Sensitivity of Modeled Soil NOx Emissions to Soil Moisture

Abstract

AbstractAs emissions of nitrogen oxides (NOx) from fossil fuel combustion decrease, the relative contribution of NOx emissions from managed and unmanaged soils (SNOx) is increasing. Modeling SNOx presents a challenge as it requires proper characterization of emission dynamics in response to environmental conditions. SNOx is often represented using the Berkeley Dalhousie Soil NOx Parameterization (BDSNP), which relies upon static relationships between soil moisture and SNOx for arid and non‐arid lands. However, soil chamber and atmospheric studies have shown that emission characteristics are more dynamic, with peak emissions often occurring at higher soil moisture content. Here, to better capture observational studies, we update BDSNP by creating a dynamic SNOx response to soil moisture based on a normalized soil moisture index. We compare the standard and updated parameterizations over the contiguous United States (U.S.) for 2011–2020 using input soil moisture data from ERA5‐Land, MERRA‐2 and NLDAS2‐Mosaic and evaluate SNOx across these different input drivers as well as between the standard and updated parameterizations. The standard parametrization exhibits strong sensitivity to different input soil moisture products, with annual U.S. SNOx differences of up to 0.28 Tg N yr−1. In contrast, the updated parameterization provides a robust representation of SNOx with reduced sensitivity to input soil moisture product with differences of at most 0.03 Tg N yr−1. The updated parameterization simulates a broad increase in SNOx in non‐arid regions, including much of the Eastern U.S., indicating that this region may be more sensitive to climatically‐driven SNOx as anthropogenic NOx emissions continue to decline.