Error-reduced digital elevation models and high-resolution land cover roughness in mapping tsunami exposure for low elevation coastal zones

This study presents a systematic exposure assessment by reconstructing the impact of the 2004 Indian Ocean Tsunami using a wide range of inundation scenarios and multiresolution exposure layers. To develop inundation and exposure models, we employed the error-reduced global digital elevation models (DEMs) and geospatially consistent multiresolution datasets: land cover roughness (LCR) models, built-up areas, and gridded population layers. We implemented three sequential validation assessments to evaluate the performance of inundation models, incorporating satellite observations, post-tsunami measurements, and the confidence level associated with inherent DEM error characteristics. The results demonstrated that the error-reduced variants of Copernicus DEM (i.e., FABDEM and DiluviumDEM) satisfied all reliability criteria. Incorporating these elevation models with LCR models improved the accuracy of inundation depth estimates; however, it reduced the agreement between simulated and observed inundation extents. We observed that applying high-resolution LCR models had a minimal impact on overland inundation extents but still influenced the exposure assessment, especially in high-density urban areas. • 11 elevation models were evaluated against ICESat-2 photon altimetry data across various land cover types. • Generated land cover roughness (LCR) models were geospatially consistent with built-up and gridded population data. • A three-step sequential validation assessment was developed to evaluate the inundation model perfomance. • High-resolution LCR models improved flow depth accuracy but reduced inundation agreement with satellite observations.