It is widely recognised that, apart from the information of the specific variables under study, EO maps of ocean scalars convey useful information about ocean dynamics. Fronts, eddies, and even submesoscale instabilities quite frequently have an apparent signature on remotely sensed maps of variables as diverse as Sea Surface Temperature, Sea Surface Salinity or Ocean Color. Using that qualitative assessment to obtain more quantitative information about the ocean state is however rather complex and not straightforward. Some approaches, such as Surface Quasigeostrophic theory, allow to calculate surface currents from scalars such as SST once the appropriate assumptions and approximations are made, what limits their range of validity to specific instances.

Singularity Analysis (SA) was introduced in the context of the Microcanonical Multifractal Formalism a decade ago as a way to characterize the structure of ocean scalars in terms of their fronts of singularity. By means of SA, each point is assigned to a value of Singularity Exponent (SE), which has no units and is a normalized measure of the local regularity of the scalar. It was shown that the SE derived from different scalars are in good correspondence, therefore being related to a structure property of the underlying flow. Therefore, SA provides access to the structural information of the flow from the knowledge of different scalars. The singularity structure has been used to develop new methods for improving the quality of a scalar using another scalar of different type and higher quality, without using any numerical model for relating them.

Very recently, a new approach to the analysis of the correspondence of ocean scalars has been introduced, that of Local Information Content or Local Entropy (LE). Using LE, we can compute the average amount of information conveyed by any point in a EO map once statistical stationarity is assumed and provided that the resolution of the data is fine enough. We have empirically verified that the maps of LE are remarkably close to those of SE, although the former provide information content (they are expressed in bits) and the later provide the degree of regularity of the scalar (equivalent to the degree of differentiability, and are thus dimensionless).

We have proved that, in fact, LE and SE are almost-linearly related, the slope of the linear component given by the available bit budget. We have verified that this theoretical correspondence in fact is well verified with high-resolution data such as OSTIA SST. Exploiting this connection, we can now produce maps of local information content from a low-resolution scalar by deriving LE from the associated SE. Going even further, we can compute the total information and the mutual information conveyed by two different scalars, providing an easy way to integrate the information of discontinuous, fragmentary, noisy observations. Given the relation between 1D and 2D SE, it is even possible to integrate in a rigourous manner the information conveyed from very different observation systems (for instance, transects, satellite tracks and satellite images).

Area C - Board 308

Numerous new satellites and sensors have arised during the past decade. The satellite constellation has never been better, providing us with a wide range of view angles to the ocean surface from the coast to the open ocean, at various scales and from physical to biological processes. Sentinel 1-2-3 program covers various sensors such as SAR, Optical, Infrared or altimeter with a repeat subcycle of only few days, yet the repeat frequency for each sensor alone is not enough for meso to submeso scale monitoring. Handling consistently these huge heterogeneous datasets in a simple, fast and convenient way is now possible using the free and open Ocean Virtual Laboratory online portal or its standalone version. Today, these tools are starting to be widely used by the scientific community to better understand and monitor oceanic processes. We will demonstrate the potential and functionalities of these tools using various test cases: - Collocating Sentinel-1 and Sentinel-3 data enables to detect oceanic fronts and eddies at meso and submesoscales, highlighting strong and energetic ocean currents. - Using jointly Sentinel-1-2 and Sentinel-3 helps to detect oil spills as well as their displacement. - Analyzing contemporaneous Sentinel-1-2-3 helps to assess ocean wave parameters and their intensification by surface currents or bottom topography. Interactive demo of the tools will also be available on the OceanDataLab booth. Online tool is available at https://ovl.oceandatalab.com and standalone version at https://seascope.oceandatalab.com

Area C - Board 309

As the offer in space data is now plethoric, space mission synergies has to be seriously thought of, not only at data level but also at space segment level. The following discusses the particular case of Sentinel 3/Meteosat Second and Third Generation synergy to propose information adapted to the coastal dynamics study around large estuary river.
As stated in preliminary End User interviews, coastal area dynamics monitoring is interested in multispectral and hyperspectral satellite images with a middle spatial resolution (∼ 300 m) and a high temporal repeatability (∼ 1 h). As no current satellite sensors have such features, we first propose a fusion method to merge images delivered by a low earth orbit (LEO) sensor with images delivered by a geostationary earth orbit (GEO) sensor. This fusion method, called spatial spectral temporal fusion (SSTF), is applied to Sentinel 3/Ocean and Land Color Instrument (OLCI) and first, real MSG/SEVIRI data, then simulated MTG/Flexible Combined Imager (FCI) images.
Regarding the use of OLCI and FCI, the OLCI bands, acquired at t0, are divided by the oversampled corresponding FCI band acquired at t0 and multiplied by the FCI bands acquired at t1. The fusion product is used for the next fusion at t1 and so on. The high temporal resolution of FCI allows its signal-to-noise ratio (SNR) to be enhanced by the means of temporal filtering.
In order to be able to evaluate the potential of those fused images regarding different type of applications, we have simulated OLCI and FCI images and compared the estimation of chlorophyll (Chl), suspended particulate matter (SPM), and colored dissolved organic matter (CDOM) maps from the fusion products with the input simulation maps. The results on the retrieval of Chl, SPM, and CDOM are presented.
Following this first results, we introduce also the preliminary methodology currently in development, of an automatic deep learning process aiming at computing “on the fly” those synergy data.

