Copernicus Evolution and Applications with Sentinel Enhancements

and Land Effluents for Shores and Seas

Date: 1 March 2019 Deliverable Number: D3.2 Due date for deliverable: 31 January 2019 Actual submission date: 6 March 2019 Leader: DHI Document Dissemination Level: PU

PU Public PP Restricted to other programme participants (including the Commission Services) RE Restricted to a group specified by the consortium (including the Commission Services) CO Confidential, only for members of the consortium (including the Commission Services)

Co-ordinator: Universitat Politècnica de Catalunya (UPC) Project Contract No: H2020-EO-2016-730030- CEASELESS Project website: http://www.ceaseless.eu

Ref. Ares(2019)1565868 - 08/03/2019

H2020-EO-2016-730030- CEASELESS

2

DOCUMENT INFORMATION

DOCUMENT HISTORY

ACKNOWLEDGEMENT This project has received funding from the European Union’s H2020 Programme for Research, Technological Development and Demonstration under Grant Agreement No: H2020-EO-2016-730030- CEASELESS. DISCLAIMER This document reflects only the authors’ views and not those of the European Community. This work may rely on data from sources external to the CEASELESS project Consortium. Members of the Consortium do not accept liability for loss or damage suffered by any third party as a result of errors or inaccuracies in such data. The information in this document is provided “as is” and no guarantee or warranty is given that the information is fit for any particular purpose. The user thereof uses the information at its sole risk and neither the European Community nor any member of the CEASELESS Consortium is liable for any use that may be made of the information.

Title Impact from bathymetric and boundary updating in a potential CMEMS extension.

Lead Author Rodolfo Bolanos Contributors Lars Boye Hansen, Manel Grifoll, Davide Bonaldo, Mikkel

Lydholm Rasmussen, Manuel Espino, Pablo Cerralbo, Agustín Sánchez-Arcilla, Sandro Carniel, Jorge Guillén

Distribution Document Reference

Date Revision Prepared by Organisation Aproved by Notes

H2020-EO-2016-730030- CEASELESS

3

1 Introduction

CEASELESS aims at improving the present level of application/knowledge for the high frequency metocean variability that dominates coastal processes but is, nevertheless, not well resolved by observations nor models that show sub-mesoscale features (modulating dynamic balances and affecting socio-economic activities) with large shifts in time/space. Bridging the gap between continuous local observations and satellite measurements or between high resolution numerical discretizations and RS fields is essential to provide a progression from oceanic domains to coastal areas where errors become larger. Smart combination of satellite, in-situ data and numerical simulations can advance existing capabilities and services for coastal operational and analytical oceanography such as CMEMS. Suspended sediment in the water column and subsequent deposition plays a critical role in coastal environment and management. High levels of suspended sediment concentration in the water column has relevant implications in aquatic ecosystem and natural habitat (Ellis et al., 2002) in particular during large exposure periods (Newcombe and Macdonald, 1991). Also, sediment supplied from rivers convey load of organic matter, pollutants and heavy metal that may be deposited in the vicinity sea bottom or transported offshore (Palanques et al., 2017). The analysis and prevention of fine sedimentation within basins and channel access is object of investigation in port engineering context in order to examine the siltation process (e.g. Ghosh et al., 2001; van Maren et al., 2015). Hydrodynamics processes such as wind-waves (Carlin et al., 2016; Grifoll et al., 2013), tides (Fan et al., 2004; Garel et al., 2009), winds (Hofmann et al., 2011; Sherwood et al., 1994), surface seiches (Jordi et al., 2008) or internal-seiches (Shteinman et al., 1997) promote the resuspension, advection and settling of sediment conditioned by the continental sediment sources. Subsequent resuspension effects due to natural causes also contributes at the reworking and final deposition of the sediment load (Grifoll et al., 2014a; Guillén et al., 2006). In this sense, anthropogenic activities such as, fishing trawling, ship propellers and waves generated by vessels may bring additional energy at the water system influencing the resuspension, transport and final sediment deposition in shallow waters (e.g. Garel et al., 2009; Hofmann et al., 2011). Bathymetric features have a strong influence on waves and currents, controlling the propagation and dissipation of flows during normal and extreme conditions and therefore an implication for sediment resuspension and transport. Thus, accurate measurements of bathymetry and its evolution is critical to the coastal marine industry. Offshore wind farms around Denmark are being developed with the objective of increasing the contribution of renewable energies to the energy national consumption. This type of activity is also being developed in other countries such as Germany, Belgium, UK, US and China among others. New challenges will arise especially during storm conditions that may directly affect the estimation of wind turbine design parameters (like extreme metocean conditions), the secure operation of the national and international electrical system (regarding e.g. the turbine cut-off speed), the fatigue and the extreme wave loads. Because of this we expect that the advent of novel Sentinel products applied to coastal zones combined with local measurements and numerical modelling will prompt an unprecedented quantum leap in oceanographic knowledge and a surge of new coastal applications with direct benefit for Copernicus services. The new Sentinel-2 multispectral sensor, containing all relevant spectral bands, is expected to provide an image selection where distortion is minimal and therefore allows regular updating of the domain geometry. The sediment distribution resulting from the bed mapping can be used as a proxy indicator of the integrated hydrodynamic fields since bed geometry responds the overlying water fluxes and turbulence levels. The frequent coverage of Sentinel-2 with respect to previous satellites

H2020-EO-2016-730030- CEASELESS

4

may allow a periodic updating of the bathymetry commensurate with the variability of metocean drivers in coastal areas. Three case studies have been included in the deliverable covering three very different physical conditions where we have combined satellite data with in-situ measurements and numerical modelling:

• North sea. Danish coast (Section 4.1) • Adriatic Sea (Section 4.2) • NW Mediterranean coast. Ebro Delta (Section 5)

2 Satellite derived bathymetry retrieval

The bathymetry focus under CEASELESS is to examine a Sentinel-2 approach for deriving accurate and high resolution satellite-derived bathymetry (SDB). There are multiple approaches to deriving bathymetry from satellite imagery. In this case, a sensor independent (can be applied to any data from any optical sensor) DHI GRAS proprietary physical radiative transfer model (extended version of Guzinski et al. (2016), Klonowski et al, (2007), Lee et al. (1998, 1999, 2001)) was used. Other methods of note, used for deriving SDB, are the empirical models based on the early work of Lyzenga(1978), added upon in Stumpf, 2003, and other radiative transfer based models, such as the SAMBUCA model developed by CSIRO (Brando and Dekker, 2003; Wettle and Brando, 2006). The fundamental difference between the empirical models, and the radiative transfer models, is that the empirical models calculate an index that correlates with depths, but has to be transformed into water depths, while the radiative transfer methods model the water column as a function of, amongst other parameters, the water depth. The physical approach allows bathymetry retrieval without input from ancillary data (known depth values) but performance can be optimized by inclusion of a small amount of calibration data. The radiative transfer model used in CEASELESS minimizes the differences between an observed satellite image and a modelled satellite image, which is created as a function of six parameters: depth, bottom type, backscattering, chlorophyll-A, gelbstoff, and the slope of the backscattering function. The bottom type is further split into fractional cover of two different bottom types, specifically unconsolidated sand and submerged aquatic vegetation. Through minimizing the difference between the observed and the modelled satellite images, accurate water depths can be retrieved in optically shallow waters, meaning light reflecting from the seabed is observed. When utilizing a radiative transfer model, a key component is to make sure that the observed satellite image is correctly pre-processed, through accurate atmospheric correction and deglinting. The atmospheric correction for this study is done using the industry standard 6S package (Vermote et al., 1997) and the deglinting is a variation of the one outlined in Lee et al. (1999). In Figure 1, the interactions that light goes through, both in the atmosphere and in the water column can be seen. Note that the interactions in the water column corresponds to the six parameters in the radiative transfer model.

