Thermomechanical stress–strain numerical modelling of deglaciation since the Last Glacial Maximum in the Adamello Group (Rhaetian Alps, Italy)

Highlights

• We developed an evolutionary model of valley deglaciation in the Alps.

• Geomorphologic analysis and stress–strain simulations model the release after LGM.

• Slope morphology controls thermal variations peaked after the last deglaciation.

• Outputs of thermomechanical modelling are not consistent with DSGSDs in Adamè Valley.

• We infer LIA negligible effects on thermomechanical trend over the next 1000 years.

Abstract

Deglaciated areas in the valleys of the Italian Alps have recently exhibited a high potential for geomorphologic hazards from diffuse deep-seated gravitational deformations (DSGSDs), caused by rock-mass creep processes that can evolve into rock falls and rock avalanches. A multidisciplinary approach that integrated geomorphologic surveys and stress–strain numerical simulations was used to investigate the effects of the stress release that was induced by the Last Glacial Maximum (LGM) in the Adamè Valley (Italian Alps). No evidence of DSGSDs has been found in this area; however, data are available to develop a well-constrained evolutionary model of the valley deglaciation. The extent and thickness of the southern extension of the Adamello glacier (the widest glacier in Italy) during its primary growth phases, including those of the LGM, three Late Glacial stages and the Little Ice Age (LIA) were reconstructed using glacial geological surveys and were reproduced using a finite-difference stress–strain sequential model spanning from the LGM to the present. A thermomechanical numerical configuration was developed based on geomechanical field measurements that were collected from rock outcrops and data from laboratory tests, specifically performed on intact rock samples. The numerical simulations demonstrate that thermomechanical viscous behaviour cannot be neglected because the resulting strains on the slopes are significantly higher than those resulting from conventional elastoplastic behaviour. The last major thermal effect in the rock masses reproduced using the model occurred from 11.5 ka until the LIA; the resulting displacement rates are as much as several tens of mm/ka, which are consistent with the absence of DSGSDs in the Adamè Valley, as these rates are significantly lower than those previously obtained from DSGSDs and other large landslides in the Alps. Based on thermomechanical numerical solutions, these rates should persist for the next 1000 years assuming that no variations in current climatic conditions occur.

Introduction

Mountain slope deformations depend on many factors, including the geological setting, properties of the rock mass jointing conditions, elevation, dip of the slope geometry, earthquakes, intense rainfalls, snowmelt, and human activities. In most areas, the current slope deformation rate can also be strongly affected by deglaciation processes, including those that continued from the end of the Late Glacial stages (Cossart et al., 2008) until the early Holocene and those that occurred since the end of the Little Ice Age (LIA, ca. A.D. 1850), which have recently accelerated because of global climate change (IPCC, 2007). In the Alps, evidence of glacial retreat since the Last Glacial Maximum (LGM) is well constrained. Furthermore, since the end of the LIA, Alpine glaciers experienced significant retreat and thinning only interrupted by very brief and weak advances (Haeberli and Beniston, 1998, Oerlemans, 2005). Over the last two decades, the decrease in the size and thickness of Alpine glaciers has accelerated owing to increasing global temperatures, which have been documented by annual surveys of glaciers in the Italian Alps (e.g., Haeberli et al., 2007, WGMS, 2008, WGMS, 2012, Baroni et al., 2011, Baroni et al., 2012). As a result, ice-free areas in the vicinity of shrinking glaciers are widening and are progressively converting to a paraglacial environment where rock-slope failures and rock-mass deformations can develop (Cruden and Hu, 1993, Ballantyne, 2002).

More than 900 deep-seated gravitational slope deformations (DSGSDs; sackungen sensu Zischinsky, 1969) and ~ 790 large landslides (Agliardi et al., 2009b, Agliardi et al., 2012) have been recognised in the Alps. These features are most abundant where foliated metamorphic rocks are exposed or where strong structural constraints are present (Agliardi et al., 2009a). Glaciers covered most of the areas affected by rock slope deformations during the LGM, which reached elevations of up to 3000 m in the major valleys (Kelly et al., 2004, Bini et al., 2009). The past presence of glaciers suggests that the stress release induced by post-glacial debuttressing has played a substantial role in triggering and controlling these deformations. Many case studies have investigated the role of mass rock creep and rock mass strength degradation in the evolution of slope deformations into slope collapse events (Chigira, 1992, Evans et al., 2006). The propagation of joints and new fracturing induced by ongoing deformation can increase the displacement rate and reduce the rock mass viscosity, leading to tertiary creep conditions and ultimately to slope failure (Emery, 1978, Eberhardt et al., 2004, Eberhardt, 2006, Alonso and Pinyol, 2010).

The rock mass temperature can affect slope deformations on various time scales: (i) a seasonal scale, during which the slope stability depends strongly on the temperature, as seasonal warming and cooling (Gischig et al., 2011a, Loew et al., 2012) coupled with weathering and snowmelt (Jaboyedoff et al., 2004) cause relative offset across active fractures (Gischig et al., 2011b), which can also drive rock slope deformation at depths below the annual thermal active layer (Gischig et al., 2011a); and (ii) a 1000-year time scale, in which the rock mass temperature depends on deglaciation processes, such as the duration and the primary phases of glacial retreat from the valleys.

The primary results and conceptual elements of thermomechanical modelling performed so far (Agliardi et al., 2012, Jaboyedoff et al., 2012, Loew et al., 2012) have been limited to investigating simplified slopes and the cyclic thermal variability within the active rock layer (i.e., temperature variations in the upper 10 m caused by seasonal changes) with respect to (i) stress–strain effects along joints in the rock mass (Gischig et al., 2011a) and (ii) landslides involving jointed rock masses (i.e., falls, slides, and topples). As it results from numerical models implemented so far by various authors (Gischig et al., 2011a, Gischig et al., 2011b) the thermomechanical processes that affect the rock masses down to depths as great as 100 m can be interpreted as micro- to mesoscale fatigue processes that involve incremental slip along critically stressed discontinuities and hysteresis driven by periodic thermomechanical loading.

The main aim of this study was to model the post LGM stress–strain evolution of the Adamè Valley slopes by experiencing a thermomechanical behaviour coupled with a creep rheology of the jointed rock masses. Geomechanical field data as well as laboratory tests on intact rock samples were collected to better constrain the assumed rheology. Because no DSGSDs or large landslides are present on the slopes of the Adamè Valley, the numerical model was validated by comparing the resulting displacement rates with those available from other parts of the Alps in areas affected by ongoing gravitational processes; in this manner, the suitability of the obtained results with respect to the measured data was verified. Based on the validated model a provisional simulation was carried out by simulating the next 1000 years' stress–strain evolution of the slopes, assuming that no significant changes will exist in the climate conditions.