Evaluation of thermal and combined methods in determining the water requirement of plants using moving correlation

Introduction

Drought and limited water resources in Iran, especially in Azerbaijan, require preventing the wastage of existing water resources and making optimal use of them. One of the main indicators for estimating the water needed by plants (crops) and optimal design of irrigation and drainage networks is determining the amount of water evaporated and transpired from the desired area. The more accurately the estimation of this factor is done, the more the efficiency of agricultural products and the reduction of irrigation system costs and the reduction of excess water consumption will increase, so determining the amount of ETP should always be considered by researchers as one of the important and effective factors in agricultural studies. Based on the climate and weather conditions of the regions, several relationships have been presented by different researchers to estimate the amount of potential evapotranspiration, and in each of these relationships, a number of climatic factors have been used as the most important effective factors in this process. Considering the recent droughts in the northwest of the country and considering that this region is one of the agricultural poles of the country, it is inevitable to conduct such a research in a part of this region. The purpose of this research is to estimate and localize the potential evapotranspiration estimation models in Pars Abad Moghan synoptic station and finally determine the most accurate method for the above location.

Materials and Methods

The area studied in this research is Pars Abad Moghan synoptic station, this station is located at 39 degrees and 39 minutes north latitude and 47 degrees and 55 minutes east longitude from the Greenwich meridian. The height of this station is about 45.4 meters above sea level. Parsabad Mughan is located in Moghan plain. As the northernmost city of Iran, this city is located on the southern banks of Aras River and borders with the Republic of Azerbaijan. In this regard, several climatic factors were used in the calculation of potential evapotranspiration with the combined methods of Penman, Penman Monteith, Penman Wright and Van Bavel, as well as thermal methods including Blaney Criddle, Thornthwaite, Linacre and Hargreaves. Visual Basic programming language was used to calculate potential evapotranspiration with eight methods and also to determine moving correlation coefficients. Among the climatic factors, monthly evaporation rate from the evaporation pan, monthly average temperature, monthly maximum relative humidity, monthly average relative humidity, monthly minimum relative humidity, air pressure, wind speed, percentage of sunny hours, extraterrestrial radiation and of synoptic meteorology of Parsabad Moghan, has been used in the calculation of monthly potential evapotranspiration, as well as their correction. Statistical methods of linear correlation and power were used to modify and regionalize potential evapotranspiration estimation models. The performance of each used method was evaluated using correlation coefficient, root mean square error and t test.

Results and Discussion

The monthly potential evapotranspiration was estimated from the eight mentioned methods. The results showed that in the combined methods, the potential evapotranspiration calculated from April to June and in some of them until July have higher values than the evaporation pan method, and after that it is the opposite until December. Potential evapotranspiration values obtained from the Linacre method indicate that in all months these values are proportionally higher than the values obtained from the evaporation pan method, which shows a high correlation with the evaporation pan method. The results of the Thornthwaite method for all months have almost uniformly lower values than the values of the evaporation pan, which indicates the high agreement of the above values with the values of the evaporation pan method. Correlation coefficients in all thermal methods were higher than combined methods. The methods of Linacre with a correlation coefficient of 0.883 and Thornthwaite with a correlation coefficient of 0.874 have the highest correlation coefficient with the values obtained from the evaporation pan method. Its climate and the fact that thermal methods work better in areas with high temperatures are not far from expected. Linacre's method has been confirmed in several researches, including Benzaghna and Aniadik in calculating potential evapotranspiration in Libya and West Africa.

Conclusion

According to the statistical analysis table, the lowest RMSE value is related to the Penman Monteith method, which indicates the lower error of the results of this method compared to the values obtained from the evaporation pan method. After Penman Monteith method with RMSE 48.71 mm per month, Penman method with RMSEof 49.65 mm per month, Penman Wright with 51.39 mm per month and Thornthwaite with 53.42 mm per month, had the lowest root mean square error value. In all methods, at the confidence level of 99%, hypothesis one is rejected (H1: µ1≠µ2) and null hypothesis (H0: µ1=µ2) is confirmed. The lowest correlation coefficient is related to the combined method of Van Bavel. Potential evaporation and transpiration values calculated by the Hargreaves method do not have a particular trend and agreement with the values of the evaporation pan method, which shows the lower correlation of the above method with the evaporation pan method. Also, due to the high correlation coefficients of the methods of Linacre, Thornthwaite, these methods are proposed after applying correction coefficients to estimate the potential evapotranspiration of Parsabad Moghan and these methods are suitable and valid alternatives for potential evaporation and transpiration values for the evaporation pan method.