H2020-EO-2016-730030- CEASELESS

5

Figure 1 Overview of the interactions that light goes through in the atmosphere and the water column, which all have to be accounted for when deriving bathymetry from satellite imagery.

One of the key challenges in aquatic radiative transfer modelling is the strong attenuation of light in the water column, which limits the part of the electromagnetic spectrum that is available for reliable in-water modelling. In Figure 2, the attenuation of light per meter of clear water as a function of electromagnetic wavelength can be seen. The logarithmic y-axis should be noted. As can be seen, the depth penetration of light with a wavelength of higher than 600nm is severely reduced, which in turn limits the parts of the spectrum usable for modelling. As a result of this, when using multispectral satellites , it is common to only have two or three spectral bands that contain information at larger depths (nearing the secchi depth).

Figure 2 Attenuation of light per meter of clear water as a function of electromagnetic wavelength.

H2020-EO-2016-730030- CEASELESS

6

In the light of the above considerations and imagery preprocessing procedures, the radiative transfer modelling of SDB can be reduced to the following series of pseudo equations:

����������� = (�ℎ���� ℎ��, �������, �� , �� , �, �) (1) Where ����������� is the modelled satellite image, bb is the backscattering, bx is the slope of the backscattering function, ρ is the summed-up bottom reflectance, and H is the depth. The next step is to minimize the difference between the modelled and observed satellite images:

�2 = 1� ��(����������� − ����������� )� (2)

With N being the number of spectral bands, where the satellite image provides information, and ����������� is the observed satellite image.

Through the two pseudo equations above, and the equations that define the relationship between the parameters in equation one, the depths can be retrieved in a reliable manner, with the main challenge being tied to the limited availability of information due to the attenuation seen in Figure 2. The primary strengths of SDB are the quick delivery time, frequent repeatability possibilities, low cost, and easy accessibility to remote or dangerous shallow waters. The spatial resolution of the Sentinel-2 imagery also allows for very detailed bathymetry mapping, typically order of magnitude better than otherwise available bathymetric datasets which could be easily integrated into CMEMS. Since the approach is based on optical data there is a limit to the depth penetration of around one secchi depth, so clear waters are a requirement for optimal SDB retrieval. The Sentinel-2 data allows for much more detailed SDB retrieval than what is typically available. For modelling purposes, a coarse resolution bathymetry is fine for deeper waters but for shallow and dynamic coastal areas up-to-date and high-resolution bathymetry is required for reliable high-resolution modelling activities. Sentinel-2 data allow for both a high-resolution bathymetry retrieval and with the free and open Copernicus data policy the cost for bathymetry creation can be kept to a minimum, making it a very cost-efficient solution accessible for a much wider range of activities. Combining the cost-effectiveness with the frequent overpass of Sentinel-2 it also becomes possible to start looking at the seabed dynamics (changing bathymetry over time) and quantifying volume changes etc. During CEASELESS several combined bathymetry data models have been created where the high resolution SDB layer was merged into the existing coarser resolution bathymetry data covering deeper waters, and the new updated model used in the evaluation of the impact of a changing bathymetry on the numerical hydrodynamical modelling. Bathymetry retrieval has been ongoing for three sites in the project:

• Ebro Delta, Spain: The Sentinel-2 archives have been thoroughly searched and screened for suitable input data for SDB retrieval. The area is dominated by either turbid conditions (large outflow of turbidity from the delta + significant amount of resuspension), wave contamination and/or sun glint in the data. As a result, only parts of the area of interest has been retrievable.

• Adriatic Sea, Italy: The northern part of the Adriatic Sea has been analysed and SDB retrieved where possible. The nature of the coast (turbidity, quick deepening of the waters) means that only a narrow strip SDB can be retrieved.

H2020-EO-2016-730030- CEASELESS

7

• North Sea/Horns Rev, Denmark: Horns Rev, the shallow area off the western coast of Denmark with significant wind farm activity, has been produced and made available to CEASELESS partners. Two coverages has been produced to analyse sea bed dynamics and quantify movement rates of large-scale morphology.

3 Water quality from Sentinel 2 (S2)

During the first year of CEASELESS, an automated download and pre-processing chain for Sentinel-2 data was set up and subsequently used for routinely deriving water quality information from all available S2 A and B data (L1). The data cover from February 2017 for Fangar Bay (Spain) and Wadden Sea (Denmark). For the further processing of the L1 data into water quality products, two different retrieval methods are in place. These were used to retrieve Suspended Particulate Matter (SPM) and Chlorophyll-a (Chl-a), which are two important water quality parameters. Finally, an improved customized cloud masking approach has been applied because the quality flags provided with the satellite images are often not sufficient to remove all pixels with bad quality. The evaluation of the results has been focused around the Ebro Delta site in Spain where cross-comparisons have been conducted for selected points along transects. The primary objective of the S2 missions is not to observe ocean colour, however, due to the 665 nm (red) and 705 nm (red-edge) bands, the Chl-a absorption and hence concentration can still be derived, yet with known limitations. On the other hand, retrieval of sediments usually works well. The evaluation of the results from the two processors and the two water quality parameters will provide more insights into which method to choose. Following the download of the relevant Sentinel-2 imagery, the Case-2 Regional CoastColour (C2RCC) processor was used for the retrieval of chlorophyll and suspended matter, which is achieved by inversion of the water leaving reflectance spectrum (Brockmann et al., 2016). As a first step, the processor applies a correction for atmospheric effects. The retrieval procedure itself is based on neural networks technology. C2RCC is widely used for instance for standard ESA Case–2 water product processing. A second approach has also been implemented in the cloud processing environment based on the ACOLITE software package developed by Royal Belgian Institute of Natural Sciences. Both chlorophyll and suspended matter can be derived using a suite of different algorithms described in literature. Cloud masking is an integrated step of the processing chain. However, despite the apparent simplicity of the problem, clouds come in a wide variety of sizes, shapes, altitudes, and optical properties. For example, haze or cloud shadows over water are often not captured accurately by the algorithm. We added therefore an extra step and screened the imagery also manually for atmospheric disturbances. Another problem occurs on days with strong winds which can impact the water surface roughness and lead to wrong values in the retrieval process, especially for S2. Hence, the resulting data may still contain some flawed values. Both retrieval schemes can with little effort be adjusted and tuned for local optimized constituent retrieval based on established local in-situ tuning. In-situ measurements from 2017 on sediment concentrations from Fangar Bay has been made available from UPC to DHI GRAS. Unfortunately, there were no precise match-ups (satellite overpass/measurement coincidence in time within +/- 2 hours) so a more pragmatic approach has been followed to evaluate the performance and agreement. The results show generally good agreement with the anticipated comparable seasonality patterns found in the time-series plots for all stations.

H2020-EO-2016-730030- CEASELESS

8

Figure 3 Time series of in-situ (red) and two different retrieval methods of Sentine-2 based sediment concentrations (light + dark grey) for six different stations in Fangar bay. Location of the stations are shown on the top panel.

The water quality product from CEASELESS has also been included in a research activity led by UPC (see Section 5), where it was examined which effect different wind conditions in the Ebro Delta area has on the resuspension rates inside the enclosed areas of the delta. The sentinel-2 products confirmed results from the coupled wave-current numerical model, showing a strong spatial variability of the potential resuspension locations within the area and that strong wind events are responsible mechanisms for the resuspension of fine sediment within the bay.

