data collected from the weekly sampling campaigns  with the ocean colour data obtained from the WISP-3 [7], and to add further in situ data received from the recently installed oceanographic buoy at  the Finisterra aquaculture, together with any remote sensing data that might be available during early 2016  from the recently launched Sentinels 2 and 3.

[1]          FAO, The State of World Fisheries and Aquaculture, Opportunities and challenges. FOOD AND AGRICULTURE ORGANIZATION OF THE UNITED NATIONS, ROME, 2014.

[2]          Decreto-lei nº 81/2008 de 16 de Maio 2008, “Estabelece o enquadramento nacional dos apoios a conceder ao sector da pesca no âmbito do Programa Operacional Pesca 2007-2013 (PROMAR),” Diário da República, 1ª Série, nº95, 2008.

[3]          Despacho nº 4223/2014 de 21 de Março 2014 Diário da República, 2ª Série nº57, 2014.

[4]          J. Icely, G. Moore, S. Danchenko, P. Goela, S. Cristina, M. Zacarias, and A. Newton, “Contribution of remote sensing products to the management of offshore aquaculture at sagres, sw portugal.,” Ed. H. Ouwehand Proceedings of ESA Sentinel-3 OLCI/SLSTR and MERIS/(A)ATSR Workshop European Space Agency, p. 6, 2013.

[5]          S. Cristina, J. Icely, P. C. Goela, T. A. DelValls, and A. Newton, “Using remote sensing as a support to the implementation of the European Marine Strategy Framework Directive in SW Portugal,” Continental Shelf Research, pp. –, 2015.

[6]          A. Hommersom, S. Kratzer, M. Laanen, I. Ansko, M. Ligi, M. Bresciani, C. Giardino, J. M. Beltrán-Abaunza, G. Moore, M. Wernand, and S. Peters, “Intercomparison in the field between the new WISP-3 and other radiometers (TriOS Ramses, ASD FieldSpec, and TACCS),” Journal of Applied Remote Sensing, vol. 6, no. 1, pp. 063615–063615, 2012.

[7]          B. D. D. Fragoso, S. Ghebrehiwot, M. Laanen, and J. D. Icely, “Comparison between estimates for chlorophyll a and total suspended matter from optical readings with the WISP-3 handheld radiometer and measurements from water samples at an offshore concession for aquaculture off the SW coast of PORTUGAL.,” ESA Sentinel-3 for Science workshop, 2015.

In the last 15 years the detection of oil spills by satellite means has been moved from experimental to operational. Actually what is really changed is the satellite image availability. From the late 1990’s, in the age of “no data” we have moved forward 15 years to the age of “Sentinels” with abundance of data. Either large accident related to offshore oil exploration and production activity or illegal discharges from tankers, oil on the sea surface is or can be now regularly monitored, over European Waters. National and transnational organizations (i.e. European Maritime Safety Agency’s ‘CleanSeaNet’ Service) are routinely using SAR imagery to detect oil due to it’s all weather, day and night imaging capability. However, all these years the scientific methodology on the detection remains relatively constant. From manual analysis to fully automatic detection methodologies, no significant contribution has been published in the last years and certainly none has dramatically changed the rules of the detection. On the contrary, although the overall accuracy of experimental methodology is questioned, the four main classification steps (dark area detection, features extraction, statistic database creation and classification) are continuously improving. In recent years researchers came up with the use of polarimetric SAR data for oil spill detection and characterizations, although utilization of Pol-SAR data for this purpose still remains questionable due to lack of verified dataset and low spatial coverage of Pol-SAR data. The present paper is trying to point out the drawbacks of the oil spill detection in the last years and focus on the bottlenecks of the oil spill detection methodologies. Also, solutions on the basis of data availability, management and analysis is proposed. Moreover an ideal detection system is discussed in terms of satellite image and in situ observations using different scales and sensors.

The paper discusses our work on satellite survey of the Black Sea coastal zone. The main focus of the observations was put on detecting anthropogenic and natural surface oil-containing films.

Our main data source was high-resolution radar imagery obtained by synthetic aperture radars onboard Envisat satellite (till the spring of 2012) and onboard Sentinel-1 satellite (starting from October, 2014). We used data in visual and IR bands taken by Envisat MERIS, Terra/Aqua MODIS and by scanning radiometers of Landsat-5, 7, 8 satellites nearly simultaneously with the SAR images, in order to facilitate the differentiation between different types of surface pollutants, to reveal meteorological and hydrodynamic processes in test areas.

Main types of sea surface pollution films were investigated, specifically, the ones caused by oily wastewaters discharged by watercraft and natural marine hydrocarbon emissions.

 In our paper we show that in case of the Black Sea, the detection of oil spills caused by ship discharges is more complicated as compared to other seas due to intensive phytoplankton bloom and to natural hydrocarbons seeps that can be detected in various areas of the Black Sea.

Our findings suggest that certain types slicks detected in the Black Sea surface radar imagery are caused by natural hydrocarbon seeps at the Black Sea bottom. Geographical distribution of these slicks is characterized by permanent location and correlates with geographical spreading of natural hydrocarbons showings in the Black Sea.

The general schematic map of oil spills in the Black Sea aquatic area was generated on the base of satellite data for the years 2009-2015. The areas of the Black Sea most frequently affected by surface pollutions are outlined. Some statistics on oil spills in different parts of the Black Sea revealed from satellite data is drawn. These are year-to year and average monthly numbers of spills detected, numbers of spill revealed in different regions of the Black Sea over the period of observations, annual size distribution of total area of the oil pollution, and total polluted areas for various regions of most intense pollution.  A specific attention was paid to a comparison of the sea pollution level observed several years ago with the present time situation.

In summary, our long-term observations show that illegal waste discharges have become very common in the Black Sea and in aggregate it is a great threat to the ecology of the sea.

The work was supported by the Russian Science Foundation under the project # 14-17-00555. Envisat ASAR data were provided by the European Space Agency in the framework of the C1P.6342, Bear 2775 and C1P.1027 projects.

Starting from 2009 up to the present moment a satellite survey of the Central and Southern Caspian is carried out by the Space Radar Laboratory of the Space Research Institute of RAS.

Our main data source is high-resolution radar imagery obtained by synthetic aperture radars onboard Envisat satellite (till the spring of 2012) and onboard Sentinel-1 satellite (starting from October, 2014). The basic radar data is complimented by other satellite data on the sea surface and water condition, such as sea surface temperature, suspended matter and chlorophyll-a concentration, meso-scale water dynamics etc. We used data in visual and IR bands taken by Envisat MERIS, Terra/Aqua MODIS and by scanning radiometers of Landsat-5, 7, 8 satellites nearly simultaneously with the SAR images, in order to facilitate the differentiation between various types of surface pollutants, to reveal meteorological and hydrodynamic processes in test areas, and to determine factors governing pollutants’ spread and drift.

The main attention is focused on the detection of surface pollution in oil-producing areas and in areas of natural hydrocarbons showings at the sea bottom as well as on the anthropogenic pollution caused by oil-containing waste-water discharges from ships. A considerable amount of data analyzed allowed us to get statistically significant results on the spatial and temporal variations of different types of surface pollution manifestations in satellite images of the sea surface.

The patterns of surface oil pollution of the Caspian Sea are described and analysed, which considerably differs from those observed in other seas (e. g. in the Black and Baltic Seas). In many respects it is determined by particular natural properties of the Caspian Sea, namely the presence of big oil and natural gas fields on its bottom.

A comprehensive map of sea surface pollution by oil films is presented. According to this map two areas of the heaviest pollution are identified. Those are the area of Absheron and Baku archipelagos as well as the Western edge of South-Caspian depression. And the main sources of the Caspian Sea surface pollution are considered to be offshore oil production as well as the natural oil seeps at sea bottom. The illegal discharges in the Caspian Sea are not the main source of sea surface film pollutions, but unfortunately their amount is increased from year to year.

The impactsof various factors: small-scale and meso-scale water dynamics, local winds and surface currents, active atmospheric processes resulting in a considerable variation of the near-surface wind field, which influence the size, extent and spreading of oil films on the sea surface are revealed and discussed.

The work was supported by the Russian Science Foundation under the project # 14-17-00555. Envisat ASAR data were provided by the European Space Agency in the framework of the C1P.6342, Bear 2775 and C1P.1027 projects.

A natural indicator of the water cycle, salinity variations in recent decades have been linked to the Atlantic Meridional Overturning Circulation (MOC) variability. The advent of high-resolution, spatially dense sea surface salinity (SSS) measurements from the ESA’s SMOS and NASA/Argentina’s Aquarius/DAC-D missions has led to a revolutionary change in the sampling and resolution of the SSS field, particularly in the Tropics, enabling a more thorough understanding of the processes determining its spatial distribution and variability. Focusing on the Tropical/Subtropical Atlantic 30°N-30°S in view of its important role to the MOC and to the European weather and climate forecasting on seasonal to decadal time scales, this work reveals new aspects of the spatial and temporal variability of SSS in the region on a sub-annual to interannual basis.

 

Initially, the first analysis of the spatial and temporal decorrelation scales of SSS variability from satellite observations is presented, by exploring how quickly consistent SSS changes evolve over the different regimes of the basin, and pointing out regions with persistent in time SSS variations. Likewise, regions with spatially homogeneous SSS changes are identified. By examining different Level 3 and Level 4 SSS products from ESA’s SMOS and NASA/Argentina’s Aquarius/DAC-D missions are examined, with varying resolution, correction and averaging characteristics, the impact of these factors on the scales estimates and the variability observed from satellite SSS fields are considered. Overall, there is reasonable agreement between the time and space scales estimated from the different satellite datasets, despite the differences in their individual characteristics. However, Aquarius suggests longer spatial scales, which dominate the Southern Atlantic (~10°-25°S). The improved description of the characteristic scales of SSS from space represents a powerful, new investigative tool for the examination of regional variations of the controlling mechanisms influencing the SSS in the Tropical Atlantic and is similarly applicable to other geophysical properties of the ocean and over other oceanic basins. Equally important, it contributes to the evaluation of satellite SSS observations in assimilation systems, the development of optimally interpolated products, and for the definition of appropriate validation procedures among the various satellite SSS products.

 

Furthermore, this work examines the seasonal to interannual variability of the Tropical Atlantic SSS measured by the two satellite missions. On interannual time-scales, the SSS seasonal cycle in the region is dominated by pronounced, out-of-phase variations close to the major river systems of the Amazon/Orinoco and Congo/Niger river systems, first seen by SMOS in 2010 (Tzortzi et al., 2013). In contrast, the variability in SSS over the Tropical Atlantic basin as a whole continues to be little during the following years. An improved understanding on the sub-annual to interannual SSS variability in the Tropical Atlantic offers insights into the associated controlling mechanisms, and serves as valuable information for oceanographic, modelling and validation studies. Finally, by pointing to any differences in the SSS variations measured by the SMOS and Aquarius satellite missions, this work contributes further to our better interpretation of the salinity variability observed from space.

Optical satellite measurements provide knowledge about the interaction between physical and biological processes in the upper ocean. The dynamics of chlorophyll-a concentration is estimated from satellite multispectral data. This is controlled by phytoplankton growth and decay rates as well as the horizontal advection of water masses. Data from ocean color spectral imagers such as SeaWIFS, MODIS, MERIS and Sentinel-3 OLCI is merged into a single product by the OC CCI team and provided as daily global snapshots of chlorophyll-a and other water quality parameters in 4x4 km pixels [http://www.esa-oceancolour-cci.org/]. However, frequent occurrence of cloud screening is a major obstacle in using this data to study surface ocean processes on a day-to-day basis. For example, only about 15% of the optical satellite data is not obscured by clouds in the North Atlantic. This leads to a reduction in the number of surface observations to only one measurement every 3 days. In highly dynamic areas (e.g. Gulfstream in the North Atlantic), the limited ocean color data does not allow characterization of the underlying causes in changing chlorophyll-a distributions, e.g., growth / decay or also advection. A method for fusion of the OC CCI data and surface current data from the GLOBCURRENT project is proposed. This helps to provide significantly improved ocean color data coverage, which in turn may help to better describe the upper ocean dynamic and biological processes.

The GLOBCURRENT project provides satellite data based ocean surface Eulerian displacement of water due to geostrophic, Ekman and tidal currents [http://www.globcurrent.org/]. The geostrophic component of the current is calculated from satellite altimetry measurements, the Ekman current is calculated from scatterometry wind observations and the tidal component is calculated using a global tidal model. The product is publicly available at 10 km spatial resolution every 3 hours for the three years period from 2009 to 2011.

It is assumed that each pixel in the OC CCI dataset represents a parcel of water that can be considered as a passive Lagrangian drifter with an associated accuracy. Under this assumption we take a field of an OC product as input data and use the GLOBCURRENT product as a forward operator. For each pixel in the region of interest we predict coordinates of this pixel at the next time step (i.e. 3 hours after actual observation). For calculating the coordinates we integrate the time and space dependent field of eastward and westward surface water velocities (U and V) using the Runge-Kutta 4th order scheme. In other words, we spatially distort the field of OC CCI chlorophyll-a using GLOBCURRENT U, V. An ensemble of realisation are produced by using the surface current uncertainties provided in the GLOBCURRENT product and perturbing the OC estimate using the provided uncertainties. This produces a probabilistic but model-free forecast of the OC fields on the next day.

We then use flow-dependent linear quadratic estimation to optimally update the OC. This ensemble Kalman filter approach is employed repeatedly to produce a stochastic high-resolution dataset with improved spatial coverage and improved accuracy.

Performing this operation at daily time steps, we obtain OC maps characterized by:

very high temporal resolution (3 hours);

more complete spatial coverage;

ability to partition horizontal advection effects from biological chlorophyll-a growth and decay

Intertidal flats are coastal areas that fall dry once during each tidal cycle. In Europe large intertidal flats can be found on the Dutch, German, and Danish North Sea coasts, on the U.K. east and west coasts, and along the French Atlantic coast, but also at other places worldwide, e.g. in South Korea and northwest Africa. Adopting the Dutch name those areas are often referred to as Wadden Seas. Since 2009 the German Wadden Sea is a UNESCO World Natural Heritage, and according to national and international laws and regulations a frequent surveillance of the entire area is mandatory.

Remote sensing techniques are ideally suited for the surveillance of areas that are difficult to access. Optical sensors are already being used for a classification of different sediment types and for the detection of seagrass, bivalves, and macrophytes; however, their applicability suffer from the usually high cloud coverage on the German North Sea coast. In this respect, Synthetic Aperture Radar (SAR) sensors, because of their all-weather capabilities and their independence of daylight, may be a valuable addition, but the radar imaging of bare soils is rather complex, and the very processes responsible for the backscattering of microwaves from exposed intertidal flats are still subject to ongoing research.

We analysed a great deal of high-resolution SAR data of dry-fallen intertidal flats on the German North Sea coast with respect to the imaging of sediments, macrophytes, and bivalves. TerraSAR-X and Radarsat-2 images of four test areas acquired between 2008 and 2013 form the basis for the present investigation and are used to demonstrate that pairs of SAR images, if combined through basic algebraic operations, can already provide useful indicators for morphological changes and for bivalve (oyster and mussel) beds.

Multi-temporal analyses of series of SAR images allow for the detection not only of morphological changes (particularly in areas of high morphodynamics), but also of bivalve beds. The radar backscattering from those beds is generally high; whereas that from bare sediments depends to some extend on the imaging geometry, environmental conditions, and on the tidal phase. Therefore, a comparison of the temporal mean and standard deviation yields indicators for bivalve beds, which is demonstrated for three of our test sites.