A discussion onto synergy use of Sentinel 3 and MSG/MTG data will conclude this presentation. What is possible at space segment level ? what can be done at ground segment level ?

Area C - Board 310

The Somali Current is an annually reversing boundary current that flows along the coast of Somalia. During the summer Southwest monsoon, the current flows parallel to the monsoon winds with a speed that can exceed 2 m/s. Part of the northeastward flowing current diverts at about 3°N to form the Southern Wedge. The remaining flow bifurcates around 10°N and either continues past the Gulf of Aden or separates to form the Great Whirl. The width of the current between these features is highly variable, with a typical width of 50 km. The Southwest monsoon wind directions are steady throughout the summer, with speeds that frequently exceed 15 m/s.
Satellite observations have shown that there is significant covariability between wind stress and sea surface temperature (SST). In the presence of fronts, the marine atmospheric boundary layer is more stable over the cool side and more unstable over the warm side. The wind stress is reduced over the cold water where the surface is decoupled from stronger upper layer winds. In contrast, the decreased stability over the warm side draws momentum from upper layer winds and increases the surface wind speed. Surface wind speeds are also affected by ocean currents since satellite radars measure relative motion between the air and the ocean.
The effects of SST fronts on near surface winds has been demonstrated in multiple studies. These previous studies used relatively low resolution satellite scatterometer and microwave radiometer data along with temporal averaging and high pass filtering to facilitate the identification of the covariability of wind and SST. These previous satellite observations of wind and SST variability focused on temporally averaged mesoscale features. We use higher resolution measurements (~1 km) winds from synthetic aperture radar (SAR) and SST from the Moderate Resolution Imaging Spectroradiometer to analyze the sub-mesoscale air-sea coupling over the Somali Current. Sea surface roughness images have the ability to show ocean surface features at high resolution and provide measurements through clouds. High resolution SST measurements are readily acquired, but are limited to clear sky conditions. Persistent clouds along the Somali coast frequently mask observations that produce a bias in the SST in this region.
SAR data from Radarsat-2, TerraSAR-X and Sentinel-1 were acquired over the Somali current during the 2016 – 2018 summer Southwest monsoons. The Somali Current, the Southern Wedge and the Great Whirl are well developed during these summer acquisitions. Wind stress and curl computed from the SAR data clearly show mesoscale and sub-mesoscale features along the Somali Current and the Great Whirl. Over the Somali Current, between the Southern Wedge and Great Whirl, the high resolution SAR winds show strong linear covariability at sub-mesoscales. At these scales, variability in this relationship is shown to be due to spatial and temporal variability of the SST along the Somali Current front.

Area C - Board 311

Satellite infrared radiometers have unveiled the existence of distinct turbulence regimes in the upper ocean that can evolve very fast and strongly depend on the environmental conditions. Some of such regimes bring associated changes on how the energy injected at larger scales is transferred to smaller scales, i.e. the energy cascade. In this study, we use Sea Surface Temperatures (SST) provided by the (A)ATSR/SLSTR family of sensors to investigate a fundamental property of the energy cascade: its intermittency. Contrary to the standard approach based on the analysis of spectra and structure functions, we base our study on the geometrical properties of SST fronts. Our results show that turbulence regimes dominated by Mixed Layer Instabilities (MLI) at submesoscales are more intermittent than turbulence dominated by mesoscale vortices and the associated submesoscale filaments. Furthermore, our results suggest that intermittency can be fully characterized by the intensity of the strongest thermal fronts. The analysis of the fractal dimension of oceanic fronts also unveils that the standard model of intermittency, the log-normal model, does not fit to observations and should be discarded in favor of other models such as the log-Poisson model. Beyond the insights on the energy cascade and its intermittency, our study point to new approaches to improve the existing cloud masks applied to SST images. Moreover, the multimission approach we have followed also provide insight on the calibration of long time-series of SST for climate studies.

Area C - Board 312

Internal waves (IW) have recently been shown to significantly contribute to sea level variability at scales lower than about 100 km [Richman et al., 2012, Rocha et al., 2016, Savage et al., 2017]. The contamination of IW at theses scales limits our ability to predict ocean currents from altimetric data because IW currents are not related to sea level via geostrophy unlike slower/mesoscale balanced structures [Arbic et al., 2015]. Some IW are of tidal origin and stationary with respect to tidal forcings such that the signatures of these waves on altimetric data can be predicted and can eventually be removed from altimetric data [Ray and Zaron, 2016]. Unfortunately most IW are either 1) from tidal origin but have lost their phase relationship with tidal forcing after interaction with the slower oceanic turbulence (non stationary internal tides, Zaron [2016]) or 2) from non tidal origin (e.g. windforced, lee-waves). In both cases, it is difficult to these waves nowadays.
A direction of research that has been poorly explored yet deals with the signature of IW on satellite data of nonaltimetric nature (e.g. sea surface temperature, color) and potential synergies in order to disentangle IW and balanced signatures in altimetric data. We recently proposed in an idealized context a tentative method in order to carry such synergies [Ponte et al., 2017]. While it is highly premature to attempt to apply such method with real data, the objective is here to determine the amplitude of IW signature on SST and verify its smallness compared to mesoscale/submesoscale signatures. This can be achieved given an a priori knowledge of IW currents (available for stationary internal tides from existing numerical model predictions and/or from existing harmonic analyses of altimetric datasets) and maps of the SST field. In this work, we will show a first attempt to quantify the IW signature on SST fields through two different case studies that use infrared SST images captured by orbiting satellites (Metop,Viirs and ENVISAT).

