From Signatures to Structures: Comparing Dominant Hydrologic Processes and Models Across North America

Understanding how dominant hydrological processes vary across hydroclimatic gradients remains a central challenge in hydrology, particularly because model structures embed assumptions about which processes matter and how they interact, limiting transferability beyond individual catchments. Intensively instrumented research basins provide a unique opportunity to link long-term observations, hydrologic signatures, and model behaviour in a controlled yet climatically diverse setting. Here, we report progress on a comparative analysis of a broad set of intensively monitored research basins that represent considerable hydroclimatic diversity, including arid, semi-arid, temperate rainforest, humid continental, snow-dominated, and humid subtropical environments. These basins are characterized by long-term records of discharge, snow, soil moisture, groundwater, and surface–atmosphere fluxes, as well as open data policies that facilitate inter-comparison.

For each basin, we combine a synthesis of existing process understanding from the literature with a data-driven analysis of long-term observations and a benchmark modelling experiment. Hydrologic signatures derived from hydro-meteorological records are used as proxies for dominant processes, enabling characterization of long-term fluxes and storages, including snow dynamics, evapotranspiration patterns, soil and groundwater storage, baseflow, and streamflow behaviour. In parallel, we implement a common set of model structures within the physically-based SUMMA framework, complemented where appropriate by conceptual models (e.g. FUSE). Model setups are harmonized as far as feasible, including meteorological forcing and spatial discretization, to isolate the influence of process representation rather than data availability and model configuration choices. Calibration results are evaluated using multi-state diagnostics and hydrologic signatures.

The anticipated outcomes are (i) an empirical synthesis of how dominant hydrological processes vary across well-instrumented basins, and (ii) evidence of systematic differences (or lack of such differences) in which model structures and process representations are most suitable under contrasting hydro-climatic conditions. These results are a step on the path towards targeted model development and testing within flexible modelling frameworks, supporting more transferable and process-consistent hydrological models.