Research On Crop Yield Difference And Climate Impact Based On Remote Sensing Big Data And Model Simulatio

Today,the rapid growth of the world population is increasing the global demand for food.The agricultural system is an important guarantee for food security.The North China Plain is one of the main production areas of winter wheat in China,and its output accounts for more than 50%of the total output of winter wheat in my country.Both the expansion of farmland and the improvement of crop yields can ensure food security.With increasing urbanization and increased competition between farmland and other land uses,it is difficult to increase crop yields by increasing the area under cultivation.Sustainable yield growth must be achieved on existing arable land.Therefore,accurate estimation of crop yield and yield gap is of great significance to improve crop yield.This paper takes the North China Plain as the research area and winter wheat as the research object,and uses various methods to explore the yield gap of winter wheat in this region.The main research contents are as follows:(1)This thesis uses the Process-based and Remote sensing driven crop Yield Model for winter wheat(PRYM-wheat)to simulate the North China Plain using multi-source remote sensing data,meteorological data,soil attribute data,and land use types.Production of winter wheat in 2015-2019.Yield verification through statistical yield data showed that the PRYM-wheat model performed well in simulating the actual yield of winter wheat on a regional scale,showing a decreasing trend from southeast to northwest spatially.In addition,using remote sensing data,meteorological data combined with winter wheat yield data to establish a prediction model based on four algorithms:K Nearest neighbor(KNN),Support Vector Regression(SVR),e Xtreme Gradient Boosting(XGBoost)and Random Forest(RF)to achieve wheat yield forecast.According to the comprehensive results of the evaluation index analysis,the RF model has the best effect,the R~2 value is 0.53,and the RMSE is 630.81 kg/hm~2.The average value of the whole study area is 5763.2 kg/hm~2,which is slightly different from the statistical yield.Compared with the PRYM-wheat model,the RF model has slightly better simulation accuracy,but the PRYM-wheat model can describe the spatial distribution of winter wheat yield in more detail.(2)The PRYM-wheat model is driven by simulating the leaf area index of winter wheat under potential conditions to estimate the potential yield of winter wheat in the North China Plain.The potential yield is mainly distributed around 12000kg/hm~2,and the spatial distribution presents the feature of gradually decreasing from southeast to northwest.Based on the actual yield and potential yield of winter wheat simulated by the random forest model and the remote sensing process model in the North China Plain,the winter wheat yield gap in the study area based on the simulation of the two models was calculated.For the PRYM-wheat model,the average yield gap across the study area was 6400.6 kg/hm~2.For the RF model,the average yield gap across the study area was 6170.0 kg/hm~2.Overall,there is still significant room for increased production in all provinces and cities,and this study provides important support for prioritizing the reduction of yield gap in the future.(3)We divide the entire North China Plain region into four parts,each with different priorities,to increase future production.In general,Jiangsu Province is ranked first in the priority of the two model simulation regions,indicating that narrowing the existing gap can obtain more output.In addition,this study analyzed the response of winter wheat yield simulated by both models to cumulus temperature,precipitation and altitude,showing that cumulus temperature,precipitation and altitude are all important limiting factors for yield.