Irrigation of biomass plantations may globally increase water stress more than climate change

The study addresses a key dilemma at the nexus of climate mitigation, water resources, and sustainable development: large-scale deployment of bioenergy with carbon capture and storage (BECCS) to achieve stringent climate targets (e.g., limiting warming to 1.5 °C) may require irrigation of biomass plantations, potentially increasing freshwater withdrawals and exacerbating water stress. Water stress already affects an estimated 1.4–4 billion people, and climate change is projected to further increase exposure. While climate mitigation reduces pressure on freshwater resources, current pledges may be insufficient to meet Paris targets, motivating consideration of negative emissions technologies. The authors investigate how irrigation for BECCS compares to unabated climate change in shaping global water stress, using a quantitative global modeling framework and a widely used water stress metric (ratio of human withdrawals to available discharge). They compare a strong mitigation scenario with widespread (partially irrigated) biomass plantations (BECCS; ~600 Mha by 2095) to a higher-warming scenario with marginal bioenergy (CC; ~30 Mha by 2095) and also test whether sustainable water management could offset adverse effects while delivering the same biomass.

The paper builds on extensive literature documenting current and future water scarcity, including estimates of population exposed and the role of climate change and socioeconomic drivers. Prior studies have suggested that mitigation pathways with substantial bioenergy can increase water stress, including a U.S. regional analysis indicating mitigation-driven water stress could exceed climate change impacts (Hejazi et al., 2015). Earlier global assessments explored water scarcity under different climate and mitigation policies, and BECCS water demand studies highlighted potential reliance on irrigation and associated trade-offs. The authors advance prior work by conducting a spatially explicit, global comparison of water stress and its drivers between a strong climate change scenario and one with BECCS-driven mitigation, incorporating transient, scenario-consistent land use for both food and bioenergy, and explicit environmental flow protections and water management assumptions. They also note that previous BECCS water demand studies often lacked such transient, SSP/RCP-consistent land-use dynamics and did not consistently account for environmental flow requirements or improved on-farm water management.

• Modeling framework: Process-based Dynamic Global Vegetation and Water Balance Model LPJmL (0.5° grid; 67,420 land cells). Simulates vegetation dynamics, crop and bioenergy yields, hydrology (green/blue water), and river routing with dams/reservoirs. Includes 12 crop functional types plus other crops, pastures, and two second-generation bioenergy crop groups (woody, herbaceous). Domestic and industrial water use from ISIMIP2b multimodel means. Irrigation techniques (surface, sprinkler, drip) and on-field management (mulching, local storage, conservation tillage) represented. Renewable groundwater is part of river discharge; fossil groundwater not included; return flows routed back to rivers.
• Environmental flow requirements (EFRs): Protection implemented via Variable Monthly Flow (VMF) method with 60%/45%/30% of discharge withheld in low/intermediate/high flow months, based on pristine (1670–1699) mean monthly flows. In scenarios with EFR protection, withdrawals for irrigation can be restricted to maintain EFRs.
• Scenarios and forcing: ISIMIP2b inputs under SSP2. Climate from four GCMs (HadGEM2-ES, MIROC5, GFDL-ESM2M, IPSL-CM5A-LR). Land use patterns for RCP2.6 and RCP6.0 from MAgPIE, ensuring food demand is met via trade and technological change. Periods: historical/transient spinup (1700–2006), analysis for Today (2006–2015) and future (2090–2099). • CC: RCP6.0 climate; marginal bioenergy (~30 Mha in 2090–2099); no sustainable water management; baseline irrigation only on agricultural land per scenario; no BECCS irrigation. • BECCS: RCP2.6 climate; extensive bioenergy (~600 Mha in 2090–2099); 30% of bioenergy area equipped for irrigation; no sustainable water management. • BECCS+SWM: Same as BECCS but with sustainable water management (EFR protection, local water storage, improved on-field irrigation efficiency on cropland and bioenergy) and 45% of bioenergy area equipped for irrigation (higher to compensate for EFR withdrawal limits).
• Determining bioenergy irrigation extent: ISIMIP2b land use assumes yield increases from technological change; LPJmL does not implement such direct physiological technological change, so to match scenario-consistent bioenergy harvests, authors performed a sensitivity analysis with 0–60% of local bioenergy area irrigated (steps of 15%). They selected irrigation levels that reproduce about 50% of the assumed productivity increase over the century (remaining 50% attributed to non-water-affecting technological improvement). This yielded 30% (BECCS) and, under EFR constraints, 45% (BECCS+SWM). The global uniform fraction may introduce irrigation in some water-limited cells where EFRs may effectively restrict to rainfed.
• Water Stress Index (WSI): Computed monthly per grid cell as withdrawals (domestic + industrial + irrigation for crops and bioenergy) divided by total discharge (including renewable groundwater). Annual mean WSI derived from monthly averages. High water stress defined as yearly mean WSI > 40%. Aggregated global area (Mha) and population under high stress computed. Also assessed maximum monthly WSI in supplementary analyses.
• Attribution of drivers: Factorial simulations to isolate contributions of (i) climate (RCP6.0 vs RCP2.6, same RCP2.6 land use, no BECCS irrigation), (ii) land use (RCP6.0 vs RCP2.6 land use, same RCP2.6 climate, no BECCS irrigation), and (iii) irrigated bioenergy (with vs without bioenergy irrigation under RCP2.6 climate and land use). Grid cells classified by dominant driver if one differential exceeds the others by >20%; otherwise undetermined.
• Ensemble approach: All scenarios run under four GCMs to capture climate projection spread; report means and ranges across GCMs.
• Population: SSP2 population trajectories (increase from ~7 billion in 2010 to ~9 billion in 2100) used to calculate exposed population.
• Assumptions and constraints: No fossil groundwater use; land use downscaling from MAgPIE to grid not optimized for local water availability; EFR protection potentially reduces irrigation allocations and crop yields, which are partially compensated by improved water management in SWM scenarios.