Further, our results show evidence that also single-acquisition, dual-polarization SAR imagery can be used to identify oyster and blue mussel beds. Defining the polarization coefficient (i.e., the normalized difference polarization ratio of SAR images acquired at both like-polarizations) and, again, applying basic statistical operations, we were able to infer indicators for bivalve (oyster and blue mussel) beds. Those indicators nicely agree with results from in-situ campaigns, and they are already being included in an existing classification scheme for Wadden Sea surface types.

P. Schaeffer (1), M-I. Pujol (1), Y. Faugere (1), A. Guillot (2), & N. Picot (2)

 

(1) CLS/Space Oceanography Div., Ramonville, France

(2) CNES, Toulouse, France.

 

A new mean sea surface (MSS) has been determined focusing on the accuracy along Exact Repeat Mission profiles, in order to offer a centimetric level of precision for the forthcoming Sentinel-3 and Jason-3 altimetric missions.

Particular attention was paid to the homogeneity of the ocean content of this MSS, and specific processings were also carried out on data from geodetic missions. For instance, CryoSat-2, Jason-1 GM, and ERS-1 GM data were corrected from oceanic variability using results of optimal analysis of sea level anomalies (SLA).

This MSS CNES_CLS 2015 is referenced to the 20 years of T/P/Jason-1/Jason-2 mean profile (1993-2012).

The MSS CNES_CLS 2015 has been determined using a local least square collocation technique which provides an estimation of calibrated error. Our technique takes into account of altimetric noise, ocean variability noise and also along-track biases which are determined independently for each observation. Moreover, variable cross-covariance models are fitted locally for a more precise determination of the shortest wavelengths lower than 50 km. A specific data processings were used near the coast, the aim has been to preserve better continuity of the MSS accuracy from the open ocean to the coasts.

The CNES_CLS 2015 MSS can thus be used for cal/val activities for Jason-3 and Sentinel-3, and especially for computation of SLA for DUACS project. Goal was also a first step to access to shortest wavelengths in preparation of very high resolution topographic surface for the upcoming mission such as SWOT.  

Abstract Text:

 

Abstract Title: Sentinel-1A SAR Observed Doppler vs Geometrical Doppler, current status  

 

Abstract Text:

 

Sentinel-1A was launched in April 2014 in order to provide radar vision for Europe's Copernicus programme.  Sentinel-1A carries a 12 m-long advanced synthetic aperture radar (SAR), working at C-band.

The Doppler shift of the return signal can be exploited to monitor ocean surface velocities, as discovered and utilised using Envisat C-band SAR data.

The difference between the observed Doppler and the expected Doppler is attributed to ocean surface velocities. Sentinel-1 has implemented the Zero-Doppler attitude steering of the platform, so the expected Doppler for the received return signal would be 0. – under the assumption of no surface velocity –

The actual attitude of the Sentinel platform will however induce a small Doppler signal. For Sentinel-1 the expected Doppler based on attitude data, the observed Doppler of the C-band signal, and the expected surface velocities, currently do not match.

Therefore this data cannot yet be exploited to monitor these ocean surface velocities.

 

The reason for this difference is not yet understood, but recent analysis by Ifremer shows that there is an apparent relation between this data.

The paper will present the status at the time of LP16 of the investigation and possible solution to the observed difference between expected and measured Doppler of Sentinel-1 C-band SAR data.

  

 

Keywords: Ocean Circulation, Ocean Currents retrieval and applications

 

Satellite and data used: Sentinel-1

As it is well known, one of the most important issues in coastal altimetry is the effect of land contamination. As the altimeter approaches the coast, radar echoes from the altimeter are likely to impinge on the coastline. In that case, if the land surface scattering is much higher than the one of the ocean, such as in the case of wet and flat terrain, a strong coherent component from these areas can highly distort the ocean-like waveform shape on which retrackers are designed to work. Hence, the combination of the effects of several scatters with different backscattering intensities on the surface can lead to random waveform variations, which inhibits conventional retracking techniques from retrieving accurate geophysical parameters.    

In SAR mode altimetry, the shrinking of the radar footprint allows to reduce the land contamination effect, expecting to obtain measurements of SSH, SWH, and WS close to the shore (several hundred meters). Hence, previous studies have already shown the capabilities of a delay-Doppler Altimeter to obtain useful radar echoes at a distance about 300 m from the coast [1].

However, in a delay-Doppler altimeter the resolution improvement occurs only on the along-track direction, while the across-track direction remains pulse-limited. Thus, the DDA’s response in coastal regions depends on the relative orientation between the coastline and the spacecraft orbital plane.

Here, is proposed an algorithm to disregard the range bins of the SAR waveform, which are likely to be affected by land contamination. This is achieved by geo-locating the delay and Doppler pairs of the SAR altimeter stack, obtaining a 2D map of the surface backscattering response. From it, it is possible to identify the SAR waveform range bins that are potentially affected by land contamination, and then a mask is applied above these points. Hence, this algorithm allows to determine which of the range gates of the SAR waveform, can be used reliably in the retracking without land contamination.  

Once the contaminated bins have been identified and masked, the multilook is applied to the non-corrupted data, and then the SAR retracking algorithm is applied. Previous work done for Cryosat-2 will be taken into account and adapted to Sentinel-3, where some aspects as the validity of this approach (as the altimeter approaches the coast, more waveform range gates are susceptible to have land contamination, and therefore to be discarded for the later retracking process), will be considered.

 Preliminary results obtained, and main advances in this topic will be presented at the conference.

This work is carried out as part of the SCOOP (SAR altimetry Coastal and Open Ocean Performance) project, supported by the European Space Agency within the SEOM (Scientific Exploitation of Operational Missions) program.

The CCI Ocean Colour dataset (version 2) is based on MERIS, MODIS-A and SeaWifS observations, which are merged into a consistent water leaving reflectance time series. This time series covers a 16-year period between 1997 and 2013 on a global scale. On a 4km grid, daily, 8-day and monthly averages are available. A wide range of ocean colour products is delivered within the project, which includesEssential Climate Variables(ECVs) likechlorophyll concentration and remote sensing reflectance and more complex products like a class membership in spectrally defined water types.

The complexity of the North Sea water bodies is represented in its optical properties as well as in the hydrographically influenced properties. While e.g. the definitions of subdivisions in the aera by the International Council for the Exploration of the Sea (ICES) originate mainly in the latter, these subdivisions are also found in ecosystem models like ERSEM.

Starting from optically derived data alone, the time series of chlorophyll concentration and the water class membership from OC-CCI are subjected to a cluster analysis, in order to identify areas of similar climatological behaviour. In addition, sea surface temperature from SST-CCI is introduced as a secondary variable in the analysis. In preliminary results, the hierachical clustering of chlorophyll time series suggests a spatial pattern which has a strong resemblance to the ICES subdivisions of the North Sea area.  

Sea oil pollution is a matter of great concern since it affects both the environment and human health. The performance of the radar SAR make it a useful tool in the control and surveillance of oil spill pollution on sea surfaces.

SAR oil slick detection is physically possible since an oil slick, damping the short resonant Bragg waves, generates a low backscattering return, that is a dark area in SAR images. However, other physical phenomena, such as biogenic films, low winds etc. can generate low scattering areas (look-alikes) in SAR images. Thus, oil slick detection and classification techniques are still an open issue.

Recent studies demonstrate that SAR polarimetry is able to provide additional information to environmental applications, in this context different approaches based on the polarimetric features are used for oil spill observation.

In this study, a dual-polarimetric feature ‘the modulus of the complex correlation coefficient between the co-polarized channels ’ is used to discriminate between oil spill and lookalikes on sea surfaces using polarimetric SAR images.

The co-polarized channels are often HH and VV. The correlation coefficient is complex and is computed as the average of the product between the complex amplitude of the HH channel and the conjugate of the complex amplitude of the VV channel. It is normalized by the square root of the product of the powers in the HH and VV channels.

If the magnitude of the correlation coefficient is unity, the received signals from the two channels are linearly related. And if it is less than one, it means that the backscattering at HH and VV are not directly related.

Calculated values of the modulus of the complex correlation between the co-polarized channels vary between 0 and 1 and we can then distinguish between slick free sea surface and slick sea surface.

Experiment results show that the modulus of the complex correlation between the co-polarized channels, can be successfully exploited for both observing sea oil slicks and distinguishing them from weak-damping properties look-alikes.

Results of multiyear research on determining zones of persistent ecological risk in the Baltic Sea are presented. They are areas most affected by oil pollution, harmful algae bloom and raised concentration of suspended particulate matter.

The research was based on satellite remote sensing data obtained over the Baltic Sea in the period from 2004 to 2015. Data from radar sensors ERS-2 SAR, Envisat ASAR, Sentinel-1, Radarsat 1,2, as well as visible and infrared data from Envisat  Meris, Landsat-5 TM, Landsat-7 ETM+, Landsat-8 OLI,  and Terra/Aqua Modis were used.

Analysis of radar data indicated areas worst affected by ship spills of bilge waters. The area sizes were estimated.  Maximal pollution area sizes were found along the main ship routes to the southeast of Gotland Island.  Increased oil pollution was also observed along ship routes to the north of Hel Peninsula, near ports of Klaipēda, Liepāja, Baltijsk, at the mouth of the Gulf of Finland, in the Neva Bay. Our maps of oil pollution for different years were compared with data of other researchers.  In recent years, a significant decrease of anthropogenic pollution associated with ship spills of oil and oil products has been observed.  Apparently, this is the result of dedicated satellite monitoring campaigns conducted by various international organizations, for instance, in the framework of the CleanSeaNet initiative.

Meanwhile, eutrophication has recently become a very important problem worldwide. It is especially true for the Baltic Sea as blooming cyanobacteria (blue-green algae) can be observed almost in any part of it. As a rule, intense algae bloom is detected using satellite data in visible range and chlorophyll-a maps. We suggest an indirect method based on satellite radar data. In the course of multiyear satellite monitoring of the sea surface we have established a relation between phytoplankton bloom intensity and duration and manifestation of ship wakes of a particular type in radar images. These are long-lasting wakes behind moving ships in the shape of bright bands of backscattered signal with lengths of tens and even hundreds of kilometers. Such wakes can be viewed as indicators of phytoplankton bloom area locations and blooming duration.

Areas characterized by increased concentration of suspended particulate matter were revealed using maps compiled from Envisat  MERIS data and Landsat  color composites.  Among the most remarkable ones are certain regions of the Gulf of Finland and various river plume zones, first of all, those of the Vistula and Neva Rivers.

The work was accomplished with partial financial support by the Russian Science Foundation, grant # 14-17-00555. The Envisat ASAR and MERIS data were obtained in the framework of European Space Agency projects.

Physical, chemical and biological characteristics of seawaters are primary descriptors to assess  the spatial and temporal dimensions  of  ecopotential productivity performances in terms of harvest from marine ecosystems. Among these characteristics, sea surface temperature (SST) can be considered the key descriptor as an Essential Variable (EV), able to characterize both fish vitality and climate changes and it is one of the most successful oceanographic measures from satellite data.

We developed a method to identify sites scenario with feasible aquaculture fish growth conditions in the Mediterranean Sea and, among these sites, to estimate the potential fish harvesting suitability.

For fish growth condition and potential fish harvesting, in addition to SST from MODIS  Terra satellite and in-situ measurements, we consider the spatial and temporal distribution of water quality parameters, such as Chl-a, Total Suspended Matter (TSM) and Colored Dissolved Organic Matter (CDOM) using full resolution products, acquired by MERIS sensor onboard ENVISAT satellite, and modeled currents collected from Copernicus Marine Environment Monitoring Service (CMEMS).

The final aim is to built a toolbox oriented in the direction of EO downstreaming services. This will provide support to the management of marine resources and the maritime uses planning. In fact, the scenario generated, as new added value EO products, can be used to describe current marine status or simulate the future ones by including forecasting of selected variables. 

OSTIA is the Met Office Operational SST (sea surface temperature) and Ice Analysis system. It uses an OI-type assimilation method to produce a daily, global estimate of foundation (pre-dawn, or free of diurnal variability) SST with a 1/20° grid resolution. A combination of satellite and in situ SST observations are used for the analysis. The product is freely available through the Copernicus marine environment monitoring service (http://marine.copernicus.eu).
 
The effect on the current SST analysis of including further satellite SST products from GOES-W, Himawari 7/8, VIIRS and AMSR2 instruments, and AVHRR and IASI on MetOp-B, will be demonstrated. A combination of methods are used for validation, including comparison to independent near-surface Argo observations and use of observation-minus-background statistics from the analysis.

Thanks to its current accuracy and maturity, altimetry is considered as a fully operational observing system dedicated to scientific and operational applications. In order to access the targeted ocean signal, altimeter measurements are corrected for several geophysical parameters among which the ocean tide correction is one of the most critical. The accuracy of tidal models has been much improved during the last 20 years. Still, significant errors remain mainly in shelf seas and in polar regions.

 

A new global tidal model, FES 2014, has been developed taking advantage of long altimeter time series, new altimeter standards (instrumental and geophysical corrections, orbits), improved modelling and data assimilation techniques. The grid resolution has been increased in areas of interest like shallow waters and on the slope of the continental shelves. The error of the pure hydrodynamic model has been divided by a factor of 2 compared to previous version FE2012. Moreover a larger assimilation dataset has been considered, including tidal gauges.

 

FES2014 performances have been compared to other global tidal models available (GOT4v10, DTU10, TPXO8) and using both tidal gauges and altimeter measurements. They show significant improvement particularly in shallow waters and in some parts of the Arctic region. Final validation results of FES2014 are shown here. We also present the perspectives of improvement of the global tidal model at a global and regional scale.

Within the SEOM Sentinel 3 for Science Ocean Colour program, a robust, complete, user friendly radiative products, including daily PAR from MERIS and OLCI will be produced.

The basis of the project is a daily PAR algorithm that has is being operated at NASA/OBPG and whose principle is described in Frouin et al., 2003, Frouin and Murakami, 2007, Frouin et al. (2012), and Frouin and McPherson (2013). This algorithm has been chosen because it is robust and sensor generic.

According to the Statement of Work Requirements and after synthesizing users wishes gathered through a on-line survey, several innovations will be implemented:

- A climatological correction for diurnal variability of cloudiness

- Daily PAR uncertainty per pixel

- Secondary experimental products:

PAR just below ocean surface

Spherical (Scalar) PAR

Spectral PAR by bins of 10 nm.

Absorbed fraction of PAR by live algae

An example of application of these products to the Primary Production modelling is foreseen. A Demonstration Data Set (DDS) covering the full year 2011 from MERIS will be distributed in the 2^nd quarter of 2016 for evaluation by the users.

Users are encouraged to use this DDS and to interact with the project (user's requirements and product evaluation) with a milestone event being a specific session of the CLIO Workshop (Colour and Light in the Ocean) that will be held in ESA/ESRIN on 5-9 of September 2016.

The project and the DDS will be presented during the symposium.

Ocean color sensors have been used since the Coastal Zone Color Scanner Experiment in the late 70s to retrieve chlorophyll-a concentration in the global oceans. To improve our understanding of marine phytoplankton's role in the global marine ecosystem and biogeochemical cycles, interest has grown in retrieving additional information about the phytoplankton communities such as the size structure, the taxonomic composition, or the functional type of a phytoplankton community. A definition of phytoplankton functional type (PFT) refers to a group of organisms having common functions in the biogeochemical cycle such as calcification, silicification, or nitrogen fixation.

