Exploring Wavelet–Based Approaches to Characterize Piezometric Variability in the Doñana Aquifer

Understanding the relative contribution of climatic recharge and groundwater abstraction to piezometric variability is essential for managing the stressed Doñana/Almonte-Marismas aquifer system. Traditional correlation-based approaches (Pearson, Spearman, cross-correlation) have shown limited ability to represent the intuitive differences observed among piezometric series. This work proposes the use of wavelet analysis to characterize the temporal–frequency relationships between rainfall and piezometry and to explore if these results can qualitatively separate climatic and pumping influences.

We analyze long-term piezometric records (1999-2025) together with monthly rainfall data. This time window was chosen because agricultural groundwater exploitation intensified markedly from the late 1990s, defining the modern hydrodynamic behavior of the aquifer. Time series are standardized and detrended, and raw rainfall is used to exploit the ability of wavelets to isolate frequency-dependent relationships. Selected piezometers represent contrasting hydrogeological conditions, including recharge-dominated areas or zones affected by stable intensive pumping.

Two wavelet-based metrics are computed: (i) wavelet coherence, which measures the shared variance between rainfall and groundwater levels across time and scale; and (ii) contribution, defined as the percentage of piezometric signal energy attributable to rainfall within selected temporal bands (1 year, 1–5 years, > 5 years).

Future work includes calculating contributions across all long series in the aquifer and evaluating the feasibility of spatializing these metrics for both the unconfined and confined sectors. By quantifying rainfall-driven variability across scales, wavelet analysis provides a robust framework to distinguish natural recharge signals from non-natural dynamics (e.g., pumping effects) in complex, non-stationary aquifer systems.