Towards practice-based greenhouse gas emissions accounting: impacts of US irrigation

Abstract

Agriculture is a key driver of climate change while also being particularly vulnerable to its impacts. Food systems produce approximately one third of total global GHG emissions, with two thirds of those emissions originating from on-farm management and land use change for agricultural use (Crippa et al., 2021; Tubiello et al., 2021). Simultaneously, adverse weather conditions that reduce crop yields are increasing in many areas due to climate change (Lobell & Di Tommaso, 2025; Vogel et al., 2019). Yield losses can further drive increased land use change emissions, as cropland area expands to compensate for reduced production (You et al., 2025). Thus, climate change adaptation and mitigation in the agricultural sector are coupled objectives. In this dissertation, I evaluate the greenhouse gas impacts of irrigation, which is a widely used management practice with strong adaptive benefits. By quantifying both direct emissions associated with irrigation – from pump energy use, increased denitrification, and groundwater degassing – as well as avoided emissions associated with land sparing, we find that irrigation provides synergistic benefits for both adaptation and mitigation goals. Moreover, we identify geographically resolved leverage points for targeted emissions reductions, further boosting the mitigation benefits of irrigation. Following a brief introduction, Chapters Two and Three of this dissertation quantify direct emissions from irrigation (Driscoll, Conant, et al., 2024; Driscoll, Marston, et al., 2024a). Chapter Two estimates direct emissions from on-farm pump energy use at the county level for the US by coupling data on producer expenditures for pump fuels with fuel prices, emissions factors, and irrigation water use. This analysis indicates that on-farm pump energy use in the US produces 12.6 Mt CO2e annually, predominantly from electricity use (69%) and groundwater pumping (85%). Using projections of electrical grid emissions factors under the 2022 Inflation Reduction Act, we demonstrate that on-farm pumping emissions are projected to fall by 46% by 2050 under current policy, with reductions of 70% achievable via modest 'pump electrification' scenario, in which 5% of remaining non-electric pump energy use is electrified each year. Chapter 3 quantifies emissions from three key remaining pathways: 1) energy use for interbasin water transfers using data collected directly from transfer operators, 2) increased N2O emissions due to increased soil moisture using a statistical metamodel of DayCent-based N2O estimates developed for the US GHG Inventory, and 3) CO2 emissions from degassing of supersaturated groundwater using USGS data on well water alkalinity and salinity. We find that these additional pathways are smaller in magnitude, with 1.1 Mt CO2e from interbasin transfers, 2.9 Mt CO2e from elevated denitrification, and 2.4 Mt CO2e from groundwater degassing. However, they are highly spatially heterogeneous and, in the case of N2O and groundwater degassing, much more difficult to mitigate. This work represents the first effort to comprehensively quantify irrigation-related GHG emissions at the national scale. Chapter 4 quantifies the emissions avoided via land sparing attributable to irrigation. First, I estimate the production benefits associated with irrigation in the United States. Using co-located, county-level observations of rainfed and irrigated yields for ten crops, the ratio of rainfed to irrigated yields is modeled as a function of weather and soil conditions. These spatially explicit modeled yield ratios are coupled with crop-specific irrigated area extent to calculate the percent loss of production that would occur in the absence of irrigation, revealing that production losses range from 3 – 39% across crops. These production losses were used as inputs to a global economic model coupled with an empirically driven land use change downscaling algorithm. This model represents consumption, production, and trade responses to economic shocks, and we used it to project global land use change in response to the loss of irrigated crop production in the US. We find that the productivity benefits of irrigation avoid substantial global conversion of forests (-32 Mha) and other natural vegetation (-7 Mha) to grasslands (+35 Mha) and croplands (+4 Mha). Using high resolution data on biomass and soil carbon stocks and responses to land use change, we estimate that a total of 6.85 Gt CO2e emissions are avoided due to irrigation. This is equivalent to 363 years' worth of direct emissions from irrigation, suggesting that irrigation provides substantial net benefits for climate mitigation.