Decline in Iran's groundwater recharge

Securing groundwater is critical globally for irrigation and domestic use, and reliance on groundwater is expected to increase as surface water scarcity grows. In Iran—a (semi)arid country and a global hotspot of aquifer depletion—the balance between groundwater withdrawal and recharge has been disrupted for decades, particularly following the replacement of traditional qanats with wells. While withdrawals are extensively monitored across more than one million extraction points, ground-truth nationwide recharge data are scarce due to site-specific processes and complex hydroclimatic controls. Hydrologic models provide large-scale estimates but can be highly uncertain in (semi)arid regions. Building on prior work highlighting widespread depletion and the lack of recharge data, this study estimates groundwater recharge across all 30 hydrological basins in Iran from 2002 to 2017 to quantify trends, identify drivers (natural vs. anthropogenic), and assess implications for water management. The purpose is to provide a comprehensive, data-driven evaluation of recharge decline to inform urgent policy and management interventions in a region experiencing increasing hydroclimatic fluctuations.

The paper situates its work within literature documenting global and Iranian groundwater depletion, uncertainties in large-scale recharge modeling in (semi)arid regions, and the importance of recharge as an index of groundwater availability. Prior studies in Iran reported unsustainable extraction rates, declines in river discharge, shrinking surface water bodies, and aquifer-specific recharge estimates using varied methods. Evidence shows limited or insignificant long-term trends in national precipitation, while changes in snowfall and land subsidence have been reported, suggesting non-precipitation factors and human interventions are pivotal in recharge decline. This study addresses the documented gap in nationwide recharge observations by deriving recharge from national monitoring and withdrawal datasets.

Study area and data: Iran’s groundwater monitoring network, initiated in the 1970s, covers 492 of 609 plains and 666 aquifers, with 12,293 active piezometers measuring monthly groundwater levels. By 2017, >96% of the ~272,000 km² aquifer area was covered. A comprehensive national groundwater database has existed since water year 2001–2002, when ~78% of aquifer areas were covered. The study period (2002–2017) was selected for data availability and adequate coverage.

Recharge estimation: Groundwater recharge for each aquifer was estimated using a water budget approach: Recharge = Qoff + ΔS, where Qoff is egressed water (positive flux out of an aquifer) including anthropogenic withdrawals and natural spring discharges, and ΔS is annual groundwater storage change. ΔS was computed from representative aquifer hydrographs using aquifer area, specific yield, and annual average groundwater level change. Recharge for 666 aquifers was aggregated to basin and national scales.

Components and assumptions: Anthropogenic withdrawals and natural spring discharges are measured via nationwide sampling campaigns and inventories (approximately five-year cycles, plus semiannual checks at representative sites). Evapotranspiration losses from the saturated zone were assumed negligible due to groundwater depths >3 m in almost all aquifers. Groundwater discharge to surface waters was assumed negligible due to >20 m declines in groundwater levels over five decades and widespread river/wetland drying. Lateral outgoing groundwater fluxes were not included (data-limited), acknowledged as a source of uncertainty at aquifer scale, but considered negligible for national aggregation. Only 6 of 666 aquifers cross drainage-basin boundaries, minimizing mismatch impacts between hydrogeologic and surface-water basin delineations.

Data processing and trend analysis: Data aggregation and analysis used Excel (Pivot Tables), ArcGIS, and MATLAB; figures were produced using ArcGIS, MATLAB, Microsoft Office, and Grapher. Temporal trends in recharge and precipitation were quantified using Sen’s slope estimator and tested for significance with the Mann–Kendall test (MAKESENS 1.0).

Uncertainty analysis: 95% confidence intervals for Sen’s slope were computed to quantify trend uncertainty at basin and national scales. Additional uncertainties stem from limited information on aquifer type, specific yield, and unmodeled lateral fluxes; details are referenced to prior documentation of the national database.