4 Satellite derived bathymetry case studies. Perspectives and limitations

The goal of this section is twofold. First, the use of a state-of-the-art cost-effective method for bathymetric mapping based on high resolution satellite images to characterize a coastal areas (North Sea and Adriatic) and second, to assess the added value of such high resolution data when

H2020-EO-2016-730030- CEASELESS

9

performing wave modelling (North Sea). Additionally an assessment of bottom friction maps derived from the EU project EMODnet (http://www.emodnet.eu/seabed-habitats) habitat map is performed.

4.1 North Sea case study

For the North Sea case, the area called Horns Rev (HR) is used to obtain bathymetry from the satellite and assess its use in high resolution wave modelling. HR is located in the North Sea, west of Denmark, the area is along the Danish North Sea coast, characterized by the presence of offshore wind farms (Figure 4) as well as a complex bathymetry that feature sand bars and a channel. Additionally, the area is exposed to storm events from the North Sea producing large waves coming from the northwest that dissipate on the sand bars. A particularly intense event occurred on the 6th November 1985 when waves approaching from northwest reached a significant wave height offshore from HR of approximately 7m. HR is a wave dissipation area where waves are reduced significantly due to their interactions with the seabed via bottom friction and depth induced breaking.

Figure 4 Location of the Horns Rev area in the North sea. Detailed figure on the right indicates positions of the three wind farms HR1, HR2 and HR3. (image from Vattenfall https://corporate.vattenfall.com/globalassets/corporate/about_energy/3c_map_horns_rev_123.png).

4.1.1 Bathymetry data

For this study, satellite derived bathymetry (SDB) was produced using imagery from the Sentinel-2A satellite. A 10m resolution SDB was derived twice, once in 2016 and once in 2018, to quantify the movement rates of largescale sand bank in the Horns Rev area as illustrated in Figure 5. This example clearly highlights the need for high resolution and recent bathymetry to accurately resolve hydrodynamic processes in the highly dynamic coastal zone in high spatial resolution. As can be seen on the two illustrations, the sand bank moved between 100m and 200m during a period of 23 months. This sediment transport might be due to a combination of wave impact and tidal current transport. Figure 6 shows the bathymetry obtained from the satellite for the Horns Rev area, showing the shallower sand banks and water depth lower than 7m.

H2020-EO-2016-730030- CEASELESS

10

Figure 5 Top panel shows a crest of a sand wave as observed on 8th of May 2016, indicated with yellow line. Bottom panel shows bathymetry obtained on 2nd of June 2018, illustrated by a red line, along with a yellow line showing its position in 2016.

H2020-EO-2016-730030- CEASELESS

11

Figure 6 Bathymetry data obtained from satellite at the Horns Rev area.

4.1.2 Modelling. Phase resolving wave model.

In order to assess the impact of sand waves on individual waves, a non-hydrostatic wave model was implemented. The model (MIKE3 Wave FM (DHI, 2019)) is based on the incompressible Navier-Stokes equations subject to the assumptions of Boussinesq and with free surface described by a height function. The turbulence is modelled using an eddy viscosity concept (κ-ε) and the vertical discretization uses sigma or combined sigma/z-levels. As with other MIKE models, MIKE3 Wave FM has the capability of using horizontal unstructured meshes. Additionally, to the expected better model for individual waves over bed features or structures compared to a spectral model, another advantage of the phase resolving model is that wave kinematics can be obtained which is fundamental information to study wave-structure interactions and sediment transport. Further details on the model can be found in DHI (2019). MIKE 3 Wave FM was implemented for a domain of approximately 1 km2 covering one of the sand bars with satellite bathymetry data and containing clear sand waves on top of it. Several model runs were performed considering regular and irregular waves (taking directional spectra from spectral model) and a flat and real bathymetry in order to outline the impact of the sand waves. A small area was selected in order to make the influence of wind input and whitecapping negligible. The area was selected to include the presence of the bedforms (see Figure 7). The horizontal spatial resolution was 2m with a quadrangular mesh using 5 sigma vertical levels. The forcing at the boundaries were selected from the output of a spectral wave model (see further details at Bolaños et al., 2018) at the site during the peak of the storm which corresponded to a significant wave height (Hm0) of 2.3m peak period (Tp) of 12s and peak wave direction of 315o these values were used for simulation of regular wave using stream function theory, but the full directional spectra obtained from a local spectral wave model was also used for the cases of irregular waves. Two bathymetries were used, a flat bed with a depth of 7m and the bathymetry from the satellite for the real case. Since the main direction of waves was around 315o for the case study a rectangular domain aligned with this direction was chosen in order to reduce the impact of lateral boundaries and the sponge layer. For all the tests the model was run to simulate a 20 minutes period. The wave generation boundary was defined to be the northwest boundary and the sponge boundary was defined as the southeast. The domain spatial units were in meters using a UTM32 projection. Table 1 summarizes the model settings for the runs performed.

H2020-EO-2016-730030- CEASELESS

12

Figure 7 MIKE 3 Wave FM model location (top panel), domain and bathymetry (bottom panel, UTM32 coordinate system). Colorbar represent water depth in meters

Table 1 Summary of MIKE 3 Wave FM model runs setup

Run BC Bathymetry R1 Regular wave (H=2.3. T=12s)

Flat (7m)

R2 Regular wave (H=2.3m T=12s)

Real

R3 Directional spectra

Flat (7m)

R4 Directional spectra Real

H2020-EO-2016-730030- CEASELESS

13

4.1.3 Results and Discussion

Figure 8 shows the result of the MIKE3 Wave FM for the case of regular waves propagating over a flat bed and the real bathymetry containing the bed forms. For the case of the flat bed (Figure 8 left panel) it can be seen that waves propagation is steady and no impact of boundaries (e.g. through reflection) is noticeable. For the case of the real bathymetry (Figure 8 left panel), wave breaking occurs when waves reach the sand wave crests which significantly reduce wave energy and produce some refraction of waves. This wave breaking process is expected to be of importance for sediment transport and for the study of wave-structure interactions. Figure 9 presents the results of the irregular case tests where the first column is the results from the flat bed after 20-minute simulation, central column is for the real bathymetry and right column is the difference between the real bathymetry result minus the flat bed case. Each row shows different variables, from top to bottom are Hm0, maximum wave height (Hmax), mean surface elevation, mean current u component of velocity and mean current v component of velocity. Similarly to the regular case, bed features reduce waves when they first reach the sand bars, this is noticeable in the reduction of Hm0 and maximum surface elevation. Hm0 is reduced approximately 60% as waves propagate through the domain (<1km). A clear impact of the sand bars is the modification of the velocities which are significantly increased due to the presence of sand bars. The mean of the u and v components during the simulation period shows an increase in the first section of the domain in agreement with sand wave orientation and the locations where waves lose energy.

Figure 8 Surface elevation result for the case of regular waves for a flat (left panel) and real (right panel) bathymetry

H2020-EO-2016-730030- CEASELESS

14

Figure 9 Surface elevation result for the case of regular waves for a flat (left column) and real (central column) bathymetry. Right column shows differences. Each row corresponds to different variable, from top to bottom: Hm0, Hmax, Mean elevation, mean current speed u-component and mean current speed v-component.

H2020-EO-2016-730030- CEASELESS

15

A satellite-based bathymetry was successfully obtained at the Horns Rev area for water depth shallower than 7m. This outlined a limitation of this approach for deeper areas as it requires very clear water conditions for the light to penetrate. Considering that offshore wind farms are typically deployed in deeper water, the satellite derived bathymetry might not be relevant for these applications. However, a more coastal use of such data which, as demonstrated here, was capable of register large bedform movement would be relevant where sediment transport studies, and coastal waters problems would benefit of access of such data. This type of data could also support seabed mobility studies which are required for cabling and pipelines projects with landfalls and where wave and currents produce significant sediment transport near the coast. The identification of large sand bedforms was possible and could help in setting the optimal model resolution in order to properly model wave dissipation and transformation. The combination of a spectral wave model with a phase resolving model is a natural approach for such type of environment where resolution lower than hundreds of meters play an important role and become somewhat restrictive for a spectral model.

