Global energy use and carbon emissions from irrigated agriculture

Irrigated agriculture accounts for about 70% of global water withdrawals and 80–90% of consumptive water use, producing roughly 40% of global food on 22% of cultivated land. As climate change and population growth intensify water and food security challenges, irrigation is increasingly critical for climate adaptation and food production. Yet, irrigation often depends on fossil energy for pumping, contributing to agricultural greenhouse gas emissions. While many studies have quantified emissions from land use, fertilizers, livestock, and food systems, direct energy-related emissions from irrigation are less resolved globally. Existing estimates focus on specific countries or regions. This study addresses the global knowledge gap by quantifying, with spatial detail, the energy use and CO2 emissions of irrigation, comparing surface versus groundwater sources, diesel versus electric pumping, and irrigation technologies (surface, sprinkler, drip). It also evaluates CO2 emissions from groundwater degassing during irrigation, projects future energy and emissions under sustainable irrigation expansion in a warmer climate, and assesses mitigation options and their feasibility to support pathways toward net-zero agriculture.

Prior research quantified agricultural GHG emissions broadly (land-use change, synthetic fertilizer production and use, livestock, and food supply chains) and produced global/regional datasets of agricultural emissions. Several studies estimated irrigation-related energy and GHGs at national or regional scales (e.g., China, India, Mediterranean region, Pakistan, United States). Indirect irrigation-related emissions (e.g., methane from reservoirs, canals, rice paddies, and nitrous oxide under varying nitrogen use efficiencies) have been examined. However, a global, spatially explicit assessment of energy-related GHG emissions exclusively from irrigation and pumping systems was lacking, leaving uncertainty about irrigation’s contribution to total agricultural emissions and mitigation potential.

- Scope and spatial resolution: Global, spatially explicit assessment at 10×10 km (5 arc-minute) resolution for 2000–2010. - Systems covered: Irrigation systems (surface, sprinkler, drip); pumping systems (diesel, electric); water sources (surface water, groundwater). Includes direct energy for pumping and delivery and associated CO2 emissions. Also quantifies CO2 emissions from groundwater degassing during irrigation. - Datasets: Groundwater table depth (Fan et al.); share of surface vs groundwater for irrigation (Siebert et al.); national fractions of drip, sprinkler, surface irrigation (Jägermeyr et al.); reconstructed irrigation water withdrawals (LPJmL-based; Huang et al., calibrated to FAO AQUASTAT and USGS); national electricity carbon intensities (IEA, Our World in Data, including electricity trade); diesel emission factor (320.21 g CO2/kWh). Groundwater pumping volumes from WaterGAP 2.2 (Döll et al.) and sectoral allocations. - Energy requirement calculation: Energy equals irrigation water volume multiplied by total pressure head divided by efficiency (pump and prime mover). Total head includes lift (groundwater table depth), drawdown (cone of depression and well inefficiency), operating pressure (assumed typical pressures: surface ~0.41 bar, set 0 for surface sources; drip 1 bar; sprinkler 3 bar), and pipe friction losses (~0.69 bar). Pump and motor efficiencies and drawdown estimated from literature and supplementary methods; spatial resampling aligned datasets. - Emissions from energy use: CO2 emissions computed as energy consumption times energy-specific emission factors. Electricity carbon intensity reflects domestic generation mix adjusted for electricity trade. Diesel CO2 per kWh as above. - Groundwater degassing: CO2 emissions computed as pumped groundwater volume for irrigation times the irrigation share of total groundwater withdrawals times groundwater CO2 concentration. Groundwater bicarbonate concentrations (US aquifer survey) used to estimate a global range of CO2 concentration, converted via stoichiometry; resulting CO2 concentration range ~34–106 mg/L. - Future sustainable irrigation expansion: Additional irrigation is allowed only where local water availability can meet irrigation water demand in a 3 °C warmer climate (circa 2050), holding current country-level irrigation technology shares constant. Energy and CO2 for added withdrawals are computed; for emissions, electric pumping and projected regional 2050 electricity carbon intensities are assumed. - Mitigation scenarios: (1) Irrigation system upgrades: replacing surface with sprinkler or drip according to global water-saving efficiencies (drip saves 68% vs surface; sprinkler saves 44% vs surface; drip saves 43% vs sprinkler). Sensitivity ±5% on savings. (2) Electrification and low-carbon electricity: convert diesel to electric pumps; supply electricity from 2000–2010 mix, solar, wind, nuclear, hydropower, or a 2050 low-carbon mix (life-cycle carbon intensities: solar 44, wind 11, nuclear 12, hydropower 23 g CO2/kWh; 2050 mix ~25 g CO2/kWh). Regional targets used for low-carbon electricity feasibility by 2050. - Feasibility analysis: Quantifies country-level feasibility of drip adoption (based on crop shares that can use drip and current drip share) and of low-carbon electricity (increase from current low-carbon share to regional 2050 targets). Potential contribution to CO2 reduction from each option and their combined effect computed. - Comparison with other farm operations: Energy and CO2 intensities for fertilizers (N, P2O5, K2O), machinery manufacture and fuel use estimated using FAOSTAT activity data and literature conversion/emission factors, to contextualize irrigation’s share of farm energy and emissions. - Uncertainty and caveats: Spatial mismatch between water source shares and withdrawals reduces covered volume; pump type shares inferred from grid coverage (except US county data); life-cycle carbon intensity ranges for low-carbon power; pre-2010 irrigation withdrawal maps limit temporal currency; pesticides not included.

