Evaluation of the Decision Tree Model in Precipitation Prediction (Case study: Yazd Synoptic Station)

Undesirable effects of droughts on the agricultural and economical sectors and especially on the natural resources are intense. Different methods have been presented to predict the main factors of drought such as precipitation and during the recent decades some new computer based models have been developed for drought prediction. In most cases these models have presented quite satisfactory results. Decision tree, as one of these models, produces rules based on evaluation of the parameters from portion (component) to the whole, and finally reaches understandable knowledge from the existing statistical data. In this research, decision tree model has been used as a data mining method to predict precipitation and evaluation of drought in Yazd synoptic meteorological station. Simulations were carried out in four different conditions. Related variables including previous monthly precipitation, mean temperature, maximum temperature, humidity, wind speed, wind direction, and evaporation were used as independent input variables for all these four conditions and the amount of precipitation was predicted 12 months in advance. Finally for evaluation of the model performance in different conditions, statistical criteria were employed. Results indicated that the decision tree model is able to presents suitable prediction of precipitation especially when 5-year moving average of data is used. Precise prediction of precipitation and the accurate evaluation of drought conditions are of great importance for a better management and planning for drought damages reduction.