References
B. K. Arbic, F. Lyard, A. Ponte, R. Ray, J. G. Richman, J. F. Shriver, E. Zaron, and Z. Zhao.Tides and the swot mission: Transition from science definition team to science team. Technical report, CNES-NASA, 2015.
A. L. Ponte, P. Klein, M. Dunphy, and S. Le Gentil. Low-mode internal tides and balanced dynamics disentanglement in altimetric observations: Synergy with surface density observations. J. Geophys. Res., Oceans, 2017. ISSN 2169-9291. doi: 10.1002/2016JC012214. URL http://dx.doi.org/10.1002/2016JC012214.

R. D. Ray and E. D. Zaron. M 2 internal tides and their observed wavenumber spectra from satellite altimetry. J.
Phys. Oceanogr., 46:3–22, 2016.

James G. Richman, Brian K. Arbic, Jay F. Shriver, and E. Joseph Metzger. Inferring dynamics from the
wavenumber spectra of an eddying global ocean model with embedded tides. J. Geophys. Res., 117(C12012),
2012. doi: 10.1029/2012JC008364.

Cesar B Rocha, Sarah T Gille, Teresa K Chereskin, and Dimitris Menemenlis. Seasonality of submesoscale dynamics in the kuroshio extension. Geophys. Res. Lett., 2016.

C. Savage, B. K. Arbic, J. G. Richman, J. F. Shriver, M. H. Alford, M. C. Buijsman, J. T. Farrar, H. Sharma, G.
Voet, A. J. Wallcraft, and L. Zamudio. Frequency content of sea surface height variability from internal gravity
waves to mesoscale eddies. J. Geophys. Res., Oceans, pages n/a–n/a, 2017. ISSN 2169-9291. doi: 10.1002/2016JC012331.

E. D. Zaron. Mapping the non-stationary internal tide with satellite altimetry. Journal of Geophysical Research:
Oceans, 2016.

Area C - Board 313

Accurate knowledge of spatial and temporal ocean surface currents at high resolution is essential for both navigation and operational applications. Therefore, a key problem in oceanography is the estimation of the synoptic velocity field. Currently, global ocean surface velocities are routinely estimated from Sea Surface Height (SSH) measurements provided by altimeters. Along-track altimetric measurements of SSH are very well suited to quantify across-track currents, but the spatial resolution of the derived 2D velocities from conventional altimeters is limited to scales above 100-150 km. We present here an approach that takes advantage of the higher temporal and spatial resolution provided by Sea Surface Temperature images (SST). It basically consists on the characterization of a transfer function between SST and SSH that allows to retrieve a synoptic ocean surface current field when applied to a single snapshot of SST.
In this work, we first show that under certain circumstances, the information contained along the track may be enough to characterize this transfer function based on model outputs from the Mercator-Ocean system. Then the method is applied to real and simultaneous SST and along track SSH observations provided by ENVISAT and Sentinels 3 satellites. The quality of the method is assessed through comparisons to other surface current estimates, such as HF radar derived currents and in-situ surface currents provided by drifter buoys. The results show the capability of the proposed approach to reconstruct the velocity field associated to coastal eddies with diameters of ~10 km, but also allow to unveil the different limitations of our approach. One of such limitations is the lack of Sea Surface Salinity (SSS) measurements. Indeed, in spite that SST and SSS gradients tend to be aligned, their sign may be inhomogeneous. To overcome this limitation, we plan to exploit the synergy between SSH and the new generation of SMOS-derived SSS provided by the SMOS the Barcelona Expert Centre (bec.icm.csic.es).

