简介：Background:Overthelastdecadesinteresthasgrownonhowclimatechangeimpactsforestresources.However,oneofthemainconstraintsisthatmeteorologicalstationsarefiddledwithmissingclimaticdata.Thisstudycomparedfiveapproachesforestimatingmonthlyprecipitationrecords:inversedistanceweighting(IDW),amodificationofIDWthatincludeselevationdifferencesbetweentargetandneighboringstations(IDWm),correlationcoefficientweighting(CCW),multiplelinearregression(MLR)andartificialneuralnetworks(ANN).Methods:Acompleteseriesofmonthlyprecipitationrecords(199.5-2012)fromtwentymeteorologicalstationslocatedincentralChilewereused.Twotargetstationswereselectedandtheirneighboringstations,locatedwithinaradiusof25km(3stations)and50km(9stations),wereidentified.Cross-validationwasusedforevaluatingtheaccuracyoftheestimationapproaches.Theperformanceandpredictivecapabilityoftheapproacheswereevaluatedusingtheratiooftherootmeansquareerrortothestandarddeviationofmeasureddata(RSR),thepercentbias(PBIAS),andtheNash-Sutcliffeefficiency(NSE).Fortestingthemainandinteractiveeffectsoftheradiusofinfluenceandestimationapproaches,atwo-levelfactorialdesignconsideringthetargetstationastheblockingfactorwasused.Results:ANNandMLRshowedthebeststatisticsforallthestationsandradiusofinfluence.However,theseapproacheswerenotsignificantlydifferentwithIDWm.InclusionofelevationdifferencesintoIDWsignificantlyimprovedIDWmestimates.Intermsofprecision,similarestimateswereobtainedwhenapplyingANN,MLRorIDWm,andtheradiusofinfluencehadasignificantinfluenceontheirestimates,weconcludethatestimatesbasedonnineneighboringstationslocatedwithinaradiusof50kmareneededforcompletingmissingmonthlyprecipitationdatainregionswithcomplextopography.Conclusions:ItisconcludedthatapproachesbasedonANN,MLRandIDWmhadthebestperformanceintwosectorslocatedinso