In this study we are aiming to evaluate the response of phytoplankton to the coastal upwelling events in the South-Eastern (SE) Baltic Sea and the Curonian Lagoon in terms of chlorophyll-a (Chl-a) concentration and phytoplankton biomass data. 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 by bringing cold, nutrient rich waters from deeper layers to the euphotic layer. Brought-up nutrient-rich waters primarily affect the pelagic communities. Particularly during the summer time, nutrients and the exposure of upwelled phytoplankton to surface radiation enhances the primary production during upwelling events.
In this work, we used Terra/Aqua Moderate Resolution Imaging Spectroradiometer (MODIS) SST maps (covering the study site with a spatial resolution of about 1 km) to characterize SE Baltic Sea coastal upwelling and its influence to the Curonian Lagoon from the distribution and evolution of its surface thermal signatures. Two different approaches were used to retrieve Chl-a concentration from Medium Resolution Imaging Spectrometer (MERIS) on board of Envisat. In the SE Baltic Sea, for the estimation of the Chl-a concentration, a FUB processor (1.2.4 version, plug-in in BEAM/VISAT software) was applied. In the Curonian Lagoon, the Chl-a concentration was estimated using a semi-empirical band ratio algorithm applied to the images after atmospheric correction with the Satellite Signal in the Solar Spectrum (6SV) code. Additionally, in situ measured Chl-a concentration and phytoplankton biomass during upwelling events was analysed in both the SE Baltic Sea and the Curonian Lagoon (in situ data originated from Department of Environment Research, Environmental Protection Agency of Lithuania).
As an example to illustrate upwelling induced Chl-a changes in the coastal waters an upwelling taking place in July-August 2008 was chosen having a sequential satellite-derived images of SST and Chl-a concentration. A detailed spatio-temporal analysis reviled the evident influence of upwelling on the Chl-a concentration, with relatively lower values in the upwelling zone compared to the ambient waters.
We also show that intensive coastal upwelling events strongly influence not only the SE Baltic coastal zone, but also impact the hydrological and ecological regime of the Curonian Lagoon. Here, a strong SST and Chl-a gradient resulting from the inflow of the upwelled water masses was identified in the northern part. It is recorded, that the Chl-a concentration in the lagoon might drop down by an order of magnitude i.e., even more drastically than in the Baltic Sea coast.
The study outcomes suggest that the satellite-derived SST and Chl-a products could be used to analyze the intensity of the coastal upwelling and its biological response, and, moreover, yield significant information about sea–lagoon interaction. Due to time difference between overpasses of MERIS and MODIS, slight mismatch between Chl-a and SST was determined. With Sentinel-3A/B that caries both instrument OLCI for Ocean Colour and SLSRT – for SST mapping, a comprehensive and more detailed analysis of coastal upwelling impact on the Baltic Sea environment is foreseen in the near future.

Area D - Board 396

Unprecedented massive landings of Sargassum are regularly registered since 2011 along the shorelines of a huge area encompassing French Guyana, the Antilles and Caribbean Sea. The phenomenon affects widely the West Indies (Guadeloupe, Martinique, Barbados ...), Dominican Republic, Mexico, etc. and many sightings have been reported. Algae arrive from the open sea as large rafts (tenths of km) after drifting over long distances in the Central Atlantic NERR, and accumulating in consolidation areas in the Brazil retroflexion current and probably the Gulf of Guinea. Compared to previous years, 2018 is far the worst year in terms of Sargassum landings on Caribbean islands shorelines. Washing-ashore has tremendous negative impacts on local populations, coastal marine ecosystems and the economy sector, especially tourism and fisheries that are severely affected.
The use of SAR technology to detect sargassum was successfully assessed using Sentinel-1 in 2015 in a project funded by the CNES. The signature of Sargassum in SAR images is an increase of the local backscattering due to an increase of the surface roughness. SAR can detect the submerged sargassum rafts with adequate wind conditions. The detection of sargassum is done visually by an SAR analyst, who brings his expertise for feature identification and can reject false alarms.
Floating macroalgae like sargassum can be detected on optical images using satellite image algorithms that utilize the spectral “red-edge” behavior of vegetation (low reflectance at red wavelengths, high reflectance at near-infrared wavelengths).
In 2018, the ESA (European Space Agency) is funding a project to implement an innovative automated service based on Earth Observation (EO) data to monitor floating Sargassum algae in the Caribbean area. The ESA project will allow the integration of observations from Europe’s Copernicus Sentinel-2 and Sentinel-3 satellites, to complement the data from ocean color sensors MODIS-Aqua for the calculation of a daily Sargassum Index.
Using synergy of optical imagery in the open sea: MSI/Sentinel-2, OLCI/Sentinel-3 and MODIS/Aqua, CLS computes a new Sargassum index called NFAI (Normalized Floating Algae Index) at the highest pixel resolution, namely 250 m for MODIS, 300 m for OLCI, and 20 m for MSI. The NFAI, inherited from Hu (2009), uses different wave lengths, better atmospheric corrections and an improved restitution of the cloud mask to reduce false alarms.
Although SAR can only detect sargassum rafts when floating at the surface, it has the capacity to observe at night and under the clouds, which is a great added value in the Caribbean region especially. Both types of sensors (SAR and Optical) are used as complementing means of observation.
The detected rafts are then advected using a drift model to predict their trajectories and give a probability of beaching on the coasts.
Sargassum invasion is becoming a regular hazard in the Caribbean and Gulf region. 2018 has been from far the most intense sargassum season in terms of sargassum detection in the area and reported beaching on the whole Caribbean and Gulf of Mexico region. The use of satellite data shall bring valuable information in support to operations & management of the various sectors impacted by sargassum landings.

Area D - Board 397

The present work is framed into the “Commercial service platform for user-relevant coastal water monitoring services based on Earth observation” (CoastObs) project whose objectives are coastal water environment monitoring based on Earth observation and distribution of validated products to end-users. One of the innovative products considered into CoastObs is seagrass mapping. Seagrass meadows are ecosystems considered as blue-carbon sequesters and reservoirs that also play other important roles for the environment, acting as nursery and habitat for a variety of marine fauna, providing sediment stabilization, regulating nutrient cycles and water turbidity and being used as coastal ecosystem health indicators. Despite their ecological importance and their recognized ecosystemic value, there are still high uncertainties in the quantification of global seagrass areas and there is a lack in comprehension of their temporal dynamics.
Nowadays, satellite data provides great opportunities for ecosystem assessment at synoptic scales in a more time and cost -efficiently way than field campaigns. For example, Sentinel-2 collects images of the Earth surface at high spatial resolution (10-60 m) with a revisiting time of ~5 days or even better in some areas. Here, we explored the capabilities of Sentinel-2 images to accurately map sub-tidal seagrass beds in the Glenan Archipelago (France). Seagrass beds in this area are monospecific, formed by Zostera marina. Other substrate types also compose the seafloor, like sand and different species of benthic macroalgae. Whatever the bottom type is, submersed substrates imply that the signal received by a remote sensor includes information not only from the bottom but also from the water column above it. To properly extract bottom information, satellite images over shallow waters have to be corrected from the influence of the water column in a way to isolate both contributions.
In this study, we used a Sentinel-2 cloud-free image acquired on 17 May 2018, 12 days before a field trip. Field sampling was performed to record at 7 calibration/validation stations used: (i) for water column characterization in terms of IOPS (absorption and backscattering coefficients) and biological information (pigment concentration and total suspended matter); (ii) to obtain radiometric measurements to further evaluate atmospheric correction; (iii) to measure percentage of coverage of underwater bottom; and (iv) to provide radiometric measurements of representative bottom types to serve as end-members during water column correction processing. The Sentinel-2 scene was downloaded at Level-1C and processed by using atmospheric corrections routines. Different atmospheric processing processors were applied at their default settings: iCOR, ACOLITE, C2X, C2RCC and POLYMER. We evaluated their performance over deep waters. Among them, we selected C2RCC as it showed the best performance at the lowest computational time of processing. For the water column correction we applied the Shallow Water Semi-Analytical Model (SWAM), a publicly available tool distributed as part of the Sen2Coral plug-in in ESA’s SNAP software. SWAM implements the inversion of a bio-optical model and make use of optimization techniques to assess bottom properties from Sentinel-2 images. In particular, SWAM generates maps of bottom depth and bottom distribution of three substrate classes. As SWAM is parameterized based on literature data collected in tropical coral reefs, here, we adapted SWAM parameterization to the benthic classes found in our study areas using in situ Reflectances collected in the Glénan Archipelago: seagrass, brown macroalgae and sand. SWAM retrieves at the same time and in each pixel: bottom depth (z), substratum detectability index (SDI), error reached during optimization, diffuse attenuation coefficient at 550 nm (Kd(550)), bottom reflectance at 550 nm (Rb(550)), and percentage of coverage of the three end-members. Pixels with SDI<5 were masked out considering that bottom contribution in those pixels was negligible.
In general, the results show that seagrass coverage had higher concentration in the center of the archipelago, with coverage of ~50%, with some patches surrounded by the bigger islands where abundances reach ~70-80%. These denser patches coincided with very low abundance of macroalgae. Higher abundance of macroalgae was found Northway the archipelago with abundance 50-80%. Sand, on the contrary, was predominant outer the archipelago, where hydrodynamics were higher. In terms of z, the whole area looked homogeneous (0-8 m), but deepest areas reached up to 20 m. However, such high depths (z > 10 m) were found around the islands, coinciding also with 100% percentage of macroalgae coverage and clear waters (in terms of low Kd(550)). Even when in situ macroalgae distribution in those areas agreed with our estimative from satellite, it is important to consider that Sentinel-2 is a Multiresolution satellite and the SWAM inputs included information from bands at 10 m (Bands at 490, 560 and 665 nm), 20 m (705 nm) and 60 m (443 nm) of spatial resolution. It means that thin boundaries around the islands could include some contamination from terrestrial targets, leading to wrong retrievals. To overcome to this multiresolution issue, next runs should be performed considering only bands with 10 m of spatial resolution and/or improvements in land mask application.

Area D - Board 398

The Estonian coastal areas are shallow and have a soft bottom, thus the optical properties of these waters are influenced by particles coming from the rivers and from resuspension by storms. The Baltic Sea is with brackish water, resulting with quicker mixing of sea water with incoming fresh water. The Pärnu River, the largest river in Estonia, flows into the Pärnu Bay in the Baltic Sea. River inputs induce a sediment plume visible from satellite data. Rivers also transport large amount of coloured dissolved organic matter (CDOM) into the sea which brings the need to use different approaches for retrieval of suspended matter concentrations from optical signal.
The sediment outflow may have an important impact on the ecosystem and for the management of the bay. The characteristics of the optically active substances (OAS) in Pärnu Bay are highly variable and there is a lack of investigation concerning their dynamics in CDOM dominated ROFI (Region of freshwater influence).
The Pärnu Bay has been part of monthly monitoring program where samples are usually taken at 5 sampling stations that haven’t changed in years. Although these stations provide relatively good overview about spatial variations in the bay, it is not the best option currently available and it is not providing smaller scale temporal changes.
Copernicus mission satellites Sentinel 2 and Sentinel 3 are both suitable to collect optical data from the bay and can provide details about spatial and temporal changes. To test different previously published algorithms for retrieving OAS one must trust the absorption (a) and backscattering (bb) values retrieved from the reflectance.
During this study Quasi-Analytical Algorithm (QAA) model by Lee et. al (2009) was applied to in situ measured reflectances. There were two different set-ups for radiometric measurements: One set-up used three TriOS RAMSES approach, where irradiance sensor was looking directly upwards and radiance sensors were aimed at the sky and water at 40 degree angle; the second was a handheld instrument Water Insight WISP-3 that uses the same viewing geometry but smaller radiance viewing angle. Prior to the fieldwork season, all of the instruments were calibrated in the same lab by the same person. First the QAA was run in default settings and then it was tuned with in situ inherent optical properties (IOPs) to see how big of an improvement can be achieved. The in situ IOPs were measured with WET Labs AC-S, ECO-BB3 and ECO-VSF3 and corrected following strictly the user manuals.
The tuned QAA was then used to retrieve IOPs from Sentinel 2 and 3 products to analyse the temporal and spatial variability of these parameters in Pärnu Bay.

