Thedrivingfactorsofrunoffchangescanbepidedintoprecipitationfactorandnon-precipitationfactor,andtheycanalsobepidedintonaturalfactorandhumanactivityfactor.Inthispaper,thewaysandmethodsofthesedrivingfactorsimpactingonrunoffchangesareanalyzedatfirst,andthenaccordingtotherelationshipbetweenprecipitationandrunoff,theanalyticalmethodaboutimpactsofprecipitationandnon-precipitationfactorsonbasin'snaturalrunoffisderived.Theamountandcontributionratesofthetwofactorsimpactingonnaturalrunoffbetweeneverytwoadjacentdecadesduring1956-1998arecalculatedintheYellowRiverBasin(YRB).Theresultsshowthattheamountandcontributionrateofthetwofactorsimpactingonnaturalrunoffaredifferentindifferentperiodsandregions.FortheYRB,thenon-precipitationimpactispreponderantfornaturalrunoffreductionafterthe1970s.Finally,bychoosingmainfactorsimpactingonthenaturalrunoff,oneerrorback-propagation(BP)artificialneuralnetwork(ANN)modelhasbeensetup,andtheimpactofhumanactivitiesonnaturalrunoffreductionintheYRBissimulated.Theresultshowsthatthehumanbackgroundof1956-1979.
