Quantifying Carbon Emissions From Cropland Expansion in the United States

Abstract

After decades of decline, croplands are once again expanding across the United States. A recent spatially explicit analysis, for example, mapped nearly three million hectares of US cropland expansion between 2008-2012. Quantifying carbon emitted from this land use change (LUC) is critically important for assessing the greenhouse gas intensity of crop production and identifying avenues by which emissions can be reduced. I developed a data-driven model that combines relatively high-resolution maps of cropland expansion with published maps of biomass and soil organic carbon stocks (SOC) to spatially assess and quantify C emissions from observed LUC. Unlike similar approaches used to examine emissions from tropical LUC, my model specifically emphasizes non-forest biomass C stocks and the spatial variation of SOC’s response to LUC. Non-forest sources represent major emission pathways in the US, where new croplands primarily replace grasslands. I found that observed expansion caused, on average, a release of 55.0 MgC ha-1 (SDspatial = 39.9 MgC ha-1), which resulted in total emissions of 38.7 TgC yr-1 (95% CI = 22.5 – 57.7 TgC yr-1). I also found wide geographic variation in both the size and sensitivity of affected C stocks. Grassland conversion was the primary source of emissions, with more than 90% of these emissions originating from SOC stocks. My findings are notably higher than comparable estimates made by the National Greenhouse Gas Inventory and thus highlight the importance of inter-model comparisons and observational constraint. The magnitude of estimated emissions also emphasizes the importance of avoiding LUC emissions and conserving natural C stocks. Using a priori knowledge of likely land sources and land use change drivers, I conclude by suggesting possible policy avenues by which ecosystem C stocks can be conserved.