Area D - Board 399

A number of studies expect an increase of carbon export by rivers to the Arctic Ocean due to rapidly changing climate in the Arctic (Camill, 2005; Freeman et al., 2001; Frey and Smith, 2005). One possible reason for the increase of carbon export is thawing permafrost which can lead to a mobilization of previously frozen dissolved organic matter (DOM). The Lena River delivers about one fifth of all river water to the Arctic Ocean and is the main source of DOM in the Laptev Sea shelf (Thibodeau et al., 2014). The discharge of the Lena River and the DOM concentration shows an extreme seasonality. Scarce measurements of DOC and the colored fraction of DOM (CDOM) (< 8 samples/year) were previously used to estimate fluxes to the Arctic Ocean for the whole year based on a proportional connection between DOM and discharge (Stedmon et al., 2011). Such estimations can potentially result in large errors. Longer sampling periods and higher sampling frequency are needed.
Here, we present a new sampling program and its viability to monitor export fluxes of DOM in the Lena River. The sampling takes place at the research station Samoylov in the central Lena River Delta. The Samoylov research station allows a unique chance for continuous sampling since it is operated and manned throughout the whole year. The sampling program includes measurements of several water parameters, such as temperature, conductivity, DOC, spectral CDOM absorption (aCDOM), dissolved and total nutrients, and stable isotopes every 4 days. Preliminary results based on first five month samples (April-September 2018) show a strong peak during the spring freshet in the beginning of June where DOM concentration is more than 3 times higher than during the winter flow. DOM concentration follows the trend of the discharge during the rapid increase in June. After rapid decrease of discharge in mid of June, DOM stays on an elevated level and decreases continually during summer towards a stable winter level. The resulting data is used to evaluate the potential of using Ocean Color Remote Sensing data (OLCI, Sentinel 3) to track seasonal variation by comparing remote sensing with in situ data. The comparison of aCDOM extraction from Ocean Color data shows a good correlation to the in situ data. Furthermore, optical remote sensing data are available from 2002 to present (MODIS Aqua) and an accurate retrieval of DOC concentration in the Lena River could enable accurate trend analysis of DOC fluxes for the last two decades.
References:
Camill, P.: Permafrost Thaw Accelerates in Boreal Peatlands During Late-20th Century Climate Warming, Clim. Change, 68(1–2), 135–152, doi:10.1007/s10584-005-4785-y, 2005.
Freeman, C., Evans, C. D., Monteith, D. T., Reynolds, B. and Fenner, N.: Export of organic carbon from peat soils, Nature, 412(6849), 785–785, doi:10.1038/35090628, 2001.
Frey, K. E. and Smith, L. C.: Amplified carbon release from vast West Siberian peatlands by 2100, Geophys. Res. Lett., 32(9), L09401, doi:10.1029/2004GL022025, 2005.
Stedmon, C. A., Amon, R. M. W., Rinehart, A. J. and Walker, S. A.: The supply and characteristics of colored dissolved organic matter (CDOM) in the Arctic Ocean: Pan Arctic trends and differences, Mar. Chem., 124(1–4), 108–118, doi:10.1016/J.MARCHEM.2010.12.007, 2011.
Thibodeau, B., Bauch, D., Kassens, H. and Timokhov, L. A.: Interannual variations in river water content and distribution over the Laptev Sea between 2007 and 2011: The Arctic Dipole connection, Geophys. Res. Lett., 41(20), 7237–7244, doi:10.1002/2014GL061814, 2014.

Area D - Board 400

The launch of Sentinel-2A mission of European Space Agency (ESA) in 2015 was a real innovation in monitoring coastal and estuarine water quality. With the Copernicus Program, oceanographic investigations have now significantly benefited from continuous, multispectral satellite products of increased accuracy, thanks to their high spatial-temporal resolution.
Suspended Particulate Matter (SPM) is an important water quality parameter, especially in front of riverine outflows, which can be derived from empirical or analytical algorithms by using the remotely sensed reflectance Rrs(λ) retrieved from satellite imagery.
The aim of this work is to assess the performance of ocean color algorithms in comparison with in-situ SPM and reflectance (Rrs) measurements collected in the outer Thermaikos Gulf, in Northern Greece, during low discharge period of Pineios river (June 2018). For this purpose, both a Semi-Analytical and a Multi-Band Empirical SPM Algorithm are tested, in order to estimate the SPM values by computing remotely sensed reflectance from Sentinel-2A imagery.
When comparing the satellite estimations against the field measurements, both algorithm approaches provide good correlation with in-situ Rrs (~20 % offset). As a result, a generic semi-analytical equation for Pineios estuary is developed, following algorithm calibration in low turbidity waters. The Leave-One-Out-Cross-Validation (LOOCV) method also confirms the reliability of the model, and the estimated SPM explains 67–98% of the variation in the measured SPM.
Overall, the present study suggests the applicability of Sentinel-2A for SPM retrieval in the outer Thermaikos Gulf, and the Sentinel 2A B4-B5 are recommended for low SPM loads. As a result, a generic semi-analytical equation for Pineios estuary is developed, followed by a calibration in low turbidity waters. The proposed algorithm can be then equally implemented to Sentinel-2A estimations, in order to assess its variations against previously tested models.

Area D - Board 401

We demonstrate how frequent (5 day), high spatial resolution (10 m) and multi-spectral satellite observations can be used for seasonal monitoring of bioindicators from space. Specifically focused on seagrass in South East Tasmania, we show how over a 6 month period it’s possible to use multiple Earth Observations to produce a ‘confidence map’ for targeted seagrass monitoring. By providing habitat, food sources, fish nurseries and requesting carbon, seagrasses are an important bioindicator for estuarine and coastal marine estuaries. Measuring the density of seagrass populations helps us to answer the question: How healthy are our estuaries? However, targeted and frequent monitoring programs of these important bioindicators are currently not in place for coastal marine environments in the Bruny bioregion of southeast Tasmania (Australia).

Our research shows it's possible to derive multiple seagrass maps by applying artificial neural network (ANN) classifiers to seven, cloud-free Sentinel-2A and B observations over the Pitt Water estuary over a six month time period. This study site is an internationally significant Ramsar protected wetland that has not been mapped for over a decade. It is also a complex, highly dynamic estuarine environment that serves as a robust test case for supervised classification algorithms. Specifically, we calculate a total seagrass area of 3.0 km2 ± 0.6 km2 for the time period from March to August 2018, which represents a reduction of 48% from 2005 to 2018. Of the seven classification maps produced, the highest user, producer and overall accuracies for the seagrass class were 92.21%, 98.25% and 98.50%, respectively, and the lowest were 55.21%, 75.87% and 90.03%, respectively.

These multiple earth observations were then collated to form a 'confidence map' giving targeted analysis of seagrass which takes advantage of frequent satellite observations over a particular study site and accounts for differences between monthly seagrass classification maps. This approach improves upon the more traditional method of land cover classification analysis where only the highest quality satellite image is retained for classification and other observations are discarded. By combining classification maps from multiple observations over a given time period we are able to give an assessment of the confidence or not certain areas are likely to contain seagrass. Over time, increases in the amount of data used to calculate the confidence map will only improve it’s overall robustness.

This study shows that by utilising state-of-the-science satellite technology we have the ability to frequently monitor bioindicators from space, which can inform government policy, industry actions and the general public so that agricultural farming practices and urban developments are designed to exist ‘within' environmental systems rather than adversely affecting them. By bringing frequent and accessible seagrass data to the user, we demonstrate the added value and unique information provided by space borne data for a better coastal management strategies.

Area D - Board 402

Coastal zones are of major economic, social and ecological importance. In particular, the coastal plains and the active estuaries, where rivers or streams flow into an open sea, have a high potential for agricultural production and are rich in the diversity of fish, hence, leading to increased human activities. Shorelines in estuarine areas are vulnerable to anthropogenic disturbance including urbanization, land reclamation and sediment loading. Shoreline changes may be a reflection of the degradation of coastal ecosystems because of human activities. Understanding the shoreline dynamics is therefore a topic of global concern. Earth observation (EO) data has been widely applied to analyse changes in coastal ecosystems. However, until recently, multi-temporal satellite image analysis was difficult due to limited availability and accessibility of data and process automation. Recent cloud-based solutions like Google Earth Engine (GEE) provide viable alternatives for analysis of large amounts of EO data. In this research, we used GEE to monitor and map historical shoreline dynamics in the Hangzhou Bay in China where Qiantang River flows into the East China Sea. Specifically, we aimed to capture and quantify both the spatial and temporal shoreline changes and to assess the long-term influence of anthropogenic activities on the integrity of this coastal area. We implemented tasselled cap analysis (TCA) on Landsat imagery from 1985 to 2010 in GEE to calculate the wetness coefficient. We then applied Otsu image segmentation method on the wetness coefficient to detect water bodies and shoreline changes. Further, we adopted the nighttime light data from Defense Meteorological Satellite Program’s Operational Linescan System (DMSP-OLS) from 1995 to 2010 as a proxy of human activities. The results show that the shoreline moved about 5 km and up to 350 km2 of new land was reclaimed from sea by human only around Cixi City in the southern part of Hangzhou Bay. The results of this work reveal the influence of human activities on the shoreline dynamics and can support policies that promote the sustainable use and conservation of coastal environments. Our methodology can be transferred and applied to other coastal zones in various regions and of different sizes.

Keywords: Shorelines dynamics, land reclamation, anthropogenic activities, Google Earth Engine, Earth observation

Area D - Board 403

Numerous satellite sensor systems have been launched during the last twenty years and satellite data are increasingly being used in regional or global mapping and monitoring. The observation from multiple satellites requires much effort to ensure continuity and compatibility due to differences in sensor characteristics and product generation algorithms. More recently the launch of Landsat-8 OLI and Sentinel-2 make the compatibility problem even more difficult as the very new spectral bands that both sensors use, need to be simulated and tested before proceed to any further comparison. Both sensors quantize the acquired information in 12-bit and thus they both present an increased radiometric resolution in comparison with the classic 8-bit sensors like Landsat-7 or previous. Furthermore, the new and quite narrow spectral bands of Sentinel-2 like bands five or six with a bandwidth of 15 nm are more similar to the bands of hyperspectral sesnors like Hyperion than to the classic bands of multispectral sensors.
The specific study focused on the automatic extraction of instantaneous coastline, as identified in the Sentinel-2 and Landsat-8 OLI imagery without taking into account the tide. The shoreline identification was based on grouping all the pixels from the imagery into only three classes: sea, land and vegetation. The classification results were achieved using two very well-known indices: The Normalized Difference Vegetation Index (NDVI) and the Normalized Difference Water Index (NDWI). The study has focused on the different spectral reflectance of water compared with that of the land environment and especially the man-made constructions. In the past the specific indexes (NDVI and NDWI) have been used frequently for water body mapping using medium resolution multispectral images. NDWI was introduced about twenty years ago aiming to separate land surface water and vegetation in Landsat TM images. Both NDVI and NDWI were initially designed to map surface water body in Landsat TM and ETM+ images but there are adaptations of them for use with other sensor data.
A comparison of multispectral data from Landsat 8 OLI to the respective data from Sentinel-2 for automatic coastline line extraction is performed and the results are presented in this study. The behavior of the two above mentioned indexes, although they have a similar algebraic structure, they may not produce similar results when applied to data from different sensors (like OLI or Sentinel-2), is examined in qualitative and in quantitative terms. For the broader area of Patras in Western Greece, the coastline was mapped using UAV images structure from motion photogrammetry. As the specific data set presents a very high fidelity it was used as a reference map in order to assess accuracy of the results derived from Sentinel-2 and Landsat-8 OLI image classification. The very encouraging results and the freely availability of Sentinel-2 and Landsat OLI assure that the presenting methodology can be applied in many areas around the world.

Area D - Board 404

Mangrove forests are coastal wetlands that contribute to biodiversity and act as major biogeochemical link between upland and coastal regions. For this reason, they play a major role in the costal ecosystem and sea coast conservation. Classification of mangrove species can be very important in obtaining an inventory for use in the development of conservation management plans. However, the classification, typically involves in intensive field surveys, and this can be very expensive and hard to be performed.

Within this context, microwave remote sensing, and in particular the Synthetic Aperture Radar (SAR), can be a very useful and relatively cheaper tool for mangroves observation, since its guarantees all day and almost all weather synoptic observations of the Earth’s surface.

However, mangroves observation using single-polarization SAR measurements is not at all an easy task and recently, the development of SAR sensors with polarimetric capabilities boosted the development of new and very robust information extraction methods. Hence, the main goals of this study are to develop polarimetric methods to classify scattering mechanisms that characterize mangroves using full- and dual- polarimetric SAR data.

For this purpose, methods based on polarimetric model decompositions (i.e. Cloude Pottier and Freeman Durden decomposition) are used in order to detect and classify different mangrove species together with the Pauli phase which, carrying on information on the electric properties of the observed scene, is expected to discriminate among the different mangrove types.

Experiments are performed processing a set of SAR data collected over the Mai Po Marshes Nature Reserve, located in Hong Kong, by the C-band Alos-2 mission. These data sets are complemented by optical imagery collected by Worldview-3 mission and by a land-cover map provided by the Agriculture, Fisheries and Conservation Department (AFCD). Preliminary results show that polarimetric information results in an added-value classification of the area covered by mangroves.

