引用本文:李任君,高懋芳※,李 强,李百寿.基于ANUSPLIN 的降水空间插值方法研究[J].中国农业信息,2019,31(1):48-57
【打印本页】   【HTML】   【下载PDF全文】   查看/发表评论  【EndNote】   【RefMan】   【BibTex】
←前一篇|后一篇→ 过刊浏览    高级检索
本文已被:浏览 27次   下载 8 本文二维码信息
码上扫一扫!
分享到: 微信 更多
基于ANUSPLIN 的降水空间插值方法研究
李任君1, 高懋芳※2, 李 强2,3, 李百寿1
1.桂林理工大学测绘地理信息学院,广西桂林541006;2.中国农业科学院农业资源与农业区划研究所/ 农业农村部农业遥感重点实验室,北京100081;3.江苏师范大学地理测绘与城乡规划学院,徐州221000
摘要:
【目的】对黄淮海平原点状降水数据进行空间插值,筛选最优模型,分析插值精度, 为该区域水分供给状况分析、农业干旱研究等提供科学依据与技术支撑。【方法】采用基于样 条函数插值理论的专业气象插值软件ANUSPLIN,根据黄淮海平原内457 个气象站点1981— 2010 年连续30 年的降水数据,分别以分辨率为90 m、1 km 的高程数据作为第三变量,对 降水数据进行空间插值,根据误差统计选出最优插值模型,分析不同分辨率的数字高程模型 与插值精度的关系;为比较ANUSPLIN 插值结果,随机选取29 个点的降水数据作为验证集, 同时将其与克里金插值方法进行比较。【结果】(1)对所选气象站点数据进行交叉验证发现, 相比较克里金插值,ANUSPLIN 得到的结果精度更高。冬季降水量较少时的插值精度比降 水集中的6—8 月份的插值精度高,利用ANUSPLIN 对冬季的降水数据插值的均方根误差为 0.38 mm,夏季为4.19 mm,克里金方法对冬季降水数据插值后对应的RMSE 为0.45 mm、夏 季为4.31 mm;(2)DEM 分辨率越高,对应的插值精度会有所提升,对夏季降水插值较明显, 利用90 m 分辨率的DEM 对夏季降水插值,RMSE 为4.19 mm,1 km 分辨率的DEM 插值后 对应的RMSE 为4.24 mm。【结论】通过ANUSPLIN 对黄淮海平原的降水插值方法研究,探 讨插值精度与DEM 分辨率的关系,发现提高协变量数据DEM 的分辨率可以获得更高精度的 降水栅格数据,相比较克里金方法,AUNSPLIN 获得的结果更加细致地描绘出地形因素对降 雨空间分布的影响,为黄淮海平原干旱分析、指导当地农作物灌溉生产提供重要的决策支持 信息。
关键词:  ANUSPLIN;降水;空间插值;黄淮海平原
DOI:10.12105/j.issn.1672-0423.20190105
分类号:
基金项目:国家自然科学基金项目“耦合遥感与作物生长模型的农业干旱预警研究(41871282)”
Research on rainfall spatial interpolation methodbased on ANUSPLIN
Li Renjun1, Gao Maofang※2, Li Qiang2,3, Li Baishou1
1.Guilin University of Technology,College of Geomatics and Geoinformation,Guangxi Guilin 541006,China;2.Institute of Agricultural Resources and Regional Planning,Chinese Academy of Agricultural Sciences,Beijing 100081,China;3.Jiangsu Normal University,College of Geography,Geomatics and Planning,Xuzhou 221000,China
Abstract:
[Purpose]This study provides an accurate and optimal model of rainfall spatial interpolation for crop water management and agricultural drought monitoring in Huanghuaihai Plain. [Method] In this study,we first apply ANUSPLIN,spline interpolation method,and use 30 years rainfall data of 458 meteorological stations in Huanghuaihai Plain from 1981 to 2010. In addition,we use DEMs(digital elevation model)with resolution of 90m and 1km respectively as the third variable for interpolating rainfall data. Then we apply error statistics to select the optimal interpolation model, and analyze the relationship between different DEMs to assess the interpolation accuracy. Finally,we randomly select 30 validation points for comparing the from rainfall data interpolation between Kriging and ANUSPLIN.[ Result](1)The ANUSPLIN results are in higher accuracy and smoothness in rainfall results than the ordinary Kriging results. The interpolation accuracy is higher in winter than that in summer,when the rainfall in winter is less than that in summer. The RMSE(root mean squared error)of ANUSPLIN interpolation can be reduced by 0.38 mm in winter and 4.19 mm in summer. Furthermore,the ordinary Kriging interpolation is 0.45 mm in winter and 4.31 mm in summer.(2)The finer DEM resolution(90m)performed better accuracy in rainfall interpolation with a RMSE of 4.19 mm than the coarser DEM resolution(1 km)with a RMSE of 4.24 mm in summer[. Conclusion] The ANUSPLIN rainfall interpolation method in Huanghuaihai Plain performs better than other ordinary Kriging interpolation method and provides better results with the support of finer resolution DEM. Compared to Kriging method,the results obtained by AUNSPLIN give a more detailed description of the topographic factors on the spatial distribution of precipitation. This study helps to provide important support for drought analysis and guiding local crop irrigation production in Huanghuaihai Plain.
Key words:  ANUSPLIN;rainfall;spatial interpolation;Huanghuaihai Plain