Area C - Board 314

Eastern Boundary Current Systems are zones of critical importance from a socioeconomic point of view by being among the most productive marine ecosystems on the planet. Coastal upwelling, in particular in the California Current System (CCS), has been well studied over the past decades, and lasts from Spring to early Fall, and occurs mainly in the central part of the region, from 43ºN to 33ºN. The presence of sea surface temperature (SST) fronts in the region has been documented in numerous observational studies, and is often directly associated to coastal upwelling. However, satellite observations (altimetry and infrared SST measurements from Pathfinder Advanced Very High-Resolution Radiometer) suggest that SST fronts can be seen in the CCS year-round, and that the southern region (off Baja California, from 33ºN down to 20ºN) may possess distinct frontal dynamics from the central region. In this study, we propose to investigate the variability of SST fronts in the CCS, and in particular the frontogenesis mechanisms in the southern CCS in contrast to the coastal upwelling system of the central CCS.
Our analysis relies mainly on a combination of satellite observations (altimetry and SST) and a high-resolution (1/48º) ocean numerical model, the MITgcm LLC4320 simulation, where LLC stands for latitude-longitude-polar cap. The altimetry products are obtained from the Copernicus Marine Environment Monitoring Service -CMEMS, while the daily SST fields obtained by Pathfinder at a spatial resolution of 4km were used to detect SST fronts in the CCS. The simulation produces output hourly, and on 90 vertical levels. We first derived vertical velocities, horizontal velocities, as well as strain, divergence and vorticity fields at several time scales (daily, monthly and seasonal), to understand their evolution over a year. We also computed the temperature, salinity and density fields at the same time scales, and an algorithm was developed to detect and track SST fronts in the simulation.
The modelled representation of the coastal upwelling frontal system is in good agreement with the findings from existing literature on the topic and our observations. In addition, our preliminary results seem to confirm the satellite observations: the frontal dynamics in the southern CCS differ from the rest of the region, and we suggest that it may be associated with water masses convergence involving the Subarctic waters, the Subtropical waters and the Tropical waters during Summer and Fall. The mechanisms resulting in the presence of SST fronts off Baja California appear to be drastically different than in the rest of the CCS, both in the origin of the SST gradients and in the formation of the fronts. The main characteristics of these fronts (cross-frontal SST step, SST gradients) and their seasonality also differ from coastal upwelling fronts, with potential biogeochemical implications in the region off Baja California. This study shows that Eastern Boundary Current Systems cannot be understood only as Coastal Upwelling Systems, as it is often the case in existing literature. In fact, besides coastal upwelling, other processes can give rise to submesoscale fronts in the CCS, and submesoscale activity in general is not observed exclusively in the coastal upwelling region or during the coastal upwelling season, but rather across the whole CCS and year-round.
SST frontal dynamics were explored in one of the four Eastern Boundary Current Systems by using a synergy of datasets; and new frontogenesis mechanisms were discovered, which may explain the presence of fronts outside of the coastal upwelling framework. A future study in the remaining Current Systems may bring to light similar submesoscale dynamics, which would expand our understanding of these regions.

Area C - Board 315

The ongoing temporal data “explosion” in Earth Observation imagery allows us to analyse 30+ year archives of ocean surface measurements. While multiple hyper-temporal analysis tools exist for terrestrial applications, unsupervised data-driven methods are needed to extract and identify spatio-temporal ocean surface patterns. In particular, heterogeneity mapping of the ocean surface is needed to underpin time series analysis of satellite-derived ocean parameters including sea surface temperature and chlorophyll. This approach can help optimise in-situ sampling efforts, with significant cost savings. Validating and interpreting such heterogeneity maps presents a complex challenge, and requires the examination of multi-variable records of the various contributory ocean-surface features.

This research adapts a terrestrial approach, to quantify spatio-temporal heterogeneity of ocean-surface parameters. This Ocean-surface Heterogeneity MApping (OHMA) algorithm was used to extract annual heterogeneity summaries from three hyper-temporal datasets (2011, 2012, and 2013) of GHRSST multi-sensor SST data. The OHMA algorithm produces a classification ensemble relying on optimised iterations of ISODATA classification. The ensembles highlight spatio-temporal homogeneous and heterogeneous ocean regions. The approach is entirely data-driven to meet current limitations when working with hyper-temporal ocean data, and to overcome challenges of obtaining ocean surface in-situ data which drive more supervised approaches.

The OHMA map is validated for 2011 using measures of in-situ spatio-temporal heterogeneity. These statistical measures (e.g. maximum standard deviations and mean kurtosis, amongst others) are extracted from in-situ transect measurements of surface temperatures, collected by underway seagoing vessels. Preliminary results indicate the OHMA science-based product successfully conveys ocean surface heterogeneity information. However, a range of statistical measures of the in-situ data are required to fully capture the various characteristics of ocean surface heterogeneity expressed in the surface temperature data collected along ship transects. Efforts undertaken to clarify interpretation of an OHMA map clearly showed that the team’s initial theory, that heterogeneous regions can be explained primarily by the presence and magnitude of SST fronts, was inadequate, as other mesoscale features (such as eddies and surface currents) also influence the produced map. Measuring the contribution of the various mesoscale features to a region’s heterogeneity requires a multi-feature (and therefore multi-variable) synergistic approach. Continued consultation with the oceanography community is helping to explore and identify the driving mechanisms of unexplained features, and to select suitable satellite-derived data products of different variables for further comparative analysis. Consultations have also highlighted that OHMA outputs will have a variety of uses across disciplines, with potential climatology, teleconnections, and biodiversity applications.

This presentation gives an overview of the OHMA algorithm, its strategic context, its potential utility, and transferability. It presents the characterisation and validation efforts undertaken to support the needs of the oceanographic community, and highlights the value of considering multi-variate data when quantifying the range and influence of mesoscale features on a hyper-temporal analysis product.

Acknowledgements: This research has been funded by the European Union, through the Horizon 2020 research and innovation programme under grant agreement no. 687289 (Co-ReSyF), and previously under the FP-7 programme under grant agreement no 607155 (SAFI). For more information on the Co-ReSyF project, see www.co-resyf.eu. Meanwhile, further information on the SAFI project can be found at www.safiservices.eu.

