Simulation Of Winter Wheat In Jiangsu Province With The Modified WOFOST Model Based On Remote Sensing Data

Crop yield information is crucial for a country or regions food safety warning,circulation of grain trade and management in decision making.Traditional routine regional agricultural information survey and monitoring means exist the lack of timeliness,economic cost and accuracy,which can not reveal the crop growth,development and yield information of the continous evolution of the internal mechanism and climate,soil and environment for the influence of crops.Crop growth model based on remote sensing data assimilation has become an effective method to solve this problem,remote sensing data and crop growth combined with crop model simulation is obtained by using remote sensing information that is nedded for crop model.In the region of parameter sensitivity and intial conditions to be adjusted,so that the crop model bring into full play the advantages in regional crop production estimation.This paper takes the winter wheat from 2001 to 2010 of Kunshan,Huaian and Xuzhou city in Jiangsu province as the reseach object and uses simulated annealing algorithm with WOFOST(World Food Study) model simulation of LAI(Leaf Area Index) and MODIS-LAI.The specific work is as follows:(1)This paper adopts the Savitzky-Golay(S-G) method to remove the abnormal information,because heterogeneity in pixels can cause MODIS products numerical problems.Then we adopt the actual measurement data of China Meteorological Data Sharing Service System to correct MODIS-LAI of study area.The results show that this method can eliminate the abnormal data of MODIS-LAI effectively. After using Savitzky-Golay method smoothing MODIS-LAI time series curve,we also use Logistic curve to fit MODIS-LAI data so that fitting crop growth curve accords with the actual growth status of winter wheat.(2)For remote sensing image in the actual existence of the mixed pixel problems,this paper makes a investigation. We adopt such MODIS-NDVI time series data from 2013 January 1st 2013 to December 19th 2014,reflectance image data of MODIS collected on April 23rd and image data of Landsat,to carry out the winter wheat planting area.Based on the decision tree classification and mixed pixel decomposition method,this paper extracts the winter wheat of Jiangsu province and make an assessment about the extraction the winter wheat area results.(3)This paper uses the data of materials including meteorological data,soil parameters and crop parameters and adopts the meteorological data in the study area meteorological stations from 2001 to 2011.Through the continuous adjustment of developmental parameters on crop growth and previous researches,we adjust the developmental parameters of winter wheat growth and establish model parameters in Kunshan,Huaian and Xuzhou site on behalf of the winter wheat growth assimulation model.At last,the WOFOST model can basically meet the growth and development of winter wheat in Jiangsu province.(4)Through the sensitivity analysis of parameters of WOFOST crop model,this paper determines the key crop parameters influencing the winter wheat growth development and takes crop sowing date,the leaf area and the maximum CO2 assimulation rate as the assimulating parameters for the preparetion of simulated annealing algorithm assimilation model crop parameters.This paper uses Kunshan,Huaian, Xuzhou three sites from 2001 to 2010 winter wheat production statistics and the growth period data to compare adaptation of model localization.(5)Based on crop model localization,this paper uses simulated annealing algorithm and MODIS-LAI assimilation of crop model based on WOFOST.Through the sensitive parameters of sowing date,specific leaf area and the maximum CO2 assimilation rate adjustment and modification,this paper conducts two proups of different LAI data contrast experiment that makes the model simulation of the LAI and MODIS-LAI tends to become smaller.Finally,this paper selects the appropriate model of the assimilation and evaluates the winter wheat yield in the study area. We choose Kunshan,Huaian,Xuzhou 3 stations of winter wheat yield statistics from 2001 to 2010 to compare with assimilation and without assimilation.The results show that two groups of experiments data assimilation can reflect the growth of winter wheat better and improve the accuracy of winter wheat of crop model WOFOST.