6-Stochastic modelling of rainfall at Udaipur

Abstract

Stochastic modelling for rainfall has done using 100 years (1901-2000) data. The performed statistical tests indicated that the series of the monthly rainfall data was trend free, The periodic component of monthly rainfall could be represented by third harmonic expression. The stochastic components of the monthly rainfall followed second order Markov model. Validation of generated monthly rainfall series was done by comparison of generated and measured series. The correlation coefficient between generated and measured rainfall series was found to be 0.9974. The correlation was tested by t-test and found to be highly significant at 1 per cent level. The regression equation is very near to 1:1 line. Therefore, developed model could be used for future prediction of monthly rainfall.