一种基于数字高程模型的流域边界提取修正方法

Boundary extraction of watershed is an important step in forest landscape research. The boundary of the upriver watershed of the Hunhe River in the sub-alpine Qingyuan County of eastern Liaoning Province, China was extracted by digital elevation modeling (DEM) data in Arclnfo8.1. Remote sensing image of the corresponding region was applied to help modify its copy according to Enhanced Thematic Mapper (ETM) image‘s profuse geomorphological structure information. Both the DEM-dependent boundary and modified copy were overlapped with county map and drainage network map to visually check the effects of result. Overlap of county map suggested a nice extraction of the boundary line since the two layers matched precisely,which indicated the DEM-dependent boundary by program was effective and precise. Further upload of drainage network showed discrepancies between the boundary and the drainage network. Altogether, there were three sections of the extraction result that needed to correct. Compared with this extraction boundary, the modified boundary had a better match to the drainage network as well as to the county map. Comprehensive analysis demonstrated that the program extraction has generally fine precision in position and excels the digitized result by hand. The errors of the DEM-dependant extraction are due to the fact that it is difficult for program to recognize sections of complex landform especially altered by human activities, but these errors are discernable and adjustable because the spatial resolution of ETM image is less than that of DEM. This study result proved that application of remote sensing information could help obtain better result when DEM method is used in extraction of watershed boundary.

Modified method for extraction of watershed boundary with digital elevation modeling

Boundary extraction of watershed is an important step in forest landscape research. The boundary of the upriver watershed of the Hunhe River in the sub-alpine Qingyuan County of eastern Liaoning Province, China was extracted by digital elevation modeling (DEM) data in Arclnfo8.1. Remote sensing image of the corresponding region was applied to help modify its copy according to Enhanced Thematic Mapper (ETM) image’s profuse geomorphological structure information. Both the DEM-dependent boundary and modified copy were overlapped with county map and drainage network map to visually check the effects of result. Overlap of county map suggested a nice extraction of the boundary line since the two layers matched precisely, which indicated the DEM-dependent boundary by program was effective and precise. Further upload of drainage network showed discrepancies between the boundary and the drainage network. Altogether, there were three sections of the extraction result that needed to correct. Compared with this extraction boundary, the modified boundary had a better match to the drainage network as well as to the county map. Comprehensive analysis demonstrated that the program extraction has generally fine precision in position and excels the digitized result by hand. The errors of the DEM-dependant extraction are due to the fact that it is difficult for program to recognize sections of complex landform especially altered by human activities, but these errors are discernable and adjustable because the spatial resolution of ETM image is less than that of DEM. This study result proved that application of remote sensing information could help obtain better result when DEM method is used in extraction of watershed boundary. Foundation item: This work was supported by Knowledge Innovation Program, Chinese Academy of Sciences (No. KZCX2-SW-32O-3 & KZCX3-SW-425).