Area D - Board 406

Coastal processes (tidal currents, storm surges, waves) are highly dependent on bathymetry and directly impact offshore and coastal activities and studies. Many studies and applications lie on a growing modelling effort of the ocean and the limited accuracy of bathymetry, especially on the continental shelves, contributes to degrade numerical model performance despite significant use of in-situ and satellite measurements assimilation. In particular, the tidal models are very sensitive to the bathymetry accuracy on the shelves, where the ocean tides show the largest amplitudes and are strongly non-linear. The increase in the grid resolution, together combined with local model tuning, is one of the means to improve the tidal model performance in the coastal regions and large improvements have been achieved thanks to this approach. However, increasing the resolution of the model grid implies consistent bathymetry quality and accuracy, which is today the main limiting factor to high resolution tidal modelling.
This has, for example, a direct impact on the quality of the satellite altimetry sea surface heights as the tide correction is one of the largest corrections on the shelves, ranging from several centimetres to several metres. With new and future altimetry missions reaching ever more coastal areas and resolving the ocean dynamics at ever finer scales, the needs for high resolution accurate tidal model solutions are more challenging than ever.
Moreover, better knowledge of the tides not only improves the quality of the altimeter sea surface heights, but also of all derived products, such as the altimetry-derived geostrophic currents, the mean sea surface, the mean dynamic topography, the bathymetry and the geoid. It is also of particular interest for boundary conditions of high resolution ocean circulation modelling on the shelves and accurate tidal models are highly strategic information for ever-growing maritime and industrial activities in the coastal regions.
Various sources of bathymetry data exist but many regions remain not well known because of too sparse measurements, data access limitation or large temporal variability of the seabed dynamics. In this context, CNES funded a project that aims to improve the bathymetry in the North-East Atlantic Ocean and on the Maritime Continent (South-East Asian Seas) continental shelves. The work was divided in several steps: 1) an inventory of existing datasets and methods to derive the bathymetry on the shelves; 2) the integration of the collected datasets into a reference global bathymetry dataset; 3) the evaluation of this new bathymetry dataset through hydrodynamic modelling and the production of regional tidal models.
This paper will present the main results obtained within this project.

Area D - Board 408

Detailed and updated information about coastal morphology and bathymetry is essential in a broad range of applications such as management, research, environment, economy or policy. Because of coastal areas are subject to continuous pressures (e.g. human-induced alterations or storms), there is a need of monitoring techniques that allow to measure the impact of these changes in a fast and accurate manner. Bathymetry estimated from optical satellite imagery has been increasingly implemented as an alternative to traditional bathymetric survey techniques. The availability of new sensors such as Sentinel-2 with improved spatial and temporal resolution, in comparison with previous optical sensors, offers innovative capabilities for bathymetry derivation. This study presents an assessment of several methodologies to derive bathymetry from Sentinel-2 data such as 1) empirical algorithms and 2) physics-based approaches. These methodologies have been tested in several study areas of the Irish coast with different morphological and environmental conditions. Results showed the potential of Sentinel-2 to derive bathymetry with coefficients of determination, R², ranging between 0.65 and 0.88. Depth values could be efficiently derived between 5 m and 10 m depending on the study area. Depth estimations varied in the different ranges considered (0 m - 2 m; 2 m - 4 m; 4 m - 6 m, 6 m - 8 m; 8 m -10 m) being shallow waters, not influenced by the bottom type, where the highest accuracy was obtained.
The evaluation of several atmospheric correction processors: Sen2Cor version 2.4, ACOLITE, iCOR and C2RCC was also performed through their ability to produce linear relationships between ratios of log transformed bands and depth, as the degree of light attenuation in the water column is wavelength dependent. The processor that showed the highest and the most consistent R² values in all the images was the C2RCC processor. The important reduction of sun glint by this processor could be the final responsible of these high values.
Atmospheric correction, bottom type influence and water column conditions proved to be key factors in the bathymetric derivation using these satellite datasets.

Area D - Board 409

As water depth greatly affects a range of coastal phenomena such as waves or tides, its precise knowledge is crucial for a better understanding of coastal dynamics in general. The paper presents a novel method that computes high-coverage and high-resolution (~50m) coastal bathymetry using satellite imagery alone. The approach consists of the fusion of two well-known methods, both of which rely on linear wave theory. The first one, called Wavenumber Transformation Algorithm (WTA), uses wave tracking to estimate both the wave period and the wavelength from the open ocean to the coast, from which the coastal bathymetry is then derived [Danilo, Melgani, 2016]. The second one, Wave Kinematics Bathymetry (WKB), exploits the short time interval (a few seconds) between two subsequent acquisitions [Abileah, 2006].
Compared to satellite-derived bathymetry based on reflectance, these two methods are completely independent of water turbidity, which constitutes a major advantage. However, the resolution of each method alone is comparatively lower. Our objective is to combine the advantages of WTA and WKB in order to increase both resolution and spatial coverage.
The idea of a fusion approach is justified by the complementarity of the two methods. The advantage of WTA is the ability to produce an estimate close to the shore using a dynamic approach following the waves. These estimates, however, are only obtained along wave traces, which results in a reduced spatial coverage. Besides, for the calibration of the method it is crucial for wave traces to cover deep sea areas. WKB, in contrast, estimates bathymetry on a regular grid. However, due to its dependence on the availability of two consecutive acquisitions, the resolution of the result (i.e., the lowest observed wavelength) is determined by the lower-resolution image. For example, when using two channels of Sentinel-2 at resolutions resolutions of 10 and 20m, wavelengths under 40m will not be observed. In consequence, the method is not able to retrieve water depths as close to the coast as WTA.
We test the fusion method using Sentinel-2 acquisitions of four different areas: Georgetown, Guyana; Waimanalo, Hawaii; Perth, Australia; and Newport, Oregon, all of which offer ground truth measurements for comparison.

Danilo, C. and Melgani, F., 2016. Wave period and coastal bathymetry using wave propagation on optical images. IEEE Transactions on Geoscience and Remote Sensing, 54(11), pp.6307-6319.

Abileah, R., 2006, May. Mapping shallow water depth from satellite. In Proceedings of the ASPRS Annual Conference, Reno, Nevada.

Area D - Board 410

Remotely sensed technology is considered a time-effective, low-cost, and wide-coverage solution for Satellite-Derived Bathymetry (SDB), which can be regarded as one of the most promising alternative tools to map bottom depths in order to advance many coastal and inland water applications. Bathymetric information of shallow zones is key to the world’s management of coastal environments, economies, and cultures, yet at present, we still have incomplete and spatially limited coverage, especially in environments that have turbidity. Even in optically shallow areas, where bottom features are visually detectable in imagery, water clarity has a significant and varying impact on the accuracy of SDB. Suspended particulate matter, that is the main contributor to turbidity, introduces a confounding reflectance of the light detected by satellite. As a result, turbidity plumes reduce the sensible depth and can be identified as false shoals. Additionally, waters of different turbidity levels scatter the incoming radiation differently, involving further complexity and development issues over highly dynamic coastal regions. Taking into account that water turbidity is the most challenging constraint to the routine application of SDB worldwide, further research is required to evaluate, and compensate for, its impact on SDB.
This study examines the high-resolution MultiSpectral Instrument (MSI) onboard Sentinel-2A and 2B for generating bathymetric maps along the east coast of the United States over three moderately turbid environments in New England, North Carolina and Florida. The MSI has particular value for SDB as it can provide bathymetric maps with 10 m pixels every five days, with clear skies. This study aims to assess the constraints of water clarity, where inspection is addressed by means of a temporal analysis involving reflectance in the red-edge near infrared (NIR) region at 704 nm as a proxy for turbidity as well as the standard OCx algorithm as a chlorophyll surrogate. A conventional ratio transform model is applied, which only requires two calibration parameters for vertical referencing. Whereas chart soundings are used for the calibration of the model, high-resolution lidar surveys are utilized for validation and error analysis. Atmospheric correction of imagery is implemented thought ACOLITE processor providing good performance and repeatability over different Sentinel-2A/B images at the three control sites. A multi-scene strategy involving temporal analysis of imagery during one-year period is established for a turbidity correction over spatially and temporally heterogeneous waters, retrieving SDB with median errors of 0.5 m at depths ranging between 0-12 m (limit of lidar surveys). The results yield discrimination of the impact of turbidity and chlorophyll on SDB model performance, demonstrating the usefulness of the 5-day revisit of Sentinel-2 twin satellites to enhance our capability to map complex geographic regions with compensation for turbidity in a robust and reproducible way. Areas with residual uncertainty can also be identified, and that uncertainty estimated. The method is applied to regions impacted by severe hurricanes and winter storms. The findings provide insights that can help develop turbidity corrections, improve synoptic bathymetric estimation, and increase the operational use of SDB over nearshore environments with varying water transparency conditions.

Area D - Board 411

There is an increasing interest on the knowledge of costal topography and bathymetry for wide purposes, such as the assessment of natural hazards events or coastal planning and management. In fact, coastal areas are one of the most dynamic areas of Earth, in constant modification due to external factors acting in different spatial and temporal scales. Natural short period and high energy processes, such as severe storms and long term processes, such as the mean sea level rise, are the cause of progressive changes on the coastal morphology over time. Monitoring coastal areas can provide updated topo-bathymetric models and to predict further impacts on morphodynamics.
The waterline method combined with the tide height is the most used method to generate the intertidal topo-bathymetric model and can be used both with Synthetic Aperture Radar (SAR) and multispectral images. The waterline is the instantaneous boundary between the water body and the exposed land at the time of the image acquisition. The height of the waterline is predicted by a hydrodynamic tide model. The temporal variability of several external parameters, such as the winds, the waves or the amount of organic suspended particles that have a direct influence on the intertidal surface reflectance or backscattering, and consequently on the threshold, are used to derive the waterline. Therefore, the accuracy of the waterline is dependent on the weather and sea state conditions at the image acquisition time, which may be different along the time frame.
In this work, we explore a different approach, proposed by Catalao & Nico (2016), based on the exploitation of the spectral reflectance variability (backscattering for SAR images) along a time series of images and its relation with the tide height. The method relies on a pixel-based algorithm and uses a logistic function to model the temporal variability of the pixel’s backscattering SAR images or the spectral reflectance multispectral images.
As our interest resides on the intertidal area, the temporal variability of the intertidal surface is used to discriminate between water, intertidal zone and the beach berm. The three classes are identified by the temporal behavior of SAR backscattering. The intertidal zone is characterized by a high temporal variability of the backscattering coefficient. The other two classes are characterized by a low temporal variability of the backscattering coefficient when compared with the intertidal areas. The temporal average of backscattering coefficient has been used to discriminate between water and beach berm classes. In fact, the beach berm class shows very high intensity values for most of the pixels, while the water class is characterized by lower backscattered intensities.
For multispectral images the temporal variability of the Normalized Difference Water Index (NDWI) is used to discriminate between the three classes. The spectral behavior of water is characterized by the absorption at wavelengths larger than the infrared (IR). Thus, using the near-IR band (NIR – B8), it is expected that the pixels of the intertidal zone have a high temporal variability due to the tide regime. The water class is characterized by a low temporal variability and a high value of NDWI when compared with the intertidal zone. The beach berm class is also characterized by a low temporal variability, but unlike water, the value of NDWI is expected to be less than the value for the intertidal zone. The intertidal zone is characterized by a high temporal variability.
The pixel’s height estimation was made by minimizing the cost function and a saturation index has been proposed to filter out false intertidal pixels. In this work, saturation index values above 0.4 identifies pixels belonging to the intertidal zones. The output of the algorithm is a set of geolocated pixels with latitude, longitude, and terrain height referred to the local tide reference system. The 3-D cloud of points are then interpolated to generate a regular grid of elevations, creating a DEM of the intertidal zone. The main advantages of the presented method lies on the possibility of being applied to multispectral and SAR images, in a fully automatic way, without manual editing by the operator. This methodology may contribute to the operational mapping of vast intertidal areas and is useful for the regular updating of the intertidal bathymetric models and contribute for flood hazard mitigation. Comparisons between the SAR and the multispectral DEM generated for intertidal areas are statistically presented in the study. The proposed methodology has been applied to a time-series images of Sentinel-1 (C-band) and Sentinel-2 (MSI), acquired over Tagus estuary (Lisbon region) and a hydrographic survey was used for validation purposes.

Area D - Board 412

Intertidal flats lie at the land-sea interface between tide marks. These systems are highly dynamic since their morphology is shaped and affected by many factors, such as hydrography, geology, climate, and vegetation. Monitoring the topography and generation of up-to-date DEMs of these systems is fundamental but limited by the poor accessibility and financial limitations. Remote sensing techniques can overcome these limitations due to their relatively low cost, the continuous and homogeneous observations, and the reasonable accuracy. In this study, multi-sensor remote sensing data were used to measure the topography of the intertidal area in the Arcachon Bay. Radar altimetry observations intended originally to measure ocean surface topography were proven to be able to follow topographic variations of intertidal flats at low tides. By comparing the results of different altimetry missions, the evolution of performance between past and current missions was assessed. New altimetry missions operating at Ka-band or using SAR mode showed significant improvements. Very good estimates were obtained using SARAL and CryoSat-2 with RMSE lower than 0.44 m for both missions. In contrast to altimetry observations that measure bathymetry profiles along the ground-tracks, SAR and multispectral observations were used to generate gridded DEMs covering the whole bay. Using the established “waterline” method, a DEM was generated over the Arcachon Bay using Sentinel-1 and Sentinel-2 data coupled to tide gauge records. The method involves detecting the land-sea (tidal flats-sea) boundary using image processing techniques and superimposing the sea level (known from the tide gauge) to this boundary. From a series of images covering the whole tidal range, a set of leveled waterlines is obtained and then interpolated to form a DEM (Mason et al., 1995). The generated DEM is then compared to an accurate lidar-derived topography. A sensitivity study of the waterline method was performed as well for the Arcachon Bay in order to understand the impact of interpolation, the number of satellite images used, and the period of image acquisitions, on the DEM accuracy. Using SAR and optical data archives, a time series of DEMs can be generated, which is a very useful tool for sediment budget estimation, and the estimation of water volume exchange between ocean and bays.


