کاربرد توابع مفصل در تحلیل خشک‌سالی هواشناسی در شهرقائن

خشک‌سالی یکی از خسارت‌بارترین وقایع طبیعی هست که به دلیل حکم‌فرما بودن آن در یک حیطه جغرافیائی و طولانی بودن بایستی موردتوجه برنامه‌ریزان منابع آب قرار گیرد. دو مشخصه بارز که در بررسی خشک‌سالی‌ بایستی موردتوجه باشند، شدت و دوام آن هست. در این تحقیق از توابع مفصل دوبعدی برای ایجاد توزیع‌های دومتغیره از شدت و مدت خشک‌سالی در ایستگاه هواشناسی قائن طی سال‌های2018-1988 استفاده شد. برای تعیین شدت و مدت خشک‌سالی از سری زمانی شاخص بارش استانداردشده(spi) 12‌ماهه استفاده گردید. ضریب همبستگی تاو-کندال و اسپیرمن برای تعیین معنی‌دار بودن همبستگی بین شدت و مدت خشک‌سالی مورداستفاده قرار گرفت که مقدار آن‌ها به ترتیب 0/74 و0/88 تعیین گردید. نتایج بررسی توزیع احتمال‌های حاشیه‌ای نشان داد که مدت و شدت خشک‌سالی به ترتیب از توزیع احتمال پرتو و لوگ نرمال پیروی می‌نمایند که پارامترهای این توزیع‌ها مشخص گردیدند. همچنین از میان توابع مفصل موردبررسی، تابع مفصل جو با داشتن مقدار rmse برابر 0/0672 و شاخص nse برابر 0/99 مناسب‌ترین تابع برای تحلیل دومتغیره شدت و مدت خشک‌سالی می‌باشد و مقدار حداکثر لگاریتم درست نمائی آن7/87 به دست آمد. درنهایت احتمال وقوع دومتغیره توام و احتمال دوره‌های بازگشت توام محاسبه شدند که نتایج حاصله می‌تواند اطلاعات ارزشمندی را برای برنامه‌ریزی و مدیریت یکپارچه منابع آب منطقه مورد مطالعه، در اختیار برنامه‌ریزان و ذی‌نفعان قراردهد.

application of copula functions in meteorological drought analysis in qaen area

introductionconsidering that the severity and duration of drought are probabilis tic phenomena, it is more appropriate to analyze those using probabilis tic theories. the use of classical multivariate functions is the usual method for analyzing such problems. the use of detailed functions provides the possibility of linking the univariate marginal dis tribution to each other and creates the bivariate dis tribution function. detailed functions have been used in various research to evaluate the effects of drought and to model the relationship between intensity and duration of drought, the mos t important of which have been mentioned in this research. in this research, the main characteris tics of meteorological drought, including the severity and duration of drought in qaen synoptic s tation, have been inves tigated using bivariate joint functions. materials and methodsin this research, the rainfall data of the qaen observation s tation during the years 1998-2018 was used on a monthly scale. in this research, spi index and 12-month intervals were used to determine drought, which was determined using observational rainfall data. therefore, to check the correlation between the severity and duration of meteorological drought in synoptic qaen s tation, tau-kendall and spearman-roe correlation criteria were used, which were calculated in r s tatis tical software. in this research, 21 copula functions were examined to choose the bes t one. in this research, in order to determine the marginal probability dis tribution functions, the mle method was used in the es timation of the detailed dependence parameter.results and discussionconsidering that the severity and duration of drought are the discussed variables in this s tudy, therefore, after determining the spi meteorological drought index values, the run-tes t was used to determine the severity and duration of drought. the results showed that the drought intensity of 6.66 with a duration of 8 months was the mos t severe meteorological drought that happened in the period of 1998-2018. the results of the tau-kendall and spearman correlation tes ts also showed the correlation values between the intensity and duration of drought, respectively, 0.74 and 0.88, which indicates the isotropic effect of these variables. the results of determining the marginal dis tributions of drought intensity and duration variables showed that pareto marginal probability function and log normal are the mos t suitable probability dis tributions for drought duration and intensity characteris tics in qaen synoptic s tation, respectively. the values of the parameters of these dis tributions were a=0.95, b=1.00 and μ=0.39, σ =1.35 and γ=0.015. the results of the analysis of the appropriate joint function showed that the search function showed the bes t results of goodness of fit based on the aic index, and the value of this index was determined as -13.74, and the parameter value of the search function also, 84 times was obtained. also, the maximum accuracy value was determined as 7.87.