• Nationwide groundwater recharge decreased significantly by about 35% from 2002 to 2017 (Sen’s slope −3.83 mm/yr; p ≤ 0.001).
• Annual groundwater recharge trend across basins ranged from −10.30 mm/yr (Great Karoon; p ≤ 0.001) to +1.93 mm/yr (Anzali; p ≤ 0.001).
• Average national recharge during the study period was 39.6 mm/yr, slightly above the upper range reported globally for (semi)arid regions due to Iran’s wetter subregions.
• The national groundwater recharge ratio (recharge/precipitation) declined from 21% (2006) to 14% (2017), averaging ~17%.
• Average national recharge (39.6 mm/yr) exceeded average national runoff (~32 mm/yr, 2001–2016), implying surface water is a main contributor to recharge via seepage.
• No significant long-term change in annual mean precipitation nationally, and in 26 of 30 basins; similar insignificance for evapotranspiration trends, pointing to dominant anthropogenic drivers of recharge decline.
• Evidence of widespread declines in surface water resources: ~56% of rivers experienced reduced streamflow; ~20% of permanent rivers became seasonal; widespread shrinkage of lakes and wetlands, reducing recharge from surface waters.
• Additional contributors include reduced irrigation return flows (due to modernization/efficiency gains), extensive land subsidence lowering soil permeability, and land use/cover changes (deforestation, desertification, urbanization).
• Approximately 422 of 609 plains are in critical/prohibited status (new wells banned except for drinking), reflecting severe overexploitation and diminished recharge.
• Uncertainty bounds for trends provided via 95% confidence intervals (e.g., national Sen’s slope −3.83 mm/yr; 95% CI −4.24 to −2.68).

The study quantifies, for the first time at national scale in Iran, a robust decline in groundwater recharge, addressing the recognized data gap that hindered assessment of groundwater sustainability. The significant nationwide downward trend (−3.83 mm/yr) and shrinking recharge ratio demonstrate that recharge is insufficient to balance withdrawals, intensifying existing depletion. The lack of significant changes in precipitation or evapotranspiration trends implies that human interventions—overextraction, reduced surface water volumes, declining irrigation return flows, land subsidence, and LUC change—are the primary drivers. The finding that recharge exceeds measured runoff indicates substantial surface water–groundwater connectivity, and the widespread degradation of rivers and wetlands likely reduces this key recharge pathway. Basin-level variability (e.g., increases in Anzali vs. sharp declines in Great Karoon, Bakhtegan, Salt Lake, and Gavkhouni) aligns with regional hydroclimate and usage patterns, with dry central/eastern basins showing high recharge ratios but severe storage losses due to dependence on groundwater. Collectively, these results underscore the need to reform water governance, re-balance demand, and manage surface water systems to sustain groundwater recharge.

This work develops a nationwide, observation-driven estimate of groundwater recharge across 666 aquifers and 30 basins in Iran (2002–2017), revealing a statistically significant 35% decline (−3.83 mm/yr), primarily due to anthropogenic drivers rather than long-term changes in precipitation or evapotranspiration. The study highlights the critical role of surface water seepage in recharge and shows that widespread degradation of surface water systems, reduced irrigation return flows, land subsidence, and LUC changes have diminished recharge, exacerbating groundwater deficits. Key policy implications include prioritizing demand reduction over reliance on nonrenewable groundwater, improving governance (shifting from top-down to cooperative, stakeholder-inclusive approaches), and recognizing the limited capacity of engineering solutions like artificial recharge at current scales. Future research should refine recharge estimates with better aquifer property data and lateral flux measurements, integrate snow dynamics and changing precipitation phase into recharge modeling, and couple surface water–groundwater management to restore recharge pathways.

• Assumptions that evapotranspiration losses from the saturated zone and aquifer discharge to surface waters are negligible (based on groundwater depths >3 m and large historical declines) may not hold uniformly and could bias recharge estimates locally.
• Lateral outgoing groundwater fluxes were not included due to data scarcity, introducing uncertainty at aquifer scale (though likely minimal at national aggregation).
• Mismatch between surface-water basin boundaries and hydrogeologic aquifer extents can lead to unaccounted inter-basin exchanges; only 6 of 666 aquifers span adjacent basins, limiting but not eliminating this issue.
• Limited information on aquifer types (confined/unconfined), specific yield, and other properties may affect ΔS estimates.
• Trend analyses depend on the quality and representativeness of national monitoring and withdrawal/spring discharge datasets; uncertainties were partially addressed using 95% confidence intervals for Sen’s slope.