Solving groundwater depletion in India while achieving food security

The paper addresses India’s simultaneous challenges of ensuring national food security and reversing severe groundwater depletion, particularly in regions favored by the Public Distribution System (PDS) for procurement (e.g., Punjab and Haryana). Since the Green Revolution, procurement policies combined with subsidized electricity for pumping have concentrated water-intensive crops like rice in areas poorly matched to local climate and soil, prompting widespread groundwater declines (1–3 m/year), high electricity use, and political lock-in. The central research question is whether reforming the government’s procurement geography and crop choices can meet grain and nutritional targets while alleviating water stress and improving farmer incomes. The study situates this within the broader context of climate variability, limited canal irrigation reach, and socio-political barriers to crop diversification, proposing PDS restructuring as a key lever to correct distortionary incentives and improve sustainability.

Prior work has documented how India’s procurement and subsidy regimes have distorted cropping patterns and water use, with repeated calls to expand procurement in eastern India and diversify away from rice-wheat. Davis et al. highlighted national gains from optimized cropping to reduce water use, and earlier Columbia Water Center analyses suggested feasibility of meeting targets via rainfed-optimized allocations. Reviews (e.g., Bhogal and Vatta for Punjab) enumerate barriers to diversification, including risk perceptions and assured incomes under MSP. Field projects (2008–2015) by the Columbia Water Center in India showed rice–water systems as major contributors to depletion and identified three solution pathways: improving on-farm water productivity, pilots to test farmer and state willingness for subsidy reforms, and restructuring procurement. These studies motivate a national, data-driven optimization that integrates climate, agronomy, and economics to propose actionable PDS reforms.

The authors develop a district-level linear programming crop allocation model for the Kharif (monsoon) season to maximize expected net national farm revenue under the Minimum Support Price (MSP) regime. Decision variables are the fractions of current cropped area in each district allocated across 12 major PDS crops (rice, other cereals, pulses, oilseeds, etc.). Expectations are taken over historical climate variability using daily precipitation and temperature data (IMD: rainfall 1901–2009; temperature 1950–2005 with simulations to extend) to estimate potential yields and yield reductions from water deficit. Key components:

• Data: IMD gridded daily rainfall and temperature; CGWB groundwater depth; Directorate of Economics and Statistics crop yields, cropped and irrigated areas, costs of cultivation; MSPs; Census 2001 population; USDA nutrient composition; FAO recommended intakes.
• Yield and water: District-level ET0 via Hargreaves; crop coefficients and FAO methods to estimate seasonal crop water requirements, water deficits, and irrigation needs. Yields are modeled as reductions from potential due to unmet water deficits, with distinct rainfed and irrigated conditions. A uniform irrigation application efficiency (e.g., 30% for rice flood irrigation) is assumed for costs and water accounting.
• Scenarios: (1) Irrigation Zero: no irrigation permitted anywhere; (2) Irrigation Capped: irrigation allowed but limited to current district-level irrigated area (non-expansion), with remaining area rainfed.
• Objective: Maximize expected net revenue = Σ over districts and crops of (MSP minus cost of production) × expected yield × allocated area, minus irrigation costs (including pumping energy costs parameterized by groundwater depth, efficiency, etc.).
• Constraints:
  1. National crop production: Average annual national production of each crop ≥ current national production (food security target).
  2. National nutrition: Aggregate production must meet or exceed FAO-recommended intakes for energy (calories) and selected nutrients (proteins, fats, iron, niacin, folate).
  3. District water: Irrigated area (or irrigation volume) constrained to current levels (Irrigation Capped) or zero (Irrigation Zero); implicit water sustainability via yield penalties from unmet deficits under limited irrigation.
  4. District land: Total cropped area per district fixed at current levels to avoid land-use change.
• Spatial suitability filters restrict crops to districts with compatible soils/climate and existing evidence of cultivation.
• Solution method: Linear programming solved using simplex (lpSolve in R). Outputs include optimal crop spatial allocations, national revenue, production, nutrition, groundwater and energy use, and sensitivity to net unit revenues (MSP − cost). Analyses focus on Kharif-only cropping and procurement targets.

