عنوان مقاله [English]
Background and Objectives: Occurrence of extreme rainfalls specifically in short-time scales causes heavy damages to human communities, municipal crowd regions and natural ecosystems. Studying and precise identifying of extreme rainfalls is essential and crucial in different agriculture and natural resources, meteorology and hydrology, engineering and natural environment aspects. The damage intensity of extreme rainfalls does not equally act in different regions and it is essential to assess the risk extent of such hazardeous rainfalls in regions with different climatic conditions. Therefore, the aim of this study is to identify and zoning of extreme rainfall occurrence risk in different 6, 12 and 24-hour time scales for West parts of Iran, which has a noticeable diversity in terms of the climatic and topographic conditions.
Materials and Methods: to perform this research, a number of 27 synoptic stations located in five provinces including Kurdistan, Kermanshah, Hamedan, Ilam and Lorestan were selected and the maximum annual values of extreme rainfalls in three time scales of 6, 12 and 24-hours in a 25 year time period (1992-2016) were extracted and by fitting different statistically distributions to each of these time series and by adopting the Chi-square test, the statistical distributions with best fit were regoized and were used to performing propabilitistic analyses. The 30, 40 and 50 mm thresholds were used to defining the torrential extreme rainfalls in 6, 12 and 24-hours time scales, respectively and after calculating the risk extents of the mentioned torrential extreme rainfalls for all of the studied stations, the risk zoning of torrential extreme rainfalls occurrence was performed by applying the multiple linear regression models between the risk extents and geographical properties (longitude, latitude and elevation) for all of the studied region. To enhancing the models accuracy, the long-term average of the number of days per year with precipitation greater than 1 mm was employed in the structure of the regression models as auxiliary variable in some cases and to achieve higher accuracy of regression models, the studied region was divided into three distinct regions.
Results: The results showed that among different fitted statistical distributions to the time series of extreme rainfalls of 6, 12 and 24-hours in the studied region, three distributions including Log-Logistic, Pearson and Gama were recognized as the best fit distributions. In terms of the accuracy of the multiple linear regression models, the results showed the high accuracy of these models for all of the three distinct regions and whole of three time scales of 6, 12 and 24-hours. The overall results of this research showed that the risk occurrence of the torrential extreme rainfalls in West of Iran has a notable diversity so that this risk is very low in some centarl parts and very high in some west and south parts of the studied region.
Conclusion: The overall results of this research revealed that the general applied algorithm of this research to estimating spatial distribution of torrential extreme rainfalls ocuurence risk was led to obtaining the appropriate and acceptable accuracy in regional estimating and generalizing the stational point results to the regional scale. Therefore, it is essential to adopt appropriate tasks and more attention in contrast to the negative consequences of the extreme rainfalls in the parts with higher degree of risk occurence.