Global patterns and controls of baseflow recession timescales from large-sample streamflow analysis

Baseflow recession analysis provides direct insight into catchment-scale groundwater drainage behavior under low-flow conditions, when river discharge is predominantly sustained by subsurface storage. The recession time constant (K) is widely used to characterize groundwater drainage timescales and describe how catchments release stored groundwater. Although recent studies have applied this metric across regional catchments, the large-scale variability of K and the relative importance of environmental factors governing its spatial variability remain insufficiently constrained, particularly at the global scale. In this study, recession time constants were systematically estimated for global catchments spanning a broad range of climatic and geological settings, following the theoretical framework proposed by Brutsaert. Estimates were derived from baseflow-dominated recession segments extracted from long-term daily streamflow records, enabling a large-sample assessment of both the statistical properties and spatial variability of K across heterogeneous environments. The resulting distribution reveals pronounced clustering around characteristic timescales, while also exhibiting substantial spatial heterogeneity among catchments. The dominant controls on recession timescales were examined using an explainable machine-learning framework based on a LightGBM model combined with SHAP-based interpretation. Drainage porosity emerges as the most influential predictor of K, highlighting the central role of effective groundwater storage capacity in regulating baseflow recession duration. Hydraulic conductivity provides additional explanatory power, reflecting the importance of subsurface transmissivity, whereas soil thickness and drainage density exert secondary but still detectable influences through their effects on storage volume and flow-path organization. These results support a physically consistent interpretation of baseflow recession time constants as emergent properties of groundwater storage and drainage efficiency at the catchment scale. By clarifying the environmental controls on K across diverse settings, this study advances process-based understanding of groundwater–streamflow interactions and demonstrates the utility of recession analysis as a scalable approach for diagnosing subsurface hydrological behavior from widely available discharge data.