Tidal simulations are extremely sensitive to the accuracy of bathymetry in shallow seas. This accuracy is highly variable in global databases. In recent releases, and thanks to gravimetry anomalies data inversion, this accuracy did improved, mostly in terms of bias, but noisy artefacts in those type of products are quite common. We continuously try to correct those artefacts by integrating other sources of data, such as local DTMs or depth soundings. For these reasons, new efforts are put in the mapping of available bathymetry data (depth sounding, ship tracks, multi-beam data) in places where we believe it will profit tidal modeling.
In a first step data are decimated to a user-assigned resolution (typically 50 m) to avoid over-densified node set. Original data are aggregated by clusters (dimensioned by decimation resolution), computing average depth and location of clustered data for mapping purposes. Additional high resolution elevation data (such as SRTM land elevation) are used to better constrain the bathymetry along shorelines and provide a ocean and land DTM. A prior bathymetry DTM is used to optionally populate "data empty" sectors and borders of the mapping area. Final DTM is achieved by using Delauney meshing and regular DTM is computed using linear interpolation over the triangular depth mesh.
In the following sections and documents, we focused on shallow water depths only. A similar study for deeper values would be of great interest, at least to check for possible deep ocean bathymetry biases.
Regions of interest/processed DTMS
North Atlantic Ocean
European Shelf (2014, to be documented)
Seine River estuary, Gironde River estuary, France (2013, to be documented)
, preliminary processing)
Hudson Bay/Foxe Basin (2013, to be documented)
South Atlantic Ocean
Amazonian Shelf (2012, to be documented)
Douala estuary, Cameroun (2013, to be documented)
Mertz (2012, to be documented)
Kerguelen (2012, to be documented)
Crozet (2013, to be documented)
Taïwan waters (2014, to be documented)