One approach to detect PFTs from space is to use hyperspectral satellite data. Specific absorption spectra of different PFTs are used within the Phytoplankton Differential Optical Absorption Spectroscopy (PhytoDOAS) method to distinguish among phytoplankton functional types. So far, this technique has been applied to hyperspectral data from the SCIAMACHY (Scanning Imaging Absorption Spectrometer for Atmospheric Cartography) sensor to retrieve the chlorophyll-a concentration of important phytoplankton functional types such as cyanobacteria, diatoms, and coccolithophores (Bracher et al. 2009, Sadeghi et al. 2012). Here, we will present the global quantitative distribution of different PFTs derived with the PhytoDOAS method from data of other hyperspectral satellite sensors such as OMI (Ozone Monitoring Instrument) and GOME-2 (Global Ozone Monitoring Experiment-2). A comparison of the retrieval products from years with overlapping operation of the sensors will be shown. In addition, new global phytoplankton distribution products for the time period not covered by the SCIAMACHY sensor, i.e. after 2012, will be presented.

Bracher A, Dinter, T., Vountas, M., Burrows, J. P. , Röttgers, R., Peeken, I. (2009) Quantitative observation of cyanobacteria and diatoms from space using PhytoDOAS on SCIAMACHY data. Biogeosciences, 6, 751–764.

Sadeghi, A., Dinter, T., Vountas, M., Taylor, B. B., Altenburg-Soppa, M., Peeken, I., Bracher, A. (2012) Improvement to the PhytoDOAS method for identification of coccolithophores using hyper-spectral satellite data. Ocean Science, 8, 1055–1070.

Polymer is an atmospheric correction algorithm which uses a spectral matching technique in which the atmospheric and oceanic signals are retrieved simultaneously. It has been initially developed to process MERIS data, in particular in presence of intense sun glint; it is a generic algorithm which does not require the presence of particular spectral bands; it has been adapted to process other sensors, like MODIS, SeaWiFS and VIIRS. Its application to Sentinel-3 OLCI is planned, and should allow maximizing the useful ocean coverage of this mission by retrieving ocean colour in the sun glint pattern.

It was initially based on a case 1 only water reflectance model, but has been recently extended to case 2 by using a more generic water reflectance model valid for both case 1 and case 2. Improved results and coverage in case 2 waters is achieved without affecting the performance in case 1 waters, nor the robustness to the sun glint contamination. The atmospheric model has also been adjusted to improve the results at high latitudes. These modifications will be presented, and the evolution of the products will be discussed. This new version of Polymer allows a wider range of applications, with significant improvements in presence of case 2 waters or at high latitudes.

Following the work of Morel and Prieur (1977) a bipartite division formed in the field of ocean optics, where ocean waters were often defined as ‘Case I’ (optically dominated by chlorophyll and covarying material) and ‘Case II’ (where optical properties can vary independently of chlorophyll).  This has allowed the tailoring of algorithms for the retrieval of important bio-optical products from satellite-based sensors.  Unfortunately, it has also led to the creation of an artificial barrier in ocean-colour data processing.  The applicability of certain algorithms to particular water types means that they are of limited use at the global scale.  Use of multiple tailored algorithms risk introducing hard, un-natural boundaries or gradients into the products.  To process data from any region, one requires a method for the smooth blending of algorithms based on their applicability to the location under study.  A fuzzy classification of ocean reflectance spectra, developed from the work of Moore et al. (2001),  is performed as part of the ESA Ocean Colour-Climate Change Initiative processing chain and has been used to assign uncertainties to the products.  Fuzzy classification techniques have also been used to weight data across lakes and coastal waters (Moore et al 2014).  In this study we discuss the possibility of using a new set of optical water classes, representing all water types from the central gyres to coastal regions) to blend algorithms while preventing hard boundaries in the processing chain and resultant products.  We will show that the existing sets of open-ocean fuzzy water classes are insufficient at the global scale.  As such, we have created a more representative set of optical classes based on fuzzy classification . We estimate the performance of different algorithms for each of the water classes.  This new set of classes is created from millions of EO-derive ocean spectra and gives an improved class membership distribution, especially in the central gyre regions.

Satellite wave data play an important role in improving the sea state forecasting. This is a crucial task in order to build a coupled ocean/waves systems as requested in the Copernicus programme. The Sentinel-1 (S-1) and 3 (S-3) of the ESA satellite missions are part of this outlook with important challenges. The Sentinel-3 wave data will be fully retrieved by SAR mode altimetry technique. Furthermore the level 2 wave spectra provided by Sentinel-1A has an improved resolution in direction and frequency to better describe several long wave systems. The goal of this work is to assimilate the significant wave height from S-3 and directional wave spectra from S-1 conjointly in the wave model MFWAM. This work is part of the calibration/validation phase of these two missions and prepares the use of these wave data in the operational wave forecasting system of Météo-France.

A long period of wave are carried out to study the complementary use between S-3 altimeter data and the level-2 S-1 wave spectra. The results of the assimilation of the test runs are validated with independent wave data such as those provided by buoys and altimeters. The persistency of the combined assimilation in the forecast period will be examined. Significant wave heights from the model MFWAM and S-3 will be compared carefully for detecting biases induced by the presence of long swell. Cases of storms generating high sea state as the ones in indian ocean cyclone season will be discussed in this study.

The Ocean-Colour CCI(OC_CCI)project as part of ESA's Climate Change Initiativehas compiled a coherent timeseries for climate studies ofmarine Essential Climate Variables(ECV)like chlorophyll concentration from all available ocean colour sensors between 1997 and 2014.

A lot of effort was put in the development of a consistent unbiased dataset not affected by a different composition of available sensors or changing environmental conditions. The generation of these data is under continuous improvement. The OC_CCI data-package is available as a global 4km gridded daily dataset. In addition an 8 day and an monthly composite is available.

In this work we study the seasonality and trends on the chlorophyll signal as part of the OC_CCI data for selected regions and time composites. The focal point of the study is the dependence of these findings from imperfect conditions like changing effects of flagging due to quality, cloud coverages, or number of available sensors. In addition for a shorter timespan( 2007-2009) a comparison to Cloud_CCI is possible. The covariance of all parameters is presented.

The ocean's mean dynamic topography as the difference between the mean sea surface and the geoid reflects many characteristics of the general ocean circulation. Consequently, it provides valuable information for evaluating or tuning ocean circulation models. Along track altimetry data, measuring the sea surface height and satellite-based gravity field information are used to estimate a local model for the mean dynamic topography and the corresponding uncertainty information. For that purpose, an integrated approach was developed to estimate the mean dynamic topography and the gravity field in a rigorous adjustment. Complementary observation types, in this study altimetric sea surface height measurements and GOCE gravity field observables are combined to estimate the gravity field, a local mean dynamic topography model and its full covariance matrix. The approach includes the modeling of uncertainties from the raw observations (along track sea surface heights) to the parameterization of the target quantities. As the full covariance matrix of the model is available, the resulting mean dynamic topography model requires no smoothing for a further use, as long as its full covariance matrix is accounted for. Whereas the mean dynamic topography is parameterized as a set finite element base functions (linear, cubic or even higher order elements) over a predefined local grid, the gravity field is traditionally parameterized as a spherical harmonic series. Along-track altimetry observations of different altimetry missions (like Jason-1/2, Envisat, ERS2, Cryosat2, ...) are combined with GOCE observations in terms of full normal equations which follow the time-wise method. As they only depend on the observations from GOCE mission, it is independent gravity field information which supports the separation of the sea surface heights into geoid and mean dynamic topography. The implementation was ported to a high-performance computing environment which allows for more complex setups. With the current implementation, hundreds of thousands unknown gravity field and finite element parameters can be estimated from millions of observations, although the linear system to be solved is highly overdetermined and dense. Within this study we present an analysis of the altimetry residuals over the North Atlantic Ocean resulting from the joint least-squares adjustment. Different set-ups for the base functions, for instance for the finite elements but also for the resolution of the gravity field are analyzed. In addition different error models for the along track altimetry observations (covariance functions and decorrelation filters) and their impact on the target parameters, their accuracy estimates as well as the altimetry residuals are assessed.

The Group on Earth Observations (GEO) seeks to optimize the use of Earth Observation data for the benefit of society. In a recent communiqué (2015), The G7 Science Ministers resolved “We intend to continue to work together through the intergovernmental Group on Earth Observations (GEO) to enable policy makers to better address [ … ] environmental, health, and socio-economic challenges.” For the oceans, the overarching GEO initiative is Oceans and Society: Blue Planet, which is led by the Partnership for Observation of the Global Ocean (POGO). Blue Planet envisions an informed society that recognises the oceans’ crucial role in Earth's life-support system and is committed to stewardship of the oceans for a healthy, safe and prosperous future for all.

Blue Planet’s Mission is to advance and exploit synergies among the many observational programmes devoted to ocean, coastal and inland waters; to improve engagement with a variety of users for enhancing the timeliness, quality and range of services delivered; and to raise awareness of the societal benefits of ocean observations at the public and policy levels. Blue Planet works by adding value to data produced by existing observational programmes. Currently, there are nine components to Blue Planet. They are:  Developing capacity and societal awareness; Sustained ocean observations; Data access and visualization; Ocean forecasting; Healthy ecosystems and food security; Services for the coastal zone; Ocean climate and carbon; Integrated maritime services; and User engagement. Here, we present an overview of the scope and purpose of Blue Planet and discuss its plans for the future.

Oil spills detection and mapping in the Adriatic Sea have been done using the RADARSAT-2 and ENVISAT synthetic aperture radar (SAR) images, acquired in different times from the ESA archive (within the project No. 19234 (P.I M. Oluić)), and operational monitoring made in the framework of a pilot project in 2011. The SAR images were processed with routine ESA software. At the final stage, processed SAR images were introduced to on-line GeoMixer application and through this software oil spills were automatically registered, vectorised, put on the map projection and web-GIS integrated. Oil spills were detected, identified and analyzed using both SAR image characteristics and contextual information such as: ship proximity to an oil target, linear feature along the ship direction, possible diffusion of the oil patch, location of ship lines, wind speed, location of known sources of oil pollution, etc.

So far, there was no serious accident related to oil pollution in the Adriatic Sea, however a high number of oil slicks/spills have been detected on SAR images, occurring practically on a daily basis. The predominant sources of slicks are considered to be deliberate illegal releases of oil from tankers and cargo ships in the open sea. Some spills were from different oily products, caused by tank washing or routine ship operations of transport or fisheries ships mainly in the coastal zone. By visual and interactive analysis we have found that large oil spills prevailed in the narrow central part of the Adriatic Sea, within the main ship corridor. Their predominant shapes were linear, and the sizes ranged from 10-110 km², but some other shapes and sizes were also present due to oil spills from a variety of other sources. It is concluded that analysis of large archive of SAR images in web-GIS application is very useful, first, for identification of oil spill sources and second, for analyzing their spatial and temporal distribution in the particular marine basin. In the paper we also demonstrate existing pollution level of the Adriatic Sea and discuss potential damage for the Croatian coastal environment.

 

 

The NW Mediterranean Sea has distinct ocean dynamics, with relatively strong boundary currents and moderate eddy energy, but with variations at small Rossby radius which are difficult to detect with conventional satellite altimetry. In this study, we analyse alongtrack altimetry from 3 different missions and measurement systems: Jason-2 Ku-band data, SARAL Ka-band data, and Cryosat-2 SAR data. We apply spectra analysis over short, 300-500 km sections in the NW Mediterranean Sea and estimate the mesoscale capabilities of the ocean dynamics here. The « white noise » background error levels differ between the 3 missions and for different seasons. This limits our ability to detect alongtrack mesoscale structures, and we quantify the typical length scales we can detect with each mission, depending on their seasonal signal to noise levels.

In the second part of the study, we also look at the observational capabilities of gliders and HFradar to study the geostrophic variability in the NW Mediterranean Sea. The high-frequency noise of the gliders limits us to observing signals greater than the Rossby radius (10-15 km). Only a few colocated glider and alongtrack altimetry data are available and these provide good results for the larger mesoscale structures. Validation of smaller-scale ocean variations is more difficult: with weaker amplitudes they are close to the noise levels for both observation types, and their rapid evolution means only a few days of short glider tracks are usefully colocated. The final data set we use is HFradar, which provides good synoptic 2D data coverage which can be easily colocated with the altimeter passes. In calm weather, the geostrophic currents associated with boundary currents and eddies are well detected and we can estimate their spatial and temporal characteristics. However, in periods of strong wind forcing (Mistral, Tremontagne), the ageostrophic component of the surface currents is dominant and altimetric noise is higher, leading to poorer intercomparisons.

The European Space Agency's Soil Moisture and Ocean Salinity (SMOS)

mission is the first satellite mission dedicated to providing global

maps of sea surface salinity. Launched in November 2009, SMOS has

provided over six years fully polarimetric brightness temperatures at

L-band over a swath about 1000 km wide and over incidence angles

ranging from nadir to about 60 deg.

Since the launch of the SMOS satellite, two additional L-band

radiometers have been launched: the NASA Aquarius radiometer (on board

the Aquarius/SAC-D satellite), which provided brightness temperatures

from August 2011 until June 2015, and the SMAP (Soil Moisture Active

Passive) radiometer, which has provided fully polarimetric brightness

temperatures since April 2015. It is important to understand how to

combine the data from these three missions to produce global salinity

free from biases related to both the instruments and the retrieval

algorithms.

The three radiometers utilise distinct measurement strategies. The

instrument onboard the SMOS satellite, MIRAS, consists of an array of

69 low-directivity antennas arranged in a Y-shape and a correlation

unit which measures the cross-correlation between signals from pairs

of these antennas. On-ground processing transforms these

cross-correlations into brightness temperaturs images with an

effective footprint size on the order of 50 km. By contrast, Aquarius

is a conventional nonrotating real-aperature radiometer consisting of

three feedhorns and a single reflector that together provide

brightness temperatures along three distinct ground tracks at earth

incidence angles of approximately 29, 38 and 46 deg. With a seven day

repeat cycle, Aquarius provides complete global coverage every seven

days with a spatial resolution on the order of 150 km. The SMAP

radiometer is a conically scanning instrument that provides brightness

temperatures at an earth incidence angle of 40 deg over a swath about

1000 km across. This instrument has a repeat cycle of 8 days and

provides global coverage every three days.

Inferring sea surface salinity from L-band brightness temperatures

remains a challenging problem regardless of the instrument. Salinity

is retrieved via the relation between specular emission and salinity

concentration. The sensitivity of the specular emission at L-band to

changes in salinity depends somewhat upon polarization and sea surface

temperature but, in tropical latitudes (where the sensitivty is

largest), is on the order of +2 K in the first Stokes parameter per

unit decrease of salinity on the pratical salinity scale. Over the

global oceans the surface salinity varies between about 32 and 38, so

the range of L-band brightness temperatures over the open ocean

associated with salinity variations is only several kelvin. Sources of

brightness other than specular emission, such as excess emission from the

wind-roughened ocean surface, sun glint and galactic glint, can

contribute several kelvin to the total brightness, depending

upon polarization, viewing geometry and surface roughness. Corrections or

flagging strategies must be introduced to minimize the impact of these

sources on salinity retrieval. This paper will examine impacts of

these extraneous sources of radiation on the measurements from the

three radiometers and will propose methods to minimize their impacts

on retrieved salinity.

Measurement biases further complicate salinity retrieval, and the

characteristics of these biases depend strongly upon the instrument.

One of the most important biases is the so-called 'land-sea

contamination' bias in the brightness temperature images obtained from

MIRAS within about 1000 km from large land and ice masses. The bias

patterns in the images depend strongly upon the distribution of land,

ice and water over the entire field of view, and this paper will

briefly summarize the approach taken to correct this 'land-sea

contamination' bias and will examine the consistency of the corrected

SMOS salinity maps with those obtained from Aquarius and SMAP in the

vicinity (i.e., within 1000 km) of large land masses.

