Assessment of SRTM, ASTER, and TanDEM-X DEMs for Modeling Storm Surge Inundation for the Island of Hispaniola

The Island of Hispaniola, which consists of Haiti and the Dominican Republic, suffers from the frequent impact of hurricanes because it is located in the path of hurricanes that originate from West Africa and the Caribbean Sea. To avoid loss of life due to flooding, storm surge threat is forecasted using numerical models such as the Sea, Lake, and Overland Surges from Hurricanes (SLOSH) model.  Recently, the World Meteorological Organization (WMO) has initiated the Caribbean Coastal Inundation and Forecast Demonstration Project (CIFDP) to develop storm surge products for flood forecasting, planning, and warning along the coast of Hispaniola using SLOSH and a coupled parametric wave model.

One of the important factors influencing the modeling of storm surge inundation and waves is the topographic elevation in coastal areas. High quality digital elevation models (DEM) from airborne Light Detection and Ranging (LiDAR) and aerial photogrammetric stereo analysis are lacking for the development of storm surge model grids and inundation simulation for Hispaniola Island. The topographic datasets from NASA’s Shuttle Radar Topography Mission (SRTM), the Advanced Spaceborne Thermal Emission and Reflection Radiometer (ASTER), and the TanDEM-X mission by German Space Agency were compared with ICESat LiDAR and kinematic ground GPS measurements to assess the quality of DEMs.  SRTM and ASTER DEMs have a horizontal resolution of 30 m and vertical errors of 8-12 m and the TanDEM-X DEM has a horizontal resolution of 12 m and vertical errors of 2-4 m. The comparison of SRTM and ASTER DEMs, with more than 2,000 ground GPS measurements distributed across one hundred 30 m × 30 m DEM grid cells in the low-relief coastal areas at Pedernales, Samaná, and Sanchez in the Dominican Republic, showed that the root mean square errors (RMSE) of SRTM and ASTER DEM were about 3.8 m and 8.5 m, respectively. The SRTM DEM exhibited less fluctuations spatially than the ASTER DEM, indicating that SRTM is better for the development of the model grids. The comparison of SRTM, ASTER, and TanDEM-X DEMs along the Torbeck coast in Haiti demonstrated that the TanDEM-X DEM better depicted local topographic ridges and river streams, which are essential for building major subgrid-scale features that affect the inundation generated by storm surge and waves. Therefore, it is recommended to use the TanDEM-X DEM in the critical flood zones and the SRTM DEM in the remaining areas to develop the storm surge and wave models for Hispaniola.