Climatic Zoning of the Southern Coastline of the Caspian Sea Using Multivariate Statistical Methods

Abstract

AbstractIdentifying homogeneous climatic zones plays an important role in the success of regional development programs. The climate system is composed of various elements, factors, and variables that together form the climatic components of a region. Multi-characteristic and multivariate methods can combine and overlap the types of elements and variables effective in constructing the climate with appropriate weights in the climatic zoning area. In the present study, climatic zoning of the Caspian region was performed using factor analysis and cluster analysis. For this purpose, a 30*30 matrix consisting of 30 meteorological stations and 30 climatic and environmental variables was formed. The results of factor analysis showed that the climate of the region is affected by 5 factors including precipitation-humidity, temperature, wind, sunlight, and environmental factors. These factors explained a total of 92.5% of the variance of the data. Then, cluster analysis was performed by the hierarchical integration method of Ward on the five mentioned factors. The results showed four climatic zones including humid, semi-humid, semi-arid, and arid in the study area.Keywords: Climate Zoning, Factor Analysis, Multivariate Analysis, Cluster Analysis, Caspian Coastline, Iran. IntroductionKnowledge of climatic zones has long attracted the attention of many scientists and has led to the presentation of various methods of climatic classification such as De Marten, Koppen, Ivanov, Amberje, Selianinov, Hansen, and others. With significant computer advances in recent years, it has become possible to perform internal methods on large volumes of data and the use of new classification methods such as multivariate statistical methods (factor analysis and cluster analysis) have expanded to classify the interactions of a large number of climatic components. Identifying homogeneous climatic zones and the capabilities and limitations of the agricultural climate of each climatic zone can play an effective role in carrying out projects and planning. MethodologyThe study area in this research is the greenest and rainiest region of the country (i.e. the southern shores of the Caspian Sea). In this study, factor analysis with Varimax rotation has been used to identify the factors affecting the climate of the study area and the hierarchical clustering method has been used in its climatic zoning. For this purpose, out of 30 variables affecting agricultural activities, including three environmental variables of latitude, altitude, and distance from the sea, as well as 27 climatic variables including maximum, minimum, and average temperature, an average temperature of winter, spring, summer, and autumn were used. Number of days with a maximum temperature of 30 ° C and above and minimum temperature of 0 ° C and below, average sunny hours, number of full cloudy, partly cloudy, and sunny days, hours of radiation, average relative humidity and annual rainfall, average winter, spring, summer and autumn, total annual rainfall with more than 1 mm, number of days more than 1 mm, more than 5 mm, more than 10 mm and more than 20 mm, average evapotranspiration and average wind speed in 30 stations, and the synoptic meteorology of the region with a suitable statistical period between 2002 to 2018 were used on a daily time scale. ETO Calculator software and radiation amount were used using the Angstrom-Prescott function to calculate the reference evapotranspiration. DiscussionThe results of the Bartlett test showed that the data are suitable for factor analysis and the results can be generalized to the statistical population. The results also showed that the region's climate is the result of the interaction of 5 different factors and explains 92.5% of the total variance. Based on the results of factor scores of variables, variables of average annual rainfall, winter, spring, summer, and autumn, total annual rainfall on days with more than 1 mm, number of days with more than 1 and 5 mm and with heavy rainfall of more than 10 and 20 mm, the number of full and partly cloudy days and average relative humidity had the highest correlation coefficient with the first factor. Due to the fact that the naming of the factors is based on the highest values of correlation coefficients, it was named the precipitation-moisture factor. In the second factor, the variables of average minimum, maximum and average daily temperatures, average temperatures of winter, spring, summer, and autumn, and the number of days with a maximum temperature of 30 ° C and above had the highest factor load and weight. Therefore, the second factor was named the temperature factor. The third factor explains 6.5% of the total variance of the data and was named the wind factor, as the mean variable of wind speed had the highest correlation coefficient with this factor. The fourth factor explains only 5.8% of the variance of the data changes. Because the variable number of sunny days had the highest correlation coefficient with this factor, it was named the factor of sunshine. The fifth factor explains only 5.5% of the variance of the data changes. Because the variables of distance from the sea and latitude had the highest factor and weight in this factor, it was named an environmental factor.After performing factor analysis and identifying the main factors using the hierarchical clustering method by the Ward method, the studied stations are grouped into homogeneous categories and zones and climatic classification was performed. According to the cluster tree diagram obtained and the cutting location of the diagram at the interval of 8, 4 clusters were identified. According to the findings, four climatic zones including a humid climate zone located in the northern parts of Gilan province to the western and central plains of Mazandaran province, a semi-humid climate zone including the eastern and central parts of Mazandaran province to parts of the western and southern regions of Gilan province and western parts of Golestan province, semi-arid climate zone located in the southern parts of the southern shores of the Caspian Sea, and arid climate zone located in the eastern and northeastern parts of Golestan province for the region were identified. ConclusionThe output of this study was four climatic clusters for the study area, which is different from the study of Nazmafar and Goldoust (2015) who in their research on the zoning of the north and northwest of the country, identified three climatic zones for the northern region. In their study, the first climate zone with the effect of precipitation factor was located in the southwest of the Caspian Sea and the second climate zone with the effect of temperature factor was located in a part of the southern shores of the Caspian Sea and the northern slopes of Alborz Mountain range. Therefore, the present study has provided more specific and accurate climatic zones. The findings of this study are consistent with the findings of Montazeri and Bai (2012). They showed in their research that Mazandaran province was located in two humid and semi-cold climates with low rainfall, Gilan province was located in two humid and semi-humid regions, and Golestan province was located in the climate zones of humid, semi-humid, cold, low rainfall, semi-cold, and low rainfall. Also, the findings of this study were consistent with the findings of Fallah Ghaleri et al. (2015) in the field of climatic zoning in Gilan province. References- Biabiany, E., Bernard, D., Page, V., & Paugam-Moisy, H. (2020). Design of an expert distance metric for climate clustering: The case of rainfall in the Lesser Antilles. Journal of Computers and Geosciences, 145, 1-15.- Carvalho, M., Melo-Gonçalves, P., Teixeira, J., & Rocha, T. (2016). Regionalization of Europe based on a K-Means Cluster Analysis of the climate change of temperatures and precipitation. Journal of Physics and Chemistry of the Earth, 94, 22-28.- Schmidt, G. (2019). The Ecological relevance of parameter choice in describing climate. The Ecological relevance of parameter choice in describing climate, 6, 1-26.- Tapiador, F., Moreno, R., & Navarro, A. (2019). Consensus in climate classifications for present climate and global warming scenarios. Journal of Atmospheric Research, 216, 26-36.- Yang, L., Bai, L., Song, B., & Liu, N. (2020). A new approach to develop a climate classification for building energy efficiency addressing Chinese climate characteristics. Energy, 195, 1-14.- Zhao, J., Xia, B., Han, J., & Liang, K. (2020). Technological adaption zone of passive evaporative cooling of China, based on clustering analysis. Journal of Sustainable Cities and Society, 66, 1-10.

Keywords

• climate zoning
• factor analysis
• multivariate analysis
• cluster analysis
• caspian coastline
• iran