Precipitation Classification of Northwest of Iran Using Cluster Analysis and Discriminate Analysis

IntroductionPlanning and management of water resources in Iran are affected by tempo-spatial distribution of rainfall. Therefore, study of precipitation can be a vital variable for evaluating water resources in different parts of Iran. Since climatic data are increasing in bulk study the individual places takes, therefore using of multivariate regionalization methods are inevitable. Recently, multivariate methods of cluster analysis has been used widely in climatic regionalization and it provides desirable results. Another method of multivariate is polynomial analysis that is less frequently used. In the latter if a number of classes are used the result of regionalization can be tested. In present study precipitation of Northwest of Iran is classified using eight characteristics of precipitation using cluster analysis and were tested by discriminant analysis methods.Materials and MethodsFor this study rainfall profile of 260 stations Northwest are analyzed. After arranging the required data, 3360 monthly maps from elements have been made by the kriging interpolation method, after removal of outside cells, 5374 cells remain within the study area. Data matrix for the regionalization was obtained by the mentioned maps in 92 column and 5374 row. Cluster analysis used for classification. So that the distance of the cell obtained by Euclidean distance coefficient and precipitation groups linked by 'Ward' linkage method. To obtain the optimal number of classes precipitation average method and the comparison process of elements have been used. Finally, four groups of precipitation were obtained in Northwest of Iran.In order to test the accuracy of the results of cluster analysis, discriminant analysis was used. The results of discriminant analysis accepted qualifications of cluster analysis in classification. So that results obtained from discriminant analysis in 97.6 percent of the regional area conform to the results of cluster analysis. Results and DiscussionThe results indicate the interaction between tropical and polar jet streams caused cyclogenesis in the North East of the Mediterranean Sea and also they determined that the anomaly of mean sea level pressure, 500-hpa geo-potential height and sea surface temperature can be used as a tool for predicting extreme precipitation. In the case with heavy precipitation, Mean sea level pressure and 500-hpa geo-potential height, respectively 10 mb and 140 meters are less than the average. In the light precipitation case, these records are 5 and 60 respectively. Anomaly of sea surface temperature in extreme precipitation shows that the Mediterranean Sea, Black Sea, Oman Sea and Indian Ocean are much warmer than average, which causes more evaporation and proper injection of moisture to the region.ConclusionTo determine the structure of synoptically pattern of extreme precipitation in North West of Iran, some meteorological parameters such as mean sea level pressure, 500-hpa geo-potential height, 300-hpa jet stream and sea surface temperature, caused heavy rainfall in March in North West of Iran, are studied in this case study. Since the fluctuations are more revealed in maximum and minimum precipitation conditions, the extreme conditions are investigated. 14th March 2009 and 4th March 2005 are chosen as a maximum and minimum precipitation respectively. The data of NCEP / NCAR were the base for reviewing synoptically parameters and the AVHRR data were used to monitor the sea level temperature. The results show that the interaction between tropical and polar jet streams are generating cyclogenesis in the North East of the Mediterranean Sea and also they determined that the anomaly of mean sea level pressure, 500-hpa geo-potential height and sea surface temperature can be used as a tool for predicting extreme precipitation.