Abstract

Continuing degradation of urban environments of developing economies results from failure to make decisions based on scientific methods that employ accurate digital elevation models (DEM) in modelling environmental phenomena such as flooding and erosion. Data dumping and wastage occur if fresh DEM are required for every new project ignoring existing data. Every DEM to be used for modelling may be accepted as is, updated or rejected for a new one based on outlier fitness tests. High cost DEM aged only a few months should also be tested since elevation changes could occur in very short periods. Older DEMs might still be useful if changes are only in limited areas which could be mapped to update the DEM, so saving cost. A DEM of Owerri, South East Nigeria created from a 1977 1/2500 topographic map series was evaluated using elevations of random Global Navigation Satellite Systems (GNSS) points resulting in a ±1.290 m root mean square error (RMSE). Outliers of the 99% certainty were traced to a heavily cut construction site and a topographical survey of the site was used to update the DEM improving RMSE to below the meter mark.

Key words: Digital elevation model (DEM), geographic information systems (GIS), outlier, DEM updating, root mean square error (RMSE).