تحلیل و پهنه‌بندی خشکسالی اقلیمی و تاثیر soi و nao بر حوضه‌های ششگانه آبخیز ایران

روش های مختلفی برای درون یابی داده مکانی وجود دارد و هر گونه کاستی در انتخاب روش مناسب توزیع مکانی عوامل اقلیمی و بی توجهی به دقت روش های درون یابی می تواند موجب بروز خطا در برآوردهای طراحی گردد. در این پژوهش از داده‌های ماهانه بارندگی 79 ایستگاه سینوپتیک با حداقل 30 سال طول دوره آماری از 1988 تا سال 2017 استفاده گردید. پس از بررسی رژیم بارش در حوزه های آبخیز شش گانه ایران، مقدار spi در بازه‌های 3، 5 و 8 ماهه از داده‌های بارندگی ماهانه در هر سال استخراج و برای هر ایستگاه محاسبه و برای هریک از پدیده ها ( soiو nao) موردبررسی قرار گرفت و در سه گام زمانی ده ساله، سال‌هایی با خشک‌سالی‌های بسیار شدید در حوضه مشخص گردید. در تعیین توزیع بارندگی در پریودهای زمانی 10 ساله در نقاط فاقد آمار از شاخص های میانگین قدر مطلق خطای نسبی (mare)، میانگین خطای اریبی انحراف (mbe)، درصد میانگین خطای مطلق (mape) و ریشه دوم میانگین مربع خطا (rmse) استفاده گردید. در مجموع با استفاده از خطاهای فوق در دوره سه ماهه، پنج ماهه و 8 ماهه در بین روش های درون یابی قطعی، روش توابع پایه ای شعاعی (rbf) و درون یابی چندجمله ای محلی (lpi) نسبت به دو روش دیگر (idw,gpi)، (نتایج دوره 3 و 5 ماهه بالای 62 درصد بوده و 8 ماهه بالای 72 درصد) نتایج بهتری ارائه کرده است. در مورد روش های کریجینگ، روش کریجینگ ساده (درجه دوم منطقی) در درون یابی متغیرهای بارش میانگین، بارش با دوره 3 و 5 ماهه، روش کریجینگ ساده (نمایی) در درون یابی متغیرهای بارش میانگین بارش با دوره 8 ماهه بهترین روش زمین آمار بوده است.

analysis and zonation of drought and the impact of soi and nao on the six watersheds of iran

background and objectivesin order to study the spatial patterns and prepare spatial distribution maps, point data are generalized to the surface by moving away from the site in the interpolation process. the values ​​of indefinite points are estimated by using the measured values at definite points. there are different methods for spatial data interpolation and any major defect in the spatial distribution of climatic factors and slight inattention to the appropriate method of interpolation method can lead to errors in the design estimates. so far, many researches have been done in the field of climate data interpolation in the world. the aim of this paper is determining the most appropriate method of rainfall extension pattern and fitting the best distribution function to the annual average rainfall data. in addition to the performances of different interpolation methods, the choice of the optimal model for zoning rainfall amounts in different time periods is the goal of paper.methodologyiran is located in west asia and borders the caspian sea, persian gulf, and oman. with an area of 1,648,000 square kilometers (636,000 sq. mi), iran ranks seventeenth in size among the countries of the world. the study area is located between 32° 25’ 14.67 n and 53° 40’ 58.86 e. iran has a variable and continental climate in which most of the relatively scant annual precipitation falls from october through april. the average elevation of this plateau is about 900 meters (2,953 ft.), but several of the mountains that tower over the plateau exceed 3,000 meters (9,843 ft.). volcanic mount damavand, 5,610 meters (18,406 ft.), located in the center of the alborz, is not only the country’s highest peak but also the highest mountain on the eurasian landmass west of the hindu kush .the monthly precipitation data from 79 synoptic stations with at least 30 years during the period from 1988 to 2017 were used. very severe drought years and spi values identified in the intervals of 3, 5 and 8 months until the end of may were extracted from monthly rainfall data in each year and in three ten-year steps. in order to determine the distribution of rainfall in 10-year time periods in the lack of data points, the indices of mean absolute value of relative error (mare), mean bias error (mbe), mean absolute percent error (mape) and the root mean squared error (rmse) has been used.findingsin general, using the above errors in the quarterly, five-month and eight-month periods between definitive interpolation methods, radial basis functions (rbf) and local polynomial interpolation (lpi) compared to the other two methods (idw, gpi) has provided proper results that is consistent with the results of misaghi (2006) and eivazy (2012). regarding the method of radial functions, the quadratic interpolation method had the lowest correlation for all variables except rainfall with a period of 3 months. kriging methods have given better results compared to idw method, which is consistent with the results of soleimani et al. (2004) and zabihi et al. (2011). the methods of kriging, simple kriging method (second-order logic) in the interpolation parameters of mean rainfall, (3 and 5 months), simple kriging method (exponential) in the interpolation variables precipitation with a period of 8 months were the best methods. therefore, kriging method has presented relatively more accurate results, which is consistent with the results of mirmousavi et al. (2010).conclusionthe results showed that the most droughts occurred in the second decade of the third decade, the trend is still less severe drought is occurring and the lowest is in the first decade. in order to study the effect of soi and nao on drought and wet years in the country, the correlation matrix was determined to investigate the drought and wet periods. then, geostatistical modelings were performed to describe these effects. finally, based on the study of stations and zoning models, it can be stated that the lanina phenomenon has a little effect on drought and annual rainfall and the el nino phenomenon has a relatively greater effect. the results show that gpi and idw methods are not considered as good performance. these methods are based on the fit of polynomial functions on spatial data. according to the results obtained in every area and for different parameters, a single method cannot be selected as the most appropriate method and the most appropriate method can be selected. based on the results, the most appropriate choice depends on the characteristics of the regions and a single method cannot be selected for all regions. the results show that gpi and idw methods are not suitable and modeling in these two methods is based on fitting polynomial functions to spatial data. rbf and lpi methods have provided better output rather than other methods because the results of the 3 and 5 month periods were above 62% and the 8 months were above 72%.