The discipline of satellite altimetry is mature end largely developed for open oceans; in contrast, it remains less exploited for coastal areas form several problems of satellite resolution and difficulties in the corrections and issues of land contamination. In the last couple of years, significant progress has done to overcome these problems and extending the capabilities of current and future altimeters to the coastal zone, with the aim to integrate the altimeter-derived measurements of sea level, wind speed and significant wave height into coastal ocean observing systems. Nevertheless, several zones, such as the case of the English Channel, remain a tricky area.  Using the future SWOT altimeter, such areas will be observed with a 120-km swath in high resolution thus leading to a huge improvement in such measurements, needed for the modeling of coastal sea levels and energy patterns.

The potential use of SWOT satellite in the English Channel was investigated by Turki et al. (2015a) with the aim of determining the temporal variability of sampled sea level measurements and its relation with Spatial changes in sea level were also exanimated by the use of mathematical models simulating the energy conditions from the ocean to coastal zones with different resolutions (Turki et al. 2014; 2015b).

Such results, developed in the framework of the future SWOT mission, have been explored in the present research in order to investigate the extent to which SWOT measurements represent time-scale energy fluctuations at different spatial scales in French coastal zones of the English Channel and Atlantic Sea. In particular, we focus on the time variability of storm surges and the significant wave height basing on different approaches of inference statistics and wavelets. Results have shown that SWOT is able to reproduce the most of extreme storm surges and spectral wave data in coastal areas with a mean variance of 70% and 85%, respectively.

Regarding the spatial variability, tests of different scenario modeling have been carried out (stormy/moderate events, neap/spring tide cycle and low/high tides, including storm surges) using several grid resolution. Simulated high resolution mapping has suggested that SWOT resolution of 250-m is able to capture approximately 80% of coastal variability. The potential applications of SWOT altimeter is strongly related to the number of overpasses per repeat cycle and the energy conditions including the tidal regime. The main finding of this research clearly shows the utility of SWOT satellite altimetry in observing and understanding the variability of energy patterns in coastal zones, complementing tide-gauge and buoy observations for the validation and improvement of coastal models.

The East Coast of North America, from 30 to 60^oN was analyzed from1992 to 2014. Such región is under the influence of Labrador Current (40-60ºN), which flows southwards from Labrador Sea, and has an influx from Artic ice, and the northernmost portion of Gulf Current before the latter flows towards northeastern (30-40ºN). The variables used were Extended Reconstruction of Sea Surface Temperature (SST), Sea Level Anomalies (SLA) and Artic Sea Ice Extent. For SST and SLA monthly anomalies were computed, and a low-pass LAnczos filter, with a 12-month window was used to eliminate intra-annual variation. The results are shown in a Hovmöller diagram of time against latitude, indicating a rise in SST from 45^oN in 1997 (0.5ºC), extending to 35^oN a couple of years later. During the 2000’s 1oC anomalies are frequent, and from 2011 there is a steady climbing in SST anomalies (1.5-2.5ºC) until the end of the study. SLA showed a different pattern than SST, with a dominance of negative anomalies until 2006 (except 1996, 1997 to 2001 and 2003, when some positive anomalies can be seen). The anomalies from 2006 to 2008 range from 1 to 2 centimeters and usually don’t go further south from 45ºN. From 2009, positive anomalies can be found from 40-60oN, reaching up to 6-8 cm in some áreas. The SLA have a relation with the decrease or Polar Sea ice extent, which showed a significant negative trend, with the lowest ice extent values in 2007 and 2012 (2.8 and 2.15 million square km respectively). Correlations between the first principal component of SST, SLA and Sea Ice Extent indicated than the observed changes in sea level in the región are more influenced more by eustatic changes.

Aquaculture industry lacks near-real time support for their daily management decisions and for monitoring environmental parameters that may impact its activities. For that purpose, satellite remote sensing can be an effective tool as it can provide the necessary time and spatial resolution at reasonable costs. The information is however scattered, from multiple sources, and not easily accessible for the regular aquaculture manager. AQUA-USERS is a user-driven project, that intends to gather, process and analyse the satellite and in situ data available, in order to provide relevant and easily accessible information that can support aquaculture activities.

Aquaculture management actions (e.g. lowering cages, harvest, feed) change according to type of aquaculture (e.g. mussel, fish, seaweed), but, when considering monitoring purposes, there is a set of parameters that is common to the needs of the majority of users. All users are interested in having information on weather and sea state conditions, as well as water temperature and chlorophyll concentration data. The latter is referred as a relevant parameter to monitor as it can be used as a biomass proxy (i.e. food availability for mussels) or as an indicator of water quality.

Ocean Colour constitutes therefore a relevant tool for aquaculture management. The construction of time series at multiple spatial scales of phytoplankton biomass indexed as chlorophyll-a concentrations can be used to assess a variety of ecological indicators. For real time management, however, temperature and chlorophyll-a data can also indicate gradients and fluctuations in the environmental conditions for effective evaluation and decision making. It is therefore crucial to evaluate the natural variability of these parameters.

For that purpose, 10 years of satellite data were processed for all the users sites and regional-specific time-series were determined. Threshold values were established based on deviations from these baselines and should provide alert conditions to site managers. An analysis of algal bloom phenology was also performed based on historical satellite data for all the targeted regions. Here, we present the Portuguese case study.

The Portuguese study site (in Sagres, Southern Portugal) is characterized by a typical spring bloom, which can last for 60-90 days. Chlorophyll concentrations can occasionally reach 10 mg.m^-3 with a yearly mean of 1.14 mg.m^-3. Strong correlations between coastal sea surface temperature and chlorophyll concentrations were observed using satellite data, showing the overall seasonal pattern. In summer, negative correlations between in situ Chl and temperature highlight active upwelling events, with a temperature decrease of at least 2-3°C in a 1-week period. Diatoms dominate most phytoplankton blooms occurring in response to upwelling events.

This work describes the multi-temporal analysis of Landsat-8-derived products performed to investigate the suspended sediment dynamics in the Po River prodelta and the adjacent coastal zone (Northern Adriatic Sea). The study area represents a complex environment, where the five major distributaries of the Po River contribute to the freshwater input in the northern Adriatic Sea, exhibiting different and variable partitioning of discharge and sediment load. This coastal system is dominated by riverine inputs and hydrodynamic forcing, and their interaction influences the physical and biogeochemical processes of the whole basin. Understanding the spatial and temporal variability of the Po river plume is therefore of primary importance for the study of northern Adriatic Sea hydrology. The application of remote sensing techniques can provide new perspectives for the submesoscale analysis of the Po River plume and the pattern of suspended sediments, through satellite-derived products, such as turbidity, total suspended matter (TSM) concentration and diffuse attenuation coefficient. Landsat-8 OLI imagery, with their finer spatial resolution and high quality of radiometric resolution, are suitable to investigate mid to small scale turbulent structures of the buoyant flow at the surface. The analysis was performed on the Landsat-8 OLI data, available in the period 2013-2015. The images were at first converted into water-leaving radiance reflectance (ρw, dimensionless) with ACOLITE (Vanhellemont and Ruddick, 2015). The ρw data were then converted into turbidity [FNU] and TSM [mg l^-1] following Dogliotti et al., (2015) and Nechad et al., (2010) respectively. Starting from these outputs, two temporal hypercubes for TSM and turbidity maps were built. The multi-temporal analysis was performed at selected target stations along the river delta and prodelta area, which are supposed to be representative of the prevailing hydrologic and hydrodynamic conditions.

The univariate and multivariate statistical analysis was computed in order to assess the correlation between all targets. In particular, the temporal variation of a reference point, located some kilometers upstream from the delta, was compared to the trend of the coastal target stations to assess the correlations. In the presence of a positive correlation, we suppose river discharge is the prevailing forcing; if no correlation is found, we assume that other processes, as sediment resuspension and transport due to wind and wave action, are acting. Geostatistical analysis (i.e., spatial autocorrelation) was performed to determine the spatial dependency of the two datasets in the whole temporal range and for each date. With the aid of in situ data, a qualitative interpretation of the factors controlling these patterns through time was proposed (e.g. interaction among hydro-meteorological factors, coastal currents and prodelta morphology).

The Landsat-8 with a 30-m spatial resolution has shown its potentiality for synoptic observations of TSM and turbidity patterns at sub-mesoscale. A drawback of Landsat-8 imagery can be the temporal resolution (16 days track repeat), which is not sufficient for multi-temporal analysis in such a complex environment. However, when also Sentinel-2 platforms will be both operational, the availability of more frequent images will offer reliable advantages to observe and understand intricate processes operating on different space-time scales.  Indeed the combination of these ESA and NASA missions will improve the temporal analysis reducing the revisit time as well as detecting trends in order to evaluate the effects of different forcings on the suspended sediment dynamics in the river prodelta.

References

Dogliotti, A., Ruddick, K., Nechad, B., Doxaran, D., Knaeps, E., 2015. A single algorithm to retrieve turbidity from remotely-sensed data in all coastal and estuarine waters. Remote Sens. Environ. 156, 157–168.

Nechad, B., Ruddick, K., Park, Y., 2010. Calibration and validation of a generic multisensor algorithm for mapping of total suspended matter in turbid waters. Remote Sens. Environ. 114, 854–866.

Vanhellemont, Q., Ruddick, K., 2015. Advantages of high quality SWIR bands for ocean colour processing: examples from Landsat-8. Remote Sens. Environ. 161, 89–106.

The Citclops project aims at developing systems to retrieve and use data on seawater colour, transparency and fluorescence, using low-cost sensors combined with contextual information and a community-based Internet platform. Citizens participate in the system by taking photographs of the sea surface on vessels, at the open sea or from the beach. These data are automatically uploaded to the Internet, archived, processed and finally accessed by end users through a webpage or a mobile application. Policy makers are able to use the information to improve the management of the coastal zone and citizens can maximize their experience in activities in which water quality has a role.
We present how remote sensing ocean colour observations from MERIS are used for the verification of the incoming crowd-sourced seawater parameters provided by non-specialists using a smartphone app and inexpensive measurement devices: seawater colour in the Forel-Ule colour scale, attenuation coefficient. The MERIS data archive provides probable values for the measured parameters in the Citclops application areas, the Wadden sea in the North Sea, and the Ebro Delta in the Mediterranean, two protected marine areas. The advent of the OLCI/Sentinel-3 sensor in 2016 will allow including real time remote sensing measurements, and improving the synergy between the remotely and the locally sensed data.

In order to improve the information on the state of the Indonesian marine environment that is gained from satellite data, the joint German-Indonesian Pilot Study IndoNACE (Indonesian seas Numerical Assessment of the Coastal Environment) was proposed to gain information on, and to further the knowledge about, the vulnerability of dedicated marine (coastal and offshore) areas in Indonesia to marine oil pollution of two dedicated regions of interest (ROIs) in Indonesian waters, namely the Western Java Sea and the Strait of Makassar.

 

One of the case study areas in the Western Java Sea is Seribu Islands, located in the northern part of Jakarta, the Capital City of Indonesia. This region is influenced mainly by the monsoon, where during the west monsoon; ocean current flows generally from north to the east, and on the other hand, during the east monsoon the current flows mainly westward while part of this current flows northward and southwestward to the Sunda Strait.

 

In this area, there are some oil spill occurrences, such as in 2004, 2007, 2008, 2011, and 2014. Unfortunately, information about the sources of these oil spills are difficult to be obtained, while, on the other hand, Seribu Islands has some Marine Protected Areas (MPAs) as well as marine aquaculture sites that vulnerable to the oil pollution. Based on this condition, in order to trace the oil pollution sources and their trajectories as well as potential impacts to the environment, the assessment based on the in-situ measurements, numerical models, and remote sensing will be done in this study. The HAMburg Shelf Ocean Model (HAMSOM) will be run to produce ocean currents that taken into account the monsoonal variation. After that, by using a tracer model and utilizing SAR data, the trajectory of oil spill is estimated backward and forward to trace back the source and trace forward the spread, respectively.

Ship detection with Synthetic Aperture Radar (SAR) imagery is essential nowadays in observing and monitoring the maritime environment. Knowledge about vessels position and type is of great importance for a wide range of applications: maritime traffic safety, fisheries control and border surveillance. The use of Sentinel-1 TOPSAR images for vessel detection, presents an interest due to a good tradeoff between spatial resolution and coverage. All relevant coastal zones and North Atlantic shipping routes are covered by either the IW or EW modes. With these 2 modes, dual-polarized SLC data can be provided.

In this study, we aim at demonstrating the benefits of using Sentinel-1 complex and dual-polarimetric data for maritime surveillance applications. Firstly, the relevance of individual polarimetric channel is analyzed via a comprehensive assessment of SAR-AIS matching. Results are compared with ship detection capabilities conducted on Radarsat-2 and CosmoSkymed datasets. The influence of different key parameters, such as SAR imaging characteristics (polarization, incidence angle) or meteorological conditions, is also assessed (See Pelich et al. 2015).

Additionally, the fusion of detection results between the two polarimetric channels is considered. The fusion of both polarization channels before the detection step is considered. For this, several signal/image processing techniques are employed, depending on whether the data are SLC or not. The availability of dual-pol complex data from the TOPSAR mode represents an asset.

The possibility to retrieve the vessel velocity from the overlap between adjacent TOPS bursts is also analyzed. Results are duly analyzed with a large number of SAR-AIS matchups.

Reference: 

Pelich et al. “AIS-Based Evaluation of Target Detectors and SAR Sensors Characteristics for Maritime Surveillance”, JSTARS, Vol 8, issue 8, 2015.

Pulse-limited or Delay-Doppler (SAR) satellite radar altimetry was designed to operate over deep oceans, or flat terrains. However, innovative processing, in particular, radar waveform retracking, has allowed retrieval of accurate altimetry ranges over hydrologic bodies, ice surfaces, and solid Earth for interdisciplinary studies. It is well known that the accuracy of altimetric ranges quickly degenerated as the observations near coastal zones, due to the signals are contaminated by non-ocean reflection and/or improper geophysical corrections. Although several algorithms have been developed to retrack the waveforms over the coastal oceans, the performance is not as good as expected within 0 to 5 km of the coasts. Here, we demonstrate the use of a radar waveform retracking system, the CurveFit retracker, to derive accurate sea surface heights (SSHs) around Taiwan coastal oceans using Envisat RA-2 data (cycles 10 to 90). The novel strategy is fitting the altimetry subwaveforms to the Brown model with Gaussian Peak models, and removing non-ocean reflected waveforms automatically. In this particular study region, the results indicated that the CurveFit retracker is the best-performed retracker among other retrackers within 0 to 5 km offshore, yielding the maximum improvement percentages (IMP) of 92%. Around 5 to 10 km from the coastline, compared with non-retracked Envisat measurements, the CurveFit retracker improved the accuracy of retracked SSHs, reaching the IMP up to 93%. Experimental results indicated the quality of altimetry measurements and accuracy of retracked heights also depends on direction of the ground tracks, or on terrain-based tracker biases. Over undulating terrain region, the altimetry waveforms are more contaminated as the satellite reaches its target, while over flat terrains area, the accuracy and quality of retracked heights improve when the ground track pass over the targets. We applied the CurveFit retracker and other retrackers to retrieve multiple-mission altimetry heights, including Cryosat-2 Delay-Doppler (SAR) and Altika/SARAL 40-Hz Ka-band altimeters where available, over test areas such as ice surfaces, tributaries and coastal embankment regions, to measure water level and land subsidence time series over South Asia including coastal Bangladesh.

