This paper wants to be a sign of appreciation and deep friendship towards Philip Deline, who has shared with the authors, for over three decades, the fascination for slope instability process in high mountain. We thank the editor and the two anonymous reviewers for their valuable and constructive comments.

Introduction

1Mass movements are part of the natural evolution of mountains, especially in relatively young orogens such as the European Alps, and reflect disequilibria of processes and systems (Korup et al., 2010). Lithological and structural settings, slope and geometry, and the geomorphological evolution (in particular glaciations) control spatio-temporal occurrence, type and magnitude of mass movements (Blondeau et al., 2021). Some mass movements start without a specific trigger, predisposed by a progressive deterioration of mechanical properties of the slope under the action of gravity (Phillips et al., 2017), other failures are triggered by meteorological drivers (Bajni et al., 2021) or by earthquakes (Gorum et al., 2014). In recent decades, cryosphere changes due to global warming have added to the usual and well-known factors that cause mass movements (Chiarle et al., 2021). Glacier shrinkage and vanishing have changed the stress field of mountain slopes (Deline et al., 2021), exposing the bedrock to atmospheric agents, and in particular to air and rock temperature increase, daily and seasonal freeze/thaw cycles and infiltration of rainfalls and snow/ice melt into rock fractures. Accelerated snow and ice melt, permafrost degradation, and a larger fraction of precipitation falling as rain instead of snow contribute to increase water pressure within mountain slopes, especially in spring and summer, causing as well a more unpredictable dynamic of mass movements (Morino et al., 2021). While directly relating individual mass movements to specific processes might be difficult, their increase at high elevation in recent decades has been statistically demonstrated by several studies and is globally recognised as an indicator of climate change, especially with regards to rockfalls (e.g., Ravanel et al., 2017). In the Italian Alps, slope instabilities at high elevation are increasing in frequency, especially in summer (Nigrelli et al., 2023b): several events are reported in autumn and spring, while very little events are documented in winter. In a context of climate change, the interpretation of winter events is particularly challenging because of risk implications during a highly touristic season. Winter mass movements are generally large events compared to those occurring in other seasons and comprising the entire recorded range of volumes. Their occurrence seemingly denies the attribution of recent mass movements to global warming. But is it really so? In this work we review 12 mass movements in the Italian Alps at elevations above 1500 m documented from mid-December to January, henceforth during the Christmas period. We excluded snow avalanches, as they are a well-known winter hazard, and are the subject of a large number of dedicated studies. Instead, we included ice avalanches, an under-researched hazard, because we think they share the sensitivity to climatic factors, and therefore to climate change, with other types of mass movements. The aim of this work is to draw some preliminary considerations on the possible relationship between the increase in the frequency of winter events in recent years and ongoing climate and environmental changes and, more generally, to raise attention towards winter mass movements, which are currently still very little investigated.

Study Area

2Our study focuses on the entire Italian Alps, stretching 1200 km from east to west and covering about 5200 km², i.e., 27.3% of the European Alps (Figure 1). A complex geological history is at the base of the geodiversity of the Alps, a double-vergence collisional belt including litho-structural domains from different plates and geological environments. In the classic Alpine literature, three main sectors showcase different paleogeographic regions and crustal levels (Dal Piaz et al., 2003):

• the “internal” (southern and eastern) sector of the Italian Alps includes south-verging structural units from the upper plate of the collisional system (“Southalpine” domain);

• the “external” (western and northern) sector belongs to the lower plate of the collisional system (“European” foreland area);

• the axial sector is the collisional wedge bounded by two crustal scale discontinuities: the Insubric Line to the south and the Pennidic Front thrust to the north.

3The complex kinematic framework of the whole chain involves extensional, contractional and strike-slip tectonics, dominating in the internal zones, whereas a coeval contractional kinematics mainly affects the external zones; late, intense uplift and exhumation of the western side culminated with the onset of the Mont Blanc Massif (4810 m a.s.l.), top summit of the European Alps. After their formation as a collisional belt, the Alps have been modelled by several geomorphic processes interacting with tectonic ones. While the repeated Pleistocene glacial cycles modelled the Alpine valleys, Holocene gravitational and fluvial/torrential processes deeply modified glacial landforms and deposits, causing widespread instability phenomena (Soldati et al., 2006).

