AI-Assisted Monitoring and Modeling of Greenhouse Gas Emissions from Organically Amended Agricultural Systems

Organic amendments such as biochar and animal manures provide multiple agronomic and environmental benefits, including nutrient enrichment, improved soil health, and enhanced carbon sequestration. Quantifying their influence on greenhouse gas (GHG) emissions, particularly carbon dioxide (CO₂), is essential for assessing their role in climate-smart agriculture. This study integrates artificial intelligence (AI) and machine learning (ML) approaches to monitor and model CO₂ and other GHG emissions from organically amended croplands. A field experiment was conducted at the Prairie View A&M University Research Farm in the Greater Houston Area, Texas, using a factorial design with three replications. Treatments included varying levels of biochar and two types of animal manure (chicken and dairy) applied at four rates (0, low, adequate, and high) to feed corn in 2025. An automated monitoring system was deployed to collect real-time data on soil moisture, soil temperature, greenhouse gas emissions (CO2, CH4, and N2O), weather variables (air temperature, relative humidity, and solar radiation), and plant health indicators. To predict short-term CO₂ emissions, multiple ML models, including Multiple Linear Regression (MLR), Random Forest (RF), and Generalized Additive Models (GAM), were evaluated. In addition, the Classification and Regression Tree (CART) algorithm was used to identify critical thresholds and interactions among environmental variables. Preliminary findings reveal strong interactive effects of soil moisture and temperature on CO₂ fluxes, with notable differences among organic amendment treatments. These results demonstrate the potential of combining AI-driven analytics with sensor-based monitoring to enhance understanding of soil carbon dynamics and guide sustainable, climate-smart land management practices.