Estimation Of Evapotranspiration Using The Crop Canopy Temperature

Evapotranspiration(ET)is a major component of the hydrologic cycle,and accurate and timely estimation of the evapotranspiration is the basis of agricultural efficient water use and regional water resource management.There are a large number of models and methods for estimating regional evapotranspiration by remote sensing,but most of them are complex and numerous surface parameters information are required.At the same time,it is hard to provide land surface temperature information in high temporal and spatial resolution by the existing single thermal infrared remote sensing images.The estimation of regional evapotranspiration using canopy temperature or land surface temperature at different scales and meteorological factors can be employed to provide a scientific basis for the accurate and real-time irrigation management in irrigation areas.In this paper,field and regional scale experiments with maize and sunflower in the Jiefangzha Irrigation Area of Hetao Irrigation District as experiment subjects were performed from 2015-2016.In the field scale experiments,the canopy temperature of maize and sunflower,soil temperature,soil moisture and crop growth index,etc.were regularly observed,and the change rule of canopy temperature in the horizontal direction and farmland temperature in vertical direction were analyzed.Meanwhile,the Canopy Temperature and Meteorology Monitoring System(CTMS)were installed in the maize field and sunflower field respectively by which the canopy temperature of both crops and meteorological factors were continuously synchronous observed and local applicability of the Seguin and Itier simplified model(S-I model)was calibrated and validated.The Enhanced Spatial and Temporal Adaptive Reflectance Fusion Model(ESTARFM)algorithm was employed to integrate high-spatial Landsat data and high-temporal MODIS images and to generate land surface temperature data set at high temporal and spatial resolution.The S-I model was used to estimate the regional evapotranspiration of maize and sunflower,and the water balance method at regional scale was utilized to verify the total crop evapotranspiration of model estimation results,in order to evaluate the applicability of the S-I model at regional scale.The main conclusions are as follows:(1)During the whole growth period of maize and sunflower,the variation of canopy temperature in the horizontal direction of both crops was very small,and there are obvious gradient change rule in the vertical direction of farmland temperature.(2)The field observation data in 2015 were employed to calibrate the S-I model.The linear regression analysis of the hourly canopy-air temperature difference(Tc-Ta)from 10:00 to 16:00 and the difference value(ETd-Rnd)of daily evapotranspiration(ETd)and daily net radiation(Rnd)was respectively carried out,and parameters a and b in the S-1 model at different moments were obtained.The field observation data in 2016 were utilized to validate the S-I model.The change trend between the observed ETd and the estimated ETd in maize and sunflower growth period are substantially the same with the best performance at 13:00,which suggests that the S-I model performed reasonably well in predicting the ETd during the growth period of maize and sunflowers at the field scale.(3)Based on the ESTARFM algorithm,the spatial downscaling of land surface temperature is achieved,and a high temporal and spatial resolution land surface temperature data set that can be used for regional evapotranspiration estimation is constructed with which the Landsat land surface temperature and ground observation canopy temperature of crops are compared.These results indicate that the fusion algorithm has good prediction precision.(4)The total amount of evapotranspiration estimated by the S-I model was verified by water balance method at the regional scale,the relative error values between water balance ET and simulation ET in the Huangji,Qinghui(the amount of irrigation in well irrigation fields was considered),Wula River and Yangjia River sections of 2015 and 2016,were less than±10%.These results indicate that the S-I model has good prediction precision to estimate the ET in the Jiefangzha Irrigation Area.Therefore,the S-I model is a simple and convenient method to estimate the ET at regional scale.