Correcting vertical errors in a global Digital Elevation Model to derive a bare earth terrain surface for improved flood modelling in data-scarce regions

Flood modelling is crucial to understanding and preparing for future flood risk. Topography data is a fundamental input for these models to describe ground surface, ideally derived from airborne laser scanning (LiDAR), which is more common in higher-income countries but remains rare elsewhere.

In these data-scarce regions, flood modelling often relies on inacurate global Digital Elevation Models (DEMs), derived from satellite imagery and known to contain vertical artefacts that reduce efficacy in flood modelling.

This research assesses the vertical accuracy of the various global Digital Elevation Models (DEMs) and developing a method to predict vertical errors and enable error correction to obtain a bare earth terrain model to improve flood modelling. 

This project will enable more reliable hazard modelling in data-scarce regions, where impacts are often most significant.