Classification and Conceptualization of Karst Recharge Processes Through Spectral and Change Point Analysis of Drip Water Dynamics

Abstract

Karst aquifers are notoriously heterogeneous, making it challenging to characterize recharge dynamics using traditional, static indicators. Percolating (drip) water records, when analyzed at high temporal resolution, offer insights into flow mechanisms in karst vadose zones. This study utilized karst caves to observe drip rates and identify recharge behaviors. An integrated time series analysis was used combining fast Fourier transform (FFT), cross wavelet transform (XWT), change point analysis (CPA), and clustering. This study analyzes one year of daily drip water and rainfall data from 46 monitoring sites across six karst regions in southeastern Australia. Five major types of recharge emerged from this analysis, including mixed flow dominated by soil (8.7%) or epikarst (37.0%), fracture‐dominated flow (17.4%), storage‐controlled flow (13.0%), and threshold‐controlled flow (23.9%). Across regions, recharge behavior varies with climate, lithology and karst maturity, reflecting contrasting storage and flow conditions. These categories capture temporal patterns, rainfall‐drip synchrony, and abrupt shifts in recharge behavior across diverse karst environments. Notably, the FFT method highlighted overall periodicity of drip counts, while XWT and CPA revealed non‐linear responses and flow paths transitions, often linked to rainfall thresholds and seasonal shifts. These results demonstrate the complexity and heterogeneity of karst systems, where similar rainfall amounts can produce significantly different recharge behaviors depending on site‐specific flow paths and storage dynamics. This framework enables interpretation of drip data sets and provides a tool for karst hydrological research, with direct implications for water resource management under shifting climate regimes.