Continental-scale evidence for residence time-driven shifts in physicochemical versus biological control of surface pH in lakes and reservoirs.

Lake and reservoir surface pH is critical to aquatic ecosystem health, influencing carbon cycling, biogeochemical transformations, and contaminant toxicity. Yet, its hydrological controls remain poorly quantified at continental scales. Here, we analyzed an unprecedented national water-quality dataset from lakes and reservoirs across China (N = 350,000) to elucidate how water residence time (RT) regulates surface pH. Using machine-learning models stratified into different RT categories and driven solely by electrical conductivity (EC) and dissolved oxygen saturation (DOS), proxies for physicochemical buffering and biological metabolism, we show that these two variables jointly explain 78 ± 8.6 % (mean±SD across eight RT groups) of observed pH variability. Crucially, RT emerges as a hierarchical regulator that shifts the relative dominance of EC and DOS along its gradient. We observed that the relative contributions of physicochemical and biological processes undergo notable transitions at approximately 50, 170, and 2500 days, corresponding to four broadly distinguishable pH regulatory regimes. Physicochemical control dominates under short RT (<50 days). A transitional regime (50-170 days) marks the shift from physicochemical to biological dominance, while biological processes dominate across a broad intermediate range (170-2500 days), with maximal biological control occurring within 500-1000 days-the peak of the "biological window" and gives way again to physicochemical dominance in long-RT systems (>2500 days). This study establishes a unified hydrological framework for predicting and managing lake acid-base balance under ongoing environmental change.