4.1.4 Bottom friction.

The bottom surface characteristics are very important to describe the boundary layer of currents and waves, and control to a large degree the dissipation of flows. However, the bottom friction is typically a calibration parameter in numerical modelling of coastal processes. In this section a map of varying bottom friction is generated based on bottom properties to explore the impact of the spatial variability in spectral wave modelling. Results indicate a potential improvement in terms of model skills. The bottom friction dissipation in the spectral wave model (MIKE21SW), based on linear theory, can be generalized as:

Where Cf is a dissipation coefficient equal to a wave friction factor (fw) multiplied by the near-bed particle velocity. fw is obtained from Jonsson (1966) as described by Johnson and Kofoed-Hansen (2000):

The information obtained from the seabed habitat map from EMODnet (Figure 10) has been converted into a preliminary map (Figure 11) of roughness length (kN) with three different values/classes (Table 2). The spatial resolution of the map is of 0.1 x 0.1 degrees.

H2020-EO-2016-730030- CEASELESS

16

Figure 10 Habitat map from EMODnet (http://www.emodnet.eu/seabed-habitats).

Figure 11 Roughness length (kN) map based on sediment properties from Figure 10. See Table 2 for values.

H2020-EO-2016-730030- CEASELESS

17

Table 2 Sediment class and associated roughness used to generate the map in Figure 11

Sediment class kN (m) Mud/clay 0.01 (red areas in

Figure 11) Sand 0.02 (green areas in

Figure 11) Coarse sediment 0.03 (blue areas in

Figure 11) MIKE 21 SW model is used to simulate wind generated waves in the North Sea. The model has been developed for flexible meshes, ideal for coastal oceanography. A case study in the North Sea where a directional wave buoy equipped with an ADCP was deployed (RUNE project, see Bolaños and Rørbæk, 2016) off the west coast of Denmark was used for sensitivity tests. The area is characterized by water depth of less than 30 m and by the development of several offshore wind farms (e.g. Figure 4). MIKE 21 SW was run for three different cases using roughness length and parameters as described in Table 3. Hm0 of roughly 8 m were modelled in the central North Sea during the simulated period (Nov-Dec 2015). The eastern part of the North Sea presented larger waves due to winds blowing from the north west. Differences of up to 0.5 m between runs are seen in the North Sea, specifically in the southeastern area.

Figure 12 Wave model domain and measurement stations for bottom friction sensitivity runs. Timeseries plot show

significant wave height (Hm0, black line) and maximum individual wave height (Hmax, blue line)

H2020-EO-2016-730030- CEASELESS

18

Table 3 Spectral wave model setup for sensitivity tests of bottom friction

Parameter/input Description Simulation period 04/11/2015 – 31/12/2015 # Frequencies 25 (0.033Hz - 0.94Hz) # Directions 24 Wind fields CFSR Model domain See Fig. 2 Surface elevation From DHI regional hydrodynamical model Roughness length (kN) Run 1: 0.02m

Run 2: 0.03m Run 3: Map (Figure 4)

Source terms Jansen (1991), Bidlot et al (2007) Cs =2.3, δ =0.7

Boundary conditions 2D spectra from DHI global model Spatial resolution Irregular mesh from ~16.5 km to ~2.5 km

Table 4 summarizes the model statistics (Bias, Root Mean Square Error and Scatter Index) at the measurement locations (see Figure 12). The use of a constant roughness length of 0.02 m produced the worst results at all stations. The constant value of 0.03 m and the map of seabed roughness seems to produce overall similar results although the model skills vary for different stations. Table 4 Summary of Hm0 wave model validation statistics for the three runs with different roughness length.

Location Hm0 Statistic (m)

Run 1 (kN=0.02)

Run 2 (kN=0.03)

Run 3 (kN map)

Ekofisk

Bias 0.25 0.2 0.22

RMSE 0.53 0.5 0.51

SI 0.15 0.15 0.14

Fjaltring

Bias 0.06 -0.07 0.01

RMSE 0.34 0.33 0.32

SI 0.14 0.13 0.13

RUNE

Bias 0.12 0 0.07

RMSE 0.31 0.27 0.29

SI 0.12 0.11 0.12

K13

Bias 0.1 0.03 0.07

RMSE 0.33 0.32 0.32

SI 0.14 0.14 0.13

Nymindegab

Bias 0.1 -0.04 0.13

RMSE 0.32 0.31 0.33

SI 0.14 0.14 0.14

H2020-EO-2016-730030- CEASELESS

19

It is interesting to note the slight reduction in Scatter Index by using a map (Run 3) at Ekofisk and K13 and the general improvement at Fjaltring. While, at the same time, a better performance of Run 2 at RUNE and Nymindegab is obtained. Although those three stations are relatively close from each other, the area is characterized by a patches of different sediment sizes and thus abetter characterization of bed features has a potential for improving wave modelling. Additionally, model results might be dependent on bathymetry accuracy as the stations are located in water depth shallower than 16 m and close to the coast where some bottom features might not be properly resolved.

4.2 Satellite derived bathymetry for the Adriatic case

Producing satellite derived bathymetry in the Adriatic Sea provides several distinctive challenges, with the primary being the low visibility in the water column due to large sediment rich river discharges in the region and the resuspension of sediments. This results in only few satellite images being suitable for seabed observation. Based on a Sentinel-2 image acquired on the 20th of June 2017, DHI GRAS has produced SDB for the area between the Venetian Laguna and the city of Cortellazzo. Due to the before mentioned challenges with low visibility and additional noise from sunglint, the maximum retrieved depth was around 3m, which corresponds to a depth slightly higher than that observed behind a longshore bar. A subset of the satellite derived bathymetry can be seen on the illustration below.

Figure 13 Satellite derived bathymetry from the between the Venetian Leguna (to the left) and Cortellazo (to the right).

As can be seen on Figure 13, the derived bathymetry covers the shoreline a couple of hundred meters out into the Adriatic Sea and contains two alongshore bars. In July 2013 and September 2014 two morphobathymetric surveys have been carried out on a 1-km wide tract of the eastern reach of Jesolo beach, in the framework of the Italian National Flagship Project RITMARE (La Ricerca Italiana per il Mare – the Italian Research for the Sea). These surveys covered the beach from the delimiting wall (in the absence of a littoral dune) up to beyond

H2020-EO-2016-730030- CEASELESS

20

the offshore limit of the active beach, at a depth of approximately 10 m. The subaerial and intertidal beach has been described by 27 transects, measured by means of a GPS RTK Trimble 5700 positioning system. The survey of the submerged beach has been performed by means of interferometric (July 2013) and multi-beam acoustic (September 2014) instruments. The gaps between the GPS RTK transects and the interferometric and multibeam surveys were subsequently filled by means of a 200 kHz single-beam Odom Hydrotrac echo sounder, operated by a vessel compatible with very shallow water navigation with no risk for the equipment. The surveys showed a mild-sloping beach morphology, characterised by a shoreline-parallel bar and a set of relatively stable transverse bars with 250 m wavelength, protruding off the coast up to approximately the -5 m isobath. Below is a map with a transect plot of the SDB and the surveys dataset from the area.