Mason, D. C., Davenport, I. J., Robinson, G. J., Flather, R. A., & McCartney, B. S. (1995). Construction of an inter‐tidal digital elevation model by the ‘Water‐Line’Method. Geophysical Research Letters, 22(23), 3187-3190.

Area D - Board 413

The surf zone is the region extending from the seaward boundary of wave breaking to the limit of wave uprush. Within the surf zone, wave breaking is the dominant hydrodynamic process. The surf zone is characterized by rapid temporal changes, large variations over short spatial scales, and often-violent conditions. Traditional in situ sensors are poorly suited to sampling this world, but remote sensors offer potential solutions. Optical imagery at high resolution is certainly one of the remote sensing techniques capable of getting signatures of the previous processes. In this work, we use imagery from Pleiades optical satellite acquired in persistent mode along coastlines around Capbreton in Landes (France) to understand and characterize, the evolving ocean processes occurring in the surf zone, e.g., surface signatures due to sand bars, mapping of bathymetry, currents, rip tides, motion of cat paws over ocean surface, currents from the drift of fishing nets, etc. A sequence of images, separated 9-10 s can resolve smaller features. The challenge is to develop methods to extract from optical imagery valuable information to be used in nearshore prediction models for full surf zone characterization. Comparisons with Sentinel2A/2B imagery will be also done. Whereas with Pleiades we can measure displacements on the order of half meter, with Sentinel we measur displacements with resolution of 10/20 meters according to the band used.

Area D - Board 414

Monitoring ocean circulation at high resolution in both space and time is of paramount importance for understanding and modelling the ocean-atmosphere climate system, especially in coastal areas. Spaceborne radar altimeters have been used to successfully monitor ocean circulation on a global scale (>30km) in the deep ocean when the geostrophic approximation is generally valid. The ocean structures seen in high-resolution satellite measurements at meso (10-100km) and sub-mesoscale (<10km) are ubiquitous but little is known about their dynamics. During the last two decades, many studies have highlighted the key role played by the ocean submesoscale in air-sea interactions, upper-ocean mixing and ocean vertical transport and the importance of ageostrophic circulation in these processus. Understanding these smaller currents is critical to drive scientific understanding of the exchanges of gas, heat, and momentum between the atmosphere and the ocean, and have important implications for forecasting models and climate projections.
A satellite mission concept, called SEASTAR, aims to study ocean submesoscale dynamics and small-scale atmosphere-ocean processes in coastal, shelf and polar seas by providing simultaneous measurements of current and wind vectors at 1km resolution with accuracy of respectively 10cm/s and 2m/s. OSCAR (Ocean Surface Current Airborne Radar) is the demonstrator for this satellite concept, and is in development at the European Space Agency (ESA) in the frame of its preparatory activities for ocean surface current retrieval with Metasensing as a prime contractor. OSCAR system will be representative of a satellite mission concept, observation parameters and accuracies directly relate to a potential satellite mission.
OSCAR is a Ku-band (13.5 GHz) three-look direction SAR system with Along-track Interferometric (ATI) SAR and scatterometric capabilities. It is tailored to the measurement of 2D ocean surface motion and wind retrieval. The OSCAR features an along-track interferometric (ATI) baseline for two fields of view that are squinted 45° fore and aft from the broadside direction. Each of these views provide an interferometric measurement and these two views are angularly separated by 90 degrees, ensuring two independent measurements of the ocean surface move-ment velocity and eventually enabling retrieval of the 2D ocean surface movement vector. In addition, a broadside antenna will ensure measurements in the zero-Doppler direction to improve the retrieval sensitivity to wind direction, which is critical to retrieve an accurate ocean surface current. Indeed the motion sensed by the microwave radar (after correcting for navigation and geometry) has two components: the total ocean surface current – consisting of all currents contributing to actual horizontal transport of water – and an unwanted measurement bias associated with wind-waves (known as wind-wave induced artifact surface velocity — WASV; see Martin et al., 2016). The WASV is un-derstood to be mainly caused by the phase velocity of the surface scatterers responsible for the microwave backscatter (e.g. Bragg waves) and the effect of the orbital motion of longer ocean waves and is at first order function of the wind direction. Fully polarimetric observations are possible and the instrument is designed to be flexible in terms of the operational parameters and configuration, including the ATI baseline length, which is adjustable to be fully repre-sentative of the potential future spaceborne mission.
A dedicated system performance analysis tool has been implemented in order to derive all the instrument and obser-vation specifications to meet the requirements on the ocean surface observables. The key performances, which are driving the instrument are a measurement accuracy of 2D ocean surface motion of 5 cm/s (for velocities below 50cm/s and wind speed > 3m/s).
The performance tool is based only on numerical models without assumptions and calculates the reflected power and the Doppler characteristic of each backscattering cell of the illuminated sea surface in dependency on the radar properties, the acquisition geometry, and the flight parameters. The performance of the instrument for the baseline design achieves a swath width larger than 2km in case of wind speed of 3m/s. The noise equivalent sigma zero (NESZ) is between -30dB and -45dB and the radiometric resolution is better than 0.1dB. The velocity accuracy that the system can obtain is fulfilled from look angles lower than 20 degrees up to look angles above 63 degrees, resulting in more than 3km of swath available.
A functional test campaign over ocean is planned in summer 2019 in the Iroise Sea to verify the end-to-end functionality and performance of the fully integrated instrument, including internal calibration capabilities, and of the processor by means of on-ground testing and flight tests.

Area D - Board 415

The Lofoten Maelstrom is an extremely intense tidal current on the southern tip of the Lofoten peninsula in Northern Norway. The maximum current strength during the tidal cycle is estimated to be around 3 m/s, which makes this one of the strongest regular currents that can be observed in open ocean waters. The strength of the currents and its dangerous whirlpools has sometimes reached mythical proportions, such as the old tale of the magical millstone on the bottom of the sea, continuously grinding salt and making eddies on the ocean above. There is also an abundance of references to it in literature e.g. Edgar Allan Poes A Descent into the Maelstrom and in Jules Vernes Twenty Thousand Leagues Under the Sea.

The area of the Maelstrom is exposed to predominating south-westerly winds and waves. For this reason, the tidal current is often subject to waves with the opposite direction, causing steeper waves, wave breaking and a complicated and dangerous situation to navigate in. Ocean currents of the strength discussed here is expected to also have a strong impact on the local surface waves. Locally, it is well know by fishermen and sailors that sea-state in certain situations become extremely complicated and sometimes impossible for manoeuvring a vessel in. Waves propagating in the opposite direction to the currents causes local convergence of wave energy and rapidly increasing wave height in certain areas. Observations are almost exclusively eyewitness accounts and the lack of solid evidence from instrument based observations is prominent.

The strong impulse caused by varying barotropic ocean currents is believed to be a source of internal waves on the oceanic pycnocline. Such waves induces convergence and divergence zones in the upper ocean mixed layer, and hence Doppler Shifts in the capillary-gravity ripples on the ocean surface. These zones are detectable from space-borne instruments as Synthetic Aperture Radar and optical imagery. We present observations from the Copernicus Sentinel-1&2 missions showing very consistent patterns of white bands during ebb and flood tide. From the high resolution numerical ocean model simulations, these bands coincide with the location of a strong convergence area at the front of the tidal current maxima.

Area D - Board 416

The tides and tidal currents are a predominant signal in shallow and shelf regions which have critical applications and societal interests, all the more with the mean sea level rise problem.

As SKIM will fly on a sun-synchronous orbit, it will not be able to sample properly the tidal signal, leading to bad aliasing frequencies of main tidal waves, and some solar waves will not even be observable (S1 and S2). Moreover SKIM geometric measurement of currents is complex as a directional component needs to be taken into account. SKIM will have a wide swath of about 280 km allowing local multiple sampling when different swaths will overfly each other (ascending-descending passes, ascending-ascending or descending-descending configurations), which may improve the high frequency signals measurement locally .

Tidal models would be an alternative solution to correct the tidal currents from the SKIM measurements, but those models are probably not yet enough accurate in all regions, which means that a significant error might remain in SKIM measurements. With the objective of evaluating the interest of the SKIM measurements of tidal currents for the tidal community and eventually envision any assimilation of these data in global ocean tidal models, we propose to quantify the observability of tidal currents by SKIM sampling while taking into account its complex geometric measurement and the local multiple sampling allowed by the wide-swath.

The global ocean circulation model MIT-gcm is sampled using the mission specific simulator, named SKIMulator. A preliminary statistical analysis of this tide sampling is performed and then a more complete OSSE experiment allows estimating the interest of the SKIM wide-swath on the de-aliasing of tidal currents . Using MITgcm allows to take into account the tidal signal as well as other oceanic variability which can prevent a proper tide estimation from satellite measurement due to crossed aliasing issues. The analysis takes into account the baseline of SKIM orbit in October 2018.
Results of this tidal currents aliasing analysis are presented here.

Area D - Board 417

Coastal zones are very productive areas, offering many valuable habitats and ecosystems services and attracting human settlements and activities. The intensive concentration of population and excessive exploitation of natural resources puts high pressure on coastal ecosystems leading to biodiversity loss, habitat destruction, pollution as well as conflicts between potential uses and space competition. Several European directives aim at sustainable management of coastal waters, retaining or restoring a high ecological status and safeguarding ecosystem services. Increasing pressure and stricter regulations increase the need for efficient monitoring solutions. Where traditional in situ sampling is insufficient to characterise the highly dynamic coastal environments, Earth Observation (EO) provides a synoptic view and frequent coverage. With the launch of the Copernicus Sentinel satellites, operational water quality services become a business opportunity.
CoastObs is building a service platform for coastal monitoring with validated products derived from EO. Work is ongoing to develop and validate innovative EO-based products based mainly on Sentinel-2 and Sentinel-3: monitoring of sublittoral habitats and intertidal areas, phytoplankton size classes, primary production, and detection of harmful algae. A particular focus of the project lies on turning the satellite-derived geophysical data layers into higher-level information products that satisfy the users’ information needs, such as trend analysis, indicators for environmental reporting or integration with predictive models. To ensure that the products developed are of added value to the users, CoastObs is working closely with public and private organisations from various sectors (water management authorities, aquaculture producers and organisations, port authorities). CoastObs will establish sustainable supply chains that can be directly integrated into the users’ systems.
By defining user groups with common requirements and identifying different users for the same area, tailored products can be developed at highly reduced costs per user. Setup of efficient data structures (data cube) to allow for quick and easy generation of new products from the basic data is part of this ambition. The commitment of 17 users to CoastObs demonstrates the need for such user-friendly and affordable coastal water services.

Area D - Board 418

For oil spill response, one of the key parameters is detection of actionable oil. Actionable oil, which tends to be the thick, emulsified oil, refers to areas of the spill that poses a threat to the environment and can be targeted for the application of dispersants or skimmers. Previous studies by MDA of controlled oil spills in the North Sea and a persistent oil slick located at MC20 off the coast of Louisiana have shown the capability of RADARSAT-2 quad polarized data to detect emulsified oil. The detection was based on the entropy parameter derived from the Cloude-Pottier decomposition algorithm: the entropy for emulsified oil > sheen > slick-free water. Although these results were encouraging, validation was based on observations that were not contemporaneous with the RADARSAT-2 data acquisition.
As part of a study supported by GOMRI and NOAA, RADARSAT-2 data was acquired on April 25, 2017 in concert with research cruises at MC20. In addition to RADARSAT-2 imagery, NASA UAVSAR data, WorldView and ASTER imagery, and in situ measurement were acquired. The RADARSAT-2 and UAVSAR imagery were acquired simultaneously. The WorldView and ASTER images were acquired five minutes apart, but five hours after the RADARSAT-2 acquisition.
A Maximum Likelihood Classifier (MLC) was applied to the ASTER and WorldView images. The ASTER image was classified as Sheen-Rainbow-Sheen and the WorldView-2 was classified as Metallic, Transitional Dark, and Dark. In situ samples of the oil collected at the time of the ASTER and WorldView images were used to validate the MLC result. The two classified images were subsequently integrated into a single classified output.
The Cloude-Pottier entropy was extracted the from RADARSAT-2 Fine quad-polarized image acquired on April 25. There was very good agreement between the areas of high entropy derived from RADARSAT-2 and areas of high-damping backscatter ratios (oil to water) from the UAVSAR imagery: high entropy and high damping ratios were associated with emulsified oil.
The areas of emulsified oil in the RADARSAT-2 imagery were compared with the WorldView and ASTER integrated classification product. Figure 1 shows the integrated classified product and the entropy image from RADARSAT-2. Areas of high entropy on the RADARSAT-2 image are coloured black. Although there was a time difference between RADARSAT-2 and the optical imagery, once the location of the emulsified oil was corrected for ocean-current drift, there was strong correlation between the areas of thicker oil classified from the optical imagery and the areas of high entropy derived from the RADARSAT-2 image.
The results indicate thicker (emulsified) oil detected by RADARSAT-2 was confirmed by contemporaneous data collected by UAVSAR and near-contemporaneous data acquired by ASTER and WorldView-2. The positive results bode well for the use of RADARSAT-2 quad polarized data for the detection of actionable oil. Future work is planned to investigate the application of compact polarimetry data for the detection of emulsified oil using data from the forthcoming RADARSAT Constellation Mission.