4The present-day climate of the Alps is characterised by different types of regimes, mainly due to the influence of the mountain range on the general atmospheric circulation. The latter is characterised by cold dry northern air masses of Polar origin, by warm or cold wet air masses of Atlantic origin coming from northwest, by cold dry continental air masses coming from east, and by warm wet Mediterranean southern air masses (Barry, 2008). The Köppen-Geiger climates in the Alps are: Arid, warm temperate, Boreal and Alpine (Rubel et al., 2016). Over recent decades, air temperatures in the Alps show an average warming rate of 0.3 °C/decade, thereby outpacing the global warming rate of 0.2 °C/decade; warming is more pronounced in summer and spring (Hock et al., 2019). In the high-elevation environments of the Alps, warming is even more pronounced: the annual mean warming rate observed during the 1991–2020 climate normal is, in fact, of 0.5 °C/10 years (Nigrelli and Chiarle, 2023a). According to the latter work, seasonal and annual temperatures at high elevation are increasing through the whole Alpine arc, mainly in summer and autumn, and the northern Alpine sector tends to be colder than the southern one.

5In the Italian Alps, the mean monthly temperatures have a unimodal trend: the hottest months are July and/or August, the coldest ones are January and/or February (Figure 2). Significant differences in the quantitative values of temperature exist, mainly in relation to elevation and to slope aspect. The total annual precipitation shows a significant variability depending on the local climate conditions, with peaks of 3000 mm per year in some alpine sectors. However, since rain gauges are often not heated (especially at high elevation), precipitation might in some cases be underestimated (Viani et al., 2020).

Case studies

6We identified 12 major mass movements, which occurred in the middle of winter in the Italian Alps, above 1500 m a.s.l. (Figure 1). The events are briefly described below, while their main characteristics are summarised in Table 1.

Table 1. Main characteristics of the mass movements analysed in this work

Event ID	Event (name)	Date	Max. elev. (scar) (m)	Volume (m3)	Aspect (scar)	Permafrost (A/C/F/N)
1	Alleghe	11/01/1771	1860	10 x 106	ESE	N
2	Colmandro	13/12/1825	1700	n.a.	NE	N
3	Fiernaz	17/01/1982	2000	500,000	SE	N
4	Brenva	18/01/1997	3725	4 x 106	SE	A
5	Whymper	24-25/01/1997	4000	25,000	S	A
6	Rocciamelone	26/12/2006	3200	104	W	A
7	Crammont	24/12/2008	2650	535,000	N	F
8	Sass Maor	19/12/2011	2200	150,000-200,000	E	N
9	Val Senales	27/12/2011	2750	10,000	SW	C
10	Tre Amici	16/12/2015	3400	200,000	NE	A
11	Monviso	26/12/2019	3200	60,000-80,000	N	A
12	Punta San Matteo	19/12/2020	3650	80,000-90,000	N	A

For data source please refer to the text; for the complete version of the table please refer to the annex). Event ID: same as shown in Figure 1; Max. elev. (scar): maximum elevation of the detachment scar; Aspect (scar): mean slope aspect in the detachment zone; Permafrost: permafrost conditions expected in the detachment zone according to the Alpine Permafrost Index Map of Boeckli et al., 2012 (A = permafrost in nearly all conditions; C = permafrost mostly in cold conditions; F = permafrost only in very favorable conditions; N = no permafrost); n.a.: not available data.

