Understanding seasonal variability in shrink-swell clay soils and the impact on parameterisation of soil-water mass transfer models

In soils exhibiting bi-modal porosity, such as shrink-swell clays, understanding the impact of seasonal variability on water movement is crucial for stakeholders. These clay soils alter their structure based on moisture content, bulging and swelling at higher moisture levels or shrinking and cracking during drier periods.

Within soil water models, this seasonal variation of soil parameters is not captured. Saturated hydraulic conductivity and porosity are considered as fixed values. Our hypothesis is that during the winter season simpler modelling approaches, such as single porosity models, can be applied using focused parameterisation and that more complex modelling approaches may be overfitting. During summer, when cracking of these soils is more prevalent, dual permeability approaches should be more adequate and capture the system complexity required. Evaluating the adequacy of these models is vital for identifying critical system controls and for scaling up findings to broader catchment models.

To test this hypothesis, we calibrated and validated single-porosity, dual-porosity, and dual-permeability mass transfer models under dry and wet conditions across three sites using HYDRUS-2D/3D. Volumetric soil moisture data was collected using Delta-T PR2 SDI-12 probes to a depth of 1 m. Parameterization involved field sampling of intact soil cores to 60 cm depth in summer and winter, analysed using laboratory methods (KSAT and HYPROP-2, METER). Additional parameters for dual-permeability models were derived through inverse estimation modelling from field infiltration tests. To determine which model is better able to represent the physical reality under dry and wet conditions, we jointly assessed their performance and complexity/parsimony using Bayesian approaches as parameters are not independent.

This study provides insights into the behaviour of the shrink-swell clay soils in these catchments, offering guidance on their conceptualization and model adjustments to better capture seasonal variability. In catchments such as the River Beane, up to 35% of soils are classified Hanslope clay overlying chalk. These chalk aquifers are critical for drinking water supply and sustaining river baseflows. Therefore, the project outputs are essential for understanding seasonal recharge and support effective water resource management.