Study On Land Surface Evapotranspiration Based On Remote Sensing Model

In recent years,global warming has become increasingly significant,leading to increased drought and water shortage.Global food security and water security are facing serious challenges.With the increasing demand for ecological water and irrigation water in China,the threat of water shortage to national food security and water security is growing.Evapotranspiration is an important way of agricultural water consumption and water resource consumption,and also plays a key role in land surface energy and water carbon cycle.For a long time,land surface evapotranspiration is a part of the hydrological cycle that is difficult to measure directly.Accurate estimation of evapotranspiration and determination of its spatiotemporal distribution model are the key to better understand the regional water and energy balance,and are also scientific premise in scientific management of agricultural water resources and prediction of climate change impacts.The development of remote sensing technology provides an effective means for large-scale evapotranspiration simulation,but the system error,driving data error and model parameter error of various remote sensing evapotranspiration models lead to the uncertainty of evapotranspiration simulation.In addition,the remote sensing evapotranspiration model usually depends on the observation data of ground meteorological stations,which requires interpolation of meteorological station data in advance and expansion to regional scale.This process is often accompanied by spatial heterogeneity,scale matching and other scale expansion problems,which further increases the uncertainty of evapotranspiration simulation.For the above problems,the paper compared the model mechanism of three remote sensing evapotranspiration models,namely,the Surface Energy Balance Algorithm for Land(SEBAL)model,the Shuttleworth-Wallace-Hu(SWH)model,the Moderate-Resolution Imaging Spectroradiometer ET(MOD16)model,and the reference crop evapotranspiration model of the Penman-Monteith(FAO-PM)model.Based on the correlation analysis of key process variables in the model,the nonlinear relationship between sensible heat flux and incoming shortwave radiation was found,and an empirical formula for calculating sensible heat flux from incoming shortwave radiation was proposed.The improved estimation method of sensible heat flux was embedded into the remote sensing evapotranspiration model,which improved the key simulation mechanism and the simulation performance of the remote sensing evapotranspiration model.The main results of the paper are as follows:(1)The temporal and spatial evolution characteristics of main meteorological factors and the reference crop evapotranspiration in northwest China from 1980 to 2017 were revealed.During the study period,the climate in most regions tended to warm and dry.Among them,the average temperature,the maximum temperature and the minimum temperature showed an obvious upward trend,and the order of significance from high to low was the minimum temperature(P < 0.01),the average temperature(P < 0.05)and the maximum temperature.The wind speed showed a significant downward trend(P < 0.05),while relative humidity showed no significant downward trend.meanwhile,the sensitivity of reference crop evapotranspiration to main meteorological factors and the contribution rate of main meteorological factors to reference crop evapotranspiration change were clarified.The evapotranspiration of reference crops showed no significant upward trend in northwest China as a whole.The absolute value of the average sensitivity(-0.65)of reference crop evapotranspiration to the change of relative humidity in northwest China is the largest,followed by the highest temperature(0.24),sunshine hours(0.20),lowest temperature(-0.18)and wind speed(0.13).According to the results of the average contribution rate of the main meteorological factors in northwest China to the reference crop evapotranspiration changes,the absolute value of the average contribution rate of the lowest temperature(-9.74%)is the highest,followed by the relative humidity(2.70%),the highest temperature(-1.69%),wind speed(-0.87%)and sunshine hours(-0.10%).(2)The simulation performance and model mechanism of different remote sensing evapotranspiration models were analyzed.The SWH model optimizes the calculation of soil impedance and better reflects the physical characteristics of the underlying surface soil.The MOD16 model optimizes the classification computer system of dry and wet canopy and dry and wet soil,and has more specific physical theoretical basis.The SEBAL model quantifies evapotranspiration as a whole and relies less on the data of ground meteorological stations.The study showed that the evapotranspiration simulation results of the SEBAL model,the SWH model and the MOD16 model in northwest China were high in the south and low in the north.Meanwhile,the spatio-temporal self-correlation of the SEBAL model,the SWH model,the MOD16 model and the FAO-PM model was higher in the vegetation growth season and lower in winter.The spatio-temporal correlation between the SEBAL model,the SWH model and the FAO-PM model also showed that it was higher in vegetation growth season and lower in winter.(3)It was found that the sensible heat flux,a key variable in the remote sensing evapotranspiration model,had a nonlinear relationship with the incoming shortwave radiation.Based on the data of the Global Eddy Flux Alliance,according to the correlation results between the key process variables in the evapotranspiration model,the sensible heat flux had a high correlation(0.69)with the incoming shortwave radiation,in which 49.7% of the stations had a correlation coefficient greater than 0.8,while only 3.1% had a correlation coefficient less than 0.3.Other variables with high correlation were the incoming photosynthetic photon flux density(0.71),net radiation(0.67),outgoing shortwave radiation(0.59)and diffuse incoming shortwave radiation(0.53).An improved thermodynamic map was created,and it was found that the sensible heat flux,the key variable in the remote sensing evapotranspiration model,was nonlinear with the incoming shortwave radiation.meanwhile,the similar nonlinear relationship was also verified in the reanalysis dataset and the atmospheric general circulation models dataset.Based on the above findings,an empirical formula for calculating sensible heat flux from incoming shortwave radiation was proposed.(4)The SEBAL model based on improved sensible heat flux estimation method was proposed,which improved the simulation performance and reduced the dependence of remote sensing evapotranspiration model on ground meteorological site data.Compared with the SEBAL model,the improved SEBAL model had improved the performance of most verification indicators,and was generally closer to the measured values of Haibei and Inner Mongolia eddy flux sites.Its spatio-temporal self-correlation was lower than the SEBAL model,which had better applicability in northwest China.The improved SEBAL model is generally sensitive to each driver,and the sensitivity(0.68)to albedo is the highest.The evapotranspiration in northwest China simulated by the improved SEBAL model was high in the south and low in the north in space,and increased first and then decreased in time.In summary,this study taked improving large-scale remote sensing evapotranspiration model as the main line,analyzed the temporal and spatial evolution characteristics of meteorological factors in northwest China from 1980 to 2017 and their impact on reference crop evapotranspiration,providing scientific support that the meteorological spatio-temporal evolution background for simulation and mechanism comparison of remote sensing evapotranspiration model.The performance and applicability of three remote sensing evapotranspiration models were compared.The mechanisms of different remote sensing evapotranspiration models were analyzed.And the key optimization ideas of remote sensing evapotranspiration models were proposed.Based on the dataset of the 159 global eddy flux sites and the correlation results of the key process variables in the evapotranspiration model,an empirical formula for calculating sensible heat flux from incoming shortwave radiation was proposed;The improved sensible heat flux estimation method was embedded into the SEBAL model,which reduced the dependence on the ground meteorological site data,improved the key simulation mechanism,improved the simulation performance and applicability of the model in northwest China,and provided support for scientific assessment of regional evapotranspiration and scientific management of agricultural water use.