Mathematical Model of Rainfall Forecast

Rainfall is one of the most common climate phenomena. How to make the forecasts as accurate as possible is an extremely difficult problem. In this paper, we set up a statistical model using the actual data from 91 site locations. We present a distribution map of the 91 sites and 2491 grid nodes according to their latitude and longitude. First we calculate the predicted data of these 91 site locations by applying the binary cubic interpolation method. Combining the actual data, we carry out the fitting analysis by an optimal fitting method. We can find that the residual error from this fitting method between the predicted data and the actual data is small. After analyzing the quality of the result we present an appraisal of our rainfall forecast method.