逐日气象要素空间插值方法的比较3姜晓剑刘小军黄芬姜海燕曹卫星朱艳33(南京农业大学江苏省信息农业高技术研究重点实验室,南京210095)摘要采用距离反比权重法(IDW)、协克里格法(CK)和薄盘样条法(TPS)3种不同的空间插值方法,对我国1951—2005年气象数据完整的559个气象站点逐月第15日的平均基本气象要素(最高气温、最低气温、日照时数和降水量)进行了插值分析与评价.结果表明:3种插值方法中,TPS法对最高气温和最低气温插值的根均方差(RMSE)最小(1102℃和1112℃)、R2最大(019916和019913);不同季节中,TPS法对秋季最高气温、夏季最低气温进行插值的RMSE均最小(0183℃、0186℃),R2均为秋季最高.对于日照时数和降水量而言,TPS法的RMSE最小(0159h和1101mm)、R2最大(019118和018135);不同季节中,TPS法对冬季日照时数进行插值的RMSE最小(0149h)、R2最大(019293),TPS法对冬季降水量进行插值的RMSE最小(0133mm),IDW法对夏季降水量进行插值的RMSE最小(2101mm),CK法对春季降水量进行插值的R2最大(018781).TPS法可作为我国大量逐日基本气象要素的最优空间插值方法.关键词逐日气象要素空间插值方法栅格化文章编号1001-9332(2010)03-0624-07中图分类号O245;S161文献标识码AComparisonofspatialinterpolationmethodsfordailymeteorologicalelements.JIANGXiao2jian,LIUXiao2jun,HUANGFen,JIANGHai2yan,CAOWei2xing,ZHUYan(JiangsuProvinceKeyLaboratoryforInformationAgriculture,NanjingAgriculturalUniversity,Nanjing210095,Chi2na).2Chin.J.Appl.Ecol.,2010,21(3):624-630.Abstract:Acomparativestudywasmadetoevaluatethemethodsofinversedistanceweighting(IDW),co2kriging(CK),andthinplatespline(TPS)ininterpolatingtheaveragemeteorologicalelements(includingmaximumairtemperature,minimumairtemperature,sunshinehours,andpre2cipitation)ofthe15thdaypermonthfromthe1951-2005comprehensiveobservationdataof559meteorologicalstationsinChina.TheresultsshowedthattheRMSEsforthemaximumandminimumairtemperatureinayearinterpolatedbyTPSwerethesmallest(1102℃and1112℃,respective2ly),andtheR2betweentheobservedandpredictedvalueswerethehighest(019916and019913,respectively),comparedwiththoseinterpolatedbyIDWandCK.Infourseasons,thesmallestRM2SEsforthemaximumandminimumairtemperatureinterpolatedbyTPSwereobservedinautumn(0183℃)andsummer(0186℃),respectively,andtheR2betweentheobservedandpredictedvaluesinterpolatedbyTPSwerehigherinautumnthaninotherseasons.TheRMSEsforthesun2shinehoursandprecipitationinayearinterpolatedbyTPSwerethesmallest(0159hand1101mm,respectively),andtheR2betweentheobservedandpredictedvalueswerethehighest(019118and018135,respectively),comparedwiththoseinterpolatedbyIDWandCK.Infourseasons,theRMSEforthesunshinehoursinwinterinterpolatedbyTPSwasthesmallest(0149h),andtheR2betweentheobservedandpredictedsunshinehourswasthesmallest(019293).TheRMSEfortheprecipitationinwinterinterpolatedbyTPSwasthesmallest(0133mm),whiletheRMSEfortheprecipitationinsummerinterpolatedbyIDWwasthesmallest(2101mm).TheR2betweentheobservedandpredictedprecipitationinwinterinterpolatedbyCKwasthehighest(018781).ItwassuggestedthatTPScouldbetheoptimalspatialinterpolationmethodininterpolatingandrasterizingthedailymeteorologicalelementsinChina.Keywords:dailymeteorologicalelements;spatialinterpolationmethod;rasterization.3教育部新世纪优秀人才支持计划项目(NCET20820797)、国家自然科学基金项目(30871448)、国家高技术研究发展计划项目(2006AA10Z219)、国家重点基础研究发展规划项目(2009CB118608)和江苏省自然科学基金项目(BK2008330)资助.33通讯作者.E2mail:yanzhu@njau.edu.cn2009209214收稿,2010201204接受.应用生态学报2010年3月第21卷第3期ChineseJournalofAppliedEcology,Mar.2010,21(...
VIP
