Mapping Island Reefs
GSD was contracted to create satellite derived bathymetry models (SDB) and benthic habitat models for an island close to Tanzania using depth and habitat field survey data together with satellite Earth Observation (EO) imagery. Data used included freely available and atmospherically corrected Sentinel-2 (S2) imagery (10 m spatial resolution) and WorldView3 imagery from Maxar.
Satellite Derived Bathymetry (SDB) modelling is a method of surveying shallow waters using satellite imagery. This method of bathymetry can reduce the movement of staff and equipment which can save costs. This SDB was carried out using a custom Python script developed by GSD for several similar bathymetry projects but for which the parameters can be customised for a particular location, imagery or client need.
For the benthic habitat mapping, indices and transformation of image bands were computed to reduce the error that derives from water depth on spectral values. GSD utilised Scikit-Learn's Random Forest classification algorithm to perform the habitat mapping. Again a custom-developed GSD python script was used.
The outputs provided included figures showing the SDB for the study sites and figures showing the modelled benthic habitat at the study sites. The data was also provided in formats that could be used locally by our client.
3D visualisations of the bathymetry were created and these could be draped over satellite or map imagery. This enabled the client to both explore the data in a very intuitive way. The products that GSD delivered were used to inform our client’s Marine Protected Area Management Plan. The maps will help to record changes in the areas over time and provide a good visual aid to present to the wider stakeholders when discussing why it is so important to protect these marine environments.