مدلسازی رابطه تبخیر و تعرق پتانسیل سالانه و فصلی گیاه مرجع با عوامل اقلیمی در حوضه آبریز فلات مرکزی ایران

تخمین دقیق تبخیر و تعرق نقش مهمی در مدیریت منابع آب، برنامه‌ریزی آبیاری و تامین نیاز آبی گیاهان به‌ویژه در مناطق نیمه‌خشک و خشک دارد. هدف از این تحقیق ارائه‌ی یک مدل رگرسیونی است که با مبنا قرار دادن روش فائو-پنمن-مانتیث بتوان مقدارتبخیر و تعرق گیاه مرجع را با عوامل اقلیمی برآورد کرد. در این پژوهش ابتدا با استفاده از داده‌های ماهانه‌ی حداکثر و حداقل دما، متوسط رطوبت نسبی، ساعات آفتابی و متوسط سرعت باد ایستگاه‌های سینوپتیک واقع در حوضه‌ی آبریز فلات مرکزی ایران در یک دوره‌ی آماری 18 ساله (2012-1995)، مقادیر تبخیر و تعرق سالانه و فصلی به روش fao56pmمحاسبه شد. سپس با استفاده از نرم‌افزار spssرابطه‌ی بین عوامل اقلیمی فوق با مقدار تبخیر و تعرق از طریق رگرسیون خطی چندمتغیره مدلسازی گردید. صحت دقت مدل‌ها نیز با آزمودن چهار فرضیه‌ی خطی بودن رابطه، نرمال بودن باقیمانده‌ها، ثابت بودن واریانس باقیمانده‌ها و ناهمبسته بودن خطای مدل سنجیده شد. نتایج حاصل از اجرای مدل مبیّن آن است که رابطه‌ی قوی بین etoحاصل از رگرسیون چندمتغیره با عوامل اقلیمی وجود دارد؛ بطوریکه در مقطع زمانی سالانه و فصول بهار و تابستان حدود 98درصد تغییرات eto توسط این پنج متغیر تبیین می‌شود و در فصول پاییز و زمستان به ترتیب حدود 97 و 96 درصد از پراش تبخیر و تعرق با پراش پنج عامل اقلیمی مشترک است. معادلات رگرسیون استاندارد حاکی از آن است که سهم متغیرهای سرعت باد و حداکثر دما در میزان تبخیر و تعرق سالانه و فصلی بیش از سایر عوامل اقلیمی است.مقایسه‌ی نقشه‌های هم تبخیر سالانه و فصلی حوضه‌ی آبریز فلات مرکزی نشان داد که ازلحاظ مکانی نیز ارتباط نزدیک و قابل قبولی بین روش فائو-پنمن-مانتیث و مدل رگرسیون وجود دارد و شمال غرب حوضه از کمترین و مناطق جنوبی از بیشترین میزان تبخیر و تعرق سالانه و فصلی برخوردار می‌باشند.

Modeling the relationship between reference evapotranspiration annual and seasonal with climatic factors in catchment central plateau Iran