Mesoscale and sub-mesoscale variability are closely related to important transport processes and dynamical phenomena of the surface mixed boundary layer, as well as of the deeper layers of the water column. Oceanic eddies are globally distributed features appearing at ocean’s surface, in large sizes with spiral shape variability. The formation, evolution and lifetime of these phenomena, depend on several physical and weather conditions related to wind velocity and direction, surface circulation and currents, indicating different momentum exchange and instability conditions between the ocean and the atmosphere. The main objective of the present study is to provide an extensive analysis on the occurrence and statistics of sub-mesoscale oceanic eddies over the Baltic Sea region. The method employed was a combination of synthetic aperture radar (SAR), Sea Surface Temperature (SST), chlorophyll–a (MODIS), wind and current model data (BSH) assessment. The study area is characterized by special conditions due to strong stratification and other lake similar physical properties, which lead to interesting results about the nature and the generation mechanisms of the spirals. For the present study, 1329 TerraSAR-X radar images (ScanSAR and Stripmap mode) were collected and analysed for the period January 2011 to May 2014. From image segmentation and feature extraction, 38 oceanic eddies were detected. Three different categories of oceanic spiral formations were recorded, with respect to their pixel characteristics. The majority of detected eddies were black eddies, related to algal bloom and biogenic material occurrence, combined with low wind velocities. Black eddies were characterized by low backscatter coefficient values. Their diameter varied from 1 to 9 km and were located near regions characterised by irregular topography and semi-closed Gulfs. During strong Western winds, eddies with higher backscatter values, known as white eddies were detected, with diameters varying from 3.5 to 6.5 km. White eddies were related with surface circulation and were located close to regions where alteration of surface circulation takes place. On the Northern part of the basin, during Winter season, ice eddies were detected, mostly in temperature front regions. Ice eddies present high backscatter coefficient values and smoother texture than white eddies. Their diameter varied from 2 to 7 km. Most of the formations were smaller or equal to baroclinic Rossby radius, which was calculated as an average of 5 km annually for the whole basin. Consequently, the majority of the formations were considered to be non-geostrophic, showing a strong relation with the wind field force. The detected oceanic spirals were both cyclonically and anti-cyclonically rotated at the same abundance. In order to study the environmental conditions and examine the generation mechanisms of oceanic eddies; different datasets and model data were analysed. Based on the acquisition date of the SAR images, sea surface temperature (SST) and chlorophyll (chl-a) concentration images from MODIS were collected. In addition, wind speed data and current model data for surface circulation with high spatial resolution were analysed, in order to detect changes and relations in the regions of the spiral detection. This approach was quite challenging due to different spatial resolutions of the datasets. From each dataset, plots and maps were generated, showing the overall conditions in the basin, as well as in the region where eddies were detected. Assessing the combined information layers, it was found that black eddies tend to appear in areas with high temperature and high chlorophyll concentrations. In contrast, white and ice eddies appeared closer to regions with low temperature and strong wind and current velocities. Finally, the number of detected eddies in the basin was significantly lower than that presented in studies focusing on previous years.

An analysis of satellite radar and optical images of ocean surface is increasingly used nowadays for diagnosis of processes on the sea surface, in particular, for the detection of film pollutions. Film slicks appear in the sea surface optical and radar images as dark or light patches at moderate incidence angles larger than 30-35 degrees and at moderate wind velocities. Usually it is considered to be established that there are films (for example, oil spills or biogenic films) on the sea surface in the area of reduced wave intensity under some additional information, e.g.,  proximity of oil platforms, oil transportation routes, seafood farms, etc. The practice of remote sensing shows that the risk of a wrong decision is high because of some “look-alike” phenomena, for example, wind shadowing zones. Many dynamic processes in the upper layer of the ocean lead to the dark and light patches in surface images, and the same sensors are used to investigate all these phenomena. The problem of discrimination between surface pollutions and “look-alikes” on the sea surface is closely related with the general problem of identification of dynamic processes from their imagery.

            An approach for better discrimination between film slicks and “look-alikes” is developed using information from sensors measuring variations of the wind wave spectrum at several wavelengths in a centimeter and decimeter range and sensors for measurements of some integral characteristics of wind wave, e.g. slope variations. It is well-known that different phenomena can manifest themselves identically when measuring variations at one wavelength or variations of one integral characteristic, but the signatures of the processes could be quite different when comparing data from different sensors. To select the parameters of the sensors that allow distinguishing different processes, in particular surface pollution and “look-alikes”, theoretical analysis of existing models of wind wave spectrum variations due to different processes in the upper ocean is carried out. The used wind wave model takes into account the influence of near surface wind and wave damping due to elastic films. An influence of oil film thickness on wind wave damping was also analyzed.

            An experimental verification of the approach was done during some field experimental campaigns. Firstly, results of recent field experiments on radar probing of film slicks using X-band scatterometer mounted on board a research vessel are compared to the theoretical calculations. The experiments were carried out with surfactant films with pre-measured physical parameters at low to moderate wind conditions. Satisfactory agreement between model and experiment is demonstrated and the information on NRCS depression can be used for better interpretation of signatures appeared in radar imagery of the sea surface. Secondly, some results of our earlier field experiments using optical devices are re-analyzed, too.  The optical spectrum analyzer developed in IAP RAS allowed registering the dramatic difference in the spectral contrasts of the film and wind slicks in dm-cm wave range, which are in satisfactory agreement with the theoretical calculation using the model mentioned above. The results of experiments on discrimination artificial films with different physical characteristics using optical devices with low spatial resolution are described. Thus recognition capabilities of different near-surface phenomena are clarified and conditions of their discriminations are determined.  

Plastic marine pollution is a major environmental concern but a quantitative description of the dimension of this problem in the open ocean and in particular in the more complex environment of enclosed or semi-enclosed seas is lacking due to absence of direct observations. As a matter of fact only 10 per cent of discarded waste is constituted by plastic, but it represents the biggest portion of the debris accumulating on shorelines due to its longevity that is estimated to be from hundreds to thousands of years (even far longer in deep sea and non-surface polar environments).

In the present work, we present an analysis of oceanic currents derived from the ESA satellite mission GOCE aiming to model the main global accumulation areas. The study is based on the results of the STSE project MEGG-C where not only the geostrophic currents but also their estimated error variance/covariance matrix has been computed at a global level using satellite data only (altimetry and GOCE geoid). The methodology takes advantage of Monte Carlo methods to simulate the accumulation of plastic debris in the oceans. Basically debris are introduced in correspondence of the main naval routes and coastal cities and their tracks are followed until they reach an accumulation point. In doing that the non-geostrophic part of the oceanic currents has been stochastically modelled. Note that the knowledge of the error covariance allows for an estimate of the final accuracy of the accumulation region too.

The obtained results are in agreement with the current knowledge about plastic debris accumulation from literature.

Since the Mediterranean Sea is a typical oligotrophic basin, productivity of its ecosystem is mostly defined by the dynamics of the surface waters. In the present work the Mediterranean Sea surface circulation features of meso- and submesoscales are being investigated by means of satellite imagery of different physical nature. The time coverage of the study is mostly from 2009 till 2011, for which lots of Envisat data have been retrieved.

Mesoscale dynamical features – mostly eddies with a diameter exceeding 20-30 km – are being investigated using NOAA AVHRR thermal infrared images (pixel size is of 1 km), which are perfectly suited for such an application, because mesoscale features usually possess significant thermal contrast comparing to the adjacent waters.

Submesoscale vortical structures (with a diameter less than 20-30 km) are being observed with the synthetic aperture radar (SAR) and natural colour images. The SAR dataset consists of Envisat ASAR Wide Swath and ERS-2 SAR images with a pixel size of 75 m. Visualization of the surface circulation features is provided either by accumulation of biogenic surfactant films (so called “black” eddies) or by wave-current interactions (“white” eddies). These manifestations take place all over the basin including the interior offshore parts of the sea. Natural colour images used in the current research are being provided by the Envisat MERIS Level 1 products (pixel size is about 270 m). In this case some passive tracers (e.g., inorganic suspended material) are required for eddy visualization which are usually available only in the near-coastal areas.

On analysis of the imagery just mentioned it was noticed that except for the biggest eddies (so called Alboran Eddies, which are anticyclonic and appearing at the Strait of Gibraltar and moving eastward along the southern coast of the Western Mediterranean) the rest of mesoscale eddies (mostly cyclonic) are distributed quite evenly all over the sea eventhough with a tendency to be located somehow offshore. The most energetic part of submesoscale eddies tend to be found in the northern part of the Western Mediterranean and in the southern one of the Eastern. This latter tendency apparently refers to the spatial variations of wind forcing and intensity of the surface currents.

In order to trace the possible influence of circulation features on ecosystem productivity, sea surface temperature (SST) and chlorophyll-a concentration (chl-a) fields were used. SST and especially its negative anomalies can be used as an indirect sign of intensive vertical movements. The chl-a data are being retrieved from the Envisat MERIS Level 2 products. For this part of the study special focus was made on the Western Mediterranean Sea, where most of interesting dynamical events (such as influence of the northerly winds, confluence of the currents, penetration of the Atlantic waters, etc.) take place.

The SAR data were obtained under the grant of the European Space Agency # 14120 “Spiral eddy statistical analyses for the Mediterranean Sea using Envisat ASAR Imagery (SESAMeSEA)”. This research was supported by the University of Liege and the EU in the context of the FP7-PEOPLE-COFUND-BeIPD project.

Stratification is of critical importance to the circulation, mixing and productivity of the ocean, and is expected to be modified by climate change. Stratification is also understood to affect the surface aggregation of pelagic fish and hence the foraging behaviour and distribution of their predators such as seabirds and cetaceans. Hence it would be prudent to monitor the stratification of the global ocean, though this is currently only possible using in situ sampling, profiling buoys or underwater autonomous vehicles. Earth observation (EO) sensors cannot directly detect stratification, but can observe surface features related to the presence of stratification, for example shelf-sea fronts that separate tidally-mixed water from seasonally stratified water. This presentation describes a novel algorithm that accumulates evidence for stratification from a sequence of oceanic front maps, and discusses preliminary results in comparison with in situ data and simulations from 3D hydrodynamic models. In certain regions, this method can reveal the timing of the seasonal onset and breakdown of stratification.

This research is based on the composite front map approach, which is to combine the location, strength and persistence of all fronts detected on EO sea-surface temperature (SST) or ocean colour data over several days into a single map, improving interpretation of dynamic mesoscale structures (Miller, 2009). These techniques are robust and generic, and have been applied to many studies of physical oceanography, most recently in the NERC FASTNEt project, and marine animal distribution. The initial estimate of stratification is based on SST fronts at 1km resolution, though ENVISAT MERIS ocean colour data allow us to investigate the structuring effect of stratification on the surface biological productivity at higher resolution (300m). The recent launch of Sentinel-3 will enable OLCI data to contribute to this analysis of stratification in near-real time, for potential use in applications such harmful algal bloom prediction.

A dataset of 224 Medium Resolution Imaging Spectrometer (MERIS/Envisat) full resolution satellite images were processed to retrieve the concentration of coloured dissolved organic matter (CDOM) in a hypertrophic estuary (Curonian Lagoon, Lithuania and Russia). Images covered a period of 7 months, spanning from the ice melting (March) to the late summer (September) of 7 consecutive years (2005-2011). The aim of the study was to analyse the spatial and temporal variations of CDOM, by focusing on the main regulating factors (riverine discharge, sea-lagoon water exchange, water temperature, chlorophyll a, wind) in a large estuary. The working hypothesis is that CDOM distribution may reveal distinct, site specific seasonal patterns. Our results demonstrated that CDOM concentrations at the whole lagoon level were elevated (1.5-4 m–1) and slightly but significantly higher in spring (1.50 m–1 on average) compared to the summer (1.45 m–1 on average). This is due to very different flow of CDOM-rich freshwater from the main lagoon tributary in spring compared to summer. They also highlight macroscopic differences among areas within the lagoon, depending on season, suggesting a complex regulation of CDOM in this system. Significant factors explaining observed differences are the dilution of lagoon water with CDOM-poor brackish water, regeneration of large amounts of dissolved organic matter from sediments and combinations of uptake/release from phytoplankton. CDOM and its variations are understudied due to inherent methodological and analytical difficulties. However, this pool has a demonstrated relevant role in the biogeochemistry of aquatic environments. We speculate that the dissolved organic pool in the Curonian Lagoon has a mainly allochthonous origin in the high discharge period and an autochthonous origin in the summer, algal bloom period. Both positive and negative relationships between CDOM and phytoplankton suggest that pelagic microalgae may act as a source or as a sink of this pool, in particular when inorganic nutrients are limiting.

Colored dissolved organic matter (CDOM) is the most abundant DOM fraction in many natural waters, especially in forested watersheds with wetlands. CDOM attenu­ates solar radiation, with major implications for physical, chemical, and biological processes. It regulates heat transfer to water, controlling lake temperatures and stratification, and by reducing light, CDOM suppresses primary productivity. CDOM also mediates transfer of con­taminants into food webs (e.g., mercury in fish). Although not directly harmful to human health, CDOM has negative effects on produc­tion of safe drinking water; it increases the consumption of water treatment chemicals, reacts with chlorine to form harmful disinfect­ion by-products, stimulates bacterial growth, and fouls membranes, which increases treatments costs. 

Few water management agencies include CDOM in their monitoring programs, however, and the paucity of CDOM data limits our understanding its dynamics in surface waters. CDOM levels also are needed for regional and global-scale models of carbon cycling. Remote sensing by satellite imagery has the potential to fill this void and improve our understanding of CDOM distribution and the environmental factors that affect CDOM levels in surface waters.

 This paper describes ongoing efforts to measure CDOM in optically complex inland waters in northern Minnesota, U.S.A. Sites were selected to obtain wide ranges of concentrations of CDOM, chlorophyll, and suspended sediment, the primary factors affecting reflectance; extensive field and laboratory measurements of water quality and optical measurements, including in situ reflectance, are major components of the program. Our goal is to produce robust algorithms for CDOM that can work in optically complex waters across major geographic regions or at least to identify waters where reliable measurements can be obtained from satellite imagery.

As a component in our algorithm development we used clear Landsat 7 and Landsat 8 images from September 2013 acquired for northern Minnesota eight days apart to compare the capabilities of the respective ETM+ and OLI sensors for retrieval of CDOM data. We examined a variety of potential band and band ratio models and found some two-variable models that included the NIR band worked well for Landsat 8 (R² = 0.82) and reasonably well for Landsat 7 (R² = 0.74). The common green/red model had a poor fit for both sensors (R² = 0.24, 0.25), and five sites with high mineral suspended solids (MSS) were clear outliers. Exclusion of these sites yielded a less optically complex subset of 20 lakes for both images. Strong models were found for this subset for many equation forms, including the green/red model with R² = 0.79 for Landsat 7 and R² = 0.81 for _Landsat 8. The less optically complex subset may explain why the green/red model has worked well in other areas. For the full data set (including optically complex waters), models using the Landsat 8 ultra-blue and narrower NIR bands worked best indicating that the new bands, higher radiometric sensitivity and improved signal-to-noise ratios of Landsat 8 improve CDOM measurements.

We applied these algorithms to Landsat 8 images obtained for Minnesota in summer of 2015 and a calibration dataset containing multiple measurements of CDOM to evaluate how near to imagery acquisition such measurements need to be for reliable calibrations.

To develop algorithms further we used in situ reflectance spectra and simulated the bands of new and upcoming satellites (Landsat 8, Sentinel-2 and Sentinel-3). All these systems worked well for CDOM and water clarity; inclusion of the red-edge band in Sentinel-2 enabled it to distinguish inorganic sediment from chlorophyll in optically complex waters. The ability to measure these water quality variables on a regional basis will greatly enhance our understanding of spatial variability and responses to environmental change in surface waters and improve lake management.

Coccolithophores belong to nano-phytoplankton size-class and produce CaC03 scales, called coccoliths, which form the shell of the algae cell. This phytoplankton group has an ubiquist repartition in all oceans but bloom in some oceanic regions, like in the North-East Atlantic ocean and South-Western Atlantic. Coccolithophores have an important biogeochemical role in the CO2 pump of the carbon cycle in the surface waters and could be impacted by the acidification process and climate change. Today, contradictory assumptions are pointed out about the evolution of coccolithophore blooms. Under the hypothesis of an increasing acidification (pH decrease due to CO2 increase) of the upper layer of oceans, the calcification and extension of the coccolithophore blooms could be positively or negatively impacted. Thus, to be able to evaluate the feedback produced by coccolithophores in the context of the climate change, it is necessary to know the long term variability of their blooms.

