Crop Yield Prediction Based on Agrometeorological Indexes and Remote Sensing Data

Timely and reliable estimations of crop yield are essential for management successful food trade. In previous studies, remote sensing data or climate often used alone in statistical estimation models. this study, we synthetically agrometeorological indicators vegetation parameters to estimate maize Jilin Liaoning Provinces China. We applied two methods select input variables, the random forest method establish models, verified accuracy models three disaster years (1997, 2000, 2001). The results show that R2 values eight established provinces were all above 0.7, Lin’s concordance correlation coefficients 0.84, mean absolute relative errors below 0.14. error was 0.12 Province 0.13 Province. A model built using variables selected by a two-stage importance evaluation can obtain better with fewer variables. final province adopts independent nine Among 11 adopted provinces, ATT (accumulated temperature 10 °C) accounted highest proportion (54.54%). addition, GPP (gross primary production) anomaly August, NDVI (Normalized Difference Vegetation Index) standardized precipitation index two-month scale July as important modeling provinces. This study provides reference selection helpful understanding impact on potential yield.