Area D - Board 419

The water dynamics and quality of the coastal zones to a large extent depends on the input of impulse and pollutants by the rivers. As part of a long-term Russian program on the coastal oceanography, river outflows in the coastal zone of the Eastern Black Sea are under investigation. The powerful river outflows are often observed there throughout the year during intense precipitation and in spring time when snow and glaciers melt in the mountains. The river outflows penetrate the coastal environment, where they transform background sea waters. Based mainly on the analysis of high and medium resolution optical, infrared (IR) and synthetic aperture radar (SAR) satellite images (Landsat-8, Sentinel-2, Sentinel-1, Radarsat-2, SPOT, etc.), which show the river outflows in the sea coastal zone, their surface manifestations are studied. They are usually observed both in the fields of ocean color (suspended matter) and sea surface temperature (SST) in optical and IR satellite images. But they can be seen also on SAR images thanks to the modulation of small-scale surface waves generated by the wind. Joint analysis of satellite images allowed obtaining new information about this phenomenon. In particular, by analyzing multisatellite and multisensor data, it is shown that river outflows have a characteristic structure and consist of three distinct regions (namely: plume, transition and tail (far) zone) that are clearly visible in optical and IR bands, and penetrate into the open sea at distance up to 30-40 km. The actual plume is most clearly detected in the SST field rather than in ocean color, but also can be seen in SAR images. Dark patches that appeared from time to time on SAR images of the sea surface in the plume zone can be explained by difference in the temperatures in the river and the surrounding waters, and by local transformation of the atmospheric boundary layer over cold water of plume. Finally, it is shown that river outflows can be effectively studied through their manifestations in optical, IR and SAR imagery, and additional information about physical processes affecting water quality and dynamics can be extracted.

Area D - Board 420

The European Sentinel-1A and Sentinel-1B satellites equipped with synthetic aperture radars (SAR) are key tools for operational monitoring of marine oil pollution. In order to see and understand oil spill distribution in the Persian Gulf more than 500 SAR images acquired in 2017-2018 and covering the entire marine basin were collected and analyzed with respect to oil pollution. In addition to standard and special processing of SAR images, analysis of the monitoring results requires the use of additional interpretation tool. To effectively analyze the SAR images and variety of detected dark patches, a geoinformation approach, specifically, a web-based GeoMixer application platform, was used. This involves the creation of a dedicated geoportal with physical-geographical and industrial information about the water basin, offshore oil and gas infrastructure and intensity of shipping, and use of electronic nautical charts. As the result of monitoring, actual oil spill distribution maps for the entire Persian Gulf have been generated and analyzed. For example, in 2017 number of detected oil spills (4905) and total polluted area (13800 km2) were very large that indicates very high that level of oil pollution. Sources of oil spills in the gulf generally are of the same origin: oil production industry is the main source with highest pollution rate. The largest oil patches (up to 600-700 km2) were detected near the oil production sites. On the other hand, according to data of the automatic ship identification systems (AIS), ship traffic appears to cause significant degree of oil pollution in the open Persian Gulf as well. The largest spill numbers in the Persian Gulf were found during the spring-summer time, especially in low wind conditions. It is shown that routine oil spill monitoring with SAR, such as onboard the Sentinel-1 satellites, together with the GIS approach and AIS data, makes it possible to observe and easily extract the distribution of oil spills of a marine area everywhere, including calculation their statistics and source detection. Obtained results indicate that the Persian Gulf can be considered as one of heavily polluted inland basins in the world on a local and international scale due to oil production and ship-made oil pollution. Finally, such large annual oil pollution will definitely have a critical long-term impact over the gulf marine and coastal ecosystems and therefore continuous monitoring is needed.

Area D - Board 421

The sea level observation system of Catalonia - aims at providing high-quality continous measurements of sea- and land levels at the coast from tide gauges and from modern geodetic techniques for studies on long-term sea level trends, but also the calibration of radar/optical satellite altimeters.
L’Estartit calibration site:PSMSL 1764. 1990-..,ICGC precise leveling. CATNET (GNSS): CREU, CASS, ICESat: L3g (2006), L3h (2007), L3i (2007), GNSS-buoys: T/P(1999, 2000), Jason-1 (2002, 2003),LiDAR: june and october 2007.
Barcelona calibration site:PSMSL 1811. 1992-2008 and 2007- .., Data: www.puertos.es, ICGC (Instituto Cartografico y Geologico de Cataluña) precise leveling, CATNET: MARE, PLAN, GARR;LiDAR: july 2014; ICESat: L3j (2008).
The sea level infrastructure has been improved by providing the harbour of Barcelona with 3 tide gauges and a GPS station nearby. Furthermore, an airborne LiDAR campaign was carried out with two strips along two ICESat target tracks.
The validation using LiDAR may be useful to fill coastal areas where satellite altimeters are not measuring. In June 2014, an airborne LiDAR campaign flying underneath ICESat tracks had been done.
The validation using LiDAR technology may be useful to fill coastal areas where satellite radar altimeters are not measuring. The ICESat laser altimeter delivers ellipsoidal heights and can be used over land to calibrate the LiDAR . A new airborne LiDAR campaign flying underneath the tracks L3g-L3j (green and blue) of the ICESat laser altimeter has been at the Barcelona harbour in June 2014. Now, some data analysis has been carried out.
A direct comparison between ICESat GLA06 and GLA14 heights for both tracks has been done. The RMS of differences is 15 cm with a standard deviation of 5 m.In order to make comparable the ICESat and the LiDAR elevations,both sets has been referred to the same vertical datum.First, a direct comparison between ICESat GLA06 and GLA14 heights for both tracks has been done. The RMS of differences is 15 cm with a standard deviation of 5 m.
The geoid model used in Catalonia – EGM08D595 (ICGC) based on the EGM2008 which has been fitted by a constant offset of - 0.595m. The global gravity field EIGEN-6C3STAT (Förste et al. 2012) has been analysed.
In order to more closely examine the ICESat and LiDAR elevations data sets, we will compute estimates of geoid undulations and later will compare then with the existing geoid models.
A detailed geodetic infrastructure to collect altimetric data from Jason-3 (EUMETSAT) and Sentinel-3 (ESA) and and plans for a lidar aerial campaign along tracks for ICESat-2 is presented.

Area D - Board 422

This work presents the results of an analysis of the cross-polarization (VH and HV, or cross-pol) signal of the synthetic aperture radar (SAR) images of the Radarsat-2 satellite acquired in oil spill areas in the Caspian, Barents, Norwegian and other seas. It was shown that it is possible to reveal the cross-pol signal from under the noise level of the Radarsat-2. With a noise level (NESZ, the noise equivalent sigma zero) of about -35 dB for the Radarsat-2 is succeeded to work with the level of the cross-pol signal of about -40 dB (and up to -45 dB).

Comparison of the behavior of the cross-pol signal versus the co-pol signal (VV or HH) shows a high level of their correlation with each other under the conditions of their damping by oil films. The analysis of the damping of the cross-pol signal in oil slicks was carried out by comparing it with the resonance (Bragg) part corresponding to the scattering of the signal by capillary ripples of several centimeters long, and with the non-resonant part corresponding to the scattering from the wave breaking. The polarization method (Kudryavtsev et al., 2013) and co-pol (VV / HH) Radarsat-2 channels were used to determine the resonance, Bragg part, and the non-resonance, wave breaking, part for images taken in full polarization of the signal (VV / VH / HV / HH, quad-pol mode).

It was shown that the cross-polarization signal consists of two almost equal parts, both the resonant, the Bragg part, and the non-resonant wave breaking part. Their proportions depend both on the level of sea wave heights associated with the wind, and the direction of the wind and the waves. Theoretical calculations according to the two-scale Bragg model (Valenzuela, 1968) of the expected intensity level of the corresponding cross-polarization signal showed that they do not correspond to practically observed values, as in conditions of pure water, as well as conditions of oil films. The calculation underestimates the signal level by several decibels.

This work was supported by the Russian Science Foundation grant No. 18-55-20010.

References

1. Kudryavtsev, V. N., B. Chapron, A. Myasoedov, F. Collard, and J. A. Johannessen. On Dual Co-Polarized SAR Measurements of the Ocean Surface // IEEE Geoscience and Remote Sensing Letters, 2013, Vol. 10 (4).
2. Valenzuela G.R., Scattering of electromagnetic waves from a tilted slightly rough surface // Radio Science, 1968, Vol. 3, pp. 1057 - 1066.

Area D - Board 423

An adequate theoretical description or simulation of the electromagnetic scattering from the breaking short gravitational-capillary waves is still an unsolved problem, the solution of which is extremely demanded for remote sensing of the sea surface. The so-called "nonresonance" signal from wave breakings measured by a radar makes 40-60% of the total signal level even for weak winds up to 3 m/s [1, 2].

A numerical investigation of breaking gravity-capillary waves is presented. The goal of the work was to study the validity of the spectral approach for waves, which uses an assumption of uniformly randomly distributed phases. The phase coherence could result in a significant modification of higher moments of the wave spectra, such as mean square slopes (MSS) and others, which are crucial for remote sensing applications. For example, the modern not resolved yet problem is a wind wave spectrum suitable for an explanation of variety of experimental data in a wide range of bands, C-, X-, Ka- (e.g., [3]).

Open source package OpenFoam was used for two- and three-dimensional CFD (Computational Fluid Dynamics) transient simulation of breaking gravity-capillary waves. As a benchmark the Ericson et al. (1999) experiment [4] on breaking gravity-capillary waves was performed. The free-surface water channel of 20 cm length and 6 cm depth with a uniform flow about 1 m/s was modeled. Waves were breaking on a submerged stationary hydrofoil of 4 cm length. The VOF (Volume of Fluid) model together with the Navier-Stokes equations was solved. A mesh grid with size of 1 mm near the free surface was used to resolve dissipating capillary waves. Simulations were performed using as the laminar model as well LES (Large Eddy Simulation) turbulence model.

Free surface realizations with time step 0.05 s were processed. Surface elevations and slopes profiles were resemble on that for the breaking wind induced waves measured in wind channel [5] (see Fig. 1). The average curvature spectra of breaking waves showed close to k^(1/2) dependency in the range of wavenumbers 200÷2000 rad/m (Fig. 2), which correspond to experimental results obtained by Ericson et al. [4] and Jähne and Riemer [6], and theoretically predicted by Phillips [7].

Based on model calculations, statistical distributions of the curvatures of the water surface in the collapse zone were determined. Theoretical work [7] showed that the contribution from the curvature of the surface must be taken into account, both for the calculation of resonance (bragg) reflections, and nonresonant (Kirchhoff-like) ones. At the same time, the calculations of the present work showed that the statistical distributions of the curvature of the surface in the region of the "tails" are several orders of magnitude higher than the commonly used normal distributions, which should make significant corrections to the EM scattering theory of the signal from the local curvature of the surface (Local Curvature Approximation [8]) in the zone of wave breakings.

This work was supported by the Russian Foundation for Basic Research grant No. 18-55-20010.

REFERENCES

[1] V.N. Kudryavtsev, D. Hauser, G. Caudal, and B. Chapron, “A semiempirical model of the normalized radar cross-section of the sea surface: 1. Background model”, J. Geophys. Res., vol. 108, no. C3, pp. FET 2-1-FET 2-24, 2003.

[2] D.V. Ivonin, S. Skrunes, C. Brekke, A.Y. Ivanov, “Interpreting sea surface slicks on the basis of the normalized radar cross-section model using Radarsat-2 copolarization dual-channel SAR images”, Geophys. Res. Lett., vol. 43, no. 6, pp. 2748–2757, 2016.

[3] A. Bringer, B. Chapron, A. Mouche, and C.A. Guérin, “Revisiting the short-wave spectrum of the sea surface in the light of the weighted curvature approximation”, Geoscience and Remote Sensing, IEEE Transactions on, vol. 52, no. 1, pp. 679-689, 2014.

[4] E.A. Ericson, D.R. Lyzenga, D.T. Walker, “Radar backscatter from stationary breaking waves”, J. Geophys. Res.: Oceans, vol. 104, no. C12, pp. 29679-29695, 1999.

[5] G. Caulliez, “Dissipation regimes for short wind waves”, J. Geophys. Res.: Oceans, vol. 118, no. 2, pp. 672-684, 2013.

[6] B. Jähne, K.S. Riemer, “Two-dimensional wave number spectra of small-scale water surface waves”,
J. Geophys. Res.: Oceans, vol. 95, no. C7, pp. 11531-11546, 1990.