The goal of this work is to use a remote-sensing method to investigate the regional and temporal varibility of the coccolithophore blooms. The data set used is based on Level 3 data for SPM and calcite, on RGB imagery and on normalized-leaving radiance L2 data from MODIS (Moderate Resolution Imaging Spectroradiometer) and SeaWiFS (Sea-viewing Wide Field-of-view Sensor) to have a temporal range of 17 years study between 1998 and 2002 (SeaWiFS) and between 2003 and 2014 (MODIS). The method is based on the satellite-derived suspended matters concentration for quantifying the coccolith blooms which have a strong backscattering signal. We discriminate coccolith blooms by using a statistical method classification pixels, using a simplified fuzzy index based on the distance between the radiance vector and the mean reflectance of the coccolith signal.

Results are compared with in-situ observations of coccolithophores in North-East Atlantic and with additional remote-sensing methods for the validation step. The result of the fuzzy method fits better with coccoliths than coccosphere in-situ data, and have the same evolution than PHYSAT algorithm between 1998 and 2003. Although the classifier may fails in waters where optical properties diverge from those of the reference data set used for training the fuzzy-index, the quantification in coccolith biomass derived from SPM maps appears to be well related to the calcite concentration derived from the NASA algorithm. Although a regular pattern in the phenology of the blooms is observed, starting south in April in Biscay and moving northwards until July in Ireland, there is a high seasonal and interannual variability in the extent of the blooms. Results of blooms budgets all along the sudy period point out differences in the development of the blooms between sub-areas during 17 years, and a strong anomaly for the year 2014.

This work will allow to improve the phenology of coccolithophores blooms to investigate the bloom budgets at a larger scale with a comparison of the method in another parts of oceans, like in the South-West Atlantic, such as Patagonian sea where blooms of this calcareous phytoplankton are recognized. To go further in the study to understand the internannual variability of the coccolith blooms, the environmental factors have to be taken account.

Satellite ocean radar data are used to assess the flat surface reflectivity for seawater at 36 GHz by comparison to an existing model for dielectric constant variation. Sea surface temperature (SST) is the dominant control and results indicate a 15 % variation in the normalized radar cross section (NRCS) at Ka-band (35.75 GHz) that is in close agreement with model prediction. Consistent results are obtained globally using near-nadir incidence data from both the SARAL AltiKa radar altimeter and Global Precipitation Measurement (GPM) mission rain radar. The observations affirm that small but systematic SST-dependent corrections at Ka-band may require consideration prior to NRCS use in ocean surface wave investigations and applications. As an example, we demonstrate systematic improvement in AltiKa ocean wind speed inversions after such an SST adjustment. Lower frequency C and Ku-band results are also assessed to confirm general agreement with prediction and a much smaller variation due to SST.

This project aim at developing a cartographic service dedicated to the survey of pollutant biological phenomena in coastal waters, in order to guide local fishery on a day to day basis and to help local authorities to create targeted environmental regulation. This service addresses in the first place researchers working for the local economic ecosystem, allowing them to accelerate their researches by gathering all the data they need and providing them tools to test their new algorithms directly in real situation.

Main short term objective is the “Liga” phenomenon (Figure 1) as called by the fishermen of South-West of France: in this area viscous and irritating mucilage regularly appear at the end of winter and at the end of summer, making them lost their day by plugging their net. As this phenomenon trends to accelerate and persist during spring, summer and fall since the middle of the twenties, a multidisciplinary program supported by the Institut des Milieux Aquatiques (IMA) in Bayonne (France) is working on this subject since 2010, leading in-situ biological campaigns (Figure 2) off-shore the Basque country (Susperregui et al. 2010). But these biological studies need now to drastically extend their geographical area if an explanation to the Liga phenomenon is to find rapidly, and more data from diverse sources need to be gathered in order to put these campaigns into a complete environmental context. 

In order to achieve these goals, IMA and Capgemini are associating their efforts to build an operational platform which will be used for the acceleration of research in coastal areas. More particularly, the Capgemini scientific office, which gather physicists in Earth Observation in various domain (passive and active remote sensing, imaging and sounding), is working with IMA biologists in order to gather all the pre-processed data that are needed to model the appearance of the Liga phenomenon, and that could be used for other related subjects. In priority, these data include optical (visible and infrared) satellite imagery at medium resolution (Sentinel 2, Landsat OLI and TIRS) and in-situ data coming from campaigns and buoys. From satellite data, the following parameters are extracted: ocean colour, chlorophyll-a and CDOM (Colored Dissolved Organic Matter) water contents, water absorption and backscatter coefficients, SST (Sea Surface Temperature). From in-situ data and buoys, are extracted in surface: chlorophyll-a water content, amounts of phytoplankton by species, nutrients, water salinity, turbidity and temperature. Meteorological products (from NOAA, WMO and MeteoFrance) will be also gathered, together with MODIS SST (that will be replaced by Sentinel 3 SLTR product) and SMOS salinity retrievals. In a second time, the project aims at adding current cartography retrieval, PAR (Photosynthetically Available Radiation) estimated from satellite data, sediment cartography retrieval, and a probabilistic cartography of phytoplankton species. And finally, as the main objective is to test algorithms for the estimation of a cartographic probability of appearance of the Liga phenomenon, biologists will be able to mathematically combine all data directly on the plateform via a MMI, and test different versions of their algorithms on in–situ data.

Algorithms for processing the satellite data (atmospherical correction and SST inversion, Figure 3) are provided by the Capgemini scientific office, more precisely, the ocean colour and bio-optical relationships will be provided by the OCEAN algorithm (Ocean Colour Estimation by principal component Analysis, Gross et al. 2007a) which performs well in coastal areas and can deal with glitter and specular reflection on waves, which are inherent to medium and high resolution imagery (Gross et al. 2007b).

The SAS developed by Capgemini includes GIS data manipulation and visualization, automatic data gathering and match-ups between in-situ and satellite data, automatic bio-optical empirical relationships improvement via machine learning techniques, functionalities that allow researchers to download data, upload data and algorithms for rapid validation tests on real conditions, visualization of validation results (Figure 4). These functionalities are developed using “big data” solutions, because of the data volume that will be manipulated (hadoop, spark), reusing the Tech4Earth solution that has been implemented for the FAAPS project (https://artes-apps.esa.int/projects/faaps).

Susperregui N., D’Elbée J., Maton V., Rihouey D., Maneux E., Etcheber H., Sautour B., Othéguy P., Maron P., Monperrus M., Soulier L., Gallet F., 2010. Etude d’une substance appelée Liga sur le littoral basque : identification, origine et facteurs influençant son apparition, rapport d’étude FEDER/IMA : 88p. + annexes

Gross-Colzy L., S. Colzy, R. Frouin, P. Henry.: « A general ocean color atmospheric correction scheme based on principal components analysis – Part I: Performance on Case 1 and Case 2 waters», Proceedings of SPIE Optics and Photonics, San Diego, 26-30 (August 2007).  

Gross L., S. Colzy, R Frouin, P. Henry.: « A general ocean color atmospheric correction scheme based on principal components analysis – Part II: Level 4 merging capabilities, Proceedings of SPIE Optics and Photonics, San Diego, (26-30 August 2007).

The ongoing launches of the Sentinel satellite infrastructure under the Copernicus programme, represents a significant opportunity to monitor and explore the function and nature of our ocean and near-shore regions, upon which so much of our activities are reliant. Our understanding of the structure and functioning of fisheries, for example, and how marine ecosystems are exploited, are but one example, forming the foundations of our fisheries and aquaculture sectors. To ensure optimal decision making, viable planning, and sustainable exploitation, these sectors need tailored Decision Support Services, much of whose information can be derived from Earth Observation (EO) data.

The EU-FP7 funded project SAFI (“Supporting our Aquaculture and Fisheries Industries”) is primary focused on the exploitation of such Earth Observation (EO) data resources, specifically targeting the support of fisheries and aquaculture industries in our ocean and near-shore areas. The initiative aims at providing information assistance for marine aquaculture deployment (optimization of cage location with respect to environmental and ecological context) and environmental monitoring during operations. It also targets supporting fisheries providing informative indicator products developed through exploitation of Ocean Colour, Sea Surface Temperature, and Sea Surface Height data to derive correlations between the environmental situation and the biological status of the fish and shellfish studied. The results are synthetic indicators covering the recruitments, abundance and maturation rates of key species of critical importance to regional fisheries and aquaculture sectors.

The SAFI services, and products underpinning them, have been developed in close collaboration with target users, with extensive stakeholder consultation and review forming the foundation of the development process. The capacity for exportation and acceptance of the developed services and products, is being evaluated on several pilot sites in Europe (Spain, Portugal, Ireland, France), and shall finally be demonstrated in Morocco.

Four service lines are being developed, making best use of emerging EO products, and in particular adapting product derivation methodologies to exploit ESA Sentinel-2 and Sentinel-3 missions.

Finally, SAFI aims to build a reliable and marketable offer. To do so, the project has iteratively developed an integrated Decision Support tool with users. This service tool is centred on a web-GIS data visualisation system, broadcasting a variety of SAFI products to the users. Both the tools and products that have been developed shall be presented to the symposium.

The SAFI project does not address links between environmental parameters (as retrieved from EO) with ecological modelling – such complementary topic is one of the main objectives of the DUE SMART project supported by ESA (presented in another abstract).

Submerse aquatic macrophytes are used as long-term indicator plants characterizing the trophic state of freshwater lakes. In the context of the European Water Framework Directive, the appearance and the composition of the macrophytes species are investigated in routine monitoring cycles every three years. This conducted monitoring is still based on in-situ measurements. Due to climate change induced increasing water temperatures, some macrophytes species show high invasion dynamics. Their partly areal spread affects the lake as well as the complete ecosystem. To detect first changes, a more-frequented monitoring method is suggested.

The presented research is about a subtask of an integral concept for a remote sensing based inventory and monitoring system for freshwater lakes. One component of this system are combined growth and reflection models based on in-situ data collections. For comparing the in-situ measurements almost the same parameters are necessary. One limiting factor for continuous in-situ measurement series throughout the vegetation period is the precondition of a cloudless sky. Therefore, an interpolation model to calculate the remote sensing reflection intensities for selected submersed macrophytes was developed. Currently the spectral signatures of four different macrophytes species are integrated in this model. Species-specific remote sensing reflectances of the investigated macrophytes were collected in the years 2011 and 2015. The reflectance signal is greatly affected by different proportions of sediment and macrophytes coverage across a vegetation period. To get the spectral variations during the growing season, systematic and multi-temporal measurements covering the complete growing season from May to October were done. In both years the same study sites at Lake Starnberg in Southern Germany were investigated. For data acquisition a hyperspectral, submersible RAMSES spectroradiometer-system was used. Additionally, the phenology of the observed populations of submerse macrophytes was recorded by underwater cameras. So the seasonal variability of the spectral signatures can be linked to the corresponding plant phenology.

Beside the seasonal variability the spectral response is affected by diverse meteorologic conditions. Due to differing air and water temperatures the duration of the vegetation period as well as the length of the four defined phenological stages are varying among the two years. The difference in phenology in the investigated years can be showed in the combined growth and reflection models. Phenological developments within each year as well as shifts in phenological stages attributed to the differences in weather conditions are well reproduced by the models. The development progress of the aquatic plants and the shift within the vegetation period are analysed and compared.

Since 1950, Tonga islands have been submitted to real effects of climate change; the temperature have increased at a rate of 0.1°C per decade, the annual rainfall have been decreasing (especially during the wet season), the sea levels are rising by about 6mm per year and the ocean acidification is increasing (PCCSP, 2011). These variations in environmental conditions have direct impacts on susceptible ecosystems, on top of which, mangroves and coral reefs. In addition to the environmental pressure attributed to climate evolution, the anthropic pressure on the natural ecosystem is increasing, as an illustration, the frequency, diversity and the rate at which humans are impacting coral reefs are increasing to the extent that some of them are threatened globally.

Mangrove health and thus sustainability is very sensible to salinity change (Venus, 2012). Mangroves fields do not only provide valuable goods (e.g. wood and fish) but are also known as vital ecological and other services (e.g. sequestration of carbon, navigable waterways, habitat for a unique biodiversity and prevention of coastal erosion). Their value and importance are not enough recognized and need to be promoted. The protection of mangroves has to be a priority.

The need for protection of coral reefs is very well known and documented. Indeed, it is estimated that 25% of all marine species inhabit coral reefs, where the number of individual species may be as high as one million (Davidson, O. G. 1998). Assessments to late 2000 are that 27% of the world’s reefs have been effectively lost. According to the World Atlas of Coral Reefs (Spalding, 2001), Tongan coral cover is around 1500 square kilometres (sq. km), which is a 1/200 of the global coral cover. Even if this one is reputed as healthy, there is a crucial need for close monitoring.

In such a context and in the frame of an action by the Asian Development Bank (ADB) about enhancement of climate resilience in Tonga, two services based on exploitation of Earth Observation have been developed:

A mangrove habitat assessment and site suitability analysis

A coastal habitat oceanographic stressors map (for coral reefs)

These two services are making best use of Earth Observation. 

For the service 1, High and Very High resolution data (QuickBird, Ikonos, Komsat, WorldView, Rapideye, Pléiades) have been used to monitor the mangrove mapping evolution in 2005-2010-2015. More specifically, the land-use characterisation, the delineation of hydrological networks and the coastal habitat assessment and delineation have been derived from imagery. Also the mangrove habitat extent and status have been mapped.

For the service 2, extensive use of High Resolution (Landsat/Sentinel-2) and Medium Resolution (SeaWiFs, MODIS, MERIS, VIIRS, OLCI) data have been exploited to derive SSH, SSS, SST, Currents, turbidity, Chl-a, PAR, Wind, UV-radiation, waves from 2000 to 2015 in order to make appear trends of the physical forcing and biogeochemical response of the coastal and marine waters of the Tonga islands.  Results will be presented at the Symposium.

ESA has been committed to reprocess its past and present Earth Observation Missions of ERS and ENVISAT using the latest available algorithms and applications to produce longer and rather homogeneous time histories. Some of the reprocessing campaigns have already concluded while some are still ongoing. Along this line, the full-mission reprocessing of the ERS altimetry products (REAPER Project) has already been completed. The data covering the ERS-1 and ERS-2 active lives (1991-2003) have been made available by ESA.

At the same time, ECMWF has been carrying out its infamous reanalysis project under the name of ERA-Interim which aims also towards producing a long-term rather homogeneous model data set using the best available model and data assimilation techniques at the start of the project.

The REAPER ERS-1/2 wind and wave products are assessed mainly against ERA-Interim (and related model runs) as well as available in situ data. The impact of assimilating REAPER significant wave height in the ECMWF wave model is also assessed. This work supports the newest ECMWF reanalysis ERA-5. The results will be shown and discussed.

Internal solitary waves (ISWs) are important dynamical features significantly impacting the hydrology of the upper ocean through transferring the energy from tides to turbulent mixing. In the Arctic Ocean they are particularly important for sub-marine navigation and construction, formation of the water structure and the maintenance of the life activity in marine ecosystems. However, the Arctic region still remains largely unexplored and internal solitary waves there are still poorly investigated.