Area C - Board 316

In the paper, we present results of studies of submesoscale fronts in enclosed seas on the base of remote sensing data. Fronts and frontal zones represent one of the essential elements of water dynamics, namely the natural boundaries between waters with different thermodynamic parameters. Fronts in the seas are the zones where properties of water masses change sufficiently at a relatively short distance. They are notable for having a complex structure and a high spatial-temporal variability. Considerable attention is given to studying these phenomena by scientists worldwide. The rapid development of these studies has been greatly stimulated by the improvement of satellite remote sensing methods in the infra-red (IR) and visible (VIS) ranges of the electromagnetic spectrum. However, to the present day, mostly large-scale oceanic fronts in climatic and synoptic frontal zones remain the subject of numerous experimental and theoretical studies
A significant limitation for observing the sea surface from space in the optical range is the presence of clouds and insufficient illumination. We show that we can significantly broaden our insights into the physics and geography of small-scale fronts in inland seas by employing satellite synthetic aperture radar (SAR). The Synthetic Aperture Radar (SAR) is a very informative tool because of its high spatial resolution and its insusceptibility to cloud and solar light conditions. Use of satellite SAR data in the research of small- and mesoscale fronts in inland seas allowed to significantly improve our understanding of their generation, propagation, and decay.
During our long-term monitoring of the Black Sea and the south-eastern Baltic Sea, we came across a significant number of surface manifestations of various fronts in satellite imagery. In our presentation we: (1) demonstrate the capability of SAR as a tool for fine-scale marine surface analyses and for detecting of sub-mesoscale fronts of various types; (2) present examples of fronts commonly detected in the SAR data of the study region; (3) explore the SAR signatures of fronts of various origin.
The front dynamics survey is based on the combined use of satellite imagery obtained by sensors installed on various satellites and operating in different bands of electromagnetic waves. Data in visible, infrared and microwave ranges provided the primary material for the analysis. We use high-resolution radar imagery data obtained by SARs onboard Envisat, Sentinel-1A, B, TerraSAR-X, and RADARSAR-2 satellites. Processes occurring on the fronts and in the frontal zones have the most diverse characteristics and scales, and it is not always possible to pinpoint typical features that would allow us to detect them in satellite images with 100% confidence. We validate our interpretations of front features in SAR images via the combined analysis of data on the sea surface and marine atmosphere condition, sea surface temperature, and mesoscale water dynamics. For more robust interpretation, we analyze the SAR data jointly with other satellite data on the sea surface and water condition, sea surface temperature, and mesoscale water dynamics. We use data in visible and IR bands taken by Envisat MERIS, Terra/Aqua MODIS, by scanning radiometers of Landsat-5, 7, 8 satellites. New Sentinel-2 MSI and Sentinel-3 OLCI data are also used. Meteorological data, required for the analysis of satellite data, are provided by coastal stations.
We examine SAR manifestations of fronts in the sea, which are made visible in SAR data due to various mechanisms of front effects on the disturbed sea surface. We focus on the small-scale fronts, which are often short-lived and have narrow frontal interfaces. By their very nature, oceanic fronts are associated with a convergent flow, which brings fluid of different properties into the frontal region. Fronts can be visible at the sea surface as delimitation lines visualized by (i) rough water or slicks (smooth surface); (ii) accumulation of debris or foam; (iii) abrupt change of sea surface temperature or color. Rough water is frequently observed when there is a significant current shear across the front. Slicks may be caused by the damping effect of surface films accumulated in the convergence zones. Accumulation of debris and foam arises from the convergence of surface currents at a front.
Fronts are visible in SAR images due to a change of the small-scale sea surface roughness responsible for the radar backscattering. We discuss several hydro-physical mechanisms connected with oceanic fronts that result in a modulation of the sea surface roughness: (i) wave-current interaction; (ii) damping of the gravity-capillary component of surface waves by surface films; (iii) changes in the stability of the air-sea boundary layer.
We focus on the small-scale fronts, which are often short-lived and have narrow frontal interfaces. We examine SAR signatures of sea surface fronts that differ in types, such as river plume fronts, small-scale thermal fronts, shelfbreak fronts. The SAR signatures specific for different frontal types are outlined that can be used for the identification of fronts by their kind.
Basing on satellite data, we estimate spatial-time scales of a front formation. We examine hydrological front generated by the outflow of Danube waters into the Black Sea and estimate the impact of its frontal instability on the formation of vortex structures and internal waves.
The work was supported by the Russian Foundation of Basic Research under the project #17-05-00715 and by the Program of RAS Presidium #1.2.50.

Area C - Board 317

Marginal ice zone (MIZ) is an area of effective interaction between sea ice, ocean and atmosphere. The exchange of energy and material in MIZ occurs on rather short time- and spatial scales, which makes challenging the collection of data possible to use for analysis. Combination of satellite and in situ data helps to obtain more complete image of the mesoscale dynamics in MIZ. It should be noted that the term “mesoscale” used in this work refers to the Rossby radius of deformation, which is about 4-8 km for Eastern Arctic, our main area of study [1].

