Spatial estimates of the mean percentage of incoming wave height reaching areas of marsh and the standard deviation of those values are provided as raster layers (ED_to_MeanPercWave_30m1.tif and ED_to_SDPercWave_30m1.tif). The products represent a simplified model for the relationship between distance from the edge of a marsh parcel and the effect of the marsh surface on wave heights. Values are changes in wave heights expressed as percentages of the incident wave height at the marsh edge. To support this model X-beach was run in 1D transect mode over idealised bathymetry and vegetation extents. Water levels, incident wave heights and cross-shore slopes were varied to represent the variety of such conditions expected across the domain. Vegetation representations in the model were taken from field data collected at Tillingham, UK, while the Cd term was calibrated against wave measurements made at Hellegat, NL. Model configuration was largely the same as that used for the EUFAST educational version except that the grid was refined to 5m cells to allow resolution of changes even within narrow fringing marshes. Parameter values (see X-Beach documentation for definitions) used are summarised in the table below: Parameter Value(s) Wavecon (wave heights Hm0) 1, 1.5, 2, 2.5, 3, 3.5, 4 WaveconS (steepness H/L) 7, 15, 30 waterLevelCon (above MSL) 0.25, 0.5, 1, 1.5, 2, 2.5, 3, 3.5, 4, 4.5, 5 Coastalslopecon (1/n) 100, 150, 200, 250, 300, 400, 600, 800, 1000, 2000 Dir0 270 m 100 taper 0.01 wavint 30 gamma 0.5 nsec 1 ah 0.3 bv 0.00125 N 1125 Cd 0.19 Exhaustive combination of parameters resulted in 2310 model runs. From each run the significant wave height was extracted at distances along the vegetated transect (every 5 metres up to 100m, then at 250, 500, 1000 and 2000m). The effect of vegetation on significant wave height with distance (and its variance resulting from different topographic and hydrodynamics conditions) was then characterised by calculating the percentage of incoming wave height (Hm0) remaining at each distance interval. The mean and standard deviations for the percentage wave height remaining at each interval were then calculated. Two-term exponential functions were fitted to describe the variation with distance of the mean percentage wave height and the mean plus one standard deviation. The functions, where x is the distance along the vegetated transect, take the form: PercHsRemaining=a*exp(b*x) + c*exp(d*x) Coefficient values and goodness of fit are reported below. Function a b c d R-square Adj. R-square RMSE SSE Mean 28.6049 -0.0576 69.9979 -0.00183 0.9969 0.9965 1.259 33.31 Mean+1Sigma 9.8323 -0.1053 90.6793 -0.00093 0.9973 0.9969 1.046 22.96 Figure 1: Aggregated results of XBeach runs represented as percentage of incoming wave height remaining with distance along a vegetated transect. Exponential functions fitted to data are also shown. The UK Environment Agency saltmarsh extents layer (Phelan, 2011) was generalised to exclude small creeks and pools using a 10m outer buffer followed by dissolving overlapping features and buffering inwards by 10m. Landward margins of marsh parcels were removed by deleting any lines that intersected a 10m buffer from the UK Shoreline Management Plan vector layer, which typically describes the location of the sea wall or other line of defence. The resulting layer was then used to calculate a surface based on Euclidean distance of the cell centroid to the nearest margin of a marsh parcel. The analysis was conducted on a 30m grid. The exponential functions derived from the X-Beach modelling were applied to the 30m Euclidean distance surface to generate maps of mean percentage of incoming wave height and the standard deviation of that value at each cell. No attempt is made to model the spatial variability in incoming wave heights across the entire domain. The products supplied are Mean residual wave percentage and standard deviation of residual wave percentage. The script used to generate these results is FAST_MorphoChange_WaveAtten.py.