[7] O.M. Phillips, “Spectral and statistical properties of the equilibrium range in wind-generated gravity waves”, Journal of Fluid Mechanics, vol. 156, pp. 505-531, 1985.

[8] T.M. Elfouhaily, C. Bourlier, and J.T. Johnson, “Two families of non-local scattering models and the weighted curvature approximation”, Waves in random media, vol. 14, no. 4, pp. 563-580, 2004.

Area D - Board 424

The project ”Determinations on the absolute sea-level rise on the German North Sea and Baltic Sea coasts“ funded by the Federal Ministry of Education and Research (BMBF) developed a model for vertical land movements in Northern Germany. The project partner at the Universität Siegen uses this model to improve about one hundred tidal records of gauge stations to estimate the absolute sea-level rise.
A major problem associated with detecting absolute sea level changes is the relative character of tidal records. Water level data describe the changes of the mean sea level (MSL), but they also include the impact of vertical movements at the gauging stations and the hinterland. To minimize these two effects on MSL estimation a model was developed to describe subsidence and uplift effects in Northern Germany. Therefore, a network of 180 GNSS stations including IGS, EUREF, GREF and SAPOS and gauge stations with GNSS from January 2010 until end of 2016 was processed and combined with Sentinel-1A and -1B data processed with Persistent Scatterer Interferometry (PSI).
In this presentation we will show the results from PS interferometry processing for a 50 km wide strip along the 1200 km long German coast. We processed the whole timespan of two tracks from Sentinel-1A and -1B and generated a combined areal solution for the estimation of vertical land movement. The results show strong movements in the area of Emden through hydrocarbon extraction and especially in Wilhelmshaven movements from cavern storages are detected. A further study with a combined solution from ascending and descending orbit gives us more insights into the spatial distribution of this subsiding phenomenon.
In the further evolution of the project those times series will in detail be analyzed to calibrate gauge benchmarks at the stations and as well to refine vertical movements of the gauge reference points in the hinterland to improve the tidal records.

Area D - Board 425

The Special Priority Program (SPP-1889) ‘Regional Sea Level Change and Society - SeaLevel’ (2016-2021), funded by the German Research Foundation (DFG), performs a comprehensive, interdisciplinary sea level analysis to advance our knowledge on regional and coastal sea level change (SLC), while accounting for the human-environment interactions. During its first funding phase (2016-2019), SeaLevel consists of 20 projects, bringing together over 80 natural and social scientists from 23 German research institutions and a wide range of disciplines, such as physical oceanography, geophysics, geodesy, hydrology, marine geology, coastal engineering, geography, sociology, economics and environmental management. By combining diverse modern methodologies, observations and models, natural and social scientists jointly aim to create a scientific knowledge base for quantitative, integrated coastal zone management (ICZM), which can be applicable to many endangered places globally and essential for safety, coastal/land use planning, and economic development.

The SeaLevel program performs global and regional studies and focus its impact studies on the North and Baltic Seas with potential impacts on Germany, and the South-East Asia/Indonesia region, encompassing coastal megacities, low-lying islands and delta regions, in order to understand how coastal vulnerability, adaptation and response strategies towards SLC vary in distinctly different socio-politico-economic and cultural contexts.

The main research activities of SeaLevel are: a) Contributions to global and regional sea level changes, b) Regional biophysical and social impacts in North Europe and S.E. Asia/Indonesia, and c) Adaptation, decision analysis and governance. These research objectives include to improve the physical knowledge of SLC and regional-to-local scale projections, investigate which socio-institutional factors enable/hinder coastal societies to cope with SLC, determine the natural and social coastal systems’ responses to future SLC, and assess adaptation and risk governance strategies under given technical, cultural, socio-politico-economic constraints. Such integrated analyses require SLC information (local SL projections, storm surges, waves and extremes), uncertainty and risk measures to be provided at the coastlines.

In this presentation, we will describe the goals and status of the SeaLevel program, and will review the use of remove sending data in interdisciplinary sea level research.

Area D - Board 426

Planning for future sea level rise, it becomes of increasing importance to understand the sea level of the past, especially in a highly populated area as the Baltic Sea. Several questions appear, among these: i) There are many long tide gauge records in the Baltic Sea, but also areas with no long records. Can the past sea level be reconstructed for all sections of the coast? ii) Satellite altimetry is widely used to calculate sea level trends, but can these products be used in regions close to the coast? iii) How does the trends from satellite compare with in-situ measurements?
This study presents a reconstruction of the 2D sea level field based on statistical modelling of monthly tide gauge observations and model reanalysis. The reconstruction provides monthly mean sea level of the Baltic Sea in 10 km resolution during 1900¬‐2014, and can reconstruct the variability of the Baltic Sea with an average correlation of 96% and RMS error of 3.8 cm to 56 independent tide gauges.
The ESA Sea Level CCI product constitutes high quality monthly sea level variability and trend analysis for the open ocean. However, it is commonly used in the coastal zone. Here, we assess the quality in the coastal zone of the Baltic Sea by comparing the monthly variability of the CCI with that of the statistical model and of independent tide gauges from PSMSL, taking land rise information into account.
The validation against this independent source of sea level information allows assessing the quality of the altimetry product approaching the coast, and can therefore be used to determine how close to the coast the Sea Level CCI can be considered reliable. The CCI sea level anomalies match reconstructed sea level variability of the southern and open parts of the Baltic Proper with correlation above 90% and RMS difference below 6 cm. However, some coastal zones, areas with sea ice, and areas of high natural variability need special treatment. Finally, the trends of the statistical reconstruction, the independent tide gauges, and the altimetry is compared.

The figure shows the sea level CCI trend map of the Baltic Sea. Black dots denote areas with increased validation errors.

Area D - Board 427

Climate change is likely to exacerbate many of the problems that the coastal environments is already facing: shoreline erosion, coastal flooding, stress and damage of coastal biodiversity. Sea level rise is one of the major threats for coastal zones; its monitoring is essential to better understand and predict the response of the coastal zones. Tide gauges offer a source of historical, long-term sea level records but they have a sparse distribution. Altimetry provides a unique long term and almost global observational dataset able to inform on how sea level evolves from the open ocean to the coastal zone.

For more than 10 years, the CTOH at LEGOS is developing the X-TRACK processing chain in order to recover as many coastal sea level observations as possible from the standard GDR altimetry products. Today , X-TRACK is a mature L3 product usable for many types of coastal applications. Its editing and processing strategy allows to obtain more accurate data closer to the coast. The new version of the processing chain we present here, integrates progress made by the altimetry community over the past few years, both in terms of retracking and the geophysical corrections applied to the data. The ALES retracker has proven its efficiency in retrieving more coastal sea level observations than other retrackers, particularly when using high-rate (i.e. 20 Hz) altimeter measurements instead of the standard (1 Hz) data. On the other hand, the geophysical corrections, in particular the wet tropospheric and tidal corrections, have also been improved in the coastal domain. In the context of the Bridging Phase of the ESA’s Climate Change Initiative sea-level project (SL_cci) and acknowledging user needs, we develop a new L3 multi-mission product combining better spatial resolution provided by high-rate data, the post-processing strategy of X-TRACK and the advantage of the ALES retracker. This product increases the number of useful near shore sea level data. As an example, for Jason-2 we obtain 80% of valid sea level data at a distance of less than 2 km in average in the Mediterranean Sea, instead of the 7 km for the X-TRACK’s previous version. X-TRACK/ALES 20 Hz product is available over the 15 years period of Jason-1&2 missions, over three coastal areas (Mediterranean Sea, North East Atlantic and West Africa). Here we will present the first results of this new product in the Mediterranean Sea region as well as its potential for coastal applications

Area D - Board 428

Microwave radiometers are instruments designed for a broad number of Earth Observation (EO) applications. However, due to their low spatial resolution, they are very limited when dealing with regional-scale studies, so that resolution enhancement techniques must be invoked [1]. From a mathematical point of view, the enhancement of the spatial resolution is obtained solving a linear inverse problem. In literature such problem is mainly addressed by solving an lp norm minimization problem with p=2 (Hilbert space) and p=1.2 (Banach space) [2].
In this work we present a novel approach, based on the idea of using a variable p exponent, ranging in the interval [1.2,2]. The proposed algorithm adapts the exponent p in correspondence to the portion of the image to be reconstructed in order to exploit the benefits and avoid the drawbacks of both the constant exponent p=2 and p=1.2 methods. In Fig. 1, the reconstructed fields at enhanced spatial resolution corresponding to a simulated rect-like-signal (blue dotted line) are shown. The variable exponent p method (purple line) allows to wipe out the oscillations due to the Gibbs phenomenon (that typically apply when p=2, shown as a red line, is adopted), and allows to reconstruct fairly the top of the rect-like signal where the p=1.2 methods (yellow line) fail. Preliminary results obtained applying our algorithm to mono-dimensional actual radiometer data (to be presented at conference time) show a significant improvement with respect to constant exponent p methods, making the proposed approach a very promising technique.

[1] D. G. Long, “Reconstruction and resolution enhancement techniques for microwave sensors,” in Frontiers of Remote Sensing Information Processing, C. H. Chen, Ed. Singapore: World Scientific, 2003, ch. 11.
[2] F. Lenti, F. Nunziata, C. Estatico, and M. Migliaccio, “Conjugate Gradient Method in Hilbert and Banach Spaces to Enhance the Spatial Resolution of Radiometer Data,” IEEE Trans. Geosci. Remote Sens., vol. 54, no. 1, Jan. 2016.

Area D - Board 429

Climate-related disasters are highly demanding in terms of life land economic losses. Among them, storms are the costliest type of disaster: their impact on coastal zones strongly depend on the relative sea level, and on its changes, that thus deserve special attention particularly where people, industrial settlements and cultural heritage coexist. In the Mediterranean basin the most exposed area to this risk is the Adriatic Sea, and in particular Venice, whose high vulnerability has determined the need to realize one of the most impressive hydraulic engineering works: the MOSE barriers. In this context, the long term implications of rising sea levels have not been exhaustively investigated yet, and the contributions from the various mechanisms impacting on sea level have not been clearly quantified. The European Space Agency (ESA) Climate Change Initiative (CCI) project on “Sea Level” has produced an improved set of satellite-based sea level gridded products by reprocessing altimeter data over 1993-2015. Recently, CTOH within the project has released a new version of the dataset which includes along track high resolution (i.e., around 300 m spaced) Jason-1 and Jason-2 data. Our contribution aims at assessing the quality of the current Sea Level CCI products in the Adriatic Sea - in particular around the cities of Venice and Trieste, how close to the coast the actual Sea Level CCI products can be considered reliable, and what are the main differences between gridded and along track data. Near Trieste we also can rely on Jason-1 and Jason-2 track data reprocessed at 20 Hz. For a better understanding of the role played by subsidence on the local sea level rise, the latest CCI gridded products and the new along-track high resolution observations at the nearest suitable measurement points are compared with tide gauge data, thanks to the very short distance between the tide gauges and the coast in those two cities (4 km in Venice and 0 km in Trieste). From in-situ measurements a relative sea level rise trend of about +6.3 mm/yr has been determined in Venice, while for Trieste we got +4.7 mm/yr. The difference between in-situ data and altimetry observations supplies an estimate of the subsidence contribution in both cities. A partial validation of these estimates has been made against GPS-derived values, in order to distinguish the contributions of subsidence and eustatism, and in view of the possible closure of the sea level budget in those two cities. Subsidence data, when available, are derived from local permanent GPS stations. We investigate also the role of the varying surface pressure in the relative sea level rise local budget: pressure data are taken from affordable time series of meteorological data at Venice and Trieste, as well as from the ECMWF ERA INTERIM reanalysis project. This work will contribute to identify problems and challenges to extend the sea level climate record to the coastal zone with quality comparable to the open ocean, and also to assess the suitability of altimeter-derived absolute sea levels as a tool to estimate subsidence from tide gauge measurement in places where permanent GPS receivers are not available.

Area D - Board 431

Coastal areas are strongly affected by storm surges. Combined with the high tide and large waves, storm surges have devastating effects along the coasts. Sea surface height variations are measured by tide gauges along the coasts. These sensors are dampened to minimize the effects of wave on the measurements of the tide. Recently, Global Navigation Satellite System-Reflectometry (GNSS-R) demonstrated a strong potential for the monitoring changes in the Erath surface. Thanks to the deployment in many countries of permanent GNSS stations, this technique can be applied when a permanent geodetic GNSS station is located on the shore. An approach, based on the analysis of the reflected component of the signal to noise ratio (SNR) measured by the geodetic antenna can be used for the monitoring of the level variations of the sea surface.
In this study, we used 3 months of records (January-March 2010) from a GNSS geodetic from the GNSS permanent GNSS network (RGP for Réseau GNSS Permanent), deployed by the French Geographic Institute (IGN for Institut de l’Information Géographique et Forestière), located in Socoa, in the south of the Bay of Biscay, Basque Country, to determine the SSH variations. During this period, the Xynthia storm hit the French Atlantic coast the 28th of February 2010 causing large floods from the Gironde to the Loire estuaries and around 50 casualties.
Using both a singular spectrum analysis (SSA) and a continuous wavelet transform (CWT), we demonstrated that we are able blindly separate the tide from the signature of the atmospheric pressure of the storm. After decomposing the signal, we found a correlation of 0.98/0.97 and RMSE of 0.21/0.28 m between the tide gauge records of Socoa and our estimates of the SSH using the SSA and the CWT respectively and a correlation of 0.7 between atmospheric pressure from a meteorological station and a mode of the SSA.

