Crop Water Footprint Measurement Based On Remote Sensing Products Of Terrestrial Evapotranspiration

Agriculture is the world’s largest consumer of water resources,and scientific evaluation of regional agricultural water use characteristics is the prerequisite for control of agricultural water usage.The crop water footprint can provide a scientific basis for assessing agricultural production water characteristics and guiding water resource management.The estimation of regional crop water footprint requires large amounts of ground data,and remote sensing technology provides an effective means for scientifically estimating the crop water footprint at a large scale.This paper proposes a method based on the remote sensing evapotranspiration to address the limitations of traditional crop water footprint calculations,which are based on the FAO Penman-Monteith formula and do not consider the actual soil water status in farmlands.The paper assesses the accuracy of remote sensing model simulated evapotranspiration data by comparing the results of remotely sensed evapotranspiration data with those measured by eight flux observation towers based on eddy correlation techniques,taking into account the implications and limitations of different indicators in statistical analysis.The paper assesses the limitations of different methods(based on FAO-PM,MOD16,and GLDAS-Noah)for estimating the water footprint of maize crops,and analyzes and discusses the differences and characteristics of different methods in different regions,to provide scientific support for quantifying the water footprint of crops at large scales.The main conclusions of the paper are as follows:(1)This study constructed a framework for estimating the regional crop water footprint based on remote sensing evapotranspiration.The paper explores the limitations of estimating crop water footprint based on FAO Penman-Monteith equation without considering the actual soil moisture conditions,which leads to large discrepancies between the calculated results and regional reality,especially in arid areas,where the crop water footprint calculation results are significantly overestimated.Regional crop evapotranspiration data were obtained based on MOD16 or GLDAS-Noah,and statistical analyses(R,IOA,Bias,and RMSE)of remote sensing evapotranspiration model simulation results combined with multi-year flux data at different locations,revealed that the simulation results of regional evapotranspiration data can better reflect the actual soil moisture coditions in the region.Compared with the FAO-PM-based method,the crop water footprint estimated by the remote sensing-based method is closer to the actual regional situation,especially in arid or semi-arid regions.(2)This study identified the regional distribution characteristics of crop water footprint estimation by different methods.The mean value of crop water footprint estimated based on FAO-PM was the largest in the maize growing areas of China,where the crop water footprint values were higher in arid or semi-arid areas than in wet areas.The crop water footprint values estimated based on remotely sensed evapotranspiration data were lower in arid areas than in humid areas.Compared with MOD16,the estimated crop water footprint values based on MOD16＿B were significantly higher in wet areas.Crop water footprints were estimated based on different methods in maize-growing areas in China.The mean value of crop water footprint based on FAO-PM was 0.78 m~3/kg,0.82 m~3/kg in arid or semi-arid regions,and 0.75 m~3/kg in southern regions.The mean crop water footprint based on GLDAS-Noah is 0.61 m~3/kg,0.38 m~3/kg in arid or semi-arid regions,and 0.73 m~3/kg in southern regions.the mean crop water footprint based on the filled MOD16(MOD16＿B)is 0.44 m~3/kg,and 0.28 m~3/kg in arid or semi-arid regions.The mean value is 0.28 m~3/kg in arid or semi-arid areas and 0.59 m~3/kg in southern areas.(3)This study investigated the reasons for the differences in the estimation of crop water footprints by different methods in different climatic zones.The difference between FAO-PM-based and MOD16-or GLDAS-Noah-based crop water footprint values being smaller in the humid or semi-humid northeast and some southern regions,probably due to the assumption that these regions are closer to adequate soil water supply conditions than arid regions.The remote sensing-based method has better applicability than the FAO-PM-based method in the dry northwest where the soil is not fully watered.The GDLAS-Noah is based on land surface model combined with satellite and ground observation data to estimate land surface evapotranspiration,which includes soil evaporation,canopy interception rain evaporation and canopy transpiration.The remote sensing-based methods consider the resistance of vegetation or soil to water vapor transport.In the soil evaporation calculation,MOD16 divides soil into saturated and moist soil,and GLDAS-Noah considers the evaporation process from saturated to unsaturated soil,and the crop water footprint estimated based on remote sensing methods in arid or semi-arid areas is close to the actual regional situation.Considering the influence of clouds or aerosols on MOD16 data in different regions,it is found that the crop water footprint estimated based on the filled MOD16significantly improves the value in wet or semi-humid regions,which is closer to the actual situation in these regions.The paper combined remotely sensed evapotranspiration data with the crop water footprint estimation process,explored the differences and linkages between different crop water footprint quantification methods in different regions in China’s maize growing areas,and examined MOD16 and GLDAS-Noah based on flux site data.The results highlight that the crop water footprints estimated based on remotely sensed evapotranspiration data are closer to the regional reality than those estimated based on FAO-PM,especially in arid or semi-arid regions,and can provide a scientific basis for large-scale crop water footprint quantification studies.