Understanding the hydrological response of groundwater discharge from freezing soils to a warming climate

Objective: Estimate the error introduced by the misrepresentation of soil freezing characteristic curves (SFCCs) in hydrological models and propose an improved method for modelling freezing soils.Key FindingsMost SFCCs used in numerical modelling studies are chosen based on convergence behaviour as opposed to physical soil properties. The choice of SFCC affects model outcomes, including governing the ice content of soils, which in turn controls the permeability and the discharge from a hydrogeologic model. AbstractMore than half of the global terrestrial surface is subject to freezing processes, either as seasonally frozen soils or as permafrost. Soil freezing processes are represented by the soil freezing characteristic curve (SFCC) that relates soil temperature to its unfrozen water content. Unfortunately, SFCCs are frequently misrepresented in models, and often chosen based on ease of model convergence behavior as opposed to physical soil properties. With climate change and increased frequency of midwinter melt, SFCCs are becoming increasingly important in accurately predicting the hydrological response of catchments.Two synthetic hillslopes, one for a permafrost system and one for a permafrost-free system affected by seasonal freezing, are simulated using SUTRA-ice and a selection of widely accepted SFCCs. SFCCs are drawn from literature values as well as a repository of collected SFCC data: "A Repository of 100+ Years of Measured Soil Freezing Characteristic Curves". The resulting discharge is compared for each simulation, showing that the choice of SFCC is important in controlling streamflow generation in these landscapes, and the choice of SFCC may be a previously overlooked controlling process in the hydrological behaviour of catchments with freezing soils. Further work upscaling these results to catchment and larger scales is needed.