Research On Field Scale Evapotranspiration Model

Surface evapotranspiration is a major part of water consumption in agricultural ecosystems.The study of crop evapotranspiration models is an important way to grasp the water consumption pattern of crops and improve the efficiency of irrigation and regional agricultural water management.The application of existing mainstream evapotranspiration models at the field scale is limited by high resolution remote sensing data and spatial scale effect.There is still much to improve in data characterization and model accuracy.In view of the characteristics of surface parameters at the field scale,it is necessary to break through the demand for thermal infrared remote sensing data in the existing energy balance evapotranspiration models,to carry out research on high-resolution field-scale evapotranspiration models from the evapotranspiration mechanism,to establish a complete and high-precision theory of field evapotranspiration estimation system,and to enhance the capability of cropland water resources management.In this study,two field evapotranspiration monitoring models are constructed for the characteristics of field evapotranspiration,from the influence mechanism of evapotranspiration,to investigate the differences of evapotranspiration change drivers at different scales,and to combine the spatial variability of surface parameters at the field scale.Several studies were conducted as follows.(1)The spatial heterogeneity of meteorological factors and surface covariates at field scale.The results showed that the spatial distribution variability of meteorological factors was low at the 1 km image scale and sub pixel scale in the cropland area.When constructing the field-scale evapotranspiration model,the influence of the variability of meteorological conditions at the km scale in the distribution factors representing the variability of field evapotranspiration can be ignored.The results of spatial variability analysis of high-resolution remotely sensed surface parameters showed that the spatial variability of different remotely sensed parameters differed.NDVI spatial variability is high in vegetation-soil mixed region and low in vegetation covered region.Albedo spatial variability is low.Land surface water index has larger spatial variability than NDVI and albedo and still high in vegetation covered region.LSWI spatial variability is close to NDVI.Without direct access to high-resolution LST data,the LST-based evapotranspiration calculation introduces large uncertainties.(2)A field-scale evapotranspiration estimation model for cropland that does not depend on thermal infrared surface temperature is constructed.Based on the physiological activity mechanism of vegetation,the conductivity model couples the photosynthetic carbon sequestration and transpiration processes of vegetation,and the high-resolution optical remote sensing data obtained from Sentinel 2 satellite and meteorological data are used to construct a field-scale evapotranspiration estimation model.The evapotranspiration mechanism is revealed and simulated through the interaction between vegetation physiological processes and environmental variables,which can be used to study the ecosystem function and vegetation carbon and water consumption strategies under the context of climate change.Based on the physiological characteristics of vegetation and surface radiation conditions,NDVI and albedo,which are insensitive to temporal changes at the daily scale,are used as remote sensing input data for the model avoiding the uncertainty caused by the multi-scale data fusion process of most models.(3)Field evapotranspiration allocation algorithm.The study clarified that the main factors causing the spatial heterogeneity of evapotranspiration at the field scale are the vegetation factor and soil water content factor of different fields,which are described by the fraction of vegetation cover and land surface water index,respectively.Based on numerical simulations of crop physiological processes and surface water content,high-resolution spatial allocation factors of evapotranspiration were developed using Sentinel 2 satellite 10 m resolution data.The model was validated with good results in the Haihe and Heihe river basins.The comprehensive results show that the accuracy of the field evapotranspiration allocation model can be consistent with and improved from the evapotranspiration products used as input data,and accurate field evapotranspiration results can be obtained with the guaranteed accuracy of the input data.(4)Sensitivity analysis of the field-scale evapotranspiration model.The sensitivity analysis of the evapotranspiration model for coupled carbon and water processes shows that evapotranspiration is most sensitive to changes in temperature,followed by changes in CO₂ concentration and changes in available solar shortwave radiation of vegetation.The model scale effects were evaluated based on the differences in the accuracy of the calculation results of the coupled carbon-water model based on the input data of remote sensing sources at different scales,and the comprehensive validation results showed that the size of the study area affected the scale effects of the evapotranspiration calculation.The model mechanism comparison proves that the coupled carbon-water model is suitable for estimating evapotranspiration in single vegetation areas and bare ground with poor distribution,and the model calibration needs to consider the physiological and ecological processes of different vegetation types when applied to complex land surfaces.The field evapotranspiration distribution model can meet the demand of spatial scale expansion of evapotranspiration for complex substrates in flat areas,but when applied to areas with rugged surfaces and obvious elevation changes,the evapotranspiration distribution factor needs to be adjusted and the evapotranspiration influence factor on spatial distribution heterogeneity needs to be added.