Research On The Prediction Of Corn Yield In Daqing City Based On The Coupling Of Remote Sensing And WOFOST Model

Food security in agricultural production occupies its central position and is also related to the national economic development,and the precision of crop estimation has great significance to food security and policy formulation in a region.With the continuous development of remote sensing technology and various crop models,the research becomes more important.Although the research can be used alone,with the rapid development and optimization and the combination of the two,the method has changed and can improve the precision.Remote sensing information has rapid,objective and large coverage over traditional survey methods Potential,but also only through the external way to reflect the crop state.Nowadays,scholars choose more to combine it with the model to improve the accuracy.And most of the assimilation algorithm is used in the selection of combined methods.By collecting remote sensing information and related inversion methods and understanding the crop model,HJ-CCD image data,MODIS LAI data,vegetation index,vegetation index inversion method and WOFOST crop model and OAT(One-At-a-Time)sensitivity analysis method were adopted to predict the regional corn yield level.The research work and main conclusions are as follows:(1)Based on Daqing City,Heilongjiang Province,two key growth periods of July28,2018 and September 15,2018 were selected and the image data of quasi-synchronous environmental satellite HJ-CCD were obtained.This paper uses MODIS LAI instead of actual LAI,to process image data by using ENVI software.(2)Through the understanding of various vegetation indices,four vegetation indices with strong ability to characterize maize population characteristics and the most widely used ones were selected.The linear regression inversion model was established with the LAI values extracted from the MODIS LAI.The results showed that the model established by the ratio vegetation index(RVI)had the strongest ability(Tasseling and silking stage period R~2=0.7796,matuniity period R~2=0.7456).Therefore,this paper takes RVI as the vegetation index LAI remote sensing inversion.(3)Using OAT methods for sensitivity analysis of parameters in WOFOST models,The relative sensitivity is calculated by the relevant formula and the parameter with value greater than 0.10 is determined as the sensitivity factor,Maximum CO2 assimilation rate(AMAXTB),Conversion rate of assimilates in leaves(CVL),Conversion efficiency of the same substance in storage organs(CVO),Relative rate of root respiration(RMR),Conversion efficiency of stem assimilates(CVS),Root dry matter distribution coefficient(FRTB),Maintaining respiration rate of stem(RMS),Life cycle of leaf area at 35?(SPAN),Specific leaf area in growth period(SLATB),Cumulative temperature from emergence to flowering(TSMU1),Leaf maintenance respiration rate(RML).We assimilate the LAI obtained from remote sensing with the model simulation,and obtain the parameter optimal set of the simulated LAI using the lookup table optimization algorithm.We simulated the optimal yield of Daqing corn in 2018,realized the purpose of assimilation,and completed the yield per unit area simulation.The final results show the combination of the two,and select a suitable optimization algorithm to improve the accuracy of corn yield.