(IJCSIS) International Journal of Computer Science and Information Security, Vol. XXX, No. XXX, 2010 An Adaptive Neuro-Fuzzy Inference System based on Vorticity and Divergence for Rainfall forecasting Kavita Pabreja Research Scholar, Birla Institute of Technology and Science, Pilani, Rajasthan, India Assistant Professor, Maharaja Surajmal Institute (an affiliate of GGSIP University), New Delhi, India kavita_pabreja@rediffmail.com Abstract— A new rainfall forecasting model based on Adaptive Neuro-Fuzzy Inference System is proposed in this paper. A neuro-fuzzy model inherits the interpretability of fuzzy models and learning capability of neural networks in a single system. It has got wide acceptance for modelling many real world problems because it provides a systematic and directed approach for model building and gives the best possible design parameters in minimum time. The datasets used in this paper for the training of Adaptive Neuro-Fuzzy Inference System (ANFIS) are the European Center for Medium-range Weather Forecasting (ECMWF) model output products and the gridded rainfall datasets, provided by Indian Meteorological Department (IMD). To determine the characteristics of ANFIS that best suited the target rainfall forecasting system, several ANFIS models were trained, tested and compared. Different training and checking data, type and number of membership functions and techniques to generate the initial Fuzzy Inference Systems were analyzed. Comparisons of the different models were performed and the results showed that the model generated by grid partitioning using gbellmf membership functions provided the smallest errors for rainfall forecasting. Keywords- NWP model forecast, ECMWF model, rainfall, vorticity, divergence, ANFIS I. INTRODUCTION Weather is not just an environmental issue; it is a major economic factor. Economic value of weather for Agriculture, Fishery, Energy, Transportation, Aviation and health area is immeasurable. With its huge and growing population and low- lying coastline and an economy that is closely tied to its natural resource base, India is considerably sensitive to weather and climate. One failure of monsoon can totally upset the economic performance of our country. But timely forecasting can help to considerably minimize the adverse effect. Analysis and forecast of weather data created through Numerical Weather Prediction (NWP) models offers an unprecedented opportunity for predicting weather events, provide information and warning of extreme weather events for minimizing losses both to human and property. Such data consists of a sequence of global snapshots of the Earth, typically available at various spatial and temporal intervals including atmospheric parameters over land and ocean (such as temperature, pressure, wind speed, wind direction, sea surface temperature, etc.). The NWP models do not produce forecast of rainfall directly. Forecast of weather elements like rain/snow, sky conditions etc. at a place are derived through statistical relation popularly known as Model Output Statistics (MOS) proposed by National Weather Service [1]. General experience is that MOS products show improved skills over the raw model output. Basis of MOS is statistical relationship which requires long term consistent series of NWP products. Since NWP models get upgraded regularly[2], the series does not remain consistent. In view of above limitation of MOS, it has been proposed to explore other Intelligent techniques like ANFIS so as to forecast rainfall based on NWP model output products. In past, Artificial Neural Networks (ANN) has been applied [3] to predict the average rainfall over India during summer- monsoon i.e. the months of June, July, and August, by exploring the rainfall data corresponding to the summer monsoon months of years 1871-1999. It has been found that the prediction error in case of ANN is 10.2% whereas the prediction error in the case of persistence forecast is 18.3%. A neural network, using input from the Eta Model and upper air soundings, has been developed [4] for the probability of precipitation (PoP) and quantitative precipitation forecast (QPF) for the Dallas–Fort Worth, Texas, area. Forecasts from two years were verified against a network of 36 rain gauges. The resulting forecasts were remarkably sharp, with over 70% of the PoP forecasts being less than 5% or greater than 95%. Of the 436 days with forecasts of less than 5% PoP, no rain occurred on 435 days. Of the 111 days with forecasts of greater than 95% PoP, rain always occurred. The application of ANFIS for forecasting of meteorological parameters is very rare and particularly rainfall forecasting has not been