干旱气象 ›› 2025, Vol. 43 ›› Issue (5): 790-798.DOI: 10.11755/j.issn.1006-7639-2025-05-0790

• 研究论文 • 上一篇    下一篇

基于ANUSPLIN软件的黑河流域降水空间插值方法研究

唐娜1,2(), 王静1, 董国涛1(), 陈子豪3, 孔令林4   

  1. 1.黑河水资源与生态保护研究中心,甘肃 兰州 730030
    2.黄河水利委员会黑河流域管理局,甘肃 兰州 730030
    3.黑河黄藏寺水利枢纽工程建设管理局,甘肃 兰州 730030
    4.昆明市生态环境监控中心,云南 昆明 650000
  • 收稿日期:2025-02-20 修回日期:2025-04-08 出版日期:2025-10-31 发布日期:2025-11-09
  • 通讯作者: 董国涛(1982—),男,正高级工程师,主要从事水文水资源遥感方面研究。E-mail:dongguotao001@163.com
  • 作者简介:唐娜(1987—),女,工程师,主要从事水资源管理与研究工作。E-mail:tangnahh@163.com
  • 基金资助:
    国家社会科学基金项目(23&ZD104)

Research on precipitation spatial interpolation methods in Heihe River Basin based on ANUSPLIN software

TANG Na1,2(), WANG Jing1, DONG Guotao1(), CHEN Zihao3, KONG Linglin4   

  1. 1. Heihe Water Resources and Ecological Protection Research Center, Lanzhou 730030, China
    2. Heihe River Bureau of the Yellow River Conservancy Commision, Lanzhou 730030, China
    3. Heihe Huangzangsi Water Control Project Construction & Management Bureau, Lanzhou 730030, China
    4. Kunming Municipal Ecological Environment Surveillance Center, Kunming 650000, China
  • Received:2025-02-20 Revised:2025-04-08 Online:2025-10-31 Published:2025-11-09

摘要:

西北内陆河地区地形地貌复杂,气象台站较少,为提高该地区降水量空间化的精确度,支撑区域水文模拟及水资源管理,以西北内陆河黑河流域为研究区,利用19个气象台站与26个水文站点数据,基于ANUSPLIN软件构建局部薄盘光滑样条插值模型,分析模型参数、站点数量与数字高程模型(Digital Elevation Model,DEM)分辨率对降水量空间插值精度的影响。结果表明:1)独立变量为经度与纬度,协变量为海拔,样条次数为3,变量转换方式为平方根时,V2C1S3_RT模型的插值精度最优,其广义交叉验证的平方根(Square Roots of Generalized Cross Validation, RTGCV)和期望真实均方根误差(Square Roots of Mean Square Error, RTMSE)分别为6.10和4.82 mm;2)站点数量影响插值精度,当站点数增加至40个,RTGCV和RTMSE均达到最小值,进一步增加站点数对精度提升有限;3)不同分辨率DEM对降水量插值结果影响不大;4)2019年黑河流域平均降水量为211.39 mm,呈现西南向东北递减趋势,上游西南地区及下游边界地区降水量标准误差较相邻地区大,主要原因是该区域站点数量较少。

关键词: 西北内陆河, 薄盘光滑样条, ANUSPLIN, 降水量插值, DEM

Abstract:

The terrain in northwest inland river of China is complex, where there are few meteorological stations. In order to improve the accuracy of precipitation spatialization, support hydrological modeling and water resource management in arid area, based on the datasets of 19 meteorological stations and 26 hydrological stations in the Heihe River Basin, the partial thin plate smoothing splines function models are constructed using ANUSPLIN. The influence of model parameters, station numbers and the resolution of Digital Elevation Model (DEM) on precipitation spatial interpolation accuracy is analyzed. The results are as follows: 1) The V2C1S3_RT model, with longitude and latitude as independent spline variables, elevation as the independent covariate, the spline order of 3, and square root as the variable transform method, achieves the highest interpolation accuracy, with square roots of generalized cross validation (RTGCV) and square roots of mean square error (RTMSE) values of 6.10 and 4.82 mm, respectively. 2) The number of stations significantly affects interpolation accuracy. When the number of stations increases to 40, the RTGCV and RTMSE are the minimum, and further increasing the number of stations has limited improvement on accuracy. 3) Different resolution DEM has little effect on precipitation interpolation results. 4) The regional mean precipitation in the Heihe River Basin in 2019 was 211.39 mm, decreasing from southwest to northeast. Due to sparse station distribution, the standard error of precipitation in the southwestern upstream and downstream boundary areas is significantly higher than that in other regions.

Key words: northwest inland river, thin plate smoothing splines, ANUSPLIN, precipitation interpolation, DEM

中图分类号: