DEM分辨率及子流域划分对AnnAGNPS模型模拟的影响

 针对大中尺度流域非点源污染模拟精度与运算效率的突出矛盾,选取数据精度与空间离散化2个关键影响因素,讨论数字高程模型(DEM)分辨率和临界源面积(CSA)取值对AnnAGNPS模型模拟结果的影响.分别比较了7组CSA取值及8组不同DEM分辨率对华北潮河流域大阁子流域AnnAGNPS模型参数提取及模拟结果的影响.研究发现,CSA取值对汇水单元的数量、地表径流量、泥沙以及总氮的模拟有很大的影响.随着CSA取值的增大,地表径流量变化不大;泥沙先逐渐增大,后呈减少趋势;总磷负荷的变化趋势不明显;总氮负荷呈逐渐减少的趋势.选择合适的CSA是流域AGNPS模型模拟的基础.DEM是分布式水文模型的基础数据,研究表明:DEM分辨率对坡度有较大的影响.随着DEM分辨率的降低,径流量变化不大;土壤侵蚀和总氮负荷变化却十分明显,当DEM分辨率从30m逐渐降低至60m时,泥沙输出及总氮负荷迅速减小,总磷呈现先减小后逐渐稳定的趋势,但变化幅度小于总氮和泥沙.当DEM分辨率小于60m,泥沙和总氮的输出随分辨率降低下降缓慢.以30m和50m DEM为例,通过不同精度的DEM坡度回归分析,建立了低精度DEM坡度修正的方法.用修正后的坡度进行AnnAGNPS模型模拟计算,结果表明坡度修正对泥沙、总氮、总磷的模拟有不同程度的改善.

The effect of DEM resolution and subwatershed delineation on AnnAGNPS model simulation

The impact of critical source area (CSA) and DEM resolution on hydrological attributes was examined by comparing seven sets of CSA (0.5, 1, 2, 3, 4, 6, and 8km2) and eight DEMs (30, 40, 50, 60, 70, 80, 90, and 100m) using the case study of Dage sub-watershed of Chaohe watershed of North China. The accuracy of agricultural nonpoint source pollution models (AnnAGNPS) stimulation on runoff, sediment and nutrient loads on these different sets of CSA and DEM was further suggested. Variation of CSA would lead to the uncertainty of average slope which increase the generalization of land characteristics. At the range of CSA from 0.5km2to 1km2, there was little impact of CSA on slope. Runoff amount did not vary so much with the variation of CSA whereas soil erosion and TN load change prominently. An increase of sediment yield was observed firstly then a decrease following later. Evident decrease of TN load occurred, especially when CSA was bigger than 6km2. TP load showed little variation with the change of CSA. So the proper CSA was the basis of AnnAGNPS Model simulation. Regarding to the impact of DEM resolution within an AnnAGNPS Model simulation, there was significant impact on slope gradient which can be corrected by cubic model with the correction of 99.80% based on the liner regression equation of average slope and DEM mesh size. Runoff did not vary much with decrease of DEM resolution whereas soil erosion and total nitrogen (TN) load change prominently. There was little effect on runoff simulation of AnnAGNPS modeling by the amended slope using an adjusted 50m DEM. A decrease of sediment yield and TN load was observed with an increase of DEM mesh size from 30m to 60m; a slight decrease of sediment and TN load with the DEM resolution bigger than 60m. There were similar trends for total phosphorus (TP) variation, but with less range of variation than the simulation of sediment, TN, and TP. In all, the amended simulation still had a large difference relative to the results using 30m DEM. Results for Dage sub-watershed showed that fine meshes were desired to avoid large underestimates of sediment and total nitrogen loads and moderate underestimates of total phosphorus loads. Decreasing the mesh size beyond this threshold did not substantially affect the computed runoff flux but generated prediction errors for nitrogen and sediment yields.