Spatial adaptation pathways to reconcile future water and food security in the Indus River basin

The Indus plains of India and Pakistan constitute one of the world’s most productive agricultural regions, relying heavily on irrigation for winter wheat—a staple underpinning regional food security and self-sufficiency goals. However, groundwater depletion, environmental degradation from dry-season withdrawals, and increasing upstream and intersectoral demands threaten water security. Future wheat production faces uncertainty from rising temperatures, altered hydrology, and, critically, continued population growth that increases demand. This study asks how irrigated wheat production in the Indus basin can adapt over time and space to jointly meet food security (SDG2) and water security (SDG6) objectives under uncertain climate and demographic futures. The purpose is to design spatiotemporal adaptation pathways that sequence, time, and locate measures to maintain or improve wheat availability while minimizing irrigation water demand, thereby informing robust, policy-relevant strategies for a water-stressed, transboundary basin.

Prior Indus-focused studies assessed large sets of adaptation measures and their potential but did not specify the magnitude, timing, sequencing, or location of actions required to achieve explicit water and food objectives through time. Adaptation pathways methods have been applied to climate adaptation (e.g., global wheat or flood defenses), emphasizing sequential decisions under uncertainty. Yet, quantitative pathways integrating multiple competing societal drivers (e.g., population growth) and objectives (water and food), especially with explicit spatial dimensions, remain limited. Existing one-dimensional pathways lack the capacity to bridge regional policy objectives and diverse local conditions, creating a scale gap for actionable planning in complex basins like the Indus.

The study develops a spatial, multi-objective adaptation pathways approach combining biophysical modeling and an algorithmic planner. - Study area and scenarios: Lower Indus basin, focusing on irrigated wheat on the Indus plains. Two integrated futures were used: SSP1-RCP4.5 (moderate climate change; population stabilizes by ~2050) and SSP3-RCP8.5 (extreme climate change; rapid population growth). Each climate scenario used four GCMs. - Crop-hydrology simulations: The LPJmL model (5 x 5 arcmin; daily; 1950–2080) simulated wheat yields (rainfed and irrigated) and irrigation water demand. Historical land use (to 2015) was used and then fixed for 2016–2080. The model was calibrated to state/province-level wheat yield statistics and validated against observed production trajectories and published blue water footprints. - Adaptation measures and options: Three measures were considered: (1) Best practices for crop/farm management (BSPR) to achieve management akin to Indian Punjab upper-limit yields; (2) Laser land leveling (LLLV) to reduce irrigation demand and modestly increase yields (soil-type based corrections applied); (3) Sustainable expansion of irrigated area by re-appropriating water savings either partially (PART) or fully up to the 2015 water budget (FULL). Five basin-wide adaptation option datasets were produced per GCM: BSPR; LLLV; BSPR+LLLV; BSPR+LLLV+PART; BSPR+LLLV+FULL. - Spatial Pathways Algorithm: For each year from 2015 onward, the algorithm selects cell-specific adaptation options to meet objectives subject to constraints on wheat production, irrigated area, and water budget. Steps: (1) Evaluate projected basin-level production vs. threshold (using a 5-year rolling average) to determine gaps or overproduction; (2) For each grid cell, identify the most beneficial option (e.g., water footprint reduction or yield increase) against current status, excluding options violating constraints; (3) Rank cells/options by efficiency (e.g., water footprint reduction or yield gain) and iteratively select until the production gap is met (or, if no gap, select options that minimize water footprint); (4) If overproduction, reduce irrigated area in least efficient cells; (5) Update the spatial adaptation map and iterate annually. This yields spatiotemporal pathways with yearly steps. - Pathways configurations: Five configurations (objectives and constraints): Reference (no adaptation); ClimateProof (maintain 2015 wheat levels; prioritize yield increases; no expansion); WaterSaver (minimize water footprint; maintain ≥150 kg/cap/yr; no net area expansion beyond 2015); FoodPrint (maintain ≥175 kg/cap/yr; minimize footprint; allow expansion only after exhausting footprint-beneficial options); FoodSec (maintain ≥200 kg/cap/yr; prioritize yield; allow expansion up to 2015 cell water demand; no area reduction). Population projections consistent with SSP1 and SSP3 set per-capita thresholds. - Pathways ensemble: 5 configurations x 2 scenario pairs x 4 GCMs = 40 unique pathways.

