Method for Dynamic Forecast of Corn Yield Based on Climatic Suitability

Summer corn is one of the major grain crops in Hebei Province.Its growing development and yield formation are influenced by weather conditions during the growing and maturity seasons. So it is of great significance to forecast the yield dynamically for agricultural production and food security of Hebei Province. Operational weather forecast for crop yield has a history over 20 years in China, laying the foundation for the application of yield forecast. How ever, many of the crop yield forecasting methods are carried out at fixed time. Dynamical forecast that tracks the whole growth period of crops and considers the three typical climatic factors (sunshine, temperature and precipitation), is rarely carried out.There are some attempts that establish comprehensive climate models based on related analysis between crops and weather conditions considering the whole growth period of crops affected by the three factors. But it hasn't been done much to use the climate suitability for yield forecast.Meteorological data from the Hebei Meteorological Bureau are adopted, including temperature, precipitation and sunlight hours per ten-day period from 1972 to 2005 in the eight representative summer corn producing cities (Tangshan, Langfang, Baoding, Shijiazhuang, Cangzhou, Hengshui, Xingtai and Handan). The annual corn yield per unit data from the Hebei Province Statistics Bureau are used for research. Based on the physiological characteristics of summer corn, temperature suitability model, precipitation suitability model, sunshine suitability model and general climatic suitability model are established. Then the climatic suitability of every ten days during the corn growing period over the eight cities is calculated for the years of 1972 -2005. Correlation analysis results betw een climatic suitability and the corresponding yield fluctuation quantity indicate that the relationship betw een them is remarkable, and the climatic suitability model of summer corn can reflect the climate and its dynamic changes in Hebei Province objectively. Taking the climatic suitability of every ten days during the growing period as the basis, the dynamical forecasting models of eight regions from 1972 to 2005 in Hebei Province have been established at different stages using the statistical analysis software SPSS.The average accuracy of the forecasting model by yield fitting validation is 88.8% for historical forecasting during 1972-2005 and 96.8% for a rolling yield forecasting during 2006-2007 respectively, proving the model applicable to operational service. The model should be further optimized in future work owing to imperfection in the precipitation suitability model, and the lagged effect of pre-precipitation should be brought into consideration.