scieee AI-readable full text Open interactive document viewer

On the potential of modeling thermal diffusivity to identify water fluxes in permafrost rock slopes

Weber, Samuel; Cicoira, Alessandro

Abstract

Purpose Permafrost is warming and thawing at a global scale because of climate change and this has consequences for slope stability. Despite numerous studies focusing on the evolution of permafrost, knowledge of the physical properties of frozen ground is based on a few in-situ measurements and laboratory experiments. There is a paucity of observations on water fluxes in permafrost, which are rapidly changing, due to active layer thickening, ground ice melt, talik formation and modified permeability. Particular attention should be given to changes in thermal regime, an indicator of water-induced permafrost degradation, which are currently inducing increasingly deep-seated slope instabilities. Methods In this study, we identify non-conductive heat flow in mountain permafrost as a potential proxy for water fluxes, using borehole temperature data. We quantify thermal diffusivity based on a linear regression between d2T/dz2 and dT/dt (Nicholson & Benn, 2012) and examine the temporal evolution of thermal diffusivity in mountain permafrost boreholes, using the largest mountain permafrost database worldwide, the Swiss Permafrost Monitoring Network PERMOS. Deviations from the regression line can be used as a qualitative indication of non-conductive heat fluxes. Following Petersen et al. (2022), the empirical estimation of thermal diffusivity based on multiple linear regression, with additional consideration of dk/dz. In addition to the described approach, we calculate conductive heat fluxes using analytical and numerical modeling. Given the one-dimensional heat conservation equation, the non-conductive heat flux is quantified using the difference between the observed and modeled temporal temperature change. Conclusions The systematic analysis of the PERMOS borehole temperature data, with three independent methods, allows us to derive a well-constrained range for the thermal properties of mountain permafrost with different substrates such as talus slopes, ice-rich rock glaciers, and bedrock. From these preliminary results, we establish the possibility of further investigating non-conductive processes governed by thawing and/or water advection. Once concluded, this analysis will represent the basis for many other studies investigating the thermal and mechanical behaviour of mountain permafrost slopes.

Full text

© Marcia Phillips Piz Cengalo, 23 August 2017 (9:30 am) ❯ 3.15 Mio m3 permafrost rock slope collapse ❯ Visual evidence of the presence of ice in the failure plane ❯ Unfortunately, no in-situ data available Borehole temperature data Samuel Weber - @Rock_n_Ice [email protected] or http://www.samuelweber.ch WSL Institute for Snow and Avalanche Research SLF Contact Acknowledgment We thank A. Bast for fruitful discussions and J. Nötzli for the additional explanation to the Swiss Permafrost Monitoring Network (PERMOS) (meta-)data. The borehole temperature data used in this study originate from PERMOS, the PERMOS office and its six partner institutions ETH Zurich (ETHZ), the Universities of Fribourg (UniFR), Lausanne (UniL) and Zurich (UZH), the University of Applied Sciences of Southern Switzerland (SUPSI) and the WSL Institute for Snow and Avalanche Research SLF (SLF) - supported by MeteoSwiss (as part of GCOS Switzerland), BAFU and SCNAT. Challenge Approach Findings ❯ Investigate frozen ground's thermal regime at depth and over time based on PERMOS borehole data. ❯ Model thermal diffusivity (statistically & numerically) and investigate the effect of morphology. ❯ Identify periods with non-conductive heat fluxes. ❯ Permafrost is often the first suspect in rockfalls. ❯ Permafrost warming is considered one of the factors leading to periglacial rock slope instabilities. ❯ Liquid water creates non-linear feedback close to 0°C. ❯ Process understanding hindered by limited field data. Permafrost and the role of water in rock slope failures ❯ Nowadays, permafrost is often quickly suspected of triggering slope failures ❯ Speculation about the role of water in permafrost ... affecting hydraulic properties, e.g. permeability ... advective heat transfer through percolating water ... increased water pressure due to local damming effects very limited field data On the potential of modeling thermal diffusivity to identify water fluxes in permafrost rock slopes S. Weber & A. Cicoira ❯ Typical values for thermal diffusivity in permafrost rock slopes are in the range of 0.5 to 4.5 mm2 s-1. ❯ Thermal diffusivity depends on morphology and thermal conditions, and shows seasonal variability. ❯ Successful identification of days with water fluxes. Permafrost rock slope failures in a warming climate Thermal diffusivity at different sites and with various landformsTemporal evolu�on of thermal diffusivity Iden�fy periods with non-conduc�ve heat flux Swiss Permafrost Monitoring Network PERMOS Permafrost temperatures are measured ❯ in 29 boreholes of 14–100 m depth at 15 sites ❯ with three morphologies: bedrock, talus slope and rock glacier Murtèl-Corvatsch borehole: drilled in 2015, 60 m deep Murtèl-Corvatsch borehole ❯ Thermal diffusivity estimated with LRM approach ❯ Seasonal variations of thermal diffusivity at 5m depth Combination of all PERMOS borehole temperature timeseries ❯ Morphology classification according PERMOS documentation ❯ Ground condition classification according site/depth specific thermal conditions © PERMOS Hillshade: SRTM © Marcia Phillips 0 Temperature (°C) -5 5 10 15 Cicoira et al., The Cryosphere, 2019. Nicholson & Benn, Earth Surf Proc Land, 2012. PERMOS, Database, 2023. Petersen et al., JGR Earth Surf, 2022. Reference Statistical modeling ❯ Single linear regression model (LRM) on dT/dt vs d2T/dz2 ❯ Sliding 2-month time window with daily iteration ❯ Proof the LRM assumption ❯ Select the valid time windows ❯ Non-conductive heat flux can be isolated, when the LRM prediction can not explain the observations. Invert and validate thermal diffusivity through joint sta�s�cal and numerical modeling using three consecu�ve thermistors Numerical modeling ❯ Finite difference method for 1D heat conduction ❯ Minimize RMSE to optimize thermal diffusivity raw LRM output 🠖 invalid LRM output with verified assumptions 🠖 valid valid LRM output with (p < 0.01) & (R2 > 0.5) 🠖 selected for analysis statistical model numerical model ❯ Thermal diffusivity for each station illustrated with violin and box plots ❯ Thermal diffusivity grouped by thermal conditions and classified by morphology ❯ Thermal diffusivity grouped by morphology and classified by thermal conditions "negligible" 2.01 mm2 s-1