论文部分内容阅读
随着数据同化方法的不断发展,数据同化已被广泛应用于遥感数据与作物生长模型的结合之中,但在关键物候期遥感数据缺失条件下的同化方法还有待加强研究。以黑龙江省红星农场为研究区,以玉米为研究对象,利用遥感数据与WOFOST模型开展同化方法研究。结果表明:经改进后的集合卡尔曼滤波算法同化,明显改善了误差较大的遥感影像对叶面积指数时序曲线的影响,同时减弱了曲线的锯齿状波动;在田块尺度上,和原始算法同化产量结果相比,R~2提高到0.67,RMSE减少到92.23kg/hm~2;在农场尺度上,R~2提高至0.61,RMSE减少至122.44kg/hm~2。
With the continuous development of data assimilation methods, data assimilation has been widely used in the combination of remote sensing data and crop growth models. However, the assimilation methods in the absence of remote sensing data of key phenophase remains to be further studied. Taking Hongxing Farm in Heilongjiang Province as the research area and corn as the research object, the assimilation method was studied by using remote sensing data and WOFOST model. The results show that the improved ensemble Kalman filter assimilation significantly improves the effect of the remote sensing image with larger error on the leaf area index curve and reduces the jagged fluctuation of the curve. On the field scale, compared with the original algorithm R ^ 2 increased to 0.67 and RMSE decreased to 92.23 kg / hm ~ 2 compared with the same yield. On the farm scale, R ~ 2 increased to 0.61 and RMSE decreased to 122.44 kg / hm ~ 2.