A Copernicus Sentinel-2 image was atmospherically corrected using Sen2Cor in SNAP 4, and then used to extract NDVI. Dike line was used to mask any area outside of the intertidal and subtidal zone. Coordinate system: WGS_84_UTM. Attribution: This product is developed by NIOZ for EU FAST project (Foreshore Assessment Using Space Technology). Contains modified Copernicus Sentinel data (2015/2016). See also https://sentinel.esa.int/documents/247904/690755/Sentinel_Data_Legal_Notice

A Copernicus Sentinel-2 image was atmospherically corrected using Sen2Cor in SNAP 4, and then used to extract NDVI. Dike line was used to mask any area outside of the intertidal and subtidal zone. Coordinate system: WGS_84_UTM. Attribution: This product is developed by NIOZ for EU FAST project (Foreshore Assessment Using Space Technology). Contains modified Copernicus Sentinel data (2015/2016). See also https://sentinel.esa.int/documents/247904/690755/Sentinel_Data_Legal_Notice

In this simulation the hinterland of Tillingham gets inundated as a result of a storm with a return period of 1/100 years. This is with taking into account the dissipating effects of vegetation. The amount of vegetation is quantified using on satellite-derived leaf area index (LAI) values with techniques derived within the FAST project. For this simulation two software packages have been used. The overtopping over the dike is calculated with XBeach-VEG. The inundation of the hinterland is carried out with LISFLOOD. For more information about the possibilities of XBeach-VEG see https://publicwiki.deltares.nl/display/VegMod/XBeach-VEG. For more information about LISFLOOD see http://www.bristol.ac.uk/geography/research/hydrology/models/lisflood/.

A Sentinel-2 image was atmospherically corrected using Sen2Cor in SNAP 4, and then used to extract Leaf Area Index (LAI), with the proviso that NDVI is larger than 0.3 to include marsh only. Dike line was used to mask any area outside of the intertidal and subtidal zone. Coordinate system: WGS_84_UTM. Attribution: This product is developed by NIOZ for EU FAST project (Foreshore Assessment Using Space Technology). Contains modified Copernicus Sentinel data (2015/2016). See also https://sentinel.esa.int/documents/247904/690755/Sentinel_Data_Legal_Notice

A Sentinel-2 image was atmospherically corrected using Sen2Cor in SNAP 4, and then used to extract Leaf Area Index (LAI), with the proviso that NDVI is larger than 0.3 to include marsh only. Dike line was used to mask any area outside of the intertidal and subtidal zone. Coordinate system: WGS_84_UTM. Attribution: This product is developed by NIOZ for EU FAST project (Foreshore Assessment Using Space Technology). Contains modified Copernicus Sentinel data (2015/2016). See also https://sentinel.esa.int/documents/247904/690755/Sentinel_Data_Legal_Notice

Coastal digital terrain model's (DTMs) are a critial component of MiSAFE. They are the geospatial data over which all wave propagation calculations are made and may form an important component of classification schemes; by defining elevation ranges over which different types of intertidal vegetation occur. EU/ESA's Copernicus, high resolution Sentinels (S1 and S2) and the NASA/USGS Landsat missions can potentially help create inter-tidal elevation maps. Elevation maps were developed using the Google Earth Maps. If you have any questions concerning the data, please contact Edward Morris.

