تغییر توزیع فراوانی بارش‌های فرین شهر زنجان

نوع مقاله: مقاله پژوهشی

نویسنده

چکیده

  تغییر عناصر اقلیمی به اشکال مختلفی قابل ردیابی است. در بسیاری مواقع این تغییرات نهان و از طریق مطالعه روند، آشکار نمی‌شوند. در بسیاری مواقع تغییر عناصر اقلیمی در شکل توزیع فراوانی و عمدتاً در دنباله توزیع فراوانی (فرین‌ها) نمایان می‌شوند. در گذشته ردیابی تغییرات از طریق مطالعه تغییر مرکز توزیع فراوانی (میانگین) صورت می‌گرفت اما اخیراً توجه دانشمندان به دنبالۀ شکل توزیع فراوانی عناصر اقلیمی معطوف شده است. بنابراین بررسی شکل توزیع فراوانی عناصر اقلیمی، نکات بسیار مهمی در خصوص رفتار بلند مدت این عناصر نشان می‌دهد. شناخت رفتار عناصر اقلیمی براساس شکل توزیع فراوانی آن‌ها به لحاظ علمی و عملی از اهمیت شایان توجهی برخوردار است. دراین تحقیق با استفاده از مشاهدات روزانه چهار نمایه فرین بارش زنجان ( شامل بیشینه بارش، پنج بارش بزرگ، صدک پنجم و نودو پنجم) و نیز سهم آن‌ها در بارش سالانه طی دوره آماری 2006-1961 و با استفاده از توزیع فرین تعمیم یافته ( GEV )، تغییرات بارش‌های فرین در معرض تحلیل قرار گرفت. هرچند تحلیل روند حاکی از ایستایی فرین‌ها بود اما معلوم شد که چند نوع ناایستایی در مشاهدات رخ داده است. این ناایستایی‌ها به صورت چرخه‌ها، جهش و نیز تغییر توزیع فراوانی قابل رویت است. برای مثال مجموع بارش‌های حاصل از پنج بارش بزرگ حاوی چرخه‌های حدوداً 3 ساله هستند. از سال 1980 به بعد یک تغییر سطح کاهشی در مجموع بارش‌های حاصل از صدک پنجم و سهم آن‌ها در بارش سالانه رخ داده است. شکل توزیع فراوانی بارش‌های حاصل از صدک پنجم در نیمه اول و دوم باهم متفاوت بوده و در نیمه اول از توزیع چوله به چپ به توزیع چوله به راست در نیمه دوم تغییر یافته است. این تغییر توام با کاهش میانگین و چولگی این توزیع بوده است. همچنین تغییرات حادث شده بر صدک پنجم بارش با شدت کم تری در بارش‌های حاصل از صدک نود و پنجم قابل مشاهده است. با تمامی این اوصاف بارش بیشینه و پنج بارش بزرگ از ایستایی بیش تری نسبت به شاخص‌های دیگر برخوردارند. بنابراین می‌توان استنباط نمود که فراوانی و مقدار بارش‌های سنگین و سبک به سمت کاهش میل نموده‌اند.

کلیدواژه‌ها


عنوان مقاله [English]

Frequency Distribution Change of Extreme precipitation in Zanjan City

نویسنده [English]

  • H. Asakereh
چکیده [English]



Extended Abstract
1- Introduction
Recently the experts have paid their attentions to tail of frequency distribution of climate elements more than averages, because the changes in these characters have significant effects on intense and continuity of events. So that variations in climate elements could be followed through many qualifications such as extreme events or the tail of statistical distribution. Accordingly the weather extreme events are definite as rare and abnormal events (Bartolin et al 2008: 1752) which are far from the centre of distribution.
Due to frequency, continuity and density of the climatic and weather extremes consequences on ecosystems, natural processes (weathering and erosion), and human society and because of their day to day increasing have gotten too much attention.
 
 
 
The weather and climate extremes are one of serious challenges in human society during recent century. Accordingly many experts call them “social challenges” and they believe on the economic sustainable developments in providing managing the extremes (WMO 2009). Consequently study the extreme events make us powerful to understanding their tempo-spatial patterns as well as managing them to protect ourselves from negative consequences.
 
2- Methodology
The Extreme analyses in climatology are based on analyses of statistical distributions tails of daily observation of climatic elements. In other words to climatic extreme investigations it has to pay attention to low and height values.  So it is necessary to investigate the daily extremes. Accordingly in this research using percentile index to determine extremes as well as absolute values, four extreme indices have been described for Zanjan precipitation during 1961-2006. So that the precipitation equal or less than 5th percentile, precipitation equal or more than 95th percentile, total of maximum precipitation and total precipitation of five greatest precipitations have been calculated for every year as indices of extreme precipitation. Finally the Generalized Extreme Distribution (GED) has been fitted on them.
In this research has been attended on the maximum () and minimum () of precipitation time series (), in standpoint of change in distribution shape and its parameters. Accordingly three extreme distributions; Gumbel, Frichit and negative Weibull have been fitted on the data based on thickness, fatness and length of distribution respectively. To finding the distribution parameters it has been used the L-moment method which has introduced by Hosking and Wallis (1997).
 
3- Discussion
First of all it has introduced the general characters of extreme precipitation in Zanjan. Based on the results of this stage of research it has cleared that low values of precipitation has shown decreasing trends as well as a jump in 1980. The high values of precipitation have shown stationary behaviours in general.
Based on testing many distributions on the daily extreme precipitation in order to modelling the tail of probability distribution of extreme precipitation, it has cleared that GED is the best fitted model on the observations. The model has been fitted on four indices (precipitation equal or less than 5th percentile, precipitation equal or more than 95th percentile, total of maximum precipitation and total precipitation of five greatest precipitations) for all time series as well as for two halves of periods (1961-1983 and 1984-2006).
The parameter of shape for 5th percentile showed the right skewness. It implicates that low values in this percentile more happened than the heights. This result is true for the second half of the period, but it is not for the first. The first half of the period has left skewness in 5th percentile values of precipitation. So the second half of the period has experienced decrease in precipitation in compare with the first half of period. In addition the scale parameter has shown greatest values in compare with the second and all the time series.
All parameters of maximum precipitation for all of the period as well as for the halves of period indicated no significant changes. These results have been achieved for the five greatest precipitation amounts too.
 
4- Conclusion
In this paper in order to investigate the extreme precipitation of Zanjan city, the daily precipitation during 1961-2006 has been analyzed. Accordingly the thresholds of daily extreme precipitation have been definite based on Climate Variability and Predictability Project.
In spite of stationary in trend, there have been many non- stationary characters as follow:
- There is a 3 year cycle in the greatest 5-day rainfall total
 - There is a downward jump in the fifth percentile and its ratio to total precipitation since 1980.    
- The shape parameter of frequency distribution of two parts of fifth percentile of precipitation shows a change in skewness and scale parameter.
- There is a similarity but lower change in 95 percentile of precipitation.
Accordingly there is a reduction in lower and higher percentiles precipitation in Zanjan

کلیدواژه‌ها [English]

  • Extreme Precipitation
  • Spectral Analyses
  • Frequency Distribution
  • Generalized Extreme Distribution