The most appropriate type of satellite measurements for MIZ studies is synthetic aperture radar (SAR), which allows to obtain the high-resolution information on sea ice presence and wind speed (e.g. up to 40 m with Sentinel-1/SAR-C Extra-wide mode). Optical, passive and active microwave measurements have lower spatial resolution and, in case of optical band, dependency of cloud coverage and day/night, but should also be added to analysis, when available to extend the time series of observations and provide other geophysical parameters, as SST, cyclone presence and cloud fraction, large-scale wind directions (12.5 km in case of ASCAT) and chlorophyll-a concentrations. For this work, we collected time series of Sentinel-1/SAR-C, MODIS, AMSR2, and ASCAT data. The sequence of Sentinel-1 EW GRD products was used to define the sea ice cover, calculate the ice drift and wind speed (using CMOD5 GMF).

Remote sensing data were enriched with following in situ data obtained during NABOS-2018 expedition on board R/V Akademik Tryoshnikov in Eastern Arctic Ocean: vessel mounted ADCP, CTD profiles, thermosalinograph underway continuous data of the surface water layer (equipped with temperature, conductivity, and chl-a sensors), and meteorological measurements.

Spatial distribution of temperature, salinity and chlorophyll-a from in situ measurements were combined with remote sensing estimations to follow different water mass mixing; mixed layer depth and buoyancy were calculated from CTD vertical profiles to analyze the nature of observed surface and subsurface currents. Atmospheric forcing was used to calculate wind drift for sea ice and open water. Results of analysis are presented in this work.

[1] NURSER, A. J. G. et BACON, S. The Rossby radius in the Arctic Ocean. Ocean Science, 2014, vol. 10, no 6, p. 967-975.

Area C - Board 318

Here we present the results of satellite observations of eddies over the Chukchi and Beaufort Seas. The study is based on analysis of high-resolution Envisat and Sentinel-1 synthetic aperture radar (SAR) images acquired in June-October of 2007, 2011, and 2016; complimented by low-resolution multi-mission satellite altimetry data from May 1993 to December 2016. Based on these two different data sources detailed maps of eddy observational frequency, their spatial and kinematic properties are obtained and compared with high-resolution (1 km) MITgcm model output.
Taking advantage of high-resolution SAR observations the analysis includes eddy features encountered both over the ice-free and in marginal ice zone (MIZ) regions. Most of eddy signatures were observed in August and September. The vorticity sign of 64% of all eddies (both open-water and MIZ eddies) is cyclonic. The large number of open-water eddies are found over the southern and central parts of the Chukchi Sea, along the shallow shelf of the southern Beaufort Sea and in the vicinity of MacKenzie River delta. For MIZ eddies, most of them were observed north of Chukchi Plateau and over some parts of the Beaufort Gyre.
Open-water eddies were observed over the depth range from 20 m to 3780 m, while about 80% of them were recorded over the depth not exceeding 200 m depth. MIZ eddies have the same range of depth values, yet about 80% of these eddies were found over the depths exceeding 1000 m (mean depth 2000 m). The range of eddy diameters for open-water (MIZ) eddies was 1-34 km (1-78 km). Diameter values of about 90% of open-water eddies (1294 eddies) don’t exceed 10 km, i.e. submesoscale eddies being clearly dominant in SAR observations. Mean (median) value of open-water eddy diameter is 5.8 km (4.9 km), while it is 9.2 km (7.2 km) for MIZ eddies.
Unlike SAR data, results of altimetry-based observations provide very important information about larger mesoscale eddies in the region and allow to infer their mean orbital velocities, lifetime and propagation trajectories. As observed, the mean (maximal) eddy diameters inferred from altimetry data are about 50 (100 km) and their lifetimes are about 10-20 days, while the mean eddy orbital velocities range around 10-30 cm/s.
This work is supported by RFBR grant 18-35-20078, RSF grant No. 18-77-00082. The support of FASO Russia State Assignment No. 0827-2018-0002 and Davidow Discovery Fund are also acknowledged.

Area C - Board 319

Images of sea surface roughness contain information on gradients of surface currents at very high resolution. This could complement other measurements of absolute surface currents at lower resolution. As such, it is essential to understand the information provided by the sea surface roughness.
We review here the recent progress made to understand which property of the surface current gradient (its deformation, its norm, its direction, its spatial extension) can be retrieved. We show that its depends on the characteristics of the surface roughness measurements: optical or radar, slanting or nadir look angle, single look or multi-look, low or high resolution.
We illustrate possible uses in synergy with lower resolution absolute currents. We discuss its application for the satellite upcoming mission SWOT (high resolution altimetry) and proposed mission SKIM (real aperture radar).

