Fuzzy rule-based prediction of monthly precipitation

Monthly precipitation in Hungary is modeled using the Hess-Brezowsky atmosp heric circulation pattern types and an ENSO index as forcing functions or i nputs. The weakness of the statistical dependence between these individual inputs and precipitation prevents the use of a multivariate regression anal ysis for reproducing the probability distribution function of observed prec ipitation. In order to utilize the existing relationship between forcing fu nctions and precipitation a fuzzy rule-based modeling technique is used. Th e first part of the observed input and precipitation data is used as the le arning set to identify the fuzzy rules. Then, the second part of the data i s used to validate the rules by comparing the frequency distributions of pr ecipitation calculated respectively with the fuzzy rules and observed data. Example results are presented for two different climatic regions of Hungar y. One of them represents a wetter climate while the other refers to the dr ier conditions of the Hungarian Plains. The fuzzy rule-based model reproduc es the empirical frequency distributions in every season. However, as expec ted, the statistical prediction is better in winter, spring and fall than i n the summer. The potential of the approach is important in climate change studies when the fuzzy rules obtained as described above can be used with i nput data stemming from GCM to predict regional/local precipitation reflect ing GCM scenarios. (C) 2001 Elsevier Science Ltd. All rights reserved.