摘要
以云南省1961-2010年多年平均降雨量和云南省数字高程模型为基础数据源,通过对降雨数据的分析,将数字高程模型作为影响降雨强度的主要因素,引入到降雨量空间预测的多元地理统计中,并利用协同克里格算法构建降雨强度与地形高程之间的数学关系模型,对云南省年平均降雨数据进行空间插值模拟。插值结果与距离权重法、一阶局部多项式法、二阶局部多项式法、样条函数法、普通克里格法等其他5种空间数据插值方法进行对比与分析,并通过分析各个交叉检验统计量来验证评估数据的误差,从而选取出最优的空间插值方法。结果显示,普通克里格法和协同克里格法对降雨量空间插值的效果和精度高于反距离法、局部多项式法和样条函数法,而协同克里格法由于考虑了地形高程对降雨量的影响,具有更好的插值效果,因而更适合山地降雨数据的空间插值。
Average rainfall of years 1961-2010 in Yunnan Province and digital elevation model are taken as basic data sources in this paper.Through the analysis of the rainfall data,digital elevation model as the main factors that influence rainfall intensity is introduced to the rainfall forecast of multiple geographic space in statistics,and collaborative kriging algorithm is used to construct a mathematical relationship model between rainfall intensity and terrain elevation.On that basis space interpolation simulating on average annual rainfall data in Yunnan is conducted.The results are analyzed and compared with those from distance-weighted method,first-order local polynomial method,second-order local polynomial method,splines method,and ordinary kriging method.The best spatial interpolation method is selected through the analysis of various cross test statistic to verify the error of the evaluation data.The result shows that the effect and precision of rainfall spatial interpolation of ordinary kriging and collaborative kriging are higher than those of inverse distance method,local polynomial method and splines method.Collaborative kriging has better interpolation results because of considering the influence of terrain elevation on rainfall,Therefore,collaborative kriging is more suitable for spatial interpolation of rainfall data in mountain area.
出处
《热带地理》
北大核心
2012年第4期370-377,共8页
Tropical Geography
基金
云南大学中青年骨干教师培养计划(编号:XT412003)
教育部"春晖计划"科研项目(编号:Z2008-1-65001)
关键词
空间插值
协同克里格
降雨量
地形高程
云南省
交叉验证
spatial interpolation
collaborative kriging
rainfall
terrain elevation
Yunnan Province
cross validation