• Global high water stress area (annual mean WSI > 40%): Today 1023 (982–1065) Mha; CC 1580 (1520–1613) Mha; BECCS 1928 (1901–1970) Mha. Thus, BECCS roughly doubles area relative to today and exceeds CC.
• Population under high water stress: Today 2.28 (2.23–2.32) billion; CC 4.15 (4.03–4.24) billion; BECCS 4.58 (4.46–4.71) billion. Increases include SSP2 population growth.
• Spatial patterns: CC intensifies stress in existing hotspots (Mediterranean, Middle East, India, NE China, SE and southern West Africa). BECCS expands high stress into regions otherwise not highly stressed under CC or Today, notably eastern Brazil and large parts of Sub-Saharan Africa, aligning with bioenergy plantation locations requiring irrigation.
• Scenario differences: About 2400 Mha (~16% of land) has WSI differences > ±10% between BECCS and CC; 72% of this area has higher WSI in BECCS (mainly Central/South America, Africa, Northern Europe). Areas with lower WSI in BECCS (28%) include Western US, India, SE China, and a belt from the Mediterranean to Kazakhstan.
• Robustness across GCMs: Share of area with higher WSI in BECCS vs CC ranges from 64% (IPSL) to 79% (MIROC), indicating consistent tendency for irrigated BECCS to increase water stress beyond climate change alone.
• Driver attribution: Globally, irrigated biomass plantations are the dominant driver of higher WSI in BECCS due to additional withdrawals. In regions where CC shows higher WSI, differences are driven by land use or climate (with similar extents), e.g., more irrigation on food-producing agriculture in RCP2.6 land use explains Western US patterns; climate-driven water availability differences explain parts of Asia.
• Sustainable water management potential: BECCS+SWM reduces global area under high water stress to 1224 (1167–1327) Mha and population to 3.66 (3.47–3.85) billion, both below CC levels. The globally aggregated area under increased water stress relative to CC decreases from 72% (BECCS) to 37% (BECCS+SWM). Nonetheless, some regions (eastern USA, parts of South America, Central/Southern Africa, Central Europe) remain unable to improve beyond CC across GCMs due to limited water availability.
• Climate targets context: BECCS scenario limits warming to ~1.5 °C (RCP2.6; 1.68 °C in 2100, inter-model mean), whereas CC reaches ~3 °C (RCP6.0; 3.15 °C). Avoiding irrigation would cut bioenergy productivity (e.g., ~50 GtC over the century), potentially jeopardizing 1.5 °C feasibility compared to ~2.0 °C pathways (87 GtC).

