Abstract. The highly nonlinear relationships governing climate phenomena, suggest an application of the nonlinear time series analysis. We examine the presence of nonlinearity in the climate system using surrogate data. This paper aims to investigate the existence of chaos in the time series of Lahore precipitation. The predictability of daily rainfall is the most difficult task because of the nonlinear behavior of climate system. This paper employs radial basis function and neural network for constructing the nonlinear model of summer monsoon precipitation of Lahore. We also compare both of the algorithms with respect to prediction accuracy and storage requirements.

__________________ * : Also, Department of Mathematics, University of Karachi