%0 Journal Article
%T 流域异常雨量站点检测对水文模型模拟精度影响研究
Study on the Influence of Abnormal Rainfall Station Detection on Hydrological Model Simulation Accuracy
%A 马振亮
%A 段瞳
%A 田逸飞
%A 许银山
%J Journal of Water Resources Research
%P 12-22
%@ 2166-5982
%D 2025
%I Hans Publishing
%R 10.12677/jwrr.2025.141002
%X 流域内存在异常的雨量站点会影响水文模型输入的质量，进而降低水文模型的模拟精度。为了提高水文模型模拟精度，先通过计算降雨过程缺测指数及相似指数，利用DBSCAN聚类算法对计算的指数进行聚类筛选出缺测异常站点；其次构建集总式新安江模型，对筛选异常雨量站点前后的1 h、3 h、6 h时段长系列流量资料进行模拟比较。乐安河流域的研究结果表明，剔除异常雨量站数据后，纳西效率系数提升了0.07，径流总量相对误差减少18.08%，可提高水文模型的模拟精度。
The presence of abnormal rainfall stations in the basin will affect the quality of input data, and reduce the hydrological simulation accuracy. To improve the hydrological simulation accuracy, this paper attempts to calculate the rainfall process missing index and rainfall process similarity index, and use the DBSCAN clustering algorithm to cluster the calculated index to screen out the missing abnormal stations. The lumped Xin’anjiang model is constructed to simulate the 1 h, 3 h, and 6 h long flow series before and after the screening of abnormal rainfall stations. The application results of the Le’an River basin show: after the abnormal rainfall station data are eliminated, the Nash efficiency coefficient is increased by 0.07, while the relative error of water volume is reduced 18.08%, which can improve hydrological simulation accuracy.
%K 雨量站，
%K 异常检测，
%K 降雨过程指数，
%K 新安江模型，
%K DBSCAN聚类
Rainfall Station
%K Anomaly Detection
%K Rainfall Process Index
%K Xin’
%K anjiang Model
%K DBSCAN Clustering
%U http://www.hanspub.org/journal/PaperInformation.aspx?PaperID=109373