Applying Association Rules Methods for Drought and Rainfall Monitoring Using the Sea Surface Temperature-A case study of Khoozestan Province

Iran encounters climatic variations due locating in 25-40 North latitudes. If atmospheric events like rain، drought and flood predicts correctly it would effects a lot on natural hazard control and water resources and agriculture planning. One of the most important factors in climatic variation is the sea surface temperature (SST) which is effective on the rain amount. In this research the relation between the Persian Gulf and Red sea’s SST and the fall and winter rain amount of three synoptic stations in Khuzestan province in Iran during years 1965- 2010 is detected by means of a data mining technique called association rules. Data mining is an advanced tool for organizations to do research on databases and detect the hidden patterns and procedure of events and clarify their future manners، and make better decisions. In this study drought and wet year events are defined according to Standardized Precipitation Index (SPI). The coincident and individual effects of Persian Gulf and Red sea’s SST on these events are analyzed using a time series data mining algorithm called ARSS. The new algorithm (ARSS) is developed to find the relations between atmospheric and climatic indices in a different manner from which in classic hydrology statistics. Results show that coincident use of Persian Gulf and Red sea’s SST enables us to find more relations than the separate use of them. The ARSS algorithm can also predict the neutral condition with most confidence in future though its ability to predict the moderate drought and extreme wet conditions is mentionable.