In this work taking the advantage of high resolution space-borne synthetic aperture radar (SAR) data we present the results of short-period ISWs observations in the Eurasian Arctic seas based on analysis of ENVISAT ASAR data for summer-autumn months in 2007-2011. More than 4000 ISW packets were identified in about 3000 SAR images. Detailed maps of IWs observational frequency and their spatial and kinematic properties for the Eurasian Arctic seas helped to identify hotspots of permanent ISW generation. We also point out the regions where large-scale nonlinear IW packets with wavelengths of 2-5 km and crest lengths >200 km are observed.

The work was supported by RFBR, research projects No. 14-05-31423 mol_a and15-05-04639 А. ENVISAT ASAR data were provided under ESA Cat-1 projects C1F-4483, C1F-29721.

Take 5 SPOT 5 time series have been acquired over Paranoá lake, located within Brasilia city, Brazil, from April to Septembre2015. This period matched two distinct seasons, the end of the wet period marked by mean monthly precipitation of 204 mm and the dry period, marked by near zero precipitation from June to September. The Paranoálake is located at 1000 meter a.m.s.l, covering 38 km²,its water storage capacity is of 498 x106 m³, its mean water depth is of 14 meters, and the water residence time is of 299 days. The lakes currently faces different water quality issues induced by the rapid urbanization all around the lake area, including sedimentation, eutrophication and pollution.

Twenty-two images presented cloud-free conditions making possible to assess finely the water quality temporal behavior over the whole observation period. Six field campaigns have been conducted to register remote sensing reflectance using Trios hyperspectral radiometers and main water quality parameters. For two dates, SPOT 5 and Landsat-8 images were acquired the same day. During September 5, radiometric data included SPOT 5, Landsat-8, field spectroradiometry and drone multispectral imagery acquired in clear sky conditions.

Paranoá lake presented rather clear water conditions over the whole with low inorganic matter content except near the two main rivers’ mouth where sedimentation processes were identified. Water optical properties near two water treatment plants revealed local eutrophication processes. Overall, remote sensing  reflectance level in the visible part of the spectrum varied from 0.3 to 1.2 % with contrasted conditions depending on the location and on the period.

Hyperspectral data allowed to classify water types across the lake and between different campaigns. In particular, it was possible to determine the bests regions of the visible / near infrared spectrum for suspended sediment concentration (SSC) and chlorophyll-a (chl-a) concentration retrieval. Hyperspectral data were degraded to SPOT-5 spectral resolution to analyze the radiometric accuracy of Level-2A images. Maps of SSC and chl-a concentrations were derived and validated using the ground truth samples. The remote sensing derived water quality maps allowed assessing the extent of the areas facing environmental problems, near the rivers mouth and the water treatment plants. In particular, we analyzed the performance of the different remote sensing data (hyperspectral radiometers, drone, SPOT 5, Landsat-8) as a function of spatial and spectral resolution. Also, hyperspectral data were used to simulate Sentinel-2 radiometric bands, showing that these data will provide enhance retrieval capacity in relation to previous high resolution earth observation satellites.

Thanks to a widening range of sensors available, the observation of continental water quality for lakes and reservoirs is gaining more and more consistency and accuracy. Consistency because back in 2012, the only free sensor with a sufficient resolution (30m) was Landsat-7 which has truncated data since 2003 and a 16-day revisit time. But today, Landsat-8 and Sentinel-2A are now operating so depending on the latitude of interest, the combined revisit time dropped to 2 to 4 days which is more appropriate for such a monitoring (especially considering the cloud cover). Accuracy because Landsat-7 has a poor contrast over water whereas Landsat-8 and Sentinel-2A have a better radiometric sensitivity (more bit) and moreover Sentinel-2 offers additional spectral bands in the visible which are helpful for Chlorophyll-A concentration assessment.  To sum up, with Sentinel-2, continental water quality monitoring capabilities are making a giant leap and it is important to exploit this potential the sooner. ACRI-HE has already built a strong basis to prepare Sentinel-2 by using Landsat data.

Indeed, more than 600 lakes are already constantly monitored using Landsat data and their biological statuses are available on EyeOnWater (see eyeonwater.eu). Chlorophyll-a retrieval from (fresh) water leaving reflectances is the result of research activities conducted by ACRI-HE in parallel with EDF (Electricité de France) to respond to an emerging very demanding environmental monitoring through European regulations (typically the Water Framework Directive). Two parallel and complementary algorithms have thus been derived for Chlorophyll-a retrieval.

Upstream of Eyeonwater, there is a complex and complete system automatically collecting images, extracting areas of interest around lakes, applying atmospheric correction (very sensitive part as atmosphere can contribute to 90% of the signal at sensor level) and then algorithms to retrieve water transparency (Secchi disk), turbidity and Chlorophyll-A concentration. A wide range of in-situ measurements was gathered to calibrate these algorithms. We present here a clear and operational system working with Sentinel-2-like data that retrieves water ecological quality parameters and provides quantified level of uncertainty. We believe that this system is of prime relevance to fulfil water quality monitoring duties at local, national and regional levels.

Water temperature is one of the most important parameter which is changing the concentration of ions by affect biological and chemical changes in the lake. It can be easily influenced by environmental factors such as climatic parameters, topographic height, geographical condition bathymetric properties, pollutants and etc. While temperature has positive correlation with electrical conductivity, the negative correlation is observed with dissolved oxygen. Thus, increasing water temperature threatening the aquatic life. Beyşehir Lake is the largest fresh water lake in our country with the 653 km² surface area. Lake water have used for drinking water of several settlements in the basin. Beyşehir Lake is a shallow lake and, especially in recent years its water level was dropped due to unplanned usage and effects of climate change.

 

In this study, determination of the water temperature in Lake Beyşehir is aimed using 90m resolution thermal infrared bands of ASTER (Advance Spaceborne Thermal Emission and Reflection Radiometer) satellite and 30m resolution thermal infrared bands of Landsat 8 OLI-TIRS satellite. The Normalized Water Different Index (NWDI) has been applied to ASTER and Landsat 8 OLI-TIRS satellite images to determine lake surface area. Accordingly, the lake water temperature is generally proportional to the depth and it relatively higher in the shallow area.

 

 

 

The Baltic Sea is one of the world’s most threatened marine environments. Alarming rates of eutrophication caused by excessive input of nutrients such as nitrogen and phosphorus stimulates the growth of planktonic algae leading to imbalance in the marine ecosystem. Intense algal blooms and production of excess organic matter result in decreased water transparency as well as oxygen depletion which leads to the formation of dead zones at the sea bottom, as well as additional stress on biodiversity.

The Gulf of Riga accumulates a large amount of anthropogenic and natural cause’s derived terrestrial nutrient influx. The Daugava River along with numerous smaller rivers brings a vast amount of dissolved and suspended material from the drainage basin to the coastal waters. The material is then transported across the wider Baltic Sea by the means of currents and wind activity. Riga agglomeration, recreational activities on the coast, coastal ecosystem transformation, and busy sea transport create additional stress on the marine ecosystem. Therefore, the Gulf of Riga serves both as a good model study reference area and a target for better management.

Remote sensing has proved to be an accurate and reliable tool in clear water environments like oceans or the Mediterranean Sea. However, the current algorithms and methods usually fail on optically complex waters like coastal and inland waters. The whole Baltic Sea can be considered as optically complex coastal waters.

MERIS with its 300 m spatial resolution and several bands in red part of spectrum (missing in the case of MODIS, for example) made it the best possible sensor for monitoring coastal water quality. The gap between the demise of ENVISAT and launch of Sentinel-3 with Ocean Land Colour Instrument (OLCI) on board left a serious gap in our coastal water quality monitoring capability. On the other hand, we should use this gap to develop better algorithms and methods for the future use of OLCI imagery.

The main aim of the study is to use hyperspectral airborne data from the flying laboratory developed by the Institute for Environmental Solutions as well as in-situ field measurements (11-12 August 2015) to simulate performance of Sentinel-3 OLCI and test remote sensing algorithms for calculation of marine L2 products for the Gulf of Riga. Testing MERIS processors with historic data will be also performed.

Not many years ago carbon-climate models, for instance those used by the Intergovernmental Panel for Climate Change (IPCC), ignored inland waters in their global carbon cycle models and rather treated them as inert “pipes” transporting terrigenous organic carbon into the oceans. This was likely due to the fact that inland water bodies cover just over 3% of the land area (Verpoorter et al 2014).  However recent estimates (Cole et al. 2007, Tranvik et al. 2009) show that lakes are by no means inert pipes and they really do contribute to the global carbon cycle. Smaller lakes are particularly important because there are usually enough nutrients and light and therefore primary production, settling, outgassing of methane and carbon dioxide are especially active. Current global models of carbon cycle are based on in situ data from few thousand lakes which were then extrapolated over the number of lakes and size distribution. Moreover, it was done using only statistical estimation of the size distribution and the number of lakes.

Remote sensing can be used to map lakes and their area. Previously there were no commercial satellites with sufficient spatial and in particularly radiometrical resolution for studying substances in water but after the launch of Landsat 8 in 2013 and Sentinel 2a in 2015 solved this technical problem. Combining earlier Landsat satellite images, global elevation model and GIS software the number of lakes and their size was estimated globally. Currently these results are used for estimating the total volume of the lakes.

Coloured dissolved organic matter (CDOM) is currently one of the most important research object in inland water bodies because it is considered as a critical component of the water ecosystem. Furthermore, it takes part in global carbon cycle which in turn influences environment. There are a large number of studies about carbon fluxes from land and oceans, however, inland water bodies have been previously rather ignored, therefore there are no good results and the estimates are preliminary. For carbon flux estimation from lakes dissolved organic carbon (DOC) data is needed. There is no direct way to measure it from satellites and therefore CDOM products are used because DOC and CDOM are closely related parameters in inland water bodies. When taking into account the total number of lakes, their area, volume and relating it with CDOM data we can give first estimation of carbon fluxes from inland water bodies. By increasing our knowledge about dissolved organic carbon in the spatio-temporal dimension, it helps to understand the role of inland water bodies in the global carbon cycle.

 

Cole JJ, Prairie YT, Caraco NF, McDowell WH, Tranvik LJ, Striegl RG, Duarte CM, Kortelainen P, Downing JA, Middelburg JJ, et al. 2007. Plumbing the global carbon cycle: Integrating inland waters into the terrestrial carbon budget. Ecosystems. 10(1):171–184.

Tranvik LJ, Downing JA, Cotner JB, Loiselle SA, Striegl RG, Ballatore TJ, Dillon P, Finlay K, Fortino K, Knoll LB, et al. 2009. Lakes and reservoirs as regulators of carbon cycling and climate. Limnol Oceanogr. 54(6):2298–2314.

Verpoorter, C., T. Kutser, D. A. Seekell, and L. J. Tranvik (2014), A global inventory of lakes based on high-resolution satellite imagery, Geophys. Res. Lett., 41, doi:10.1002/2014GL060641. 

Modern Active SAR advanced Polarized Satellite radar  and additiona Infrared or visible passive methods provide better discrimination and higher resolution in a wealth of complex geophysical flows. In the ocean. Even more so, in the coastal zone, with Regions of Freshw Water Influence, here the turbulent flow is generated in the ocean surface either by waves, wind or/and local currents, But buoyancy by river flows may also be important. The conditions of the medium are highly non-homogeneous, and in the presence of a pollutant the SAR also detects many topological features [1, 2]. New techniques are used for the subsequent analysis, of more than 2000 Images provided by the ESA ERS1/2, ASAR, ENVISAT, RADARSAT and other European, Canadian and Russian Satellites. We shall concentrate on a taxonomy of different dinamical features and provide some statistics, as well as describing some events detected by several satellites and even with additional cruise observations and measurements [4-9] in the North-west Mediterranean Sea area between 1996 and 2012.

The structure of the flows are presented using self-similar traces that may be used to parametrize mixing at both limits of the Rossby Deformation Radius scale. RL Results show the ability to identify different SAR signatures and at the same time provide calibrations for the different local configurations of vortices, spirals, Langmuir cells, oil spills and tensioactive slicks that eventually allow predicting the self-similar structure of the turbulence.[3,4]. Depending on the surface wind level, and also on the fech. the topology, the spiral parameters and the resolution of vortical features change. The previous description [,] did not include the new wind and buoyancy features.

It is noteworthy that such complex coastal field-dependent behavior is strongly influenced by stratification and rotation of the turbulence spectrum is observed only in the range smaller than the local Rossby deformation radius, RL. The measures of diffusivity from buoy or tracer experiments are used to calibrate the behaviour of different tracers and pollutants, both natural and man-made in the NW Mediterranean Sea [4, 5]. Thanks to different polarization and intensity levels in satellite imagery can be used to distinguish between natural and man-made sea surface features due to their distinct self-similar and fractal as a function of spill parameters, environmental conditions and history of both oil release and weather conditions[6,7].

Environmental factors determine [6,8,13] spreading, drift and weathering of oil on the sea surface, but note that coastal turbulence becomes helical influenced by the wind field. This means that the one-point correlation functions of the flow velocity. On the other hand helical and stratified turbulence is also modified by the intermittency and by the maximum fractal dimension, [5,8], which is related to the energy ( and posibly of vorticity ) spectra of the flow. Using all the available scaling information ,[12,13,14] it is possible to investigate the spatial variability of the horizontal eddy diffusivity K(x,y).[2,9] This information would be very important when trying to model numerically the behaviour in time of the oil spills [10,4,15]

[1] Platonov A., Carillo A., Matulka A., Sekula E., Grau J., Redondo J. M., Tarquis A. M. (2008) Multifractal observations of eddies, oil spills and natural slicks in the ocean surface", Il Nuovo Cimento, Vol. 31 C, N. 5‐6, pp. 861‐880.

[2] Redondo, J.; Matulka, A.M.; Carrillo, J. (2010) Vortex decay in stratified flows. Topical Problems of Fluid Mechanics 2010. Praga, AS. pp. 127-130.

[3] Castilla R., Redondo J.M., Gamez P.J. and Babiano A. (2007), Non Linear Processes in Geophysics, 14, (2007) pp. 139.

[4] J. M. Redondo, J. H. Fernando and S. Pares (1995) Cloud entrainment by internal or external turbulence, Mixing in geophysical flows, J. M. Redondo and O. Metais (Eds), CIMNE, Barcelona (1995) pp. 379-392.

[5] Sekula E., Redondo J. M. (2008)"The structure of turbulent jets, vortices and boundary layer: Laboratory and field observations", Il Nuovo Cimento, Vol. 31, N. 5, pp. 893 ‐ 907

[6] Redondo J.M. (1996) Vertical microstructure and mixing in stratified flows. Advances in Turbulence VI. Eds. S. Gavrilakis et al.(1996), pp. 605-608.

[7] Redondo J.M.(2001) Mixing efficiencies of different kinds of turbulent processes and instabilities, Applications to the environment” in Turbulent mixing in geophysical flows. Eds. Linden P.F. and Redondo J.M., pp. 131-157.

[8] Nicolleau, F.C.G.A.; Cambon, C.; Redondo, J.M.; Vassilicos, J.C.; Reeks, M.; Nowakowski, A.F. (Eds.)(2011) New Approaches in Modeling Multiphase Flows and Dispersion in Turbulence, Fractal Methods and Synthetic Turbulence. ERCOFTAC Series.

[9] Fraunie P., Berreba S. Chashechkin Yu.D., Velasco D. and Redondo J.M. (2008) Large eddy simulation and laboratory experiments on the decay of grid wakes in strongly stratified flows. Il Nuovo Cimento C 31, 909-930.

[10] Matulka, A., López, P., Redondo, J. M., and Tarquis, A.(2014) On the entrainment coefficient in a forced plume: quantitative effects of source parameters, Nonlin. Processes Geophys., 21, 269-278.

[11] Castilla R., Oñate E. and Redondo J.M. (2007) Models, Experiments and Computations in Turbulence. CIMNE, Barcelona, 255.