Figure 14 Map showing the coastline just south of Jesolo illustrating the SDB data and the single beam data. Additionally, a transect plot of the satellite derived bathymetry (in green) and a single beam transect (in blue) shows the migration of a sandbar over time. The location of the transect is marked in red

The key feature in Figure 14 is the sandbar that has migrated around 50m towards land between the two datasets. It should be noted that the single beam dataset was acquired in 2013 and as such, many cycles of sandbar migration may have occurred between those measurements and the acquisition of the satellite image used for production of the SDB dataset.

H2020-EO-2016-730030- CEASELESS

21

5 Satellite derived suspended sediment case study. Alfacs Bay, NW

Mediterranean

Alfacs Bay is a micro-tidal estuary in the NW Mediterranean Sea; Southern part of the Ebro Delta. It has been investigated extensively in the past in terms of hydrodynamics response (Cerralbo et al., 2015a, 2016; Llebot et al., 2014; Solé et al., 2009), tidal wave propagation (Cerralbo et al., 2014), biochemical processes (Llebot et al., 2010, 2011) an optical water properties (Ramírez-Pérez et al., 2017). The estuary receives freshwater discharge from the rice fields of the Ebro river. Thus, Alfacs Bay is an intensively exploited area with tourism, fishing and aquaculture activities being an ecosystem of relevant economic importance in the region. Several episodes of algal blooms (linked with the increase of nutrients and perhaps triggered by resuspension mechanisms) and presence of harmful bacterium in bivalve with negative effects on aquaculture have been reported (Loureiro et al., 2009; Roque et al., 2009). The average depth is 4 m and the maximum depth is about 6.5 in the middle of the Bay (Figure 15). The connection with the open sea is 2.5 km, with a central channel of 6.5m and shallow edges of around 1-2 m on both sides. The bay is surrounded by rice fields to the north, which spill around 10 m3·s-1 of freshwater loaded with nutrients during 9-10 months per year (April-December) distributed in several channels, and a sand beach closing it on the east side. The seabed in the central part of the bay is composed by very fine sediment (typically 65-65% silt, 30-35% clay and around 5% sand) increasing the sandy content towards the edges of the bay (Guillen and Palanques, 1997; Satta et al., 2013). The bottom sediment of Alfacs is composed by mud and sandy mud, with significant content of clay (Palacín et al., 1991). They found that the muddy sediment extended by the central part of the bays and the content of sand increased near to both spits that separate bays from open sea and also in the southern shallow edge. With the purpose to improve the knowledge in fine sediment dynamics in coastal bays, the goal of this investigation is to provide a physical interpretation of the sediment resuspension events observed within a micro-tidal bay (Alfacs Bay; NW Mediterranean Sea). Using sea-level, water currents and wind measurements supported with satellite derived information we investigate the driven mechanisms that resuspend fine bottom sediment within the bay. Then, the spatial and temporal interpretation of the resuspension mechanisms linked with the hydrodynamics is analyzed through the implementation of a wave-current coupled numerical model. The contribution aims to provide explanation of resuspension mechanisms; the knowledge of these mechanism may have an evident benefit for human activities management, for further details reader is referred to Grifoll et al. (2018). The water circulation in Alfacs Bay has been widely analyzed in previous contributions using observational data set and numerical results (Camp and Delgado, 1987; Cerralbo et al., 2014, 2015a; Llebot et al., 2014). However, fine sediment dynamics and its resuspension mechanisms has not been examined yet. Synchronous optical measurements, jointly with velocity and sea-level measurements and satellite derived total suspended matter (TSM), has entailed a good chance to advance in the interpretation of resuspension mechanisms in Alfacs Bay. This area is an example of micro-tidal estuary, thus being the wind or wind-waves posible mechanisms of fine sediment dispersal.

H2020-EO-2016-730030- CEASELESS

22

Figure 15 a: Regional location of Ebro River Delta b: Alfacs Bay in Ebro River Delta. c: map of Alfacs Bay. Triangle shows the meteorological station (M-Sc). White cross for Sant Carles de la Ràpita harbour tide gauge. Gray circles shows the ADCP and OBS mooring locations (A1 and A2). Gray arrows on the northern coast shows the freshwater drainage points considered in the simulation. Double line square indicate the domain for the hydrodynamic numerical model, which is shown in detail in image d (colorbar indicates depth in meters).

5.1 Measurement campaigns

The bulk of the observational data correspond at two months field campaign from July to mid-September 2013 that correspond to summer conditions. The data set consisted of water currents from two 2MHz Acoustic Doppler Current meter Profiler (ADCPs) moored in the mouth (A1) and inner bay (A2) (see Figure 15 for locations) configured to record 10 min averaged data from 10 registers per minute and with 25 cm vertical cells. Both devices were equipped with Optical Backscatter Sensor (Campbell Scientific OBS-3), bottom pressure meter and a temperature sensor, and they were mounted on the sea bottom at 6.5 m depth. The study area used to present a linear relation between optical signal and suspended sediment concentration (Guillén et al., 2000). The distance of the ADCPs and OBS sensor were 0.25 m above the sea bed. The ADCP has a 20 cm of blanking zone. Additional sea level data were obtained through a sea level gauge mounted in Sant Carles de la Ràpita harbor (Figure 15) and bottom pressure systems from the ADCPs. Atmospheric data (wind, atmospheric pressure, solar radiation and humidity) were obtained from a fixed land station located in Sant Carles de la Ràpita (M-Sc in Figure 15).

H2020-EO-2016-730030- CEASELESS

23

Without access to local measurement data for Total Suspended Matter for the Alfacs Bay, the Sentinel-2 approach for TSM retrieval has been applied in two different scenarios for the Alfacs Bay: NW wind and Calm conditions. It was confirmed that NW wind conditions substantially increase the TSM concentrations in the southeastern shallow edges (Figure 16). This would be a source of a subsequent advection of fine sediment towards the central Basin. In calm conditions, on the other hand, the values of TSM decrease significantly. Also, the proximity of the Ebro river mouth may increase the suspended sediment within the bay under particular circumstance. River discharge is the main driver of the Ebro River plume, followed by wind and regional oceanic circulation that tends to be southward (Fernández-Nóvoa et al., 2015; Mestres et al., 2003). Analysis of the turbid plume by remote sensing products indicate that more than 70% of the plume extension was located south of the river mouth influenced by the regional oceanic circulation (Fernández-Nóvoa et al., 2015). Others external sediment sources may be associated freshwater discharge from channels, overwash in the bar, flash flood from small rivers or aeolian transport. The study of the suspended sediment dynamics will provide objective information to address the problem of degrading water quality within the bay and how to make use of natural mechanisms to limit undesired concentrations of nutrients or pollutants. This applies in particular to harmful algae blooms prone to occur in the area under present and future conditions.

Figure 16 Total Suspended Matter (TSM in mgr•l-1) obtained from Sentinel-2 for the Alfacs Bay in two different scenarios: NW winds (left; 27th of December 2017) and calm conditions (right; 15th of February of 2018).

