Depression removal method for grid DEM based on three-direction search

The Digital Elevation Model(DEM)is the fundamental data source to extract the watershed parameters. However, the depressions and flat areas commonly exist due to the errors in the production of DEM. Various methods are thus proposed to remove the depressions and flat area in DEM. Most of the methods assume that the depressions are caused mainly by the underestimation of DEM values. In fact,the overestimation can also cause the depressions in the nearby grid cells. In this paper, a new method based on three-direction search is presented to handle the problems by both raising and reducing the DEM values. The result suggests that the proposed method is easy to implement and can eliminate the "parallel river" in the result,and makes less modification of the original DEM values only to keep the original information to the largest extent.