- Global totals (2000–2010): Irrigation consumes 1896 PJ/year of energy and emits 222 Mt CO2/year in total, of which 216 Mt CO2/year are from energy use and ~6 Mt CO2/year from groundwater degassing. - Share in agriculture: Irrigation accounts for about 15% of total agricultural energy use and GHG emissions; irrigation energy input intensity contributes 32% of on-farm energy intensity globally and >50% in several regions (Sub-Saharan Africa, Middle East and North Africa, South Asia, Latin America and the Caribbean). - Groundwater dominance: Although groundwater supplies only ~40% of irrigated area, groundwater pumping accounts for 89% (1670 PJ/year) of irrigation energy and 90% (193 Mt CO2/year) of energy-related CO2 emissions. - Technology and fuel intensities: Average energy and CO2 intensities per m3: sprinkler ~1.8 MJ/m3 and 188 g CO2/m3; drip ~1.0 MJ/m3 and 109 g CO2/m3; surface ~0.5 MJ/m3 and 58.5 g CO2/m3. Diesel pumping is more energy- and carbon-intensive (1.2 MJ/m3; 106 g CO2/m3) than electric pumping (0.5 MJ/m3; 69 g CO2/m3 on 2000–2010 mixes). - Regional and national distributions: Asia uses 72% of global irrigation energy; North America 14%. Top energy users: India 535 PJ, China 299 PJ, Pakistan 135 PJ, Iran 121 PJ, United States 205 PJ (US is top in North America). These five contribute ~68% of global irrigation energy use. Top CO2 emitters (energy-related): India 70 Mt, China 35 Mt, United States 24 Mt, Iran 13 Mt, Pakistan 12 Mt (72% combined). Groundwater degassing is largest in India (2.9 Mt) and the United States (1.4 Mt); in some hotspots (US, India, Pakistan, Iran, Saudi Arabia) degassing exceeds 20% of total irrigation CO2. - Energy/emissions intensities per area: Median energy intensity 2.655 GJ/ha and CO2 intensity 259 kg CO2/ha, with Asia highest (8.0 GJ/ha; 1063 kg CO2/ha), followed by Africa, South America, North America, Oceania, and Europe (1.9 GJ/ha; 218 kg CO2/ha). - System-level flows (Sankey): Diesel pumping provides ~74% (1234 PJ/year) of groundwater pumping energy; electric pumping 26% (436 PJ/year). Surface irrigation systems account for 75% (1400 PJ/year) of energy use and 76% (162 Mt CO2/year) of energy-related CO2. Groundwater extraction using diesel with surface irrigation contributes 57% (1065 PJ/year) of total energy and 45% of energy-related CO2. - Future sustainable expansion (3 °C warmer circa 2050): Additional irrigation would require ~536 PJ/year (+28% over 2000–2010). Regional additions exceed 50% of current use in North America (+139 PJ), Africa (+63 PJ), South America (+60 PJ), and Europe (+148 PJ, ~2× current). Top countries for added energy: United States 97 PJ, India 49 PJ, Russia 39 PJ, Brazil 39 PJ, Mexico 18 PJ (~45% of total). Assuming electrification and projected 2050 electricity carbon intensities, added energy-related CO2 ~15 Mt/year (~7% of 2000–2010 energy-related irrigation CO2), with India and Russia ~3 and ~2 Mt, respectively. - Mitigation potentials: Under technology-only scenarios, global irrigation energy can be roughly halved (drip scenario energy ~894 PJ/year; electric pumping scenario ~930 PJ/year). For CO2, drip reduces to ~100 Mt/year, sprinkler increases to ~295 Mt/year; electrification with 2000–2010 mix ~175 Mt/year; with low-carbon supplies: solar ~11, wind ~3, nuclear ~3, hydropower ~6, and 2050 low-carbon mix ~6 Mt/year. Considering country-level feasibility, combined drip plus low-carbon electricity could cut energy-related CO2 by ~55% globally by 2050, with ~82% of this reduction from low-carbon electricity and ~18% from drip. Some regions (Middle East and North Africa, North America, Western Europe) could achieve >60% reductions; Eastern Europe and Central Asia ~15%.

