Based on themonthly precipitation datameasured by 4 representativemeteorological stations in Liupanmountain area during the period of1960-2007, the spatial and temporaldistribution ofprecipitationwas analyzed by using some statisticalmethods such as regression, periods and abrupt change analysis. The seasona,l annual trend ofprecipitation there and the primary cycles ofannualpre cipitation were revealed. Results show that annualprecipitationwas decreasing and the primary cycleswere fourteen and eleven years,and abrupt change occurredmany times in annual precipitation in this period.

Developed  from  linear  StatlShCa1 anB}yS1S}  non-linear  StatlShCa1 fnreeaSt  meth0(1S  f)eenme  One  aSpeCt Of  StatlShCal  method. In
TCCent yPaCS}  non-linear  StatlShCa1  fnreeaSt  method  Ina}PS great pTOgTCSS  f)Llt SyStematlC  1ntT0(1LIChOn  1S  Still  laCl}. In  tI11S  hahPT}  Study TCSLI1tS  and  appllCatlOn  eXampleS  Stlldled  by  aLIthOT  In  TCCent  yeaTS  aCe  Slnnmed  Llp.  The  method  Of  paCanleteT  CalClllatlng  and  the StatlShCal  fnreeaSt  m0(le1S  Of  Single  independent  VaClaf)le  aCe  S110Wn  In  tI11S  paper.  1Vlany  fnreeaSt  meth0(1S  aCe  1ntT0(1LICP(1}  SLICK  aS fitting 11nCaT}  half-linear  and  CLITVe-linear  fnreeaSt  n10(1C1S}  SLIhSCChOn  appTOaCh  m0(lel  and  analog  fnreeaSt  method.  It a1S0  S110WS  the example index of non-linear statistical forecast models.