5.2 Modelling in Alfacs Bay

We use the coupled version of SWAN-ROMS models included in the COAWST system in order to simulate the hydrodynamics within the bay. The COAWST system (Warner et al., 2010) consists of several state-of-the-art numerical models that include ROMS (Regional Ocean Modeling System) for ocean and coastal circulation and SWAN (Simulating Waves Nearshore) for surface wind-wave simulation. SWAN is a third-generation numerical wave model that computes random, short-crested waves in coastal regions with shallow water and ambient currents (Booij et al., 1999). It is based on the wave action balance with sources and sinks and incorporates the state-of-the-art formulations of the processes of wave generation, dissipation and wave-wave interactions. ROMS is a three-

H2020-EO-2016-730030- CEASELESS

24

dimensional circulation model which solves the primitive variables on a sigma-level in the vertical and horizontal curvilinear grid. Numerical aspects of ROMS are described in detail in (Shchepetkin and McWilliams, 2005). In COAWST system, the wave model provides hydrodynamic parameters (i.e., significant wave height, average wave periods, wave propagation direction, near-bottom orbital velocity and wave energy dissipation rate) to the water circulation model. The ocean model provides water depth, sea surface elevation, and current velocity to the wave model. The variables exchange is made “on-line” during the simulation processes, via Model Coupling Toolkit (Jacob et al., 2005), where a multi-processes MPI protocol is used to distribute the computations among several nodes. The COAWST also include different formulations to parametrize the wave-current bottom boundary layer and the wave effect on currents (Kumar et al., 2012; Warner et al., 2008). The implementation of the COAWST system in Alfacs Bay consisted of a regular grid of 186 x 101 points with a spatial resolution of 100 m (in both x and y) and 12 sigma levels in the vertical. Details of the implementation and the skill assessment of the ROMS model in Alfacs bay is provided in (Cerralbo et al., 2015a). The same regular grid is used by the SWAN model. A two-year water circulation simulation (2012-2013) was performed in order to obtain realistic three-dimensional temperature and salinity fields. The interval time between change of variables of ROMS and SWAN was established in 3600 s. For both simulations, open boundary was forced with depth-averaged velocities and sea level measured at A1 (10’ data). The freshwater inputs are distributed on 8 points simulating the main rice channels with a total flow of 10m3s-1 (see Cerralbo et al., 2015a). The bottom boundary layer was parameterized using the combined wave-current (Styles and Glenn, 2000) adopted in ROMS and SWAN coupling in (Warner et al., 2008). The input parameters for the model are the velocitiy components near the bottom and wave characteristics near the bottom (wave period, wave direction and the wave orbital direction). For each computational step, an initial assessment of bed roughness length is estimated in function of the grain size, ripples and sediment transport.

5.3 Model results and Discussions

The bottom stress is obtained from the coupled numerical model implemented in Alfacs Bay. The Figure 17 and Figure 18 show different snapshots in order to examine the bottom stress pattern for both components (i.e. wave and current-induced bottom stresses). These snapshots correspond to different episodes identified from the previous observational analysis. The plot scale of the bottom stress is transformed in log10 for clarity. During the case E1 (3hd of August 2013; 10:00 hr) the combined bottom stresses are mainly due to the current bottom stress (Figure 17). Maximum values of 0.15 Pa for the combined bottom stress are obtained in the center of the bay and the mouth. This episode corresponds to a seiche event and the spatial variability of the bottom stress is consistent with the spatial pattern of the node/antinode position. It means that the maximum combined bottom stress (associated at maximum water currents) corresponds to the node position (minimum sea-level amplitude). In opposite, the minimum bottom stress to the antinode position (maximum sea-level amplitude). The position A2 is located near to the node, where the water currents are maximum during the seiche event (0.08 Pa for combined bottom stress). It is worth to mention the node/antinode pattern of the current-induced bottom stress, which presumably would indicate a large spatial variability on the resuspension process within the Bay. After the seiche activity (second stage of E1 where the wind intensity increases due to the sea-breeze), the current-induced bottom stress decreases significantly in particular in the center of the Bay (Figure 17). The bottom stress distribution shows how the maximum values are obtained near the shoreline (2.2 Pa) due to the contribution of the wave-induced bottom stress. In A2, the combined bottom stress is equal to 0.03 Pa (value presumably far to induce resuspension). For this

H2020-EO-2016-730030- CEASELESS

25

event, the wave field during the sea-breeze is shown in Figure 19. This figure shows how the maximum significant wave height (equal to 0.3 m) occurs near the northern and southern shallow edge consistent with the maximum wave-induced bottom stress. The bottom-stress pattern during the episode E2 (Figure 18) is similar to the second stage of the episode E1. Both wave and current bottom stress tends to be small in A2 in comparison to the seiche event. Only substantial bottom stresses are observed in the shallow edges of the Bay due to the wave action originated by the sea-breeze. During the episode E3 (NW wind, Figure 18), the combined bottom stress is dominated by both wave and current action. The southern part of the bay shows the maximum wave induced bottom stress consistent with the wave climate (Figure 19). Also, the current induced bottom stress presents non-negligible values within the bay. Focusing in A2, both mechanisms contribute in similar manner (wave and current bottom stress is 0.09 and 0.06 Pa respectively) in the combined bottom stress.

Figure 17 Distribution of the current, wave and combined wave-current bottom stresses log10(Pa) in the Alfacs Bay during the first stage of the episode E1 (i.e. seiche) and the second stage of the episode E1 (i.e. sea breeze). Magenta dot show the A2 station. Isobaths (in grey) are plotted each 3 m. Note that for clarity, the plot scale is transformed in log10 and the vertical range differs between both bottom stress distributions.

H2020-EO-2016-730030- CEASELESS

26

Figure 18 Distribution of the current, wave and combined wave-current bottom stresses log10(Pa) in the Alfacs Bay during the first stage of the episode E2 and E3. Magenta dot shows the A2 station. Isobaths (in grey) are plotted each 3 m. Note that for clarity, the plot scale is transformed in log10 and the vertical range differs between both bottom stress distributions.

Figure 19 Snapshot of the wave field for the episode E2 (sea-breeze) and E3 (NW wind). Color map represents the significant wave height and black arrows the direction of propagation. Note that the ranges of the significant wave height are different.

H2020-EO-2016-730030- CEASELESS

27

The exchange of fine sediment within the Bay and the open sea seems also evident according to remote sensing images. As a region of high-anthropogenic pressure, this contribution may contribute to develop better integrated plans in the context of sustainable aquaculture activities and the mitigation of the effects of climate change in the Ebro Delta. The observational set and the wave-current numerical results obtained for Alfacs Bay have permitted to investigate the resuspension mechanisms of fine sediment. The results evidence a clear mechanism of resuspension induced by eventual seiche events, which according to the bottom stress patterns may have a relevant spatial variability within the Bay. The wind and wind-wave mechanisms also are responsible of fine sediment resuspension during energetic wind events.

6 Impact from bathymetric and boundary updating in a potential CMEMS

extension.

Sentinel-2 data has been shown to be a valuable source of information for coastal zone processes. With Sentinel-2 the coastal zone dynamics can be captured in high spatial and temporal detail, and the three test cases presented here show relevant examples of how a potential CMEMS extension can benefit from inclusion of Sentinel-2 based products:

- Satellite derived bathymetry can provide up-to-date bathymetry in high detail for the coastal zone - highly relevant for high resolution modelling activities and coastal processes including sediment transport and re-suspension studies

- Sediment distribution can be mapped in high spatial and temporal detail for model calibration/evaluation and forcing

- Coastline changes and dynamics is relevant for detailed land/water interaction activities (flood modelling, sediment modelling etc.)

With the increased focus on high spatial resolution for CMEMS coastal services it seems obvious that Sentinel-2 based information should be an integral part of CMEMS activities going forward.

7 CEASELESS publications related to this deliverable

The work carried out with in the Task 3.2 has been subject of publication and presented at different conferences as indicated below. The North Sea case study has led to Bolaños, R., Boye-Hansen, L., Lydholm-Rasmussen, M. and Golestani, M. 2018. Coastal

bathymetry from satellite and its use on coastal modelling. To be presented at the International Conference on Coastal Engineering, July 30 – August 3, 2018, Baltimore, Maryland, US. https://journals.tdl.org/icce/index.php/icce/article/view/8789

Bolaños, R., Jensen, P., Kofoed-Hansen, H., Tornfeldt Sørensen, J. Bottom friction. A practical approach to modelling coastal oceanography, Poster presentation at European Geo. Union, Vienna, Austria, 2017.