Introduction The precise estimation of evapotranspiration plays an important role in managing water resources, planning irrigation and providing water needed for plants, especially in semiarid and dry areas. The FAOPenmanMonteith method is proposed as the only standard method for calculating reference evapotranspiration from climate data and for evaluating other methods (Allen et al., 1998). In the last few decades, multivariate regression has been used a one of the important methods for recognizing the interrelationship among variables and determining the correlation between evapotranspiration and a set of climatic factors. In this method, evapotranspiration was the dependent variable and various climatic elements were included as independent variables to the model. The best model is a model that can provide a better estimate of two or more other dependent variables. Arab Solghar et al. (2010) predicted annual evapotranspiration using multivariate regression in a number of subtropical stations in Iran. The results showed association between estimated evapotranspiration using PenmanMonteith equation and the method had a relatively small error level. Therefore, the aim of this study was to calculate evapotranspiration in different months of the year and subsequently in annual and seasonal periods for stations located in the central watershed of Iran during 19952012 using the FAOPenmanMonteith method and modeling and estimating ETo values based on five climatic factors (maximum and minimum temperature, relative humidity, suny hours, and average wind speed at 2 meters above ground level) and the FAO56PM equation.   Materials and methods In this study, first, annual and seasonal evapotranspiration were measured by the FAO56PM method using Cropwat software based on monthly data of maximum and minimum temperature, average relative humidity, sunny hours and average wind speed of synoptic stations located in the central watershed of Iran during a statistical period of 18 years (19952012). Then, SPSS software was used to model the relationship between the above climatic factors and evapotranspiration through multivariate linear regression. The accuracy of the models was evaluated by testing the four hypotheses of the linearity of the relationship, normality of the residuals distribution, constant variance of the residuals, and statistical independence of the errors. In the end, the GIS software capability was used for zoning the annual and seasonal evapotranspiration of the central plateau watershed, and preparing plans related to ETo values ​​derived from the FAOPenmanMonteith equation and multivariate regression equation. The two methods were also spatially compared and analyzed.   Results and discussion The results of the model show a strong correlation between ETo obtained from multivariate regression and the five climatic factors, so that about 98% of changes in ETo at annual scale and in the spring and summer were explained by these five variables, and approximately 97 and 96 percent of evapotranspiration changes were associated with changes in climatic factors in the autumn and winter seasons. Standard regression equations showed that the contribution of wind speed and maximum temperature variables in annual and seasonal evapotranspiration was more than other climatic factors. In summer, sunny hours and in other seasons, as well as at annual time scale, relative humidity had the least effect on the amount of evapotranspiration. Also, the highest effect of wind speed area and maximum temperature calculated in the study was 0.81 in the summer season and 0.41 in the spring. Comparison of the annual and seasonal evaporation maps of the central plateau watershed showed an acceptable correlation between the FAOPenmanMonteith method and the regression model. In spring and summer, the highest ETo observed in both FAOPenmanMonteith and regression methods occurred in the southern and middle regions of the central plateau watershed and the lowest ETo was in the northwest of the watershed area. In winter, ETo was increasing from the center of the watershed towards lower latitudes due to the relative increase in temperature and wind speed. The annual ETo value was lower in the northern regions but higher in the southern regions. Wind speed, average maximum temperature and minimum temperature had the greatest effect on increasing ETo respectively. The results of this study are consistent with studies by Azizi et al. (2009) on estimating ETo through multivariable regression in Isfahan province; Arab Solghar et al. (2011), in predicting annual ETo using meteorological data in a number of stations in semiarid areas of Iran; Sheikholeslami et al. (2014) on modeling ETo using daily data in Khorasan Razavi; and Bakhtiari et al. (2015) on the estimation of daily ETo in selected semiarid climates of Iran.    Conclusion The purpose of this study was to provide a regression model based on the FAOPenmanMonteith method that can estimate ETo using climatic factors. In this study, the maximum and minimum temperature, average relative humidity, sunny hours and average wind speed were measured monthly at 2 meter height from the ground surface of the synoptic stations via Cropwat software. ETo values in the watershed of the central plateau of Iran ​​were calculated in annual and seasonal time scales using the FAO56PM method during the statistical period of 18 years (19952012). Using SPSS 20, the relationship between the above climatic factors and ETo was modeled through multivariable regression. The results of the model showed a significant relationship between ETo obtained from multivariable regression and the climatic factors, in which the contribution of wind speed and maximum temperature variables in annual and seasonal ETo were more than other climatic factors. Comparison of annual and seasonal ETo maps of the central plateau watershed showed a close and acceptable association between the FAOPenmanMonteith method and the regression model at spatial level, so that the northwest of the watershed had the lowest and the southern regions had the highest annual and seasonal ETo. Therefore, due to the acceptable results of this study in Iran’s central plateau watershed, it is recommended that regression equations be used in other watersheds of the country with insufficient or lack of lysimeter data to predict ETo with acceptable accuracy, which plays a very important role in determining the water requirement of plants.