Background Since 2018, Ministry of Water Resources of the People's Republic of China has been doing dynamic monitoring of soil erosion across the country. Up to now, three phases of national dynamic soil erosion monitoring results have been obtained and released publicly. By obtaining seven factor data in the water erosion area, calculating the amount of soil erosion in the current year with the China Soil Loss Equation (CSLE; A=RKLSBET), we carried out a graded statistical analysis of various types of information at six levels: slight, mild, moderate, strong, extremely strong, and severe. At present, the topographic factor (slope, and slope length) data used in the dynamic monitoring of soil erosion in the provincial monitoring areas of Shaanxi province and the 1∶50 000 topographic map have certain deviation in the slope classification, especially in that of plain areas. The graded slope cannot accurately reflect the topographical features, and this will affect, to varying degrees, various data classified by the slope.

Methods Taking Chang'an district of Xi'an as an example, we took the three commonly used DEM data of SRTM1, NASADEM and ASTER GDEM to perform a spatial calibration with the grid DEM generated by the elevation points and contour of a 1∶50 000 topographic map, which was followed by a fill-in analysis on these 4 kinds of DEM data, and generation of 6 grades of slope classification maps (< 5, ≥5°-8°, ≥8°-15°, ≥15°-25°, ≥25°-35°, and ≥35°). Finally, the 3 slope classification maps of SRTM1, NASADEM, ASTER GDEM and that of the grid DEM were used for deviation analysis.

Results Compared with the grid DEM, the deviation of SRTM1 is the slightest of the three in spatial distribution, and is more accurate in reflecting the topographic information in the area. By SRTM1, the maximum deviation of the proportion of slope classification in plain areas was 1.15%, and that in mountainous areas was -4.80%. In contrast to this, the deviations of the other two DEM data in the proportion of slopes above 35° in mountainous areas were lower than -16%, which, obviously, cannot reflect accurately the distribution of high-grade slopes in mountainous areas.

Conculsions The findings of the present study show that in the observed area, the STRM1 DEM is superior to the other two types of DEM data in terms of the degree of deviation of the slope classification ratio and the accuracy of spatial distribution. The data could be used as the basic topographic data for dynamic monitoring in Shaanxi province, specifically for various statistical analysis involving slope information, the calculation of soil erosion intensity, and objective and accurate reflection of the situation of soil erosion in the area.