Area D - Board 432

The Northern Mediterranean Current is the northern branch of the cyclonic circulation of the northwestern Mediterranean Sea. Occasionally the Northern Current can penetrate on the Gulf of Lion shelf, strongly impacting the local biogeochemistry and in turn the primary production. The space-time variability of slope currents plays a key role on the across-shore transport of natural and anthropogenic elements. It is therefore of critical importance to monitor the positions of these hydrodynamical features in a synoptic way. Our paper proposes to address this issue by developing a new multi-satellite altimetric-based approach aiming at monitoring the long-term behaviour of the Northern Current.
The approach has been developed combining i) the output of a numerical simulation performed with the model SYPHONIE with a 3 km resolution spanning from 2001 to 2011, ii) coastal oriented optimized along-track altimetry dataset and iii) in-situ measurements acquired by a bottom-moored ADCP (RDI Ocean Sentinel, 300 kHz) positioned at at the JULIO (JUdicious Location for Intrusion Observation) station on the eastern side of the Gulf of Lion.

Area D - Board 433

The coupling of models is a commonly used approach when addressing the complex interactions between different components of earth system. In climate and forecasting research and activities, advanced models are needed and there is an urge towards the use of coupled modelling. This study presents the synergy between the satellite observations (e.g. Sentinel, CryoSat) with the high-resolution, coupled atmosphere, ocean and wave model system for the North Sea and the Baltic Sea, which is part of the Geestacht COAstal model SysTem GCOAST. We focus on the nonlinear feedback between strong tidal currents and wind-waves, which can no longer be ignored, in particular in the coastal zone where its role seems to be dominant. The proposed coupling parameterizations account for the feedback between of the upper ocean on the atmospheric circulation by accounting for the effects of sea surface temperature and the sea surface roughness. The performance is illustrated for the cases of several extreme events. For example, the inclusion of wave coupling leads to decreases strong winds through wave dependent surface roughness or changes sea surface temperature, the mixing and ocean circulation, leading to better agreement with in-situ and satellite measurements. The quality of the satellite data is evaluated and the focus is brought to the coastal areas, where altimeter data tend to get worse than in the open ocean. Satellite data of Semtinel-3 Jason-2 and CryoSat-2 performance is assessed, with respect to in-situ measurements and the GCOAST simulations. Analyses have been carried out, in which data from the satellite tracks are separated between onshore and offshore flights, with onshore flights passing from the ocean to the shore and offshore flights passing from the shore to the ocean. The differences have been also assessed between the satellite tracks under varying metocean conditions with focus on extremes. The prediction limits and applications for the combination of available satellite observations, in-situ data and coupled GCOAST system are also discussed.

Area D - Board 434

The use of a mean sea surface (MSS) is paramount for deriving sea level anomalies and gravity field from satellite altimetry. Sentinel-3 offers independent SAR altimetry along new ground tracks and can hence provide independent evaluation of existing mean sea surfaces. We perform an evaluation of the recent available mean sea surfaces using one year of 1 Hz Sentinel-3A data taking into account the shift in temporal average and coverage. Sentinel-3A is now available since February 2016 giving a two-year record, whereas the DTU15MSS is an average of altimetry from 1993-2012 with an averaging median year in 2003. Hence, recent models for sea level changes must be introduced in order to account for 14 years of sea level changes which can regionally contribute with a signal on the decimetre scale. Similarly, the first data from the Sentinel-3B satellite along its interleaved orbit with Sentinel-3A is presented upon availability.

Detailed coastal evaluation of the MSS models in the North Sea and Danish waters is performed by using Sentinel-3-A 1Hz data from RADS to study the impact of SAR altimetry on MSS determination and determination of short wavelength coastal signals which is likely better determined using SAR altimetry than with conventional altimetry, which is the base for most current mean sea surfaces. Our studies show that some of the most used MSS models differ with more than 10 cm in coastal areas. These discrepancies can be minimised if the new SAR data are included in future derivations of the MSS. Furthermore, differences of around 5 cm are found in the applied ocean tide corrections in coastal and arctic areas from the two ocean tide models FES2014 and GOT4.10. The disagreement between the S2 component in the tidal models is amplified by the sun-synchronous orbit of the Sentinel-3 satellites.

The short wavelength SLA can be analysed by subtracting a moving average of 15 along-track measurements of the MSS and thereby filtering out signals with wavelengths longer than roughly 100 km. We have investigated short wavelength signals valuable to gravity field prediction in the Indian Ocean and across the Aleutian Ridge where short wavelength MSS signals are particularly strong. Sentinel-3a shows a clear signal in the area, which is not visible to the same extend in averaged SARAL or Envisat data.

Area D - Board 435

The European Space Agency’s (ESA) Sea Level Climate Change Initiative (SL_cci) project in its latest phase is focusing on coastal sea level variability. While standard altimetry is often not able to extract meaningful observations in the coastal zone, where the accuracy of the measurements from satellite radar altimetry is jeopardised by land contamination in the radar echoes, the new X-TRACK/ALES 20 Hz product benefits from a better fitting of the satellite altimetry echoes and a dedicated post-processing in order to improve the quality and quantity of data in the coastal zone.
Preliminary results using the X-TRACK/ALES along-track dataset in West Africa, one of the ESA SL_cci test regions, show generally a much higher trend variability than the trends derived from gridded altimetry datasets. This variability is often hard to interpret, due to the scales of few kilometres and the variety of phenomena that can influence the sea level in this challenging environment. In special cases, some altimetry tracks are located at the estuary of a river, where local trends can be influenced by the variability of the river discharge and the river plume propagating off-shore. In this study, we want to explore whether increasing trends observed at the coast near the mouth of Sassandra River can be related to or explained by river discharge.
The Sassandra River (basin area of 70’750 km²) is a transboundary river that origins in Guinea and flows mainly in Côte d’Ivoire (East Africa). It is characterized by a Buyo reservoir, built in 1980 for hydroelectricity and located near the middle of the river length. Downstream of the dam, the river system flows south-southeast to empty into the Gulf of Guinea in the Atlantic Ocean. Despite the river is monitored along its course, the in situ measurements are difficult to obtain. For this reason, we adopt an approach that uses MODIS images to derive proxy of river discharge time series. The method is based on the different values of reflectance between the water body and the land. By considering two boxes, one in urban area and one in the river, the ratio between the average reflectance of the two boxes shows a variation consistent with the variation of river discharge. Previous studies demonstrated also that differences among the peaks flow can be distinguished in the ratio derived by MODIS. The location of the two boxes has been calibrated by considering temporal series of water level derived by Sentinel-3 (http://hydroweb.theia-land.fr/). Successively, the ratio of reflectance has been extended for the years 2003-2018.
The analysis of the 16-year time series shows an increased trend at monthly and annual aggregation according the Mann-Kendall test. The results are also confirmed by the increasing trend of the rainfall, extracted by ERA-5 dataset within the basin.
The growing trend on the proxy of river discharge can be directly compared to the growing trend of the sea level close to the Sassandra River mouth. Based on the achieved results, this preliminary study paves the way to investigate the interaction processes between the freshwater and the ocean by using the synergy among different satellite data.

Area D - Board 436

At the entrance of the Gulf of California converge the influences of the Eastern Tropical Pacific and the subarctic waters of the California Current. Sea level and upper-ocean properties (temperature, salinity) vary seasonally due to the presence of different water masses. In the summer a poleward coastal current carries tropical waters towards the interior of the Gulf of California. In the spring the surface expression of the California Current reaches its yearly peak, creating thermal fronts and mesoscale features. To study the seasonal variability we use sea level time from two tide gauges in the Baja California Peninsula and the Mexican mainland and sea level anomalies from the X-Track altimetry product generated by CTOH/Legos (France). The altimetry data correspond to the satellite tracks of the Topex/Poseidon, Jason 1 to 3 and Geosat Follow On missions. The correlation between coastal altimetry and tidal gauge SLA time series are high (R> 0.7) for distances smaller than 30 km. We discuss the seasonal and interannual sea level changes for the mainland and Baja California coasts. We put special attention on the effect on sea level of two recent extreme events: The Blob and the El Niño 2015-2016.

Area D - Board 437

The coastal zone of the northeastern part of the Black Sea is characterized by a dozen of small rivers with runoff between a dozen and 800 m3/s. Maximum values of the runoff are related to heavy rains that occasionally occur over the coastal zone of the Krasnodar Krai and the Republic of Abkhazia. River runoff forms a river plume in the coastal zone of the sea. Depending on local wind and current direction and intensity, amount of the river runoff, and density difference with the sea water a river plume may propagate in different directions from the river mouth, and be involved in a mesoscale and sub-mesoscale circulation. River plumes significantly influence quality of seawater taking in mind that the whole coastal zone from Kerch Strait to Adler and border with Abkhazia is the most important resort area in Russian Federation. The situation is aggravated by the fact that sewage systems are in unsatisfactory conditions or absent in most of small villages located along the Black Sea coast. After heavy rains, waste and feces waste very often reach the sea and represent a serious threat to people’s health. The aim of this study is to investigate the relationship between runoff of the main rivers along the Russian coast of the Black Sea and area of the river plumes.
For this study we compared the Mzymta River and Sochi River daily runoff and area of river plumes detected by high concentration of suspended matter acquired daily from MERIS (MEdium-Spectral Resolution Imaging Spectrometer, ENVISAT) from 2006 to 2011.
To calculate the concentration, we chose the developed algorithm FUB/WeW recommended by ESA and embedded in the MERIS processing software. Plume boundary was calculated by analyzing the average concentration of suspended matter outside the plume and used it as a threshold. The analysis showed that, in most cases, the concentration of 1 g/m3 could be considered as the boundary of the river plume. In the season of intensive bloom of Coccolithophore, the threshold shifted greatly.
Water fluxes for the rivers Sochi and Mzymta were taken from the hydrological stations in the City of Sochi and Krasnaya Polyana. The first results showed that based on the plume area it is possible to determine the discharge of the river. For both rivers the dependence between river runoff and plume area is linear. We continue this research using data of the MultiSpectral Instrument (MSI) of Sentinel-2 for the analysis of river plumes of Kodori and Bzyb rivers in Abkhazia.
The study was carried out with the financial support of the Russian Foundation for Basic Research project No. 19-55-40004_Abkh_a «Assessment of the impact of extreme atmospheric precipitation on the ecological state of the Black Sea coast and coastal waters of the Krasnodar Territory and the Republic of Abkhazia, based on remote sensing data».

Area D - Board 438

Oceanic coastal zone is of great interest as this part of ocean is the closest to where most of the people live, so that effort to better understand coastal sea level changes become more and more important. Coastal sea level research studies need good quality data close to the coast. The first point is then to develop techniques to get more and better data closer to the coastline to improve coastal sea level monitoring.
Satellite altimetry have allowed to give information about sea level for over 25 years. Data quality of altimetry missions is excellent over open ocean, but sea level measurements are tricky difficult to apprehend when approaching coasts. A lot of effects caused by the instrument, atmosphere, ocean and land can reduce the quality of sea level estimation from altimetry: we need to remove them as much as possible. There are partly due to heterogeneous surface reflections within the altimeter footprint (over both ocean and land surfaces), as the altimeter beam illuminates a circle of ocean or land surfaces with a 3 to 10 km radius, depending on the sea state, the wave height or the corrugated land. Radiometers are impacted from 10km (SARAL) to 40km (Jason) to coast, which can directly affect corrections that are deduced from radiometer. Geophysical corrections’ sensibility to shore distance can also impact the computed sea level, in both cases derived from model or instrument.

Conventional altimetry (Jason-3 in Ku-band, SARAL in Ka-band) and Delay Doppler Mode (Sentinel-3A SARM) methods are used in altimetry with different footprint and instrumental noise. Thanks to a reduced footprint coupled with a lower instrumental noise, SARAL and Sentinel-3 should provide improved sea level observations over coastal regions compared to Jason-3. In order to reduce the unwanted effects, that were not as accurately known or had a greater influence in the coastal zone than open ocean, the first step is to identify them and evaluate how much there are important over that particular coastal areas.
We will focus on how different sources of error (instrument and/or geophysical corrections) interact during sea level computations for these three missions.

This presentation aims overviewing the performances of different altimetry missions over coastal areas, through the analysis of various data from Sentinel-3A, Jason-3, and SARAL/AltiKa.