- Baseline pressure without adaptation: By 2080, irrigated wheat production declines ~14% (SSP1-RCP4.5) and nearly 20% (SSP3-RCP8.5) vs. 2015. Per-capita availability falls from ~200 kg (2015) to ~145 kg (SSP1) and ~60 kg (SSP3). Irrigation water demand decreases over time in both scenarios (stronger than production declines) due to climate effects (precipitation change, CO2 fertilization, shorter growing seasons), reducing the water footprint per ton of wheat, especially under SSP3-RCP8.5. - ClimateProof pathways: SSP1-RCP4.5 can maintain ≥150 kg/cap/yr with gradual intensification in lower-yield Pakistani regions until ~2050; population stabilization aids feasibility. SSP3-RCP8.5 requires continuous adaptation, including widespread LLLV in Pakistani Punjab; Indian Punjab has limited scope for further yield gains. Under SSP3, per-capita availability drops below 150 kg by ~2030, indicating climate-only adaptation is insufficient when population grows rapidly. - FoodSec pathways (priority on food): Under SSP3-RCP8.5, sustaining 200 kg/cap/yr is possible only until ~2060; by then all adaptation options are exhausted. Achieving targets requires channeling all water savings from climate and adaptation into expansion, yielding no irrigation water savings vs. 2015; nonetheless, ≥150 kg/cap/yr is retained by 2080, implying basin-level self-sufficiency. Under SSP1-RCP4.5, 200 kg/cap/yr is achieved with relatively few steps and minimal extra irrigation water compared to Reference, focusing on intensification in Pakistani Punjab. - WaterSaver pathways (priority on water): In SSP1-RCP4.5, irrigation water demand can be reduced by over 50% by 2080 while maintaining ≥150 kg/cap/yr, mainly via LLLV and focusing on high water-use areas in Indian Punjab, withdrawing lower-productivity southern areas. In SSP3-RCP8.5, constraints on options that worsen water footprint limit adaptive space; ≥150 kg/cap/yr cannot be maintained after ~2050, but this is the only pathway set that still achieves significant water demand reductions vs. Reference. - FoodPrint pathways (balanced with water-first): In SSP1-RCP4.5, trajectories resemble WaterSaver (LLLV prioritized; footprint reductions with adequate production). In SSP3-RCP8.5, water demand falls sharply initially but rises after ~2050 due to resorting to options not beneficial for footprint to meet production; 2080 demands remain below 2015, but per-capita production drops below 150 kg by ~2070. Only FoodSec preserves self-sufficiency under SSP3. - Overarching insight: Strategic combinations of intensification, LLLV, and targeted expansion can increase wheat and save irrigation water in the short term. However, under rapid population growth, long-term maintenance of high per-capita wheat within existing water budgets is infeasible; trade-offs intensify and priorities must be set.

The pathways show how spatially targeted, sequential adaptation can reconcile water and food objectives in the near term, yet under severe climate change and rapid population growth (SSP3-RCP8.5) mutually beneficial actions are insufficient to sustain per-capita wheat thresholds. Food-prioritizing pathways must emphasize yield gains in Pakistani Punjab (often at the expense of basin-level water footprint), whereas water-focused pathways target improved water productivity in already high-yield Indian Punjab, offering limited production gains and failing to meet long-term production thresholds under SSP3. In contrast, with moderate climate change and population stabilization (SSP1-RCP4.5), multiple objectives (SDG2, SDG6) can be achieved concurrently, and strategies addressing both climate and population changes perform best over the long run. The spatial dimension is crucial: it localizes where measures are most effective, bridging the scale gap between basin-level objectives and local biophysical heterogeneity, and highlighting coherent, low-risk short-term actions (e.g., LLLV in Indian Punjab; intensification in low-yield Pakistani areas). The findings underscore the need to set clear priorities or pursue transformational changes when drivers are unabated, and demonstrate the methodological value of integrating multiple drivers, constraints, and objectives in spatially explicit pathways.

This study introduces a spatial, multi-objective adaptation pathways methodology that integrates climate and demographic drivers to coordinate water and food security goals in the Indus basin. Smart combinations of best-practice intensification, laser land leveling, and targeted, water-budget-constrained expansion can boost wheat production while reducing irrigation water demand in the short term. However, under severe climate change with rapid population growth, current production practices alone cannot sustain long-term per-capita wheat thresholds within existing water budgets, necessitating explicit prioritization or broader system transformations. The approach advances pathways practice by adding spatial explicitness and competing constraints, offering both robust near-term actions and a flexible framework for long-term planning. Future research should incorporate intersectoral water demands and dynamic water security targets, explicit economic costs and feasibility constraints, and explore transformation-oriented pathways beyond technical optimization; the framework is transferable to other irrigation-dependent basins (e.g., Nile, Ganges, Mekong).

The pathways are model-based and simplify complex socio-ecological systems. Only three biophysical indicators (yield, irrigation water demand, sown area) are considered; economic costs, institutional feasibility, farmer knowledge and behavior, and broader policy constraints are not modeled. The adaptation measures are technical and aim to optimize the existing wheat system, which may be insufficient under SSP3-RCP8.5 where system transformation could be required. Land use after 2015 is fixed in simulations, and results depend on GCM selection and LPJmL assumptions (e.g., CO2 fertilization effects). Therefore, feasibility and desirability depend on societal priorities and real-world constraints beyond the technical potentials shown.