Estimating soil N 2 O emissions induced by organic and inorganic fertilizer inputs using a Tier-2, regression-based meta-analytic approach for U.S. agricultural lands

Yushu Xia,Hoyoung Kwon,Michelle Wander
DOI: https://doi.org/10.1016/j.scitotenv.2024.171930
IF: 9.8
2024-04-10
The Science of The Total Environment
Abstract:Consistent methods are essential for generating country and region-specific estimates of greenhouse gas (GHG) emissions used for reporting and policymaking. The estimates of direct N 2 O emissions from U.S. agricultural soils have primarily relied on the use of emission factors (EFs, Tier-1) and process-based models (Tier-3). However, Tier-1 estimates are relatively crude while Tier-3 calculations can be costly. This work addressed this gap by developing a Tier-2, regression-based approach by leveraging a meta-database containing 1883 field N 2 O observations together with environmental and management covariates from 139 studies. Our results estimated higher monthly soil N 2 O emissions (N 2 O m , kg N/ha) during the growing season (0.38) than the fallow period (0.15), highlighting the importance of considering measurement periods when utilizing meta-databases for analyzing N 2 O drivers. Significantly different N 2 O m were found for tillage practices (conventional > no-till: 0.42 > 0.27), fertilizer type (liquid > solid manure: 0.55 > 0.32), and soil texture (fine > coarse: 0.36 > 0.22). The comparisons of the influence of crop type and rotation, water management, and soil order on N 2 O emissions are complicated by regional data availability and interactions among different factors. Additionally, the finding that N 2 O emissions reported based on area (N 2 O m ), N input rate (EF), or yield can alter treatment rankings underscores the need to establish transparent criteria for rewarding or discouraging regionally-based management practices using N 2 O metrics. Finally, we show how General Linear Models (GLMs) can be used to estimate country and regional Tier-2 N 2 O m using a suite of covariates. Our GLMs identified tillage, water management, N input type and rate, soil properties, and elevation as the most influential covariates for the conterminous U.S. The limited accuracy of regional-scale GLMs, however, suggests the need to further improve the quality and availability of GHG and covariate data through concerted efforts in data collection.
environmental sciences
What problem does this paper attempt to address?