[12] M.O. Bezerra, M. Diez, C. Medeiros, A. Rodriguez, E. Bahia., A. Sanchez and J.M. Redondo (1998) Study on the influence of waves on coastal diffusion using image analysis. Applied Scientific Research 59,.191-204.

[13] Cuxart J., Yagüe C., Morales G., Terradellas E., Orbe J., Calvo J., Fernández A., Soler M. R., Infante C., Buenestado P., Espinalt A., Joergensen H. E., Rees J. M., Vilá J., Redondo J. M., Cantalapiedra I. R. and Conangla L.: Stable atmospheric boundary-layer experiment in Spain (SABLES98):a report, Bound-Layer Meteor., 96, 337-370, 2000.

[14] Kolmogorov  A. N.: A refinement of previous hypotheses concerning the local structure of turbulence in a viscous incompressible fluid at high Reynolds number, J. Fluid Mech. 13, 82-85, 1962.

[15] Redondo J.M., Grau J., Platonov A and Garzon G.(2008) Analisis multifractal de procesos autosimilares, Imagenes de satelite e inestabilidades baroclinas. Rev. Int. Met. Num. Calc. Dis. Ing. 24, 1, 25-48.

During the oil spills response actions, the rapid location and mapping of the oil slicks at sea surface is needed, as well as the characterization of different types of oils, avoiding false alarms whenever possible. In this way, all information provided by remote sensors in different spectral ranges can be useful to provide this type of information, improving the strategies to cleanup oceanic regions and to protect the vulnerable coastal areas, minimizing possible environmental damages.

Synthetic Aperture Radars (SAR) are widely used to detect and map oil spills at sea surface, due to their potential for systematic data acquisition covering large areas with near real time delivery capability, offering different spectral and spatial resolutions, swaths, incidence angles and polarization channels. Along the microwave spectrum, oil slicks are identified as dark spots, regions with low backscatter at sea surface. Despite SAR has been effectively used, the detection of oil spills still remains a challenge considering the ambiguities caused by different biological, meteorological and oceanographic phenomena with backscatter similar to the mineral oils, originating false alarms.

Recent studies, using several polarimetric attributes, have been investigating the potential of the polarimetric SAR (PolSAR) data in dual, quad and compact polarization modes to detect different backscatter mechanisms in biogenic and mineral oil slicks at sea surface, probably related to variations of the chemical properties, thickness and weathering.

Traditionally, the majority of the SAR data used for surveillance purposes have commonly been acquired in large modes, with wide swaths in single or dual-pol channels. Despite of the advantages offered by the full polarimetric images, this type of data are not always available for operational uses, particularly considering the unpredictable nature of the oil spills events.

In this context, the objective of this paper is to evaluate the potential and the trade-offs provided by single, dual and quad-polarized data to discriminate different types of oil slicks at sea surface. To accomplish this, a set of SAR data acquired by RADARSAT-2 [quad-polarized] and Sentinel-1 [dual-polarized] were processed and classified using a new region based classifier named PolClass.

The PolClass was designed to process SAR and optical data, using adequate statistical modeling through three processing modules: I) Multivariate Amplitude: optical and SAR images in amplitude format, assuming Gaussian distribution; II) Intensity: pair of SAR images in intensity format, considering Bivariate Gamma distribution, and; III) Polarimetric: covariance matrix in a complex format, modeled by Complex Wishart distribution. The supervised classification uses stochastic distances between statistical distributions and hypothesis tests to associate uncertainty measurements to the classification results.

This work used the Intensity and the Polarimetric modules to classify the SAR data. The Polarimetric module permits evaluate all complex information contained in the PolSAR data, and the Intensity module provides the dual and single-polarized channels assessment, making feasible to compare the advantages of using all possible polarimetric combinations.

The classifier demonstrated potential to discriminate three types of oil, being more effective to detect biogenic oils than mineral oils, returning an overall accuracy around of 95% at a significance level of 5%. Until this moment, the results were obtained considering all polarimetric information, the analyses comparing single and dual-polarized approach are in course.

The main expectation is to contribute with the scientific community to evaluate the trade-offs provided by the single, dual and quad-polarized SAR data applied to oil detection, considering as reference the needs of the operational use during both, routine monitoring and emergency situations. Finally, new controlled experiments integrating field data and different types of oils released should be done, considering different meteo-oceanographic conditions to consolidate the conclusions about the PolSAR data potential to discriminate biogenic and mineral oils.

Ocean surface winds are the main factors affecting the scatterometer backscattered radiation. In clear weather conditions, the surface wind modifies the ocean surface roughness which plays a key role in the regulation of the backscattered radiation and, in turn, in the wind estimates. However, in presence of atmospheric precipitation, the surface roughness is also modified by the impact of the raindrops imping on the surface. Therefore, modeling the ocean surface modification combining both wind and rain offers an opportunity to physically understand the contribution of the rain in the ocean backscattering coefficient. It opens the way towards a complete theoretical forward model to simulate the scatterometer observations in both rainy and non-rainy conditions once the contributions due to rain attenuation and volume backscattering are introduced. Such model allows also possible correction techniques of the wind retrievals and synergistic estimates of rain parameters as well. In this perspective, our work focuses on the development of a physical model of the ocean surface backscattering coefficient accounting for the modification of the surface roughness due to both wind and rain.

The ocean surface roughness is described by the equilibrium ocean wind wave spectrum which can be defined as the distribution of the mean waves’ energy considering the spectral sources as balanced, such as wind contributions, breaking dissipation and non-linear interactions. Several representations of the ocean surface spectrum in equilibrium range already exist and, in term of wavenumbers, they describe the ocean surface as composed by two scale of roughness such as large scale gravity waves and small scale capillary waves. Our approach consists in modifying the spectrum in the region of capillary waves in order to include the rain-induced wave damping and the generation of ring waves [Contreras and Plant, 2006]. We have modeled these effects by introducing an attenuation factor accounting for the variation of the water viscosity [Nystuen, 1990] and an additive contribution representing the increase of the surface roughness due to ring waves [Bliven et al., 1997]. The main features of our approach is that the rain model does not depend on the representation of the surface wind wave spectrum so that it can be applied to any equilibrium spectrum as long as it shows a separation between gravity and capillary waves. We have first analyze the spectrum developed by Donelan and Pierson (1987) and the preliminary results are as expected, however analysis using further spectra will follow. To compute the co-polar ocean surface backscattering coefficient, the ocean surface two scale model with the new rain affected wind wave spectrum, has been used. We have first focused on Ku band simulations in non-rainy conditions and comparisons to the empirical geophysical model function developed for the NASA scatterometer QuikSCAT show good agreement especially at vertical polarization. Then, we have focused on low-medium wind regime to analyze the ocean surface backscattering coefficient at different rain intensities. As expected, the backscattering coefficient increases when the rain rate becomes higher due to the increasing roughness and this shows that our approach is physically consistent. Analysis at C band will be also carried out. The validation in non-rainy conditions will be performed using the CMOD5.N geophysical model function developed for C band scatterometer such as European Remote Sensing Satellite-2 (ERS-2) and Advanced Scatterometer (ASCAT). Analysis of the ocean response at different rain rates as well as comparisons between C and Ku bands are planned in order to study the impact on the backscattering coefficient, at different frequencies, of the rain induced surface modification.

Monitoring suspended sediment concentration (SSC) in continental waters (river, lakes, estuaries) has been for years a question of great interest in the scientific community dealing with water resources, geomorphology, limnology or coastal environments. The development of tools brought by remote sensing allowed an increase in monitoring capacities making possible to better understand the transfer of mass through the watersheds from the mountains to the oceans. The “water color” community has developed many algorithms to retrieve SSC from water optical properties over the oceans (Case I) and coastal areas (Case II). Despite these improvements, precise optical characterization of suspended sediment is lacking, especially for the greatest watersheds in the world, such as the Amazon basin.

 

Several field campaigns have been conducted along the Amazon River, the Rio São Francisco River (Brazil) and the Volta River (Bagré’s artificial lake, Burkina Faso). Those large basins represent a great diversity of ecosystems and pedology, and offer the opportunity to cover a large range of radiometric responses and of physical / chemical characteristics of the suspended particulate matter (SPM). Field campaigns conducted in the Amazonian basin were focused on the Amazon River and its main tributaries such as Rio Solimões, Rio Negro and Rio Madeira. Apparent optical properties were measured, including remote sensing reflectance Rrs, irradiance reflectance R0-, diffuse light attenuation coefficients Kd, Ku and geometric factor Q. Simultaneously, water samples were collected in order to assess SSC, particular and dissolved organic carbon contents (POC and DOC, respectively), particles size distribution by MALVERN laser diffraction, mineralogy by scanning electron microscopy (SEM). Inherent optical properties included measurements of total absorption and dissolved absorption coefficients (a_total, a_cdom) allowing to derive particulate matter absorption coefficient a_nap using OSCAR TrioS spectrophotometer.

 

The data collected made possible to assess empirical relationships between Inherent or Apparent Optical Properties (IOPs and AOPs, respectively) and SSC, in order to better understand the optical response of the continental waters. In particular, we assessed the variability of the IOPs and AOPs between the different catchments.

 

A model based on Mie-Lorenz theory (Merhoff Mie Program) was used to support the interpretation of the field data. The model runs were parameterized using the field measurements of the optically active components (OAC) such as suspended particulate matter granulometry, mineralogy and organic content, allowing to produce realistic simulations of the river water optical properties. Sensitivity analysis was performed using the model to assess the variability of IOPs as a function of the different SPM OACs. In particular we assessed the relative importance of parameters as the imaginary part of Refraction Index (IR_imag) or the content of small particles defined with Junge’s parameter (J slope factor) describing the particles size distribution. Our results make possible to understand the AOPs and IOPs variability between the different catchments observed.

 

The methodology developed in this work, based on sampling campaigns in different large river watersheds used to parametrize an EM model represents a first step in a generalized understanding of the inland water optical properties. This approach sets the basis for developing robust retrieval algorithms that will be exploited by new spaceborne sensors such as Sentinel-2 and 3.

This paper presents the system with functions of automatic data acquisition, pre-processing and ship detection using Sentinel-1 C-band data. The module of automatice data acquisition can download automatically Sentinel-1 SAR data over the designated area through a satellite observation plan query. The data is unzipped and connected to the module of pre-processing for radiomatric calibration and geomatric correction. Then the ship detection module starts to extract a ship list with position and shape, and finally the detected targets is displyed on GoogleEarth or a ECDIS-based system.

We will apply the system to Korea and Pacific Islands countries.

An operational automated oil spill monitoring system by space-borne SAR is developed. Different from the usual workstation application systems, the system presented is a B/S architecture web application which requires the degree of automation as high as possible because most of its data processing modules silently run on the server side.

The system includes five key techniques: T1) Picking out dark targets from SAR images using a new adaptive threshold segmentation algorithm based on estimate the sea background of SAR image. The algorithm does not require SAR calibration and is applicable for SAR images of different satellites; T2) Reducing the number of dark targets by applying a filter chain; T3) Extracting features from all the remained dark targets and dynamically selecting the most useful features based on their statistical characteristic; T4) Discriminating the targets between oil spill and look-alikes by an Adaboost neural network(ANN) and Adaboost decision tree.(ADT); T5) Using an intelligent feedback with expert knowledge to continually update training data set, optimize the classifiers and therefore improve the detection rate.

Usually, there are lots of dark targets on a SAR image caused by various physical mechanisms, therefore the number of dark targets out from one wide swath ASAR image, for instance, may reach as many as 3000 even using the algorithm of T1, but most of them are obviously not oil spills. By applying the filter chain of T2, more than 90% targets are removed. All the filter rules are very weak in order to preserve the oil spill candidates. In T3, 68 features are calculated on each reserved dark targets, but only part of features are selected for training classifiers. The selecting procedure is carried out whenever the target training set is updated by T5. Recognition of oil spill is done in T4 with two classifiers, ANN and ADT. So far, the test on 6000 targets from 56 ERS/SAR, 11 Envisat/ASAR and 75 Cosmo Sky-Med/SAR images shows that the correct recognition rate by T4 can reach 88 %. The performance may be better with the continuously updating of training data set by T5 in future.

HOPE algorithm developed by Z.P. Lee was used to retrieve shallow water depth and bottom reflectance from hyperspectral data (i.e. Hyperion), which did not depend on in situ measurements. Recently, Hu and HE extended the idea of HOPE algorithm to multi-spectral data (i.e. Landsat) to retrieve shallow water depth and bottom reflectance.

Nine Hyperion L1 images between 2007 and 2011 were used to retrieve water depth and bottom reflectance around twelve reefs in the Spratly Islands. The retrieved water depth was compared with in situ measurements or C-MAP. The average percentage error is 14% in Zhongzhou Reef, Zhongy Island, Xinyi Reef and Shuanghuang Shoal.

Landsat-7/8 and MERIS full resolution images between 1999 and 2015 were used to retrieve the water depth and bottom reflectance in the same twelve reefs in the Spratly Islands. The retrieved water depth were compared with in situ measurements .The percentage error are comparable with Hyperion data.

Meanwhile, Landsat and MERIS full resolution data were used to monitor the change of water depth around the twelve reefs.

This study aims at demonstrating the ability of satellite optical images, specifically SPOT-5, as an alternative independent method for mapping ocean currents with world-wide coverage. The method for estimating currents is as follows: SPOT-5 imaging arrays are pointed forward and backward such that Panchromatic (PAN) is offset about 2 seconds from the Multispectral (MUL) images. Ocean waves propagate at speeds of several to 10 m/s (depending on wavelength). In the 2 seconds between Pan and MUL the waves have moved a measurable fraction to several pixels. The ocean current is the relative displacement between the waves in PAN and MUL after adjusting for the gravity wave dispersion velocity. Our algorithm can separate wave velocity and currents. The first SPOT-5 test image used in this study is one acquired over Le Reunion Island to illustrate preliminary results for vertical shear. Other SPOT-5 data sets are being chosen to align with availability of ground truth for verification.

Cyanobacteria form spectacular mass occurrences almost annually in the Baltic Sea. These algal blooms are the most visible consequences of marine eutrophication, driven by a surplus of anthropogenic nutrients and internal loading. A quantitative environmental indicator with set boundaries for ‘good environmental status’ (GES), enabling its utilisation in the assessment of eutrophication within the framework of the Marine Strategy Framework Directive (MSFD) of the European Union, has so far been lacking. We present a novel method for the assessment of cyanobacterial blooms by using data from Baltic Sea open sea areas. The Cyanobacterial Surface Accumulation (CSA) indicator is based on an index combining indicative bloom event characteristics, namely the length, seasonal volume, and severity of algal surface accumulations, estimated from the EO based algae bloom satellite product of the Finnish Environment Institute. A correspondence of 77% was found between the CSA index and independent observations of algal blooms, i.e. the satellite-data based Fraction with Cyanobacterial Accumulations (FCA) by Kahru and Elmgren (2014). The FCA, which constitutes the longest available spatially extensive time series on algal blooms in the Baltic Sea, was also utilized in the setting of CSA GES boundaries. The setting of GES boundaries was based on the identification of time periods in the FCA time series with similar mean values and without significant break points. The indicator assesses the status of each sea area by comparing the CSA index to the determined CSA GES boundaries. The applicability of the CSA method as a new indicator of algal blooms is discussed and the method is demonstrated to pass the general criteria set for marine indicators. A great advantage of the method is its flexibility; it is not restricted to the data used in the development, but can be complemented, or fully applied, by using different types of data sources giving information on cyanobacterial surface accumulations.

References: Kahru, M., & Elmgren, R. (2014). Satellite detection of multi-decadal time series of cyanobacteria accumulations in the Baltic Sea. Biogeosciences Discussions, 11, 3319-3364.