Landslide response to climate change in permafrost regions

Highlights

• Landslide frequency and magnitude are likely to increase as permafrost thaws.

• Landslide style will likely shift to thaw-driven processes temporarily.

• We predict that adjustment to a new equilibrium state will take 10¹–10³ years.

• Research needs include improved geographic representation and long-term monitoring.

Abstract

Rapid permafrost thaw in the high-latitude and high-elevation areas increases hillslope susceptibility to landsliding by altering geotechnical properties of hillslope materials, including reduced cohesion and increased hydraulic connectivity. This review synthesizes the fundamental processes that will increase landslide frequency and magnitude in permafrost regions in the coming decades with observational and analytical studies that document landslide regimes at high latitudes and elevations. We synthesize the available literature to address five questions of practical importance, which can be used to evaluate fundamental knowledge of landslide process and inform land management decisions to mitigate geohazards and environmental impacts. After permafrost thaws, we predict that landslides will be driven primarily by atmospheric input of moisture and freeze-thaw fracturing rather than responding to disconnected and perched groundwater, melting permafrost ice, and a plane of weakness between ground ice and the active layer. Transition between equilibrium states is likely to increase landslide frequency and magnitude, alter dominant failure styles, and mobilize carbon over timescales ranging from seasons to centuries. We also evaluate potential implications of increased landslide activity on local nutrient and sediment connectivity, atmospheric carbon feedbacks, and hazards to people and infrastructure. Last, we suggest three key areas for future research to produce primary data and analysis that will fill gaps in the existing understanding of landslide regimes in permafrost regions. These suggestions include 1) expand the geographic extent of English-language research on landslides in permafrost; 2) maintain or initiate long-term monitoring projects and aerial data collection; and 3) quantify the net effect on the terrestrial carbon budget.

Introduction

Changing climate, including increased average temperature and changing weather patterns, is particularly pronounced at high latitudes (ACIA, 2004; Intergovermental Panel on Climate Change (IPCC), 2013, Intergovermental Panel on Climate Change (IPCC), 2014; Blunden and Arndt, 2017; Francis et al., 2017). Arctic and subarctic average yearly temperatures have increased at 2–4 times the rate of global averages in the last several decades, and are likely to continue warming at faster rates in the coming century (Masson-Delmotte et al., 2006; Screen et al., 2012; Alexeev and Jackson, 2013; Snyder, 2016; Intergovermental Panel on Climate Change (IPCC), 2013). Many of the defining conditions of alpine and high-latitude tundra environments, including the presence of permafrost and permafrost ice, are likely to experience dramatic changes as average temperatures continue to rise (Harris et al., 2009; Christiansen et al., 2010; Romanovsky et al., 2010; Farbrot et al., 2013; Slater and Lawrence, 2013; Westermann et al., 2013; Gisnås et al., 2017; Westermann et al., 2017; Blunden and Arndt, 2017). One such change includes an increase in hillslope soil erosion and landslides in permafrost areas with sufficient topographic relief (Gooseff et al., 2009; Shan et al., 2014b). In this review, we analyze the influence of permafrost thaw on landslide processes and geomorphic implications of changing landslide regimes in permafrost regions (Fig. 1). In addition, we synthesize existing knowledge and identify paths forward based on pertinent research questions that have yet to be addressed.

As used throughout this review, the term “permafrost” refers to ground (soil, sediment, or rock) that remains below 0 °C for at least two consecutive years (Dobinski, 2011). The term “landslide” is used to refer to the types of slope movements in the classification scheme developed by Varnes and Cruden (Varnes, 1978; Cruden and Varnes, 1996) and updated by Hungr et al. (2014), including planar block slides, rotational slides, debris slides and avalanches, flows of various materials, and rockfall. In this review, we focus our discussion on flows and slides, which maintain contact with the bed during downslope transport (Cruden and Varnes, 1996; Hungr et al., 2014), and landslides in permafrost regions where seasonal gradients between frozen and thawed substrate influence slope failure processes (Leibman, 1995; Lewkowicz and Harris, 2005; Lewkowicz, 2007; Blais-Stevens et al., 2015). In recent literature describing the effects of permafrost degradation (e.g. Bowden et al., 2008; Gooseff et al., 2009; Dugan et al., 2012; Lafrenière and Lamoureux, 2013; Hong et al., 2014), the term “thermokarst features” often includes various types of landslides. In these cases, thermokarst subsidence occurs on hillslopes and includes a component of downslope movement. The term “cryogenic landslide” is also used in the literature to specify landslides that occur due to ice-related processes (e.g. Leibman, 1995; Leibman et al., 2014).