Area D - Board 439

The Wadden Sea is an ecological region that is sensitive to the regional sediment balance.
The sediment balance is strongly dependent on hydrodynamics, which is affected by regional sea level rise.
To predict the future state of the Wadden Sea it is required to understand the hydrodynamics and to estimate future sea level rise.
Based on climate models and tide-gauge records, sea level in the Wadden Sea is expected to closely follow that of the North Sea.
The North Sea is densily sampled by tide gauges and since 1992 its mean sea level can be estimated using satellite radar altimetry. However, to understand the hydrodynamics, instant water level estimates are required within the Wadden Sea, which can be used to validate models or for assimilation. Unfortunately, only a handful of sparsely distributed datum-controlled tide-gauge records are available near the coasts or on the islands. Next to that, conventional satellite radar altimetry is not useful in a complex environment as the Wadden Sea, since its beam-limited footprint is typically >10 km in diameter, which causes distortions in the practically every waveform due to nearby land targets. Conventional altimetry is due to its limited resolution also not able to distinguish the narrow tidal canals from the tidal flats. We aim to enhance the spatial distribution of water level estimates using full bit rate altimetry data from the Cryosat-2 mission. By applying a recently developed Fully-Focused Synthetic Aperture Radar (FF-SAR) algorithm, the along-track resolution of this altimeter is enhanced to 0.5 meter. Firstly, this enables us to better remove static land-targets that pollute near-shore waveforms. Secondly, waveforms over tidal slats are possibly distinguishable from those over open water by waveform classification techniques. Besides the benefits for hydrodynamic modelling, the complex Wadden Sea domain is a suited location to study the novel FF-SAR technique in detail. Note that the configurations of the Sentinel-3 and the future Sentinel-6 satellites also allow for the application of FF-SAR, which could further enhance the temporal and spatial sampling of the Wadden Sea.

Area D - Board 440

Coastal zones and inland waters (rivers, lakes and reservoirs) are crucial zones for the human development. Dedicated processing evolutions are needed for these areas when using altimetry data in order to retrieve the water surface height. The contribution of undesired targets is to be addressed and the retracking process is to be adapted, among others.
The Sentinel-3A&B are the most recent altimetric missions of a long historical series, and the first altimeters of the Copernicus constellation. SAR (high resolution) mode data from these satellites are used for this work.
The coastal processing has been developed over complex coastal topography areas, including cays, reefs, shallow waters and cliffs. It is focused at filtering out off-nadir contamination from the sea surface signal. This specific processing comes from the idea explained at (Garcia et al., 2018. https://doi.org/10.1016/j.asr.2018.03.015) using CryoSat-2 SARin data, and adapted to Sentinel 3 SAR data.
The Inland Waters work deals with the challenge of restricting the land contamination of inland waters signals. Medium size lakes, reservoirs and rivers are monitored. We are presenting three different retracking algorithms for retrieving water level and compared the results with the ESA L2 products.
isardSAT has developed the study as ESA Expert Support Laboratory within the Sentinel-3A Mission Performance Centre team.

Area D - Board 441

Satellite altimetry has been providing global maps of sea surface information within the last three decades with unprecedented accuracy and spatial coverage on a frequency of 10 days or longer. However, in the coastal zone within 10-50 km away from the shoreline, measurements are generally flagged as bad in the standard satellite altimetry products which limits the recovery of useful sea surface height signals in these strategic areas.

In recent years, there have been significant developments of specialised radar echo fitting (‘retracking’) algorithms in addition to the noticeable improvement of several geophysical corrections that must be applied to the altimetry data. These have enabled retrieval of more data closer to the shoreline and minimised the gap of the unmapped zones between the tide gauges and the altimetric useable measurements.

One example of the aforementioned improvements is National Oceanography Centre (NOC) retracking algorithm, the Adaptive Leading-Edge Sub-waveform (ALES). ALES is applicable to pulse-limited altimetric measurements over coastal zones and open oceans and aims at recovering coastal data better than the standard processing, maintaining the same accuracy in both open ocean and coastal areas. The current Coastal Ocean and Altimetry Processor at NOC (COAPN) is based on the ALES algorithm and uses a state-of-the-art range and geophysical corrections from the RADS database to produce sea level information such as sea surface height, significant wave height (SWH) and the backscattering coefficients.
COAPN is already producing outputs from six altimetric missions, namely Jason-1, Jason-2, Jason-3, Envisat, SARAL/ALtika and the PLRM mode of Sentinel-3A. Results from this processor have already been used in several projects such as Coastal Risk information Services (C-RISe) founded by the UK Space Agency,
Copernicus Evolution and Applications with Sentinel Enhancements and Land Effluents for Shores and Seas (CEASELESS) funded by the H2020-EU. COAPN has also been integrated on the Coastal water Research Synergy Framework (Co-ReSyF) platform as a part of H2020-EU project for the Jason series missions.

In this work, we aim to evaluate the performance of COAPN from the aforementioned altimetric missions based on the distance and orientation from the coast. The ALES retrieved range and SWH will be compared against tide gauge data. In addition, the consistency of the noise level of each altimeter data between the open ocean and the coastal zone for the sea level and SWH will be shown. The results of this research will enhance our understanding of the performance of the ALES retacker for each mission closer to the coast.

Area D - Board 442

As the CryoSat-2 altimeter (SIRAL), the Sentinel-3 altimeter (SRAL) features the novel Synthetic Aperture Radar (SAR) mode that enables higher resolution and more accurate altimeter-derived parameters in the coastal zone, thanks to the reduced along-track footprint.
Exploiting the SAR data in the recent years, many researchers have shown that the performance of SAR altimetry with specific coastal retrackers is superior to collocated Pseudo-Low Resolution Mode (PLRM) coastal altimetry but they also pointed out that residual errors due to land contamination are still present in the close proximity of land (0~3 km).
The objective of this work is to move on from this context and attempt to provide even better results exploiting some extra information provided by SAR altimeter such as the Range Integrated Power (RIP).
Indeed, one of the novel data products of a SAR altimeter is the L1B-S or Stack product. This is a collection of all the Doppler beams steered to a fixed ground cell location.
From this new product, it is relatively straightforward to build a new waveform by integration of the Doppler Beams in the range direction, while the classical SAR waveform is built by integration in the azimuth direction. This new waveform is referred to as the RIP and characterises the backscattering nature of the ground cell where all the Doppler beams are steered to.
In this new retracker, that we name SAMOSA++, the RIP, as computed from the L1B-S data, will be converted into a surface backscattering law and directly fed into the SAMOSA model as part of the model formulation. In this way, the SAMOSA model will be able to cope with the different return waveform shapes from different surface types: either diffusive or specular.
The RIP also allows for the estimation of the mean square slope, a novel output for SAR altimeter waveform retrackers. The geophysical significance of these estimates will be investigated.
The performance of the new retracker on Sentinel-3 data will be cross-compared, in the coastal zone and open ocean, against the SAR mode SAMOSA+ retracker [1], the SAMOSA 2.5 SAR and PLRM retracker used in the Sentinel-3 Ground Segment. The analysis in open ocean is necessary to ensure a seamless transition between coastal zone and open ocean.
The region of interest for this study is the North East Atlantic, an area well covered by in situ data, and challenging for radar altimetry, for its complex coastal morphology and high tide dynamics.
University of Porto has provided the GPD+ solution for the wet tropospheric correction, along the Sentinel-3 tracks [2].
Furthermore, the results of the SAMOSA++ retracker will be cross-compared against the network of tide gauges and buoys in the German Bight managed by BfG (German Federal Institute of Hydrology) and WSV (German Waterway and Shipping Administration) and versus the output of the “Bundesamt für Seeschiffahrt und hydrographie” (BSH) regional circulation model and the wave forecast system of the “Deutscher Wetterdienst” (the DWD wave model).
The mean square slope output from the SAMOSA++ retracker will be cross-compared against the mean square slope from ECMWF’s operational model.
Our goal is to assess the capacity of the SAMOSA++ SAR mode retracker to provide more accurate altimetric measurements closer to the coast than provided by SAMOSA+.

References:
[1] https://doi.org/10.1016/j.asr.2017.12.018
[2] http://dx.doi.org/10.3390/rs8100851

Area D - Board 443

Calculation of water exchange between different parts of the Caspian Sea is an important task both for knowledge of heat transfer and for environmental studies (for example, transboundary pollutant transport). Existing numerical models describe well the 3D dynamics of the Caspian Sea, however, none of them takes into account significant fluctuations in the sea level observed in recent years, which is a characteristic feature of the Caspian Sea.
Since 1993 satellite altimetry is used for satellite monitoring of the Caspian Sea level, which rose by 2.5 m from 1977 (-29.0 m a.s.l.) till 1995, and then has decreased by 1.5 m till present (-28.0 m a.s.l.). Satellite altimetry data showed uneven interannual sea level variability in different parts of the Caspian Sea, which is explained by different hydrometeorological conditions resulting in significant variations of evaporation, precipitation and river runoff over the area of the Caspian Sea, which is as long as 1200 km from North to South. Satellite altimetry data can also be used in the study of water exchange between the Northern and Middle Caspian, as well as between the Middle and Southern Caspian. For this purpose, satellite altimetry data derived from Topex/Poseidon and Jason-1, -2, -3 satellites taken along the 133 and 209 tracks were analyzed. The position of these tracks does not coincide with the geographical boundaries between the Northern, Middle and Southern Caspian, however, this methodology allowed to assess water exchange between the main parts of the Caspian Sea.
The results of calculation showed that the water flow between the Northern and Middle Caspian is strongly influenced by the Volga River runoff to the Caspian Sea. Water exchange between other parts of the sea is also substantial.
The study was funded by the Russian Science Foundation research project No 19-77-20099 «Investigation of the climatic variability of thermo-hydrodynamic regime of the Caspian Sea based on remote sensing data».

Area D - Board 444

The White Sea is a southern inlet of the Barents Sea located in the northwest of European Russia. It is surrounded by Karelia to the west, the Kola Peninsula to the north, and the Kanin Peninsula to the northeast. The climate is cold, it varies between polar and moderate continental with frequent fogs and clouds. Winds are predominantly southwestern in winter with speeds of 4–8 m/s, which bring cold air from the south, establishing in February air temperature of about −15°C. Arctic anticyclones can change winds to the northeastern ones, bringing much colder weather with air temperatures of about −25°C. Summers are cold, cloudy and relatively humid, with northeastern winds and frequent rains. Average air temperature in July is 8–10°C, however, occasional southeastern winds bring warm air masses from Europe, raising the temperature to 17–19°C and sometimes even to 30°C.
The consistency between satellite altimeter derived wind speed and measurements on coastal stations of the White Sea was checked in the present study. For different parts of the coastal zone, satellites tracks were chosen to be located close to the weather station (within 15 miles). Then, for every cycle of the selected satellite passes, corresponding values of satellite derived wind speed together with corresponding time (UT) were selected. In situ wind speed data for the time nearest to the satellite pass were picked from the databases collected at coastal weather stations. For the comparative analysis, six places on the White Sea coast were selected (Kem’ Port, Kanin Nos, Solovetsky Islands, Svyatoy Nos, Kandalaksha, Umba) and tracks of satellites TOPEX/Poseidon, ERS-1/2, Jason-1/2/3, ENVISAT and SARAL.
Improvement in correlation between meteo and altimeter data on wind speed can be obtained when using method of decomposition of the winds in four quadrants according to wind direction relative to local coastline direction.
The study was carried out with the financial support of the Russian Foundation for Basic Research project No. 18-05-01053_a «Investigation of hydrometeorological and hydrodynamic regime of the White Sea based on satellite altimetry».

Area D - Board 445

Traditional monitoring of the Arctic sea ice cover and ice edge location is based on radar (SAR), optical, infrared and microwave radiometric data. Optical and infrared methods require cloudy free conditions which are rare in the Arctic. During polar night optical method does not work as well. Microwave radiometric data have spatial resolution of about 25 km, and SAR data are expensive. Satellite altimetry data can be also used to estimate ice cover extent or position of an ice edge.
For this purpose, 34 descending tracks of ERS-1 (phase С and G), ERS-2, ENVISAT and SARAL satellites have been chosen. These ground tracks are located at the optimal angle to the mean climatic sea ice edge position in the Barents Sea. Significant sea surface height (SSH) records along the satellite tracks over the sea surface and over the ice fields strongly differ. The point of the signal change located on the track was identified as an ice edge.
From 1995 to 2018, the sea ice edge didn’t reach its mean climatic position along the track 988. For the track 932, the sea ice edge observed in winters of 1996, 2003, and 2004 was displaced southwestward by 60, 40, and 53 km, respectively, from its mean climatic position. For both tracks, winters of 2000 and 2007 were found to be very warm.
According to our calculations, interannual trends of sea ice edge propagation northeastward along the tracks 988 and 932 were 13.2±0.6 and 10.7±0.4 km/year, respectively. For summer months, the trends along the same tracks were 24.8±1.2 and 8.2±0.5 km/year, respectively. Averaged velocity of the ice edge along-track displacement northeastward of the Barents Sea for all tracks was 11.8±0.5 km/year for 1995-2018 period.
The study was funded by the RFBR research project No 18-05-80065 «Analysis and forecasting of dangerous hydrometeorological phenomena in the coastal areas of the Arctic zone of the Russian Federation».

Area D - Board 446