Area C - Board 320

Today, the global ocean surface current products (GlobCurrent, OSCAR, AVISO, CTOH…) are mainly based on the geostrophic and the Ekman contributions derived from altimetry and wind observations, respectively. The higher frequency processes, such as the ocean tides, are not included in these products, although they often represent one of the main contributors to the ocean currents on the shelves.
Australia’s marginal seas include a wide range of ocean current regimes ranging from tide-dominated in the North-West where the continental shelf is wide, to boundary current- and eddy-dominated in the South-East, where the shelf is narrow. Here, we take the opportunity to test the GlobCurrent surface ocean current products against in-situ observations in these two contrasting regimes. Observations by Acoustic Doppler Current Profiler (ADCP) instruments of Australia’s Integrated Marine Observing System (IMOS) and drifting buoys of the Global Drifter Programme (GDP) are used.
The monthly-timescale variability of the GlobCurrent alongshore current component is in moderately good agreement with the observations on the continental shelf in the South-East but neither the shorter period variability nor the long-term mean are representative of the ADCP observations. The error of the time-mean velocity, derived from the Mean Dynamic Topography (MDT) CNES-CLS13, is the dominant error because that product is not designed to represent the highly anisotropic nature of the sea level gradients over the continental shelf nor the constraints on the flow field that are imposed by the topography. While the observed tidal currents are negligibly small, the wind-driven signals are not, but these are too transient to be adequately sampled by altimetry. The inclusion of an Ekman component does not represent these signals because the interaction of the Ekman transport with the coastal boundary condition is not included. We thus conclude that the GlobCurrent product needs improvements before it can be described as very suitable for applications on this narrow continental shelf. Off the continental shelf, in contrast, the GlobCurrent products compare quite well with the trajectories of drifting buoys, confirming that the products are quite suitable for blue-water applications.
In contrast to the South-East, the tides are very strong in the North-West region of Australia. The sub-tidal variability is weak, in both relative and absolute senses. Consequently, the removal of the tidal signal from the sea level observations needs to be very complete for the residual error to be smaller than the true sub-tidal signal. Transient wind forced signals are also occasionally large so this step of the de-aliasing also needs to be very accurate. Unfortunately, it appears that more work is required before accurate estimates of sub-tidal variability are available from GlobCurrent: the magnitude of the GlobCurrent estimates of sub-tidal current variability far exceed the magnitude of, and are uncorrelated with, the detided ADCP data.

Area C - Board 321

In the context of an altimetry constellation observing the large scales (above 200km wavelength) of geostrophic currents, it is worth exploring the potential synergy with observations of the full surface current such as SKIM. This later includes geostrophic motions, but also Ekman, inertial oscillations, tidal (barotropic and baroclinic) surface currents and other ageostrophic processes.
Observing the sum of these components has an obvious added value to the single geostrophic component presently observed. In this study, we propose to quantify the contribution of the two observing systems combined to map the mesoscale (100km,10days) circulation.
Observing System Simulation Experiments (OSSEs) are conducted from very high resolution simulations resolving all components of the surface currents and topography. A multivariate optimal interpolation scheme is used to combine along-track topography with radial velocity synthetic observations unevenly distributed in time and space, following the sampling of the SKIM design (Ardhuin et al., 2017). Prescribing accurate covariances for both topography and currents, accounting for all components and their best linear dependences (cross current/topography terms) is the challenge for exploiting the synergy. In a few key regional configurations (western boundary current, low-latitudes, eastern basins, …) we will quantify the mesoscale reconstruction with each system separately and with their combination, through metrics of residual errors and effective resolution of the signal.

Area C - Board 322

The ocean is absorbing over 90% of the heat from the warming of the planet by increased greenhouse gasses, of which about 25 % the dominant carbon dioxide released to the atmosphere is absorbed by the ocean. The ocean is thus playing a critical role in climate change via the processes of upper ocean circulation, especially its vertical velocity associated with the vorticity of the circulation. These processes take place at scales shorter than the resolution of the present constellation of conventional nadir-looking satellite altimeters, which allows two-dimensional mapping of the sea surface height (SSH) to a resolution of 100-150 km in wavelength.

The proposed SWOT (Surface Water and Ocean Topography) mission will utilize the technique of radar interferometry to observe SSH in two dimensions over finite swaths. The mission goals are to study the upper ocean circulation at wavelengths shorter than about 150 km. A key question is: to what extent SWOT is able to provide information on the vorticity and vertical velocity of the upper ocean?

Because of the narrow swath (120 km with a 20 km nadir gap), it takes 21 days to complete the coverage of Earth between North and South 78 degree latitudes. The coverage of the mission is limited by the desire to sample ocean tides without aliasing to important oceanic time scales. The frequency of repeat observations varies from 2 at low latitudes to more than 6 at high latitudes. A challenge is to estimate the continuously evolving SSH from the irregular and infrequent sampling. Furthermore, in order to study the balanced quasi-geostrophic motions of climate importance, we must deal with the presence of internal tides and internal gravity waves in the under-sampled observations.

We are using a regional high-resolution (nested to 300 m resolution) modeling/assimilation system (including tides and internal gravity waves) to study the extent to which we can estimate the ocean state from the SWOT SSH observations. The system assimilates multiple observations from in-situ and spaceborne platforms: sea surface temperature and salinity, as well as SSH from existing constellations of orbiting satellites, upper ocean profiles of temperature and salinity from the Argo arrays, and other available operational observing systems.

We are conducting a series of simulation experiments with aims to first design an in-situ observing system that are able to provide information of the vorticity and vertical velocity of the upper ocean at the calibration/validation site of SWOT. Secondly, simulations will be conducted to construct the evolving ocean state from simulated SWOT data and compare the vorticity and vertical velocity derived from SWOT with the “ground truth” from the in-situ observations. The results will provide demonstration of the utility of SWOT observations to study the upper ocean circulation and methodologies for achieving the mission objectives.

