A Comparative Study On Estimation Methods Of Regional Evapotranspiration Based On Remote Sensing

Water problem is the core issue of the research on resources,environment and ecosystems in arid areas,among which evapotranspiration is the final way of water consumption in the arid inland water cycle.There are many methods for estimating regional evapotranspiration.This paper studies three common methods for estimating evapotranspiration,including SEBS model,P-M formula,and feature space.Compareing the advantage and disadvantage of these three methods and their applicable conditions can provide reference for the further development of such remote sensing methods for estimating evapotranspiration and their application in related fields.Based on MODIS data and meteorological data from April to October of 2001,2005,2010,and 2015,this paper uses SEBS model method,RS-PM model method,MOD16A2 data,FAO formula method,and feature space method to estimate the regional evapotranspiration of Yinwei Plain.It analyzes and studies the advantage and disadvantage,applicability of each estimation method,as well as the temporal and spatial variation of evapotranspiration and its influencing factors.The results show that:(1)The estimated evapotranspiration was verified by the water balance equation in the Yinwei Plain area,It was found that the results obtained by the SEBS model and the RS-PM model were close to the results obtained by the water balance method.Compared with the water balance method,the results obtained from the MOD16A2 data seriously underestimate the evapotranspiration.The FAO formula method and the feature space method are compared with the water balance,and it is found that the evapotranspiration of the Yinwei Plain is seriously overestimated.(2)The advantage of the SEBS model method is its best mechanism.The disadvantage is that due to factors such as differences in the water and heat characteristics of vegetation and soil,changes in surface roughness,etc.,it will bring some errors to the SEBS model method.It is suitable for large-scale heterogeneous surface complex terrain.(3)The advantage of the RS-PM model is that it optimizes the calculation accuracy of the P-M formula and expands the scope of application of the P-M formula.The disadvantage is that there are many formulas and required parameters,and the calculation steps are cumbersome,and the applicability on complex underlying surfaces is not ideal.It is suitable for large-scale natural surfaces with relatively simple underlying surfaces.(4)The advantage of the MOD16A2 data algorithm is that no calculation is required,as long as the downloaded MOD16 data is processed,the evapotranspiration data can be obtained.The disadvantage is that only the evapotranspiration in areas covered by vegetation is calculated.It is suitable for all or most areas covered by vegetation.(5)The advantage of the FAO formula method is that the calculation process is very simple,does not rely on remote sensing data,and can be calculated with only meteorological data.The disadvantage is that there are few types of underlying surfaces applicable to this method.(6)The advantage of the feature space method is that it is clearly defined and requires less data.The disadvantage is that it requires very high-quality remote sensing data of surface temperature and has a large locality.(7)The SEBS model method is the most suitable method for estimating the regional evapotranspiration of Yinwei Plain among the methods used in this paper.The regional evapotranspiration from April to October of 2001,2005,2010,and 2015 were353.54 mm,359.37 mm,393.56 mm,396.16 mm.(8)The change of regional evapotranspiration in Yinwei Plain is positively correlated with temperature.The annual evapotranspiration estimated by the SEBS model increases with the increase of the cultivated land area.