Redefining the watershed line and stream networks via digital resources and topographic map using GIS and remote sensing (case study: the Neka River’s watershed)

Abstract

The accurate delineation of area plays a key role in the surveying of land change detection and the classification of land covers. In a hydrologic system, the watershed delineation and the detection of the boundaries among watershed is a basic method for performing spatial analyses. After recent advances in image processing and raster-based spatial analysis in geographic information systems, and being easily accessible data via various sources especially through remote sensing, the reliable determination of topographical boundaries possible is possible. Therefore, an integrated approach of data analysis and modeling can accomplish the task of delineation. The main aim in this research is to evaluate the delineation method of watershed boundary using four different digital elevation models (DEM) including advanced spaceborne thermal emission and reflection radiometer (ASTER), Shuttle Radar Topographic Mission (SRTM), digital topography, and topographic maps. In order to determine a true reference of boundary of watershed, sample data were also obtained by field survey and using global positioning system (GPS). The comparison reference points and the results of these data showed the average distance difference between reference boundary, and the result of ASTER data was 43 m. However, the average distance between GPS reference and the other data was high; the difference between the reference data and SRTM was 307 m, and for digital topographic map, it was 269 m. The average distance between topographic map and the GPS points differed 304 m as well. For the statistical analysis of comparison, the coordinates of 230 points were determined; the paired comparisons were also performed to measure the coefficient of determination, R 2, as well as analysis of variance in SPSS software. As a result, the R 2 values for the ASTER data with the digital topography and topographic map were 0.0157 and 0.171, respectively. The results showed that there were statistically significant differences in distances among the four means of the selected models. Therefore, considering other three methods, the ASTER DEM is the most suitable applicable data to delineate the borders of watersheds, especially in rugged terrains. In addition, the calculated flow directions of stream based on ASTER are close to natural tributaries as well as real positions of streams.