• Feasibility without irrigation: Even with no irrigation (Irrigation Zero), a reallocation of Kharif crops across districts can meet or exceed current national PDS crop production targets and nutritional requirements on average while increasing national farm revenue relative to the status quo.
• Revenue gains:
  • Current national agricultural revenue (Kharif PDS crops): INR 2.90 trillion.
  • Irrigation Zero: INR 3.06 trillion (+5%).
  • Irrigation Capped: INR 3.74 trillion (~+30%).
• Water and energy:
  • Irrigation Zero saves about 146 billion m³ of irrigation water nationally (relative to current patterns).
  • Under Irrigation Capped, aggregate national agricultural energy use reported at ~2.5797 GW compared to ~2625.2 GW currently (as reported), indicating substantial pumping energy savings with optimal reallocation.
• Optimal spatial patterns:
  • Rice: Concentrates in Northern, Central-Northern, and parts of Southern India; sharp reductions (>75%) in procurement-heavy northern/eastern states (Punjab, Haryana, Indo-Gangetic Plain), despite their high yields, reflecting correction of incentive-driven overconcentration that drives groundwater depletion.
  • Other cereals (jowar, bajra, maize, ragi): Increase in Punjab, Haryana, southern/eastern Andhra Pradesh, and Chhattisgarh.
  • Pulses: Increase across Indo-Gangetic Plain districts.
  • Oilseeds: Favor Rajasthan, Gujarat, Maharashtra, Odisha, and Tamil Nadu.
  • These patterns resemble pre–Green Revolution regional crop mixes better aligned with local climates.
• Sensitivity to prices:
  • Spatial allocations for rice, other cereals, and oilseeds are robust to 10–50% reductions in net unit revenue.
  • Pulses’ allocation is sensitive to MSP; careful MSP policy is needed to ensure procurement targeting for pulses.
• State-level distributional effects:
  • Winners (Irrigation Capped net income gains): Andhra Pradesh (+INR 89b), Chhattisgarh (+INR 39b), Gujarat (+INR 141b), Madhya Pradesh (+INR 329b), Odisha (+INR 70b), Uttar Pradesh (+INR 281b).
  • Losses: Maharashtra (−INR 111b), Punjab (−INR 79b), Karnataka (−INR 75b), Haryana (−INR 21b).
  • Punjab: Despite net income losses (≈INR 79–86b), large savings in irrigation water (~9.34 billion m³) and energy (~4366 GWh) under Irrigation Zero; electricity subsidy reductions could partially offset income shortfalls, but groundwater value and depletion reversal (1–2 m/year) must be weighed.
• Nutrition: Optimal allocations meet or exceed recommended intakes for energy, proteins, fats, iron, niacin, and folate, outperforming current nutrition profiles.

The analysis demonstrates that reforming India’s PDS—specifically, where and what crops are procured—can simultaneously achieve food security and national nutrition targets, raise average farm incomes, and substantially reduce irrigation water withdrawals and pumping energy in groundwater-stressed regions. By shifting rice out of severely depleted aquifers (e.g., Punjab, Haryana) into climatically suitable regions and increasing pulses and oilseeds where appropriate, the model corrects incentive-driven distortions from the Green Revolution era. These findings directly address the research question: PDS redesign is an effective, singular policy lever capable of producing significant co-benefits across the water–energy–food nexus. Distributional outcomes are mixed across states, implying the need for transition policies. In states facing income declines (notably Punjab, Haryana, Maharashtra, Karnataka), savings in electricity subsidies and avoided groundwater depletion represent fiscal and environmental dividends that could be recycled to support farmers through price guarantees for diversified crops, investment in higher-value perishable supply chains (fruits, vegetables, dairy) with preservation and risk management, and targeted research/extension to improve yields and reduce costs in new production zones. Sensitivity results highlight the importance of MSP settings for pulses to secure adequate allocation. The authors propose multi-objective policy formulations that incorporate net revenue, equity, and water risk, along with instruments such as crop insurance enhancements and solar-powered shallow irrigation where appropriate, to improve political feasibility and resilience.

The study provides a national, district-level optimization showing that India can restructure Kharif-season procurement to meet current production and nutrition targets while improving net farm income and reducing groundwater depletion and pumping energy, even under a no-irrigation scenario. Optimal spatial reallocations reduce rice in overexploited aquifers and increase other cereals, pulses, and oilseeds in climatically suitable regions, aligning cropping patterns more closely with pre–Green Revolution agroecology. Policy implications include: using PDS as a strategic lever to correct distorted incentives; recycling electricity subsidy savings to fund transition support and income redistribution; and building robust supply chains and pricing mechanisms for diversified and higher-value crops. Future work should expand to multi-season (include Rabi), integrate explicit equity and water-risk objectives, refine energy and groundwater valuation, explore dynamic adoption and behavioral factors, and assess technological options that raise yields or reduce costs in newly targeted regions.

• Seasonal scope: Analysis is limited to Kharif-season crops; Rabi-season dynamics and annualized procurement and nutrition targets are not explicitly modeled.
• Data and parameter assumptions: Fixed irrigation efficiency (e.g., 30% for rice), district-level averages for groundwater depths, and historical climate (1901–2009) drive yield and water estimates; local heterogeneity and recent climate trends may not be fully captured.
• Economic simplifications: Revenues are based on MSP and average costs of cultivation; market price variability, transaction costs, and supply chain constraints for perishable diversification are not endogenously modeled.
• Feasibility and adoption: Socio-political acceptability, transition costs, infrastructure needs, and behavioral responses (e.g., risk aversion, labor markets) are discussed qualitatively but not modeled.
• Spatial constraints: Crop suitability filters rely on existing evidence and broad soil–climate matches; detailed agronomic constraints and technology adoption pathways are simplified.
• Energy metrics: Reported national energy figures are aggregated; unit conventions and load profiles may introduce uncertainty for policy valuation.