| This paper mainly focuses on the improvement of crop yield estimation technique, one of the critical techniques for regional food security early-warning. Huang-huai-hai Plain, which is one of the main grain crop growing regions in China, was selected as the study area. The main objectives of this research are to improve the final yield estimations precision of winter wheat, to explore the technological method based on the integration of crop growth simulation model and spatial information technology for regional crop yield estimation.Firstly, the author extensively reviewed different crop yield estimation methods in this paper. Taken the unique characters of agricultural production and the advantages and disadvantages of different methods into account, this paper finally choose the method based on crop growth simulation model to estimate crop's yield. In this paper, EPIC (Erosion Productivity Impact Calculator) model as one of these crop growth simulation models was used to estimate winter wheat's yield in Huang-huai-hai Plain, because EPIC model simulates the whole growth development more accurately and consider the effect by others aspects well. Secondly, the study areas were chose. Luancheng Ecological Experimental Research Station of Chinese Academy of Sciences was select for the local model validation because there are good relevant research work and background iformation. Hebei study area, including Shijiazhuang, Hengshui and Xingtai, embracing Luancheng, was select for regional yield estimation application area because of the availability of many spatial information data and other supporting background data.For the local model validation, EPIC model firstly was validated based on references so that it can be used locally. Then it was used to estimate winter wheat's yield at Luancheng ecological experimental station. Some parameters were calibrated and sensitivity analysis was done according to observed yield. Lastly, EPIC model was scaled up for regional crop yield estimation. For the regional yield estimation application, the feasibility that spatial EPIC model was used to estimate the regional crop yield was discussed. The main conclusions are as follow:For the local model validation, EPIC model can well simulate the growth condition and forecast crop's yield at Luancheng ecological experimental station. The yield estimation's RMSE was 0.18t/ha, which equals to 2.68ï¼…of mean observed yield. This figure is lower than normal error of EPIC model (15ï¼…). The decision coefficient between estimation yield and observation yield was 0.8433. It was useful for final estimation precision to collect more accurate crop parameters data, such as biomass-energy ratio, harvest index, potential heat unit and input files like temperature and sunshine radiation. The final accuracy was sensitive to the precision of the input management operation.For regional yield estimation application, when the spatial difference of management operation was considered, the decision coefficient between estimation yield and observation yield was 0.4596, which was higher than the decision coefficient without considering the spatial difference. When the input data of the EPIC model are observation, it could well estimate the winter wheat's yield of the Huang-huai-hai Plain.Methodology which crop growth,development and yield formation in local area by crop model and regional scale simulation by combining remote sensing information with crop model were studied and some good results were approached. However, because of complex interdisciplinary involving agronomy, geosciences and remote sensing technique, lack of data, and the limitation of methodology, some problems, such as the daily weather data simulation, winter wheat growth simulation in overwintering, match of spatial size in remote sensing data with crop model and more parameters calculation based on remote sensing data were needed to study further.
|
PDF Full Text Request