Review on algorithms of dealing with depressions in grid DEM

ABSTRACT Depressions in grid digital elevation models (DEMs) need to be dealt with before the topographic attributes (such as specific catchment area) and terrain features (such as drainage networks) related to flow directions can be derived from DEMs in a hydrologically-correct manner. Many depression-processing algorithms, which adopt different strategies and take different information under consideration for determining correct flow directions in depressions, have been proposed. However, currently, there is still no one algorithm which can satisfactorily deal with depressions in grid DEMs under various application contexts. In this paper, we review existing depression-processing algorithms based on the adopted strategies (i.e. the DEM-revising strategy and the DEM-unchanging strategy). Algorithms with the DEM-revising strategy especially are discussed in detail according to their designs relating to the revision of DEM elevations, i.e. the smoothing filter, depression filling, depression breaching (or carving), using other qualified data, and applying different algorithms to depressions with different characteristics. Existing ways of improving the computation efficiency of depression-processing algorithms are also presented, i.e. serial algorithm optimization and parallel algorithms. Lastly, we discuss a possible design for an optimal depression-processing algorithm which may be developed in the future