Decadal Variability of Infiltration Rates in Soil Aquifer Treatment Systems: Insights from STL Decomposition

Understanding long-term variability in infiltration rates (Ir) in Soil Aquifer Treatment (SAT) systems is important for effective operation, yet the relative contributions of environmental and operational drivers remain insufficiently quantified under field conditions. This study analyzes a 10-year dataset (2015–2025) from 50 recharge basins at the Shafdan SAT facility (Israel) to separate temperature-driven seasonal variability from non-periodic influences. Daily soil temperature and water viscosity were estimated using predictive relationships, and Ir time series were decomposed into seasonal, trend, and residual components using Seasonal-Trend decomposition based on LOESS (STL). The results show that temperature-driven seasonality produces consistent annual patterns but accounts for a limited portion of total variability, with a median contribution of approximately 25% across basins. The remaining variability is dominated by non-periodic components, including long-term trends and short-term fluctuations. Logistic relationships between temperature and the seasonal component describe a consistent, nonlinear temperature dependence. However, the dominance of trend and residual components indicates that operational and site-specific factors play a larger role in controlling infiltration dynamics at the system scale. These findings provide a quantitative framework for distinguishing between seasonal and non-seasonal drivers of infiltration variability in SAT systems, supporting improved interpretation of system behavior under operational conditions.