全国近岸海域水质空间插值算法精度分析  

Analysis on Simulation Precision of the Spatial Interpolation Models Application on China Coastal Seawater Environmental Quality

作　　者：李俊龙[1] 丁页[1] 刘喜惠[1] 刘方[1] 李曌[1] 陈平[1] 刘海江[1] 李铭煊[1] 高锋亮 

机构地区：[1]中国环境监测总站、国家环境保护环境监测质量控制重点实验室,北京100012

出　　处：《中国环境监测》2015年第2期135-140,共6页Environmental Monitoring in China

基　　金：环境保护部公益性行业科研专项(201309008);中国环境监测总站转型发展科研支撑项目(2011ZX-HY014)

摘　　要：空间插值技术对近海环境数据的分析和表征具有广泛应用价值,但目前针对各类模型适用性和精度比较的研究很少。基于2009年全国近岸海域299个站位的水质监测数据,利用交叉检验法,对3种插值算法13种模型的空间拟合结果进行分析和验证。结果表明,13种模型的模拟值与实测值显著相关（r=0.643～0.750,P〈0.01,n=299）,说明其均适用于近岸海域水质的空间模拟;综合比较模拟结果与实测结果的相关系数、平均误差、平均绝对误差、均方差及偏离程度,排序为IDW（p2）〈Spline-Tension〈IDW（p3）〈IDW（p4）〈IDW（p1）〈IDW（p5）〈IDW（p6）〈Kriging-Guassian〈Kriging-Spherical,Kriging-Circular〈Kriging-Linear〈Kriging-Exponential〈Spline-Regularized,提出IDW（p2）是精度最优的空间插值模型。The application prospect of spatial interpolation method will be great for coastal environmental monitoring in future, but there were few papers discussing spatial interpolation applicability and precision. Based on coastal seawater environmental monitoring data from 299 stations in 2009, we validated the simulation precision of three spatial interpolation method including thirteen models with the method of Cross-Validation. All the spatial interpolation models are suitable for the simulation of coastal seawater environmental quality, which was shown by the significant relationship between modeling results and monitoring results （ r = 0. 643 - 0. 750, P 〈 0.01, n = 299）. Furthermore, based on integrative analysis of correlation coefficient, mean error, mean absolute error and root mean square error, it shown the order as follows： IDW（p2） 〈 Spline-Tension 〈 IDW（p3 ） 〈 IDW（p4） 〈 IDW （pl） 〈 IDW （p5） 〈 IDW （p6） 〈 Kriging-Guassian 〈 Kriging-Spherical, Kriging-Circular 〈 Kriging-Linear 〈 Kriging- Exponential 〈 Spline-Regularized. Thus the Inverse Distance Weighted（p2） model is the best one in precisions for simulating China coastal seawater environmental quality.

关 键 词：近岸海域 海水水质 空间插值 精度 

分 类 号：X834[环境科学与工程—环境工程]