Harnessing Sentinel-2 imagery and AgERA5 data using Google Earth Engine for developing chickpea mechanistic growth modeling and pre-harvest empirical yield forecast

Precision Agriculture (PA) adoption by farmers is limited by costs and technological complexity. Google Earth Engine (GEE) is used in large-scale crop research but remains underutilized for PA applications. Crop yield variability is widely studied, yet research advancements increasingly widen the gap to practical use. To address this, a GEE platform was established, harnessing Sentinel-2 and AgERA5 for chickpea mechanistic daily simulation of the Total Above-Ground Dry Biomass (TAGDB) and Grain Dry Biomass (GDB). In addition, Sentinel-2 spectral reflectance was used to train an empirical Random Forest (RF) model on GEE to forecast Grain Yield (GY) two months prior to harvest. Both mechanistic and empirical models were evaluated at field scale using GY data from 68 fields (2021–2024), including sub-field evaluation from eight fields. The mechanistic and empirical RF models achieved sub-field GY performance with a coefficient of determination (R²), root mean square error (RMSE), and relative RMSE of 0.49, 1.49 t ha⁻¹, and 19.89%, and 0.24, 1.15 t ha⁻¹, and 15.35%, respectively. At the field scale, the mechanistic model resulted in 0.43, 0.9 t ha⁻¹, and 19.35%, while the RF model achieved 0.37, 0.83 t ha⁻¹, and 17.85%, respectively. The models performed similarly to studies in other crops but with a key advantage - they can be fully executed within GEE. A companion app was built to support both the mechanistic and empirical models within GEE. Chickpea farmers can use the mechanistic model to examine the spatial progression of TAGDB and GDB, both retrospectively and in a near real time manner. The RF forecast model can then be used to anticipate GY variability prior to harvest. The streamlined design of the mechanistic model, together with the empirical model implemented in GEE and the open-source scripts available on GitHub, supports efficient adaptation to additional crops.