Hameed et al. (2025) Groundwater storage changes in the United States using baseflow recession method: Comparison with GRACE and well observations

Identification

• Journal: Journal of Hydrology Regional Studies
• Year: 2025
• Date: 2025-11-20
• Authors: Mehvish Hameed, Munir Ahmad Nayak, Manzoor A. Ahangar
• DOI: 10.1016/j.ejrh.2025.102946

Research Groups

• Water Resource Engineering, Department of Civil Engineering, National Institute of Technology Srinagar, Jammu and Kashmir, India

Short Summary

This study quantifies long-term groundwater storage changes in over 1000 minimally disturbed watersheds across the contiguous United States using a novel event-based baseflow recession algorithm, demonstrating its reliability by comparing estimates with GRACE-DA and well observations.

Objective

• To examine the spatial and temporal variability of the baseflow recession parameter, known as the drainage timescale (K), and its relation to hydroclimatic regions.
• To reliably estimate long-term groundwater storage changes in minimally disturbed watersheds using an event-based baseflow recession analysis.
• To evaluate the consistency of baseflow-derived storage trends with GRACE-DA and well-based storage estimates.
• To investigate the relationship between estimated storage changes and meteorological drivers, particularly precipitation and evapotranspiration.

Study Configuration

• Spatial Scale: Over 1000 (specifically 1144) minimally disturbed watersheds across the contiguous United States (CONUS), with watershed areas greater than or equal to 150 square kilometers.
• Temporal Scale: Two decades (2001–2020) for streamflow, precipitation, and well observations; 2003–2020 for GRACE-DA data. Analysis focused on May to October each year to avoid snowmelt/freezing effects.

Methodology and Data

• Models used:
  • Novel event-based baseflow recession algorithm (exponential decay model: y = y0 * e^(-t/K)).
  • Linear storage-discharge relationship (S = K * yL7) for groundwater storage estimation.
  • Water balance approach (ΔSWB = Precipitation - Evapotranspiration - Runoff) to derive storage changes and human influence.
  • Sen’s slope for trend estimation in well-level data.
  • Linear regression for trend analysis across all datasets.
• Data sources:
  • Daily streamflow data: U.S. Geological Survey (USGS) GAGE-II watersheds (2001–2020).
  • Daily precipitation data: GPM-IMERG Final Precipitation (0.1° x 0.1° spatial resolution, 2001–2020).
  • Well level observations: USGS (~14,000 wells, 2001–2020).
  • Groundwater storage data: GLDAS 2.2 GRACE-DA (0.25° x 0.25° spatial resolution, 2003–2020).
  • Evapotranspiration estimates: NLDAS Noah Land Surface Model L4 Monthly (0.125° x 0.125° spatial resolution, V002).
  • Geospatial data: GAGE-II watershed boundaries (USGS), glacier outlines (NSIDC).
  • Storage coefficients (SC) for aquifers from existing literature (Rateb et al., 2020).

Main Results

• The baseflow recession parameter (drainage timescale K) exhibits significant spatial and temporal variability across CONUS, ranging from 4 to 128 days. Higher K values are found in regions with low permeability (e.g., Appalachians), while lower K values are in alluvial/sedimentary aquifers (e.g., Great Plains). K generally increases from mid-summer due to dry/hot conditions.
• Most watersheds (66%) show rising low flows, primarily due to increased precipitation, while warming and drought-prone regions (e.g., West Coast, Southwest) exhibit declining low flow trends.
• Baseflow-derived groundwater storage changes range from -49.84 mm to 43.04 mm over the 20-year study period, with 66% of watersheds showing increasing trends.
• Baseflow-derived storage trends show over 75% agreement in sign with GRACE-DA estimates.
• 65% of watersheds exhibit similar directional trends when comparing baseflow estimates with well-based storage data (for wells within watershed boundaries).
• Consistent trends across all three methods (baseflow, GRACE-DA, and wells) were found in 57% of watersheds.
• Precipitation is identified as the primary meteorological driver of groundwater storage changes (correlation of 0.31), with evapotranspiration playing a secondary, localized role (correlation of 0.02).
• The estimated human-induced storage variation (ΔShumans) in these minimally disturbed watersheds is relatively small, with maximum negative changes of approximately -170 to -120 mm over 20 years (≈ -6 to -8 mm per year).

Contributions

• Introduction of a novel event-based recession algorithm that effectively characterizes streamflow decline during precipitation-free periods and extracts robust recession parameters.
• First large-scale application of an event-based baseflow recession analysis to estimate long-term groundwater storage changes across nearly 1000 minimally disturbed watersheds in the contiguous United States.
• Comprehensive validation of baseflow-derived groundwater storage trends against independent satellite-based (GRACE-DA) and in-situ (well) observations, demonstrating high agreement across diverse hydroclimatic and topographic regions.
• Establishes baseflow recession analysis as a reliable, high-resolution proxy for groundwater storage monitoring, particularly valuable in regions with sparse well-level observation networks or where satellite data are too coarse for small watersheds.

Funding

• Research fellowship provided by the National Institute of Technology Srinagar.

Citation

@article{Hameed2025Groundwater,
  author = {Hameed, Mehvish and Nayak, Munir Ahmad and Ahangar, Manzoor A.},
  title = {Groundwater storage changes in the United States using baseflow recession method: Comparison with GRACE and well observations},
  journal = {Journal of Hydrology Regional Studies},
  year = {2025},
  doi = {10.1016/j.ejrh.2025.102946},
  url = {https://doi.org/10.1016/j.ejrh.2025.102946}
}

Generated by BiblioAssistant using gemini-2.5-flash (Google API)