A Weather Forecasting System using concept of Soft Computing: A new approach

Weather forecasting and warnings are the major services provided by the meteorological profession.. Many government and private agencies are working on its behavior but still it is challenging and incomplete. We propose a new technique to construct the learning set of images, which represents the actual data. We relate this data to the forthcoming weather events based on their previous records and history or whatever recognized by our system. Our work presents a new approach where the data explanation is performed with soft computing technique i.e. a neuro-fuzzy system is used to predict meteorological position on the basis of measurements by a weather system designed by us. This model can help us in making forecast of different weather conditions like rain and thunderstorm, sunshine and dry day, and perhaps a cloudy weather i.e. purpose of this model is to represent a warning system for likely adverse conditions. Our model is designed with a concept of Adaptive Forecasting Model(AFD) in the mind. The simple meaning of this term is that our model has potential to capture the complex relationships between many factors that contribute to certain weather conditions. The results are compared with actual working of meteorological department and these results confirm that our model which is based on soft computing have the potential for successful application to weather forecasting. We considered atmospheric pressure a primary key parameter and atmospheric temperature and relative humidity secondary type. We will, however, examine temperature as signature of weather conditions in some typical cases where we could observe the effect of temperature. As an example, the estimations produced by the proposed methodology were applied on different weather forecasting data provided by the Gwalior meteorology center to make the result more practical and believable. The forecasts are always current are based on neuro-fuzzy[2] systems that utilize the concept of soft computing. Real - time processing of weather data indicate that the neuro-fuzzy based weather forecast have shown improvement not only over guidance forecasts from numerical models, but over official local weather service forecasts as well.Adaptive Forecasting Model(AFD)