河北省冬小麦干旱预测技术研究

Study on drought prediction technology of winter wheat in Hebei province

本研究以河北省冬小麦干旱综合监测模型为基础,对其包含的土壤相对湿度指数、作物水分亏缺距平指数、降水量距平指数进行未来10天的预测,建立冬小麦干旱预测模型;应用2001年、2008—2010年唐山、涿州、定州、黄骅、深县、栾城、南宫、肥乡8个站气象和土壤墒情资料,以农田土壤水分平衡方程为依据,对0～50 cm、0～20 cm土壤相对湿度预测结果和冬小麦干旱预测模型模拟结果进行检验。结果表明:0～50 cm土壤相对湿度相对误差在10%以下的站点占73%,在10%～15%的占12%,在15%以上的占15%;0～20 cm土壤相对湿度相对误差在10%以下的站点占68.4%,在10%～15%的占13.7%,在15%以上的占17.9%;干旱预测模型预测准确率为77.8%,如果以预测结果与监测结果相差一个干旱等级为正确,则干旱预测与监测结果完全一致,干旱预测模型可满足业务应用的需要,但干旱预测模型预测准确率依赖于降水预报的准确率。

Based on the drought monitoring model of winter wheat in Hebei province,the drought prediction model was established to predict the soil relative moisture index,crop water deficit anomaly index and precipitation anomaly index for the next 10 days.With the meteorological and soil moisture data at the eight stations in Tangshan,Zhuozhou,Dingzhou,Huanghua,Shenxian,Luancheng,Nancheng and Feixiang in Hebei province in 2001 and from 1981 to 2010,and based on the soil water balance principle,the predicted soil relative moisture in 0～50 cm and 0～20 cm depth and the simulated results of drought prediction model of winter wheat were tested.The results showed that,as for the soil relative moisture in 0～50 cm depth,the cases with the relative error lower than 10% was 73% of the total,the cases with the relative error within 10%～15% was 12% of the total,and the cases with relative error higher than 15% was 15% of the total.As for the soil relative moisture in 0～20 cm depth,the cases with the relative error lower than 10% was 68.4% of the total,the cases with the relative error within 10%～15% was 13.7% of the total,and the cases with relative error higher than 15% was 17.9% of the total.The general accuracy of drought prediction model was 77.8%.If the prediction was thought to be correct when the difference between predicted results and monitored results was within a drought level,these two kinds of results were wholly identical with each other. The drought prediction model could meet the needs of business application,but the prediction accuracy was dependent on the accuracy of precipitation forecast.