H2020-EO-2016-730030- CEASELESS

28

Rasmus Eskerod Borgstrøm: Water depths and coastal dynamics from space - examples of satellite derived information. Danish bi-annual Marine Research meeting, 23rd January, 2019.

Rasmus Eskerod Borgstrøm: New Bathymetry data from satellites – an example on how satellite-based information can enrich MIKE models. Annual Danish MIKE user Meeting, 30th October, 2018.

The Mediterranean case study has led to the following journal paper: Grifoll, M; Cerralbo, P; Guillén, J; Espino, M; Boye Hansen, L; Sánchez-Arcilla, A. 2018.

Characterization of bottom sediment resuspension events observed in a micro-tidal bay. Ocean Sci. Discuss., https://doi.org/10.5194/os-2018-107, in review, 2018.

8 References

Bidlot, J., P. A. E. M. Janssen and S. Abdalla (2007). A revised formulation of ocean wave dissipation and its model impact. Technical Memorandum, European Centre for Medium-Range Weather Forecast: 27.

Bolaños, R. and K. Rørbæk (2016). RUNE. Metocean buoy deployment. RUNE report. Deliverable 3.1 DHI: 28.

Bolaños, R., Boye-Hansen, L., Lydholm-Rasmussen, M. and Golestani, M. 2018. Coastal bathymetry from satellite and its use on coastal modelling. To be presented at the International Conference on Coastal Engineering, July 30 – August 3, 2018, Baltimore, Maryland, US.

Booij, N., Ris, R. C. and Holthuijsen, L. H.: A third-generation wave model for coastal regions: 1. Model description and validation, J. Geophys. Res., 104(C4), 7649, doi:10.1029/98JC02622, 1999.

Brando, V. E., & Dekker, A. G. (2003). Satellite hyperspectral remote sensing for estimating estuarine and coastal water quality. IEEE Transactions on Geoscience and Remote Sensing, 41(6), 1378–1387. http://doi.org/10.1109/TGRS.2003.812907

Brockmann, C., Doerffer, R., Peters, M., Stelzer, K., Embacher, S., and Ruescas, A. (2016). Evolution of the C2RCC Neural Network for Sentinel 2 and 3 for the Retrieval of Ocean Colour Products in Normal and Extreme Optically Complex Waters. In Proceedings of the ESA Living Planet. Prague, Czech Republic: ESA.

Camp, J. and Delgado, M.: Hidrografía de las bahías del delta del Ebro, Investig. Pesq., 51(3), 351–369, 1987.

Cerralbo, P., Grifoll, M. and Espino, M.: Hydrodynamic response in a microtidal and shallow bay under energetic wind and seiche episodes, J. Mar. Syst., 149, doi:10.1016/j.jmarsys.2015.04.003, 2015a.

Cerralbo, P., Grifoll, M., Moré, J., Bravo, M., Sairouní Afif, a. and Espino, M.: Wind variability in a coastal area (Alfacs Bay, Ebro River delta), Adv. Sci. Res., 12, 11–21, doi:10.5194/asr-12-11-2015, 2015b.

Cerralbo, P., Grifoll, M., Valle-Levinson, A. and Espino, M.: Tidal transformation and resonance in a short, microtidal Mediterranean estuary (Alfacs Bay in Ebre delta), Estuar. Coast. Shelf Sci., 145, doi:10.1016/j.ecss.2014.04.020, 2014.

DHI, 2019b. MIKE 3 Wave Model FM Scientific documentation. http://manuals.mikepoweredbydhi.help/2019/Coast_and_Sea/MIKE_3_Wave_FM_Scientific_Doc.pdf

H2020-EO-2016-730030- CEASELESS

29

Ellis, J., Cummings, V., Hewitt, J., Thrush, S. and Norkko, A.: Determining effects of suspended sediment on condition of a suspension feeding bivalve (Atrina zelandica): Results of a survey, a laboratory experiment and a field transplant experiment, J. Exp. Mar. Bio. Ecol., 267(2), 147–174, doi:10.1016/S0022-0981(01)00355-0, 2002.

Fernández-Nóvoa, D., Mendes, R., deCastro, M., Dias, J. M., Sánchez-Arcilla, A. and Gómez-Gesteira, M.: Analysis of the influence of river discharge and wind on the Ebro turbid plume using MODIS-Aqua and MODIS-Terra data, J. Mar. Syst., 142, 40–46, doi:10.1016/j.jmarsys.2014.09.009, 2015.

Garel, E., Pinto, L., Santos, A. and Ferreira, Ó.: Tidal and river discharge forcing upon water and sediment circulation at a rock-bound estuary (Guadiana estuary, Portugal), Estuar. Coast. Shelf Sci., 84(2), 269–281, doi:10.1016/j.ecss.2009.07.002, 2009.

Grifoll, M., Aretxabaleta, A. L. and Espino, M.: Shelf response to intense offshore wind, J. Geophys. Res. C Ocean., 120(9), 6564–6580, doi:10.1002/2015JC010850, 2015.

Grifoll, M., Gracia, V., Aretxabaleta, A. L., Guillén, J., Espino, M. and Warner, J. C.: Formation of fine sediment deposit from a flash flood river in the Mediterranean Sea, J. Geophys. Res. Ocean., 119, 5837–5853, doi:10.1002/2014JC010187, 2014a.

Grifoll, M., Gracia, V., Aretxabaleta, A., Guillén, J., Espino, M. and Warner, J. C.: Formation of fine sediment deposit from a flash flood river in the Mediterranean Sea, J. Geophys. Res. C Ocean., 119(9), doi:10.1002/2014JC010187, 2014b.

Grifoll, M., Gracia, V., Fernandez, J. and Espino, M.: Suspended sediment observations in the Barcelona inner-shelf during storms, J. Coast. Res., (SPEC. ISSUE 65), doi:10.2112/SI65-259, 2013.

Grifoll, M., Navarro, J., Pallares, E., Ràfols, L., Espino, M. and Palomares, A.: Ocean–atmosphere–wave characterisation of a wind jet (Ebro shelf, NW Mediterranean Sea), Nonlinear Process. Geophys., 23(3), 143–158, doi:10.5194/npg-23-143-2016, 2016.

Guillen, J. and Palanques, A.: A shoreface zonation in the Ebro Delta based on grain size distribution, J. Coast. Res., 13(3), 867–878 [online] Available from: http://www.scopus.com/inward/record.url?eid=2-s2.0-0030746817&partnerID=tZOtx3y1, 1997.

Guillén, J., Palanques, a, Puig, P. and Durrieu de Madron, X.: Field calibration of optical sensors for measuring suspended sediment concentration in the western Mediterranean, Sci. Mar., 64(4), 427–435, doi:10.3989/scimar.2000.64n4427, 2000.

Guzinski, R., Spondylis, E., Michalis, M., Tusa, S., Brancato, G., Minno, L., & Hansen, L. B. (2016). Exploring the Utility of Bathymetry Maps Derived With Multispectral Satellite Observations in the Field of Underwater Archaeology. Open Archaeology, 2(1), 243–263. http://doi.org/10.1515/opar-2016-0018

Jacob, R., Larson, J. and and Ong, E.: {M}\cdotn communication and parallel interpolation in {CCSM3} using the {M}odel {C}oupling {T}oolkit, Int. J. High Perf. Comp. App., 19, 293–308, 2005.

Janssen, P. A. E. M. (1991). "Quasi-linear theory of wind generation applied to wave forecasting." Journal of Physical Oceanography 21: 1631-1642.