Estimates of changes in wave heights reaching the line of defence are provided as point features. Estimates are derived from manual digitisation of vegetated marsh margins in aerial photography for the years 1992 (t0) and 2013/14 (t1). The Digital Shoreline Analysis System (DSAS – Theiler et al. (2008)). DSAS computes shoreline change rates along approximately shore-normal transects. For this analysis a nominal 10m alongshore separation was used. Further details of accuracy will be available in forthcoming publications but the method is capable of resolving changes in the order of 0.1m/y. The marginal change estimates derived above were filtered to include only those points where a line to the nearest line of defence crossed only coherent marsh surface (i.e. no creeks or major pools/mudflats), such that changes in the width of the marsh could reasonably be expected to represent changes in hydrodynamic conditions at the sea wall. For each marginal change point, the nearest point on the UK Shoreline Management Plan vector was identified and the distance between the two measured. The marginal change rate measured using DSAS was then applied to estimate initial (t0) and end (t1) distances between the sea wall and the marsh margin. X-beach was run in 1D transect mode over idealised bathymetry and vegetation extents. Water levels, incident wave heights and cross-shore slopes were varied to represent the variety of such conditions expected across the domain. Vegetation representations in the model were taken from field data collected at Tillingham, UK, while the Cd term was calibrated against wave measurements made at Hellegat, NL. Model configuration was largely the same as that used for the EUFAST educational version except that the grid was refined to 5m cells to allow resolution of changes even within narrow fringing marshes. Parameter values (see X-Beach documentation for definitions) used are summarised in the table below: Parameter Value(s) Wavecon (wave heights Hm0) 1, 1.5, 2, 2.5, 3, 3.5, 4 WaveconS (steepness H/L) 7, 15, 30 waterLevelCon (above MSL) 0.25, 0.5, 1, 1.5, 2, 2.5, 3, 3.5, 4, 4.5, 5 Coastalslopecon (1/n) 100, 150, 200, 250, 300, 400, 600, 800, 1000, 2000 Dir0 270 m 100 taper 0.01 wavint 30 gamma 0.5 nsec 1 ah 0.3 bv 0.00125 N 1125 Cd 0.19 Exhaustive combination of parameters resulted in 2310 model runs. From each run the significant wave height was extracted at distances along the vegetated transect (every 5 metres up to 100m, then at 250, 500, 1000 and 2000m). The effect of vegetation on significant wave height with distance (and its variance resulting from different topographic and hydrodynamics conditions) was then characterised by calculating the percentage of incoming wave height (Hm0) remaining at each distance interval. The mean and standard deviations for the percentage wave height remaining at each interval were then calculated. Two-term exponential functions were fitted to describe the variation with distance of the mean percentage wave height and the mean plus one standard deviation. The functions, where x is the distance along the vegetated transect, take the form: PercHsRemaining=a*exp(b*x) + c*exp(d*x) Coefficient values and goodness of fit are reported below. Function a b c d R-square Adj. R-square RMSE SSE Mean 28.6049 -0.0576 69.9979 -0.00183 0.9969 0.9965 1.259 33.31 Mean+1Sigma 9.8323 -0.1053 90.6793 -0.00093 0.9973 0.9969 1.046 22.96 The exponential functions derived above were then applied to estimate the change in the percentage of incoming wave height reaching the sea wall as a result of morphological changes over a 22 year period (1992-2014). No attempt is made to model the spatial distribution, or changes to, incoming wave conditions. The standard deviations represent a wider range of topographic and hydrodynamic conditions than most locations are likely to experience and therefore the uncertainty that they signify around the mean is conservative in the sense that it is likely to be an overestimate. Fields are: MarMovement – Change in margin position (m) along a line between the initial DSAS point estimate and the nearest point on the line of defence. MeanPercStart – Mean percentage of incoming wave height remaining at the line of defence at t0 MeanPercEnd – Mean percentage of incoming wave height remaining at the line of defence at t1 MeanPercDelta – Difference MeanPercEnd – MeanPercStart SDstart – Standard deviation of percentage wave height remaining at t0 SDend – Standard deviation of percentage wave height remaining at t1 SDofDiff – Standard deviation of MeanPercDelta

This data consists of 33 files of global topography (SRTM v4) completed with downscaled bathymetry for the coastal zone of the world. SRTM tiles have been downloaded from http://srtm.csi.cgiar.org/SELECTION/inputCoord.asp, GEBCO originates from http://www.bodc.ac.uk/data/online_delivery/gebco/

Vegetation presence map Donna Nook, UK map based on L3A RapidEye data, image 08 Feb 2015,top of atmosphere radiance converted to surface reflectance and atmospheric correction using 6S, threshold NDVI=0.3. This is an unvalidated product for demo purposes only! For more information, please see EU FAST Internal Deliverable 3.8 (Example Geospatial products)