Abstract
The atmospheric concentration of nitrous oxide (N(2) O) has increased by 23% since the pre-industrial era, which substantially destructed the stratospheric ozone layer and changed the global climate. However, it remains uncertain about the reasons behind the increase and the spatiotemporal patterns of soil N(2) O emissions, a primary biogenic source. Here, we used an integrative land ecosystem model, Dynamic Land Ecosystem Model (DLEM), to quantify direct (i.e., emitted from local soil) and indirect (i.e., emissions related to local practices but occurring elsewhere) N(2) O emissions in the contiguous United States during 1900-2019. Newly developed geospatial data of land-use history and crop-specific agricultural management practices were used to force DLEM at a spatial resolution of 5 arc-min by 5 arc-min. The model simulation indicates that the U.S. soil N(2) O emissions totaled 0.97 ± 0.06 Tg N year(-1) during the 2010s, with 94% and 6% from direct and indirect emissions, respectively. Hot spots of soil N(2) O emission are found in the US Corn Belt and Rice Belt. We find a threefold increase in total soil N(2) O emission in the United States since 1900, 74% of which is from agricultural soil emissions, increasing by 12 times from 0.04 Tg N year(-1) in the 1900s to 0.51 Tg N year(-1) in the 2010s. More than 90% of soil N(2) O emission increase in agricultural soils is attributed to human land-use change and agricultural management practices, while increases in N deposition and climate warming are the dominant drivers for N(2) O emission increase from natural soils. Across the cropped acres, corn production stands out with a large amount of fertilizer consumption and high-emission factors, responsible for nearly two-thirds of direct agricultural soil N(2) O emission increase since 1900. Our study suggests a large N(2) O mitigation potential in cropland and the importance of exploring crop-specific mitigation strategies and prioritizing management alternatives for targeted crop types.