Zowam et al. (2026) Climate Variability and Groundwater Levels: A Correlation and Causation Analysis

Identification

• Journal: Remote Sensing
• Year: 2026
• Date: 2026-03-19
• Authors: Fabian J. Zowam, A. Milewski
• DOI: 10.3390/rs18060932

Research Groups

• Department of Geology, University of Georgia, Athens, GA 30602, USA

Short Summary

This study investigated the dynamic relationship between terrestrial water cycle intensity (WCI) and groundwater level (GWL) anomalies in arid Arizona, USA, using statistical correlation and causation analyses. It found a dominantly negative, lagged relationship where GWL changes typically precede WCI responses by 1–2 months, implying that an intensified water cycle may signal already depleting groundwater resources.

Objective

• To explore the dynamic relationship between terrestrial water cycle intensity (WCI) and groundwater level (GWL) anomalies in an arid region (Arizona, USA) using statistical correlation and causation analysis.
• Hypothesized that monthly WCI and GWL anomalies over Arizona (January 2010–December 2019) would show a strong, negative contemporaneous correlation, and that WCI anomaly would lead and improve the prediction of GWL anomaly.

Study Configuration

• Spatial Scale: Arizona, USA, focusing on 59 groundwater monitoring well locations. Data were resampled to a 0.125° × 0.125° grid size.
• Temporal Scale: Monthly data from January 2010 to December 2019 (10 years).

Methodology and Data

• Models used: Pearson correlation, Cross-correlation function (CCF), and Granger causality (GC) tests.
• Data sources:
  • Groundwater Level (GWL): Daily observations from 59 monitoring wells, downloaded from the National Groundwater Monitoring Network (NGWMN) data portal, aggregated to monthly averages.
  • Precipitation (P): Global Precipitation Measurement (GPM) Integrated Multi-Satellite Retrievals for GPM (IMERG) algorithm, 0.1° × 0.1° spatial resolution, 30 minutes temporal resolution, monthly averaged from NASA data portal.
  • Evapotranspiration (ET): Synthesized global satellite ET ensemble product, 0.01° × 0.01° resolution, monthly, from Harvard Dataverse repository.
  • Water Cycle Intensity (WCI): Derived by summing resampled monthly precipitation and evapotranspiration rasters.

Main Results

• A contemporaneous linear relationship between WCI and GWL anomalies was largely absent across the monitoring wells.
• Moderate to strong, statistically significant linear relationships were observed at various lags, with the strongest correlations being dominantly negative.
• The strongest correlations indicated that Groundwater Level Anomaly (GWLA) led Water Cycle Intensity Anomaly (WCIA), meaning WCIA responded 1–2 months after GWLA changes.
• The maximum lagged correlation coefficient observed was -0.60, occurring at a +1 month lag (GWLA leading WCIA).
• Granger causality tests confirmed that past GWLA values significantly predict future WCIA values at all six selected wells with strong lagged correlations.
• Bidirectional Granger causality (WCIA also predicting GWLA) was identified at two wells (1 and 33).
• Locations exhibiting the strongest lagged correlations and Granger causality showed a close association with Active Management Areas (AMAs), indicating hydrogeological sensitivity to water cycle shifts in these regions.

Contributions

• Provided a novel application of statistical correlation and Granger causality analysis to investigate the dynamic, lagged relationship between terrestrial water cycle intensity and groundwater levels in an arid region.
• Demonstrated that groundwater level anomalies often precede water cycle intensity anomalies, offering a "backward interpretation" where an intensified water cycle can signal prior groundwater depletion.
• Highlighted the spatial association of these significant lagged relationships with Active Management Areas (AMAs), reinforcing the importance of continuous monitoring and adaptive management strategies in groundwater-stressed regions.

Funding

• Miriam Watts-Wheeler Research Fund, Department of Geology, the University of Georgia.

Citation

@article{Zowam2026Climate,
  author = {Zowam, Fabian J. and Milewski, A.},
  title = {Climate Variability and Groundwater Levels: A Correlation and Causation Analysis},
  journal = {Remote Sensing},
  year = {2026},
  doi = {10.3390/rs18060932},
  url = {https://doi.org/10.3390/rs18060932}
}