MODEL PRAKIRAAN CURAH HUJAN BULANAN DI WILAYAH JAWA BAGIAN UTARA DENGAN PREDIKTOR SUHU MUKA LAUT (SML) DAN OUTGOING LONGWAVE RADIATION (OLR)

Abstract

Aims of this research is to analysis the influence of sea surface temperature (SST)) and Outgoing Long wave Radiation (OLR) on monthly rainfall in northern Java Island. The rainfall data are from Rainfall Type (ZOM) 30, ZOM 43, ZOM 88, and ZOM 90 as defined by Meteorology, Climatology and Geophysics Agency (MCGA). The monthly rainfall of each region were calculated by Pearson coefficient ( r) to be correlated to SST resolution 1º X 1º and OLR resolution 2.5 º X 2.5º at area domain 5º N - 20º S and 90º E – 150º E and over period 1979 – 2007. The result indicates significant correlations between the monthly rainfall and SST and OLR. Artificial Neural Network (ANN) was applied to predict monthly rainfall over the four ZOM using input SST and OLR selected base on the correlation result. The validation of ANN model was done by comparing output of the monthly predicted rainfall to its observation over period 2003 – 2007. It is found out that the output model pattern is reasonably its consistent to its observation. The value of RMSE is smallest in 2006. The evaluation result using bias indicates that the biggest error occurred during dry season