Soil shrinkage effects on variably saturated properties and thermal properties of peatland-dominated permafrost mires

Soil shrinkage significantly alters hydraulic and thermal properties in peatland-dominated permafrost regions. This study examines the impact of shrinkage on soil water characteristic curves and thermal conductivity drying curves in Storflaket Mire, Sweden. Seven peat samples were collected at three depths close to the surface. The HYPROP and WP4C devices determined the soil water characteristic curve parameters. The HYPROP device is a transient evaporation experiment that measures soil water potential heads and corresponding volumetric water content. The WP4C measures the dry-range soil water potential and the corresponding volumetric water content. The VARIOS device was used to determine the thermal conductivity drying curves of the peat samples. The shrinkage effects were accounted for by measurements taken with a vernier caliper, followed by validation using a three-dimensional structured light scanner under air-dried conditions. 

The results from the hydrological experiments showed that shrinkage effects were most pronounced in the deepest layers. Comparing cases with and without shrinkage revealed a 40% reduction in volume under air-dried conditions. The hydraulic conductivity curves showed minimal changes between the cases with and without shrinkage, assuming that tortuosity remains constant with shrinkage. Including dry-range measurements was essential for a more reliable soil water characteristic curve representation. Shrinkage alongside dry-range measurements showed that the pore size distribution shifts from macropores (300–3000 μm) to micropores (3–30 μm), indicating reduced bimodality with depth. This change likely explains the higher matric potential in the deepest layers. The results from the thermal experiments revealed near-linear thermal conductivity drying curves, with dry surface peat exhibiting lower conductivity than saturated deeper layers. Empirical models based solely on volumetric water content outperformed traditional parameter-based models in predicting thermal conductivity.