Seasonal ground temperature variation controls stress regime and rock anchor tension in warming permafrost rock slopes

Widespread permafrost degradation and glacier recession have increased rock instability in many mountain regions over recent decades. Despite these significant changes, few studies have investigated bedrock deformation and stress variation using in-situ geotechnical monitoring, primarily due to challenging measurement conditions in steep, perennially frozen rock faces. In this study, we analyze stress changes in a warming and deglaciating permafrost rock slope based on a unique five-year dataset (2016-2020) of loads recorded from three 25 m long, grouted steel anchors at the Kitzsteinhorn, Central Alps, Austria. We interpret load variation as a proxy for deformation along the 18 m free anchor length, providing valuable insights into the deep subsurface, where climate warming effects are more discernible due to the reduced influence of strongly fluctuating meteorologic conditions. Over the five-year observation period, anchor loads exhibited a pronounced seasonality of 40-125 kN, translating to strain rates of 1.3-4.1 mm. Seasonal subsurface warming and active layer formation (~June to September) led to a decrease in anchor loads, whereas seasonal subsurface freezing (~October to December) caused a load increase. During winter and spring (~ January to May), loads remained stable. From 2016-19, seasonal load declines (-80 kN on average) exceeded subsequent winter load increases (+63 kN on average), leading to a gradual, year-to-year decrease in anchor loads. We hypothesize that seasonal load cycling is directly linked to the formation and thawing of subsurface ice. Load increases in autumn and winter strongly correlate with cryosuction-inducing thermal gradients in the subsurface, suggesting that load accumulation is primarily driven by the build-up of segregated ice. Conversely, summer load losses are attributed to the melting of ground ice. Small variations in the maximum thickness of the permafrost active layer—measured in a nearby 20 m deep borehole – directly influence seasonal anchor load minima. This indicates that ice melt at the base of the active layer is a key driver of load (stress) variations and suggests that minor thermal fluctuations can shift the balance toward either net load loss (observed in 2016–2019) or net load gain (observed in 2020).​