Integrating remote sensing and mechanistic modeling for crop monitoring and yield estimation

Agricultural crop production is critical for global food security, yet accurately monitoring and estimating yields remains a challenge. Traditional empirical methods provide useful insights but fail to capture the complex interactions between genetics, environment, and management that drive crop growth. Advances in remote and proximal sensing now enable a shift toward mechanistic and physical models that better represent crop processes, offering more reliable and explanatory predictions.
This research integrates remote sensing data with mechanistic modeling to improve yield estimation. First, retrieved biophysical variables are combined with Gross Primary Production (GPP) fluxes simulated using the SCOPE model and translated into yields with a Harvest Index. Building on this, SCOPE’s Radiative Transfer Model (RTMo) is coupled with the WOFOST Crop Growth Model (CGM), directly linking reflectance to plant traits and yield while reducing reliance on empirical assumptions.
The study also explores the integration of fluorescence measurements as indicators of photosynthetic efficiency, further strengthening the connection between plant function and biomass accumulation. Finally, empirical, mechanistic, and physical approaches are compared at regional scales to assess scalability and reliability.
Overall, this work demonstrates how combining sensing, modeling, and eventually AI can enhance yield predictions and support more informed decisions for farmers and policymakers.