The Research On Evapotranspiration Simulation Of Typical Paddy Fields In East Asia Based On Different Mechanism Models And The Influence Of Climate Change

Evapotranspiration is a major component of the energy balance and hydrological cycle of the Earth,as well as a key process interconnecting atmosphere,hydrosphere,and biosphere.Accurate monitoring and simulating paddy evapotranspiration can contribute in improving crop water use efficiency,which is also of great significance to regional agricultural water resources management.At present,rice evapotranspiration research is challenged by many factors,such as regional differences,lack of unity in observation and simulation methods,and the impacts of climate change on rice evapotranspiration,growth and production,which has great uncertainty.Based on the measured rice data collected at three typical stations in East Asia,this paper compared the differences between large aperture scintillometer（LAS）and eddy covariance method（EC）in evapotranspiration observation of rice field based on the footprint model.Three crop evapotranspiration models（Si B2＿paddy,Aqua Crop,ORYZA2000）combined with the results of climate model CMIP5,and bayesian model averaging（BMA）were used to comprehensively evaluate the simulation of rice field evapotranspiration during rice growth period,and quantitatively analyze the impact of climate change on rice evapotranspiration,yield and water use efficiency.This study estimates the evapotranspiration of rice fields in different regions of East Asia,and evaluates the changes of rice productivity and water resources utilization under climate change,which is of great value for food security and sustainable development of water resources in China and its surrounding areas.The main conclusionsare as follows:（1）The evapotranspiration of red soil paddy field in southern China has obvious seasonal and interannual variation characteristics.This paper compared and analyzed the differences between LAS method and EC method by using flux footprint model.The results showed that the daily evapotranspiration of rice reached the peak in the middle of August during the rice growing season.From 2017 to 2021,the net radiation and latent heat flux in paddy field showed an increasing trend year by year,while the soil heat flux and sensible heat flux showed no significant annual variation.In the early growing season,the sensible heat flux of LAS was higher than that of EC,and the difference between LAS and EC gradually decreased in the middle and late growing season.On the contrary,the latent heat flux and evapotranspiration values observed by LAS were lower than those observed by EC in the early growing season.Combined with the footprint model,it was found that the area of LAS flux source area in the same month was about 2.6⁴.7 times that of EC observation source area,and the underlying surface type of LAS observation was also more complex than EC observation.The greater the difference between LAS and EC in the underlying surface of the flux source area,the greater the difference between the evapotranspiration values.Therefore,the direct reason for the difference of evapotranspiration observation is that the observation area of LAS and EC is not completely consistent.（2）The main parameters of SiB2 model were calibrated in the application of three rice stations in East Asia（Yingtan station in China,MSE station in Japan and IRRI station in the Philippines）,and the simulation results were also statistically evaluated.On the basis of the original Si B2 model,considering the layered water and heat transfer in rice canopy and the existence of paddy field irrigation water,the multilayer water and heat transfer module and irrigation water layer module were added to improve it respectively,and the simulation results before and after the improvement were compared.The results showed that the net radiation and soil heat flux simulated by the original Si B2 had good simulation results at the three stations.But the simulated values of sensible heat flux and latent heat flux were significantly different from the actual values,and significantly overestimated sensible heat flux and underestimate latent heat flux.The Si B2 model was improved to Si B2＿paddy model and then the rice field simulation was carried out again.It was found that the overestimation of sensible heat flux and the underestimation of latent heat flux in rice field decreased from about 20%to 8.2%¹1.2%.These results indicated that the improved Si B2＿paddy model could be better applied to the simulation of energy distribution and evapotranspiration of rice fields in East Asia.（3）The main parameters of Aqua Crop and ORYZA2000 models were calibrated and tested in three East Asian rice stations,and the simulation results were verified and analyzed.The results showed that the simulation errors of ORYZA2000 and Aqua Crop models for rice leaf area index and biomass processes in three typical paddy fields in East Asia were within an acceptable and reasonable range.In the yield simulation,the ORYZA2000 model had four season’s simulation errors less than 5%and two season’s simulation errors less than 30%in the simulation of six growing seasons,while the Aqua Crop model had relatively large rice yield errors.Therefore,for the three East Asian sites in the study,ORYZA2000 model was better than the Aqua Crop model in simulating yield and final biomass.（4）The collected data of rice field evapotranspiration and three calibrated ET models（Si B2＿paddy,ORYZA2000,Aqua Crop）were used to analyze the actual meteorological characteristics and evapotranspiration characteristics of the stations and evaluate the simulation of paddy evapotranspiration.The results showed that for Yingtan station,Si B2＿paddy model had better results in simulating rice field evapotranspiration,with correlation coefficient R=0.95.In the evapotranspiration simulation of MSE station and IRRI station,the Aqua Crop model had better simulation results,with R values of 0.95 and 0.97,respectively,and the root mean square error was relatively low,0.39mm/d and 0.34mm/d,respectively.Considering that the parameters of Aqua Crop model are easier to debug,we believe that Aqua Crop model has more advantages in the application of rice evapotranspiration simulation.（5）The effects of climate change on evapotranspiration,yield and water use efficiency of paddy fields in East Asia were quantitatively analyzed by using three modified rice models combined with the output data of climate model CMIP5 and BMA method.The results showed that from 2020s to 2080s,cumulative evapotranspiration during the rice growing season at Yingtan station and IRRI station would gradually increase with climate warming under the three emission scenarios.Under the RCP8.5 emission scenario,the maximum increase of evapotranspiration at both stations occurred in the 2080s,which was 25.8%and 33.7%,respectively.However,the MSE station was different from the other two stations.In RCP2.6and 4.5,the cumulative evapotranspiration of rice growth period decreased in the 2020s,which was caused by the shortening of the growing season days in the MSE station.Under different emission conditions,the rice yield and water use efficiency decreased gradually at the three sites from 2020s to 2080s,especially without considering the fertilization effect of CO₂.In general,considering the effect of CO₂ fertilization,the evapotranspiration of rice showed little change,and rice yield and water use efficiency decreased unsignificantly,indicating that sufficient CO₂ could improve the water use efficiency of rice.