Science1 publisher2 min readPublished
A PNAS study weighs irrigation's pumping emissions against 6.86 gigatons of avoided land clearing
Colorado State researchers modeled a United States where every field went rainfed, and found the replacement cropland would release 6.86 gigatons of carbon, or 363 years of what irrigation's own pumping emits in a year.
The Scientist · Science desk
What happened
- Colorado State University researchers, writing in PNAS, estimate that replacing the yields lost to a sudden shutoff of U.S. irrigation would emit 363 times what irrigation itself emits.
- Lead author Avery Driscoll puts the avoided land-conversion emissions at 6.86 gigatons, above the 5.91 gigatons the United States emitted in 2024 and about 13 percent of global emissions that year.
- The estimate comes from a global economic model of production, consumption and trade flows, run under a single scenario in which all U.S. crop production is rainfed.
- Most irrigation-related emissions come from the energy used to pump water, both on farms and for interbasin transfers.
- The study excludes other farm emission sources, including nitrogen applied to crops and methane from livestock.
Compiled by The ScientistSomething wrong?How this is made
Why it matters
- decision If the direct emissions sit in pump energy, the choice in front of irrigation districts is which pumps run on electricity; the authors want incentives for that swap written into climate-smart agriculture programs.
- precedent This is the first paper to net irrigation's field emissions against its indirect land-use effects. Anyone scoring an agricultural climate program now has a published two-sided baseline to argue with or against.
- constraint With land use and conversion for food, fiber and fuel near a quarter of global emissions, a program scored only on what happens inside the field is measuring the smaller of the two terms.
Irrigation's own annual emissions fall out of the two published figures. 6.86 gigatons spread over 363 years works out to roughly 19 million tonnes of carbon dioxide equivalent a year [1], about 0.3 percent of what the United States emitted in 2024 [2]. Beyond pump energy, the direct side is small: nitrous oxide released through microbial respiration, and dissolved carbon dioxide dispersing when groundwater is sprayed on fields [8].
Each side of the comparison is a different kind of quantity. Pumping emits every year, while the land saving comes from growing more food on fewer acres [20], and converting grassland or forest to cropland releases the carbon stored in soil and vegetation once [4]. The paper's own phrasing keeps that distinction, putting the avoided emissions at 363 years' worth of today's U.S. irrigation emissions [2].
Before the carbon estimate, the team worked out what going rainfed would cost in yield, county by county, as a ratio of rainfed to irrigated yields built from existing survey data and a machine learning model [10]. Emissions then came off maps of carbon held in biomass and soils, combined with data on how those stocks change when land is converted [13].
Supply, demand and the suitability of different crops in different regions settled where the replacement acres appeared, so the answer is national and international bookkeeping [12]. The scenario removes U.S. irrigation entirely, so it yields no coefficient for a partial cut in a single basin [11].
"The avoided emissions from reductions in indirect land-use change were much greater than the direct emissions, and those could decline further with electrification," said lead author Avery Driscoll, who did the work as a CSU doctoral student and is now a postdoctoral researcher at Purdue [15][16].
Co-author Nathan Mueller, a CSU associate professor in the departments of Ecosystem Science and Sustainability and Soil and Crop Sciences [19], said the team could "show this large benefit of U.S. irrigation to greenhouse gas emissions from the food system" [17]. "Yet, when we talk about water use, particularly in the western U.S., there are trade-offs with every use and trade-offs beyond food and beyond greenhouse gas emissions," Mueller said [18].
What to watch
- Whether the authors publish a partial-curtailment version that prices a basin-scale delivery cut in carbon terms.
- Whether pump electrification incentives show up in the next round of climate-smart agriculture program rules.
- How sensitive the 6.86 gigatons is to where the economic model sends replacement cropland, and whether the county yield ratios hold up against independent yield data.