Modeling of carbon sequestration and productivity for maize and oats crops using artificial neural network

Abstract

Climate change presents a critical challenge to global food security, especially as the global population continues to rise. A major driver of this phenomenon is the accumulation of greenhouse gases, particularly CO₂, which intensifies Earth's warming. Key contributors to elevated CO₂ levels include fossil fuel combustion and agricultural activities. However, agricultural systems have the potential to mitigate this effect by capturing atmospheric CO₂. Notably, few models account for the net CO₂ flux in agricultural systems, which is critical for understanding their true carbon sequestration potential. This study introduces a machine learning-based approach to model CO₂ sequestration and productivity in two forage crops, a variety of maize (Zea mays) and oats (Avena sativa), under diverse environmental conditions. The model leverages critical variables such as degree days, NDVI, and water balance. Using an artificial neural network (ANN), the study achieved robust predictive accuracy for both crops, with determination coefficients (R) of 0.95 for maize and 0.96 for oats, and low mean squared errors (MSE = 0.02). These results highlight the model’s high performance and reliability, offering a valuable tool for predicting carbon sequestration and productivity in forage crops while addressing a key gap in net CO₂ flux modeling.