Enabling Adoption of Regenerative Agriculture through Soil Carbon Copilots

TL;DR

An AI-driven Soil Organic Carbon Copilot is introduced that automates the ingestion of complex multi-resolution, multi-modal data to provide large-scale insights into soil health and regenerative practices and finds that integrating public data and specialized models enables large-scale, localized analysis for sustainable agriculture.

Abstract

Mitigating climate change requires transforming agriculture to minimize environ mental impact and build climate resilience. Regenerative agricultural practices enhance soil organic carbon (SOC) levels, thus improving soil health and sequestering carbon. A challenge to increasing regenerative agriculture practices is cheaply measuring SOC over time and understanding how SOC is affected by regenerative agricultural practices and other environmental factors and farm management practices. To address this challenge, we introduce an AI-driven Soil Organic Carbon Copilot that automates the ingestion of complex multi-resolution, multi-modal data to provide large-scale insights into soil health and regenerative practices. Our data includes extreme weather event data (e.g., drought and wildfire incidents), farm management data (e.g., cropland information and tillage predictions), and SOC predictions. We find that integrating public data and specialized models enables large-scale, localized analysis for sustainable agriculture. In comparisons of agricultural practices across California counties, we find evidence that diverse agricultural activity may mitigate the negative effects of tillage; and that while extreme weather conditions heavily affect SOC, composting may mitigate SOC loss. Finally, implementing role-specific personas empowers agronomists, farm consultants, policymakers, and other stakeholders to implement evidence-based strategies that promote sustainable agriculture and build climate resilience.

Figures (6)

• Figure 1: Tillage detection for a region of central California for a Sentinel-2 tile (11SKA) which is roughly, 109.8 km by 109.8 km area.
• Figure 2: SOC Copilot Architecture. The copilot processes user queries containing location data (e.g., county names or coordinates) and customizes responses using role-based personas. It retrieves and analyzes relevant data using multi-resolution, multi-modal tools.
• Figure 3: Comparison of SOC Copilot and GPT-4 Responses. For a query comparing regenerative practices' impact on SOC in Riverside vs Marin counties, GPT-4 offers general analysis while SOC Copilot provides nuanced data-driven insights, highlighting environmental conditions as the primary driver of SOC change. See Table \ref{['tab:soc_gpt4_responses_practice']} for complete responses.
• Figure 4: Excerpts of stakeholder-specific SOC Copilot responses. The Agronomist, Farm Consultant, and Policy Maker roles provide role-specific insights for regenerative practices' SOC impact at Orella Ranch, Santa Barbara County. See Table \ref{['tab:persona_responses']} for expanded responses.
• Figure 5: Tillage detection for 4 fields in Washington state.