Controls on saturated hydraulic conductivity in a degrading permafrost peatland complex

• Published in: Water Resources Research
• Main Authors: Fewster, Richard E., Morris, Paul J., Swindles, Graeme T., Baird, Andy J., Turner, T. Edward, Ivanovic, Ruza F.
• Format: Article in Journal/Newspaper
• Language: English
• Published: 2023
• Subjects: pedotransfer function; hydraulic conductivity; von post score; palsa; peat; permafrost; /dk/atira/pure/sustainabledevelopmentgoals/climate_action; name=SDG 13 - Climate Action; palsas
• Online Access: https://pure.qub.ac.uk/en/publications/e043a400-ea50-417b-b519-cc2e3ad6ac68; https://doi.org/10.1029/2023wr035398; https://pureadmin.qub.ac.uk/ws/files/527115959/2023WR035398.pdf

Permafrost peatlands are vulnerable to rapid structural changes under climatic warming, including vertical collapse. Peatland water budgets, and therefore peat hydraulic properties, are important determinants of vegetation and carbon fluxes. Measurements of hydraulic properties exist for only a limited number of permafrost peatland locations, primarily concentrated in North America. The impacts of thaw‐induced collapse upon properties such as horizontal saturated hydraulic conductivity (K h ), and thus lateral drainage, remain poorly understood. We made laboratory determinations of K h from 82 peat samples from a degrading Swedish palsa mire. We fitted a linear mixed‐effects model (LMM) to establish the controls on K h , which declined strongly with increasing depth, humification and dry bulk density. Depth exerted the strongest control on K h in our LMM, which demonstrated strong predictive performance (r2 = 0.605). Humification and dry bulk density were influential predictors, but the high collinearity of these two variables meant only one could be included reliably in our LMM. Surprisingly, peat K h did not differ significantly between desiccating and collapsed palsas. We compared our site‐specific LMM to an existing, multi‐site model, fitted primarily to boreal and temperate peatlands. The multi‐site model made less skillful predictions (r2 = 0.528) than our site‐specific model, possibly due to latitudinal differences in peat compaction, floristic composition and climate. Nonetheless, low bias means the multi‐site model may still be useful for estimating peat K h at high latitudes. Permafrost peatlands remain underrepresented in multi‐site models of peat hydraulic properties, and measurements such as ours could be used to improve future iterations.