To evaluate the direct effect of permafrost thaw on landslide regimes, we explore the literature on landslides in permafrost regions and the pertinent properties of permafrost that influence landslide mechanics. We include both high-elevation and high-latitude systems in this synthesis (Fig. 2) because many of the processes discussed are similar in both types of permafrost. Much of the available literature focuses geographically either in highly sensitive permafrost regions (e.g. northern Alaska) or where population density is high (e.g. the European Alps) (Fig. 3). After introducing this literature, we organize our discussion according to five pertinent questions related to the type, frequency, magnitude, and timescale of expected change to landslide patterns in permafrost regions, as well the ecological impacts of changing landslide occurrence. Most pertinent to this review are the hydrologic and physical effects of permafrost thaw and the loss of permafrost ice, described below, as these processes directly influence the shear strength of slope materials by varying pore pressure, cohesion, and internal friction, with a net effect of reducing shear strength of both bedrock, soil, and sediment.

Landslide process and climate patterns are closely linked (Soldati et al., 2004; Huscroft et al., 2004), with evidence for this correlation throughout the geologic record and apparent in modern observation. Correlative studies indicate that climate variables are controls on landslide process (Borgatti and Soldati, 2010; Gariano and Guzzetti, 2016; Moreiras, 2017; Matthews et al., 2018). Glacier retreat at the end of the last glaciation and increase in precipitation in the mid-Holocene both correlate with periods of increased landslide occurrence in Europe (Soldati et al., 2004; Borgatti and Soldati, 2010), and dates of Holocene rockslides in Norway correlate with periods of high temperature (Matthews et al., 2018). In fact, even when accounting for other factors (anthropogenic activity, seismicity, and changes in vegetation), changes in landslide frequency may serve as effective proxies to indicate climate change in past and present records (Soldati et al., 2004; Borgatti and Soldati, 2010; Leibman et al., 2014).

As human activity increases average global temperature, changing climate conclusively impacts slope stability and typical patterns of landslide occurrence (Gariano and Guzzetti, 2016). This poses a serious challenge for landslide practitioners and land managers, as landslide hazard forecasting is already difficult in a static climate and even more challenging in a changing climate (Coe and Godt, 2012). The specific impacts of climate change on landslide frequency, spatial distribution, and magnitude are poorly understood (Gariano and Guzzetti, 2016), particularly because climate predictions typically provide average conditions across regional spatial scales, while landslide hazards forecasting usually focuses on extreme weather at smaller scales (Coe and Godt, 2012). Although this review focuses on the effects of permafrost thaw on landsliding, it is important to note the diverse effects of climate on landslide occurrence.

In general, changing precipitation patterns increase subsurface saturation and pore pressure; as average temperatures rise, total precipitation is also likely to increase, as well as the frequency of high-intensity rainfall events at northern high latitudes (Orlowsky and Seneviratne, 2012; Kharin et al., 2013). Similarly, melting snow and ground ice (Huggel et al., 2010; Daanen et al., 2012) can increase local pore pressure in response to daily or seasonal fluctuations in temperature. Where rainwater or meltwater persists in the subsurface, increased pore water pressure and reduced shear strength increase the likelihood of slope failure. Landslide initiation commonly occurs when threshold values of cumulative rainfall and intensity are met (Dhakal and Sidle, 2004) during atmospheric events (e.g. Eisbacher and Clague, 1984; Coe et al., 2014; Pavlova et al., 2014; Parker et al., 2016; Patton et al., 2018). Changing precipitation patterns and rapid snow/ice melt are therefore likely to increase the frequency and magnitude of large landslides irrespective of the presence of permafrost (Huggel et al., 2012; Stoffel and Huggel, 2012). Changes in vegetation, soil, and land use that relate to climate may also influence patterns of landsliding, although multi-directional feedback processes introduce complex changes in slope stability, and are therefore difficult to evaluate (Gariano and Guzzetti, 2016).

Glacier retreat as climates warm also reduces lateral support of over-steepened valley walls (Ballantyne, 2002a, Ballantyne, 2002b) and can reduce slope stability by de-buttressing steep hillslopes (Lane et al., 2016), initiating the propagation of stress-release fractures, and contributing to steep hillslopes through crustal rebound (Evans and Clague, 1994; Deline et al., 2015; Moreiras, 2017). Paraglacial adjustment after glacier retreat therefore includes a period of heightened landslide activity (Ballantyne, 2002a; Soldati et al., 2004; Klaar et al., 2014). Destabilization of recently glaciated hillslopes in response to modern climate change has already been observed in the European Alps, Canada and Alaska following the retreat of valley glaciers (Haeberli et al., 1997; Huggel et al., 2012; Stoffel and Huggel, 2012). These effects of meltwater, atmospheric water input, and glacier retreat on landslide occurrence can be seen in both permafrost and seasonally thawed systems.

Although we briefly consider the effects of permafrost degradation on rockfall in our discussion of the impact of permafrost thaw on mechanical properties of the subsurface, the link between permafrost thaw and increased rockfall rates is well established in the literature (e.g. Gruber and Haeberli, 2007; Harris et al., 2009; Allen and Huggel, 2013; Draebing et al., 2017; Ravanel et al., 2017). Readers are directed to an existing review of the effects of climate change on rockfall for further discussion on this topic (Gruber and Haeberli, 2007).