Amiha et al. (2026) Inferring groundwater overdraft in data-scarce arid agro-ecosystems: a semi-empirical remote sensing framework applied to the Elfeija watershed, Morocco

Identification

• Journal: Frontiers in Earth Science
• Year: 2026
• Date: 2026-09-11
• Authors: Rachid Amiha, Belkacem Kabbachi, Mohamed Ait Haddou, Youssef Bouchriti, Hicham Gougueni, Mustapha Ikirri
• DOI: 10.3389/feart.2026.1914670

Research Groups

Geosciences, Environment and Geomatics Laboratory, Department of Geology, Faculty of Sciences, Ibn Zohr University, Agadir, Morocco

Short Summary

This study develops a semi-empirical remote sensing framework using Sentinel-2 and Landsat data to infer groundwater abstraction and assess overdraft in the data-scarce Elfeija watershed, Morocco. It quantifies significant groundwater pumping for agriculture, revealing a severe state of overdraft with an 82% probability of exceeding renewable aquifer recharge.

Objective

• To develop and apply a synergistic Sentinel-2 and Landsat time-series analysis (2020–2024) to detect and monitor the distinct thermal-phenological signature of small-scale irrigated agriculture.
• To quantify the volume of groundwater pumped for irrigation by translating the observed surface cooling effect into an estimated evapotranspiration flux, thereby making the “invisible pumping” visible.
• To assess the sustainability of these withdrawals by comparing the estimated pumping volumes with local aquifer recharge data and contextualizing the findings using independent climate and water stress datasets.

Study Configuration

• Spatial Scale: Elfeija watershed, Morocco (125,175 hectares). Analysis focused on 16 irrigated polygons (total 8,780 hectares) and 16 reference polygons (total 11,126 hectares). Satellite data resolutions: Sentinel-2 (10–20 meters), Landsat (30–100 meters), ERA5-Land (approximately 9 kilometers).
• Temporal Scale: 1 January 2020 to 31 December 2024 (5-year period), with monthly composite time-series.

Methodology and Data

• Models used: Semi-empirical thermal-phenological model (ΔLST to ET conversion), Monte Carlo uncertainty propagation, FAO-56 Penman-Monteith equation (for calibration), Surface Energy Balance (conceptual basis), MOD16, ERA5-Land Water Balance, and GLEAM v3.8 (for cross-validation).
• Data sources:
  • Satellite: Sentinel-2 Level-2A (COPERNICUS/S2SRHARMONIZED) for Normalized Difference Vegetation Index (NDVI). Landsat 8/9 Collection2 Level-2 Science Products (LANDSAT/LC08/C02/T1L2 and LANDSAT/LC09/C02/T1L2) for Land Surface Temperature (LST).
  • Reanalysis: ERA5-Land (ECMWF/ERA5LAND/MONTHLYAGGR) for total evaporation, 2-meter temperature, dew-point temperature, 10-meter wind speed, and surface solar radiation. TerraClimate dataset for long-term climatic water balance.
  • Other: WRI Aqueduct 3.0 global water risk atlas for baseline water stress. Local hydrogeological literature for aquifer recharge estimates (4.70 million cubic meters per year). Local agronomic benchmarks for watermelon crop water requirements.

Main Results

• A distinct and repeatable seasonal thermal-phenological anomaly was observed in irrigated plots (2020–2024), characterized by peak NDVI values between 0.20 and 0.28 and surface cooling of up to 5 °C compared to reference areas during the March-April peak growing season.
• The estimated annual groundwater abstraction for the studied plots in 2022 was 5.85 million cubic meters (Mm³) (conservative lower-bound estimate, c = 5.0 mm/month/°C) and 6.79 Mm³ (calibrated central estimate, c_cal = 5.8 mm/month/°C).
• A severe state of groundwater overdraft was identified: the calibrated central estimate (6.79 Mm³) exceeds the estimated renewable aquifer recharge (4.70 Mm³) by 45%. Probabilistic modeling indicates an 82% probability that actual annual groundwater abstraction exceeds renewable recharge.
• The framework successfully disentangled climatic triggers from anthropogenic forcing, capturing an anomalous, water-intensive late-season agricultural cycle in October 2024 (1.20 Mm³ pumping), which showed a 143% surplus evapotranspiration (12.9 millimeters per month) in irrigated plots compared to reference areas.
• The study area is classified as a "High" to "Extremely High" water risk hotspot by the WRI Aqueduct 3.0 dataset, with a mean Baseline Water Stress score of 3.76.

Contributions

• Developed a scalable, cost-effective, semi-empirical remote sensing framework for inferring groundwater abstraction in data-scarce arid agro-ecosystems using multi-sensor satellite data (Sentinel-2 and Landsat).
• Provided the first spatially explicit, multi-annual quantification of "invisible pumping" volumes for small-scale agriculture in the Elfeija basin, establishing a fundamental quantitative baseline.
• Delivered robust, multi-annual evidence of severe groundwater overdraft in a representative arid agro-ecosystem, supported by rigorous Monte Carlo uncertainty propagation and multi-model validation.
• Demonstrated the ability to successfully disentangle climatic triggers from anthropogenic forcing in agricultural water use, crucial for effective water governance.
• Offers a proactive tool for river basin agencies to identify over-extraction hotspots, evaluate compliance with agricultural expansion limits, and provide data-driven baselines for negotiating sustainable water quotas.

Funding

The author(s) declared that financial support was not received for this work and/or its publication.

Citation

@article{Amiha2026Inferring,
  author = {Amiha, Rachid and Kabbachi, Belkacem and Haddou, Mohamed Ait and Bouchriti, Youssef and Gougueni, Hicham and Ikirri, Mustapha},
  title = {Inferring groundwater overdraft in data-scarce arid agro-ecosystems: a semi-empirical remote sensing framework applied to the Elfeija watershed, Morocco},
  journal = {Frontiers in Earth Science},
  year = {2026},
  doi = {10.3389/feart.2026.1914670},
  url = {https://doi.org/10.3389/feart.2026.1914670}
}