Wheat yield prediction using an enhanced WOFOST with soil stratified hydrothermal module driven by GLASS and ERA5-land products over Yellow River Basin

文献类型： 外文期刊

第一作者： Zhang, Jing

作者： Zhang, Jing;Yang, Guijun;Gao, Meiling;Chen, Weinan;Liu, Miao;Zhang, Youming;Tang, Aohua;Li, Zhenhong;Yang, Guijun;Chen, Weinan;Zhang, Youming;Tang, Aohua;Yang, Guijun;Zhang, Lijie

作者机构：

关键词： Yield estimation; Crop growth model; E-WOFOST model; Data assimilation; Soil moisture; LAI

期刊名称：COMPUTERS AND ELECTRONICS IN AGRICULTURE （ 影响因子：8.9； 五年影响因子：9.3 ）

ISSN： 0168-1699

年卷期： 2025 年 237 卷

页码：

收录情况： SCI

摘要： Crop models are effective tools for monitoring crop growth and predicting yield. However, existing crop models rarely consider multi-layer soil moisture (SM), and overlook the SM contributions to crop growth especially in arid regions. Accurate assessment of SM profile distribution and its stress conditions is essential for improving crop growth simulation and yield forecasting, especially in water-limited agricultural regions. In this study, to better predict SM dynamics and crop yield, an enhanced WOFOST (E-WOFOST) with soil stratified hydrothermal module was developed. Firstly, a novel stress factor was proposed based on the SM contributions at different depths and growth stages. Second, the stress factor was integrated into the crop growth model. This approach enabled us to develop a comprehensive crop growth model that incorporates soil stratified hydrothermal process. In this way, leaf area index (LAI) from the Global Land Surface Satellite (GLASS) and SM from ERA5-Land products were assimilated into the E-WOFOST model using the ensemble Kalman filter (EnKF) algorithm to predict winter wheat yield in the Yellow River Basin (YRB). Results showed that the strategy incorporating weights and joint LAI and SM assimilation performed best at the point scale. When SM weights were considered, joint assimilation of LAI and SM (R2 = 0.85-0.90, RMSE = 441.24-740.99 kg/ha, MAPE = 7.83-14.62 %) outperformed LAI assimilation alone (R2 = 0.81-0.84, RMSE = 490.30-778.75 kg/ha, MAPE = 15.75-22.27 %) and SM assimilation alone (R2 = 0.72-0.84, RMSE = 583.10-1089.95 kg/ha, MAPE = 16.25-21.59 %) during 2017-2020. When without accounting for SM weights, dual-variable assimilation outperformed single-variable assimilation in 2018 and 2020, with a non-significant decrease in accuracy in 2017 and 2019. At the regional scale, joint LAI and SM assimilation with the SM weights also provided good crop yield estimates (R2 = 0.60, 0.66, 0.57 and 0.64, RMSE = 543.77, 340.66, 407.62 and 635.56 kg/ha from 2017 to 2020). This newly developed E-WOFOST model offers a valuable reference for improving the crop yield estimation precision and guiding agricultural water management.

分类号：

  • 相关文献
作者其他论文 更多>>

微信小程序