The study quantifies, for the first time at global and high spatial resolution, the direct energy use and associated CO2 emissions attributable to irrigation pumping and delivery. Findings show that irrigation is a major component of agricultural energy use and emissions, concentrated in regions dependent on groundwater and diesel pumps. This highlights key levers for mitigation: reducing pressure heads (where feasible), improving irrigation efficiency with drip systems (where agronomically appropriate), and shifting from diesel to electric pumps powered by low-carbon electricity. However, technology choice entails trade-offs: sprinkler systems, despite saving water, have higher energy and CO2 intensities than surface irrigation in many contexts. Drip systems can reduce both water and energy/CO2, but are not applicable to all crops and may increase energy use when shifting from gravity-fed systems, necessitating context-specific benefit–cost analyses that internalize carbon costs. Electrification coupled with low-carbon electricity is the dominant pathway to deep emissions reductions, although farmer-level incentives may be weak without carbon pricing or supportive policies. Groundwater management is crucial to prevent falling water tables from eroding energy efficiency gains, and where possible, prioritizing surface water or shallow groundwater can reduce energy demand. The projected energy needs of sustainable irrigation expansion imply additional strain on national energy systems (e.g., Pakistan, Bangladesh already see irrigation at ~4% of national energy supply), underscoring the importance of integrated water–energy–food planning. By comparing irrigation’s energy and emissions intensities to other on-farm operations, the study situates irrigation as a substantial share of farm energy use, reinforcing its relevance for climate-smart agriculture strategies.

This work maps and quantifies global energy consumption and CO2 emissions from irrigation, disaggregated by water source, pumping energy, and irrigation technologies. It demonstrates that groundwater pumping—especially via diesel—dominates energy use and emissions, and that targeted mitigation through drip irrigation (where suitable) and electrification with low-carbon power can halve energy use and reduce CO2 emissions by up to ~90% in idealized scenarios; when feasibility constraints are considered, global CO2 reductions of ~55% are still achievable. The analysis also projects the additional energy and emissions associated with sustainable irrigation expansion in a warmer climate and identifies national and regional hotspots. These insights provide an evidence base for prioritizing investments in irrigation technology, energy system decarbonization, and groundwater management to enhance resilience while reducing the agricultural carbon footprint. Future research should update estimates as newer, post-2010 spatial irrigation withdrawal datasets become available, refine pump-type distributions with country data, better constrain life-cycle carbon intensities of low-carbon power by region, incorporate pesticide-related energy, and integrate socioeconomic constraints and incentives affecting irrigation technology adoption.

- Data coverage and spatial mismatch: Irrigation water source shares and withdrawal datasets do not perfectly align spatially, resulting in 2451 km3 of withdrawals being modeled versus a global 2588–2673 km3 benchmark; thus, results are likely lower-bound estimates. - Pump type shares: Country-level distributions of diesel vs electric pumps are largely inferred from grid coverage in irrigated areas (except detailed US data), introducing uncertainty. - Electricity carbon intensity: Life-cycle carbon footprints for low-carbon electricity (solar, wind, nuclear, hydropower) vary widely by technology and region; using typical values introduces uncertainty in mitigation estimates. - Temporal limitations: Spatially explicit irrigation withdrawal datasets are mainly pre-2010; post-2010 changes are approximated using irrigated area updates, not full hydrological reconstructions. - Groundwater degassing: Global groundwater CO2 concentrations inferred from US aquifer bicarbonate data may not capture regional variability. - System applicability: Drip irrigation is not suitable for all crops/contexts; sprinkler systems can increase energy and CO2 despite water savings, affecting generalizability of mitigation impacts. - Other inputs: Energy and emissions from pesticides were not included due to data limitations; machinery fuel type mixes are simplified with average emission factors. - Policy assumptions: Regional 2050 low-carbon electricity targets and COP28 policy signals may change; country-specific trajectories are uncertain, affecting projected mitigation potentials.