为快速获得低等级河流的高精度河网,以松花江一级支流——阿什河为例,运用SWAT模型和河网密度相结合的方法合理计算集水面积阈值,通过最佳阈值的设定准确提取数字河网。研究结果表明:过大或过小的集水面积阈值都将导致所提取的河网与真实河网不符,当集水面积阈值〉5 km2时,河网密度及其一阶导数趋于稳定、二阶导数趋于零,说明河网变化趋于稳定,提取河网接近真实河网;当集水面积阈值分别为5、10和20 km2时,三者的模拟结果在平原区差距较小、在山区差距较大,其中当阈值为10 km2时,提取的河网与真实河网吻合度最高。采用相对误差Re、相关系数R2和纳什效率系数Ens对模拟结果进行验证和评价,结果显示,当集水面积阈值为10 km2时,Re=6%、R2=0.996、Ens=0.987,模拟结果最好。
To quickly SWAT model combining with and accurately extract digital create high-precision drainage network of low-level river, the method of drainage drainage density was used to reasonably calculate catchment area threshold network through setting the optimal threshold, taking Ash River as example, which is a tributary of the Songhua River. The results showed that there was discrepancy be- tween the extracted drainage network and the actual drainage network when the catchment area threshold was too large or too small. If the threshold was more than 5 km2 , the drainage density and its first deriva- tive tended to be stable, and its second derivative tended to zero, which indicated that changes of drain- age network tended to be stable, and the extracted drainage network was close to the actual drainage net- work. When the thresholds were 5 km2, 10 km2 and 20 km2 respectively, the simulation results had a small difference in the plain region and a large difference in the mountains region. Moreover, if thethreshold was 10 km2, the extracted drainage network fit best with the actual drainage network whether in the plain region or in the mountains region. By using Re, R2 and Ens to validate and evaluate the simula- tion results of Ash River, it was indicated that the simulation result was the best when the threshold was 10 km2, Re =6%, R2 =0.996 and Ens =0.987.