Johnson, H. K. and H. Kofoed-Hansen (2000). "Influence of bottom friction on sea surface roughness and its impact on shallow water wind wave modelling." Journal of Physical Oceanography 30: 1743-1756.

Jonsson, I. G. (1966). Wave boundary layers and friction factors. Int. Conf. on Coastal Engineering, Tokyo, Japan, ASCE.

Klonowski, W. M., Fearns, P. R. C. S., & Lynch, M. (2007). Retrieving key benthic cover types and bathymetry from hyperspectral imagery. Journal of Applied Remote Sensing, 1(1), 011505. http://doi.org/10.1117/1.2816113

H2020-EO-2016-730030- CEASELESS

30

Kumar, N., Voulgaris, G., Warner, J. C. and Olabarrieta, M.: Implementation of the vortex force formalism in the coupled ocean-atmosphere-wave-sediment transport (COAWST) modeling system for inner shelf and surf zone applications, Ocean Model., 47, 65–95, doi:10.1016/j.ocemod.2012.01.003, 2012.

Lee, Z., Carder, K. L., Chen, R. F., & Peacock, T. G. (2001). Properties of the water column and bottom derived from Airborne Visible Infrared Imaging Spectrometer. Journal of Geophysical Research, 106, 639–651.

Lee, Z., Carder, K. L., Mobley, C. D., Steward, R. G., & Patch, J. S. (1998). Hyperspectral remote sensing for shallow waters. I. A semianalytical model. Applied Optics, 37(27), 6329–38. Retrieved from http://www.ncbi.nlm.nih.gov/pubmed/18286131

Lee, Z., Carder, K. L., Mobley, C. D., Steward, R. G., & Patch, J. S. (1999). Hyperspectral Remote Sensing for Shallow Waters. 2. Deriving Bottom Depths and Water Properties by Optimization. Applied Optics, 38(18), 3831. http://doi.org/10.1364/AO.38.003831

Llebot, C., Rueda, F. J., Solé, J., Artigas, M. L. and Estrada, M.: Hydrodynamic states in a wind-driven microtidal estuary (Alfacs Bay), J. Sea Res., 85, 263–276, doi:10.1016/j.seares.2013.05.010, 2014.

Llebot, C., Solé, J., Delgado, M., Fernández-Tejedor, M., Camp, J. and Estrada, M.: Hydrographical forcing and phytoplankton variability in two semi-enclosed estuarine bays, J. Mar. Syst., 86(3–4), 69–86, doi:10.1016/j.jmarsys.2011.01.004, 2011.

Llebot, C., Spitz, Y. H., Sol??, J. and Estrada, M.: The role of inorganic nutrients and dissolved organic phosphorus in the phytoplankton dynamics of a Mediterranean bay: A modeling study, J. Mar. Syst., 83(3–4), 192–209, doi:10.1016/j.jmarsys.2010.06.009, 2010.

Loureiro, S., Garces, E., Fernandez-Tejedor, M., Vaqu??, D. and Camp, J.: Pseudo-nitzschia spp. (Bacillariophyceae) and dissolved organic matter (DOM) dynamics in the Ebro Delta (Alfacs Bay, NW Mediterranean Sea), Estuar. Coast. Shelf Sci., 83(4), 539–549, doi:10.1016/j.ecss.2009.04.029, 2009.

Lyzenga, D. R. (1978). Passive remote sensing techniques for mapping water depth and bottom features. Applied Optics, 17(3), 379–383.

Mestres, M., Sierra, J. P. A. U., Sánchez-arcilla, A., González, J., Río, D. E. L., Wolf, T. and Rodríguez, A.: Modelling of the Ebro River plume . Validation with field observations *, , 67(4), 379–391, 2003.

Newcombe, C. P. and Macdonald, D. D.: Effects of Suspended Sediments on Aquatic Ecosystems, North Am. J. Fish. Manag., 11(1), 72–82, doi:10.1577/1548-8675(1991)011<0072:EOSSOA>2.3.CO;2, 1991.

Palacín, C., Martin, D. and and Gili, J. M.: Features of spatial distribution of benthic infauna in a Mediterranean shallow-water Bay., Mar. Biol., 321, 315–321, 1991.

Palanques, A., Lopez, L., Guillén, J., Puig, P. and Masqué, P.: Decline of trace metal pollution in the bottom sediments of the Barcelona City continental shelf (NW Mediterranean), Sci. Total Environ., 579, 755–767, doi:10.1016/j.scitotenv.2016.11.031, 2017.

Ramírez-Pérez, M., Gonçalves-Araujo, R., Wiegmann, S., Torrecilla, E., Bardaji, R., Röttgers, R., Bracher, A. and Piera, J.: Towards cost-effective operational monitoring systems for complex waters: Analyzing small-scale coastal processes with optical transmissometry, PLoS One, 12(1), 1–21, doi:10.1371/journal.pone.0170706, 2017.

Roque, A., Lopez-Joven, C., Lacuesta, B., Elandaloussi, L., Wagley, S., Furones, M. D., Ruiz-Zarzuela, I., De Blas, I., Rangdale, R. and Gomez-Gil, B.: Detection and identification of tdh- And trh-positive Vibrio parahaemolyticus strains from four species of cultured bivalve molluscs on the Spanish Mediterranean coast, Appl. Environ. Microbiol., 75(23), 7574–7577, doi:10.1128/AEM.00772-09, 2009.

H2020-EO-2016-730030- CEASELESS

31

Satta, C. T., Anglès, S., Lugliè, A., Guillén, J., Sechi, N., Camp, J. and Garcés, E.: Studies on dinoflagellate cyst assemblages in two estuarine Mediterranean bays: A useful tool for the discovery and mapping of harmful algal species, Harmful Algae, 24, 65–79, doi:10.1016/j.hal.2013.01.007, 2013.

Shchepetkin, A. F. and McWilliams, J. C.: The regional oceanic modeling system (ROMS): a split-explicit, free-surface, topography-following-coordinate oceanic model, Ocean Model., 9(4), 347–404, doi:10.1016/j.ocemod.2004.08.002, 2005.

Stumpf, R. P., Holderied, K., Spring, S., & Sinclair, M. (2003). Determination of water depth with high-resolution satellite imagery over variable bottom types. Limnology & Oceanography, 48, 547–556.

Styles, R. and Glenn, S. M.: Modeling stratified wave and current bottom boundary layers on the continental shelf, J. Geophys. Res., 105(C10), 24119–24139, doi:10.1029/2000JC900115, 2000.

van Maren, D. S., van Kessel, T., Cronin, K. and Sittoni, L.: The impact of channel deepening and dredging on estuarine sediment concentration, Cont. Shelf Res., 95, 1–14, doi:10.1016/j.csr.2014.12.010, 2015.

Vermote, E. F., Tanré, D., Deuzé, J. L., Herman, M., & Morcrette, J. J. (1997). Second Simulation of the Satellite Signal in the Solar Spectrum ( 6S ). Software User Guide.

Warner, J. C., Armstrong, B., He, R. and Zambon, J. B.: Development of a Coupled Ocean–Atmosphere–Wave–Sediment Transport (COAWST) Modeling System, Ocean Model., 35(3), 230–244, doi:10.1016/j.ocemod.2010.07.010, 2010.

Warner, J. C., Sherwood, C. R., Signell, R. P., Harris, C. K. and Arango, H. G.: Development of a three-dimensional, regional, coupled wave, current, and sediment-transport model, Comput. Geosci., 34(10), 1284–1306, doi:10.1016/j.cageo.2008.02.012, 2008.

Wettle, M., & Brando, V. E. (2006). SAMBUCA Semi-Analytical Model for Bathymetry, Un-mixing, and Concentration Assessment. CSIRO Land and Water Science Report, (July).