The analysis demonstrates a prominent trade-off: achieving stringent climate mitigation via irrigated BECCS can increase water stress globally and regionally, in many areas exceeding the stress induced by unabated climate change. This finding, robust across four GCMs, underscores that large-scale irrigated bioenergy introduces substantial additional freshwater withdrawals, expanding and intensifying high water stress beyond current hotspots. The study also clarifies the mechanisms: irrigation of biomass plantations is the primary contributor to heightened WSI in BECCS, while climate and land-use differences govern areas where CC has higher stress. Importantly, the results indicate a viable pathway to mitigate this trade-off—sustainable water management (EFR protection coupled with improved on-farm water use efficiency and local storage)—which can reduce water stress under BECCS below levels seen under stronger climate change without BECCS. However, implementing such measures worldwide poses logistical, managerial, and financial challenges, and some regions remain constrained by water availability even with optimal management. The broader implication is that mitigation portfolios must integrate water constraints explicitly, as water limitations may shift the balance among negative emission options and influence the feasibility of 1.5 °C pathways. The study calls for incorporating water availability, EFRs, and irrigation decisions into integrated assessment models and for considering the regional distribution of bioenergy to minimize additional water stress.

This work provides a global, spatially explicit quantification showing that irrigating biomass plantations to deliver BECCS sufficient for ~1.5 °C mitigation can, absent additional measures, double the global area and population under high water stress relative to today and surpass the water stress from a ~3 °C climate change trajectory. It identifies irrigated biomass withdrawals as the dominant driver of the additional stress and highlights geographic hotspots. Crucially, it shows that coupling BECCS with sustainable water management—protecting environmental flows and improving irrigation efficiency—can avert much of the added stress, bringing global exposure below that of a higher-warming scenario. The study underscores the necessity of integrating water constraints into mitigation planning and land-use optimization, potentially revising the placement and extent of irrigated bioenergy and expanding reliance on residues and wastes. Future research directions include: coupling biosphere–atmosphere models to capture irrigation-induced moisture recycling and rainfall feedbacks; integrating EFRs and irrigation rules into integrated assessment models; exploring region-specific irrigation thresholds; assessing alternative negative emissions portfolios under water constraints; and evaluating socioeconomic feasibility and investment needs for global implementation of sustainable water management.

• Modeling framework relies on a single global model (LPJmL); results are influenced by external climate and land-use inputs.
• Land-use downscaling from MAgPIE to grid cells did not account for local water availability, potentially overallocating irrigation in water-scarce locations.
• Irrigation fractions for bioenergy are applied uniformly (globally constant percentages), which can introduce irrigation to cells with insufficient water; EFR constraints may effectively revert some to rainfed.
• Fossil groundwater abstraction is excluded; only renewable groundwater (as baseflow) is considered, potentially underestimating withdrawals in regions dependent on nonrenewable aquifers.
• The WSI metric uses withdrawals/discharge; alternative water scarcity metrics might yield different magnitudes or thresholds.
• Atmospheric feedbacks of irrigation and land-use change on precipitation, humidity, and remote moisture recycling are not modeled (would require coupled atmosphere–land modeling), possibly altering local water availability and stress patterns.
• Results are scenario-specific to ISIMIP2b SSP2/RCP2.6 and RCP6.0 trajectories (approx. 1.5 °C vs 3 °C by 2100); deviations in future emissions, tipping points, or socioeconomic pathways would alter outcomes.
• Implementation feasibility of sustainable water management globally (policy, governance, financing) is assumed but not analyzed; crop yield impacts from EFR constraints are balanced by assumed management improvements.