Area C - Board 323

We conduct statistical and spectral analyses of (1) observations of coastal radar-derived surface currents and altimeter-derived geostrophic currents and (2) forward regional circulation model simulations forced with realistic wind stress, tides, and boundary conditions off the Oregon coast in view of the geostrophy and ageostrophy and the balanced and unbalanced motions. The delineation of balanced and unbalanced motions was investigated with the Helmholtz decomposition in the physical and spectral spaces and wave-vortex spectral decomposition. The observed and simulated coastal surface currents do not exhibit a clear decomposition of the balanced motions and unbalanced motions because wave-vortex decomposition may not always work in the coastal regions where multiple sources of dominant variances are present. On the other hand, the stream functions obtained from the Helmholtz physical decomposition can be a proxy for the balanced motions when the variance in vertical currents is weak. The submesoscale ageostrophic currents account for up to 50$\%$ of the total variance and are primarily associated with near-inertial currents and internal tides, which can be missing components in the currents retrieved from only submesoscale sea surface heights (SSHs). As the pure SSHs may not constrain all submesoscale ageostrophic components, the missing variance in the submesoscale SSH observations is required to complement other high-resolution {\it in-situ} observations, including HFR-derived surface velocities and surface drifter-derived tracks in view of constructive data integration to resolve submesoscale currents.

Area C - Board 324

With an increasing rate of production, plastic management has become a global concern. As a majority of plastic items are single use products, mismanagement of waste leads to a lot of items ending up at sea. Thus, plastic litter monitoring have become of great importance in the aim of the ocean and water quality protection. The IFADO (Innovation in the Framework of the Atlantic Deep Ocean) project aims to create new marine services to support the Marine Strategy Framework Directive implementation with the North Atlantic Ocean as a study case. One of the objectives of this project is to use Earth Observation products in order to characterize marine litter accumulation in the North Atlantic Ocean.
The GlobCurrent ocean surface current products are used in this study. These products consist in geostrophic and Ekman contributions derived from altimetry and wind observations, respectively. These products do not include ocean tides, even though they are a major contributor to the ocean currents on the continental shelves and in the coastal regions. As about 80% of marine litter originates from land-based sources, coastal currents have an essential role on these debris pathways, in particular for stranding. Hence, NOVELTIS used its own tidal atlas to add the surface tidal current component to the GlobCurrent data and implement a more realistic surface ocean current product. As current tidal atlases only provide the barotropic component of ocean current, NOVELTIS has developed a methodology based on maritime industry guidelines to compute the surface tidal current to be added to the GlobCurrent ocean surface current products.
This study presents an assessment of the GlobCurrent products (original and with tidal currents) against drifting NOAA buoys in the Bay of Biscay. A skill score has been implemented to assess the improvement of ocean current products when the tidal component is added. Then, these surface ocean currents products were used as input of a Lagrangian modelling tool (Opendrift) in order to highlight plastic particles convergence zones, pathways and main stranding areas. Various scenarios were considered to compute the marine litter trajectories with this ocean tides contribution. The results of this study highlight the necessity for global ocean surface current products (such as GlobCurrent) to take into account the tidal current component to improve their representation of the coastal ocean dynamics.

Area C - Board 325

Understanding of surface transport plays a key component in search and rescue operations, pathways of oil spills, debris, fish lavae etc. So far however, most of our knowledge about surface currents have mostly utilised hydrodynamical models, surface drifters, current meters and satellite data. All of the above sensors whilst capable of deriving surface currents, they are in effect resulting in different components of the surface currents in the coastal area. For instance, using the satellite altimetry sea surface height data, the geostrophic current is basically derived whilst for surface drifters the ageostrophic (non-geostrophic) components (for instance forcings by the local wind) is calculated. Thus the main focus of this study is to examine and compare satellite derived surface currents (sourced from Jason-2 and Sentinel- 3A) with that of in-situ surface drifters (that follow currents in the uppermost 2 m) and results of a hydrodynamical ocean model (using sea level and surface currents) data. The study site is that of the Gulf of Finland which is known for its semi-persistent sub(mesoscale) current patterns. It is the intention that a comparison of different sensors can assist in identifying and assessing the geostrophic and ageostrophic components of the surface currents that can be used for a deeper understanding the ocean dynamics in the study area.

Using statistical methods, a comparison is made with the different sources of data. Prior to comparison two essential tasks was performed: (i) it is important to identify (spatially and temporally) the surface drifters pathways with that of the satellite passes and (ii) the satellite altimetry along track data (Jason-2 and Sentinel-3A) and hydrodynamic model sea level data are all reduced to the same reference datum (using a high resolution marine geoid). It is important to note that to ensure reliability of the hydrodynamic model data is compared to that of tide gauge, satellite sea surface temperature and ocean colour data to ensure that the dynamics observed are similar both in space and time.

Finally, by inter-comparison of the differences between data sources their geostrophic and ageostrophic components are computed. The results show that whilst some limitations exist amongst that datasets the geostrophic and ageostrophic components of surface currents may be reasonably estimated. Identification of these components (e.g. influence of waves) are very important in our understanding of surface transport.

Area C - Board 326