7On 11 January 1771, at around 23:00 local time, about 20 x 10⁶ m³ of rock suddenly fell from Mount Piz (Eastern Italian Alps; Figure 1, case 1), at an elevation of about 1750 m a.s.l. The movement started as a translational rock fall evolving into a rock avalanche. This latter blocked the Cordevole River at the valley bottom, and ran up about 150 m on the opposite side of the valley, causing the formation of a natural dam and of the Alleghe Lake (Eisbacher & Clague, 1984). The event caused 49 casualties. On May 1^st of the same year, a further 3 x 10⁶ m³ of material detached from the same slope via rock fall, generating a devastating wave which hit the village and caused 3 deaths. The resulting depositional body from the two combined events shows a length of 1500 m, a height of 880 m and a mean slope of 30° (Dykes et al., 2013). A relatively complex system of discontinuities is observed in the landslide scar, made of limestones and dolomites: the instability is thought to have been promoted by the lower part of the slope being overdipping. Fractures had been observed in the slope a few years before the event. The abundant rainfall of the preceding autumn can be deemed as preparatory factor, while the cold temperatures of January may have triggered the event, causing water springs to freeze and water pressure to build-up in the slope.

The Colmandro Rock slide, 13 December 1825

8The northeastern flank of Monte Colmandro (2079 m a.s.l., Eastern Italian Alps) is drained by the Rebrut stream, a small tributary of the Vanoi Torrent (Figure 1, case 2). This mountain slope is composed of micaschists, phyllites and paragneiss. The first record of slope instability dates back to 1748, and shortly before 1800 a fracture causing a 10 m lowering of part of the slope was observed (Filippi Gilli, 2014). In 1809, more fractures appeared after intense deforestation, which was perpetuated during the 19^th century to cultivate meadows irrigated by collecting water from nearby springs. In October 1823, intense and prolonged rains activated the instability of the slope, with movements that lasted for about six months. In November 1825, abundant precipitations lead to snow accumulation in the area. On 7 December, a sharp increase in temperature due to Scirocco winds completely melted the snow. On 13 December, a large slope failure occurred, whose mass crossed and blocked the Vanoi Torrent, raising its bed of about 15 m and causing the formation of a large lake. The slope remained unstable for several years, and breakage of the lake dam during heavy precipitations caused recurring disruptions in the valley floor until 1882, when a violent flood of the Vanoi Torrent removed the landslide dam for good.

The Fiernaz Rock Avalanche, 17 January 1982

9Early in the morning on 17 January 1982, the inhabitants of the Fiernaz Hamlet heard an explosion and saw the top of the mountain above Fiernaz rise a few metres for a second, before it subsequently collapsed (Figure 1, case 3). About 500,000 m³ of rock (gneiss and micaschists) detached along a slope section between 2000 m and 1500 m a.s.l., originating from the head of the small catchment that overlooks the hamlet, on the right slope of the middle Valtournenche (Aosta Valley, Western Alps; Mandrone, 1995). The initial rock fall rapidly turned into a rock avalanche. Most of the failed mass accumulated in the upper-middle part of the debris cone at the base of the slope. Part of the deposit reached the valley floor, blocking the main valley road and touching some houses. No one was injured, but thousands of tourists were stuck in the upper part of the valley. According to local witnesses, the rock fall occurred after a few very warm days that caused abundant snow melting, followed by a sharp drop in temperature, which would have caused the water to freeze. On 22 May 1983, heavy rains combined with snowmelt mobilised a large part of the rock avalanche accumulation, triggering a debris flow of 1 x 10⁶ m³, which destroyed some houses in Fiernaz, blocked the main road on the valley floor, and reached the main river (T. Marmore).

The Brenva Rock-Ice Avalanche, 18 January 1997

10On 18 January 1997, in the early afternoon, a catastrophic rock failure affected the Gendarme Rouge (3872 m a.s.l.), at the head of the Brenva Glacier Basin (Mont Blanc Massif, Western Alps; Figure. 1, case 4). A volume of 2 x 10⁶ m³ of rock (granites and massive crystalline schists) detached from a scar 250 m wide and 330 m high, slid along the surface of the glacier and transformed into a huge rock-ice avalanche, which entrained about 2 x 10⁶ m³ of ice (Barla et al., 2000). The avalanche crossed the glacier front and ran up the opposite side of the valley along the northwest flank of Mont Chetif (Figure 3), blocking the main river (Dora River) and invading a ski run: two skiers were killed. Signs of slope instability had already been documented in summer 1996, and two days before the main collapse. In November 1920, the Brenva Glacier had already suffered the impact of a large rock avalanche, and the right-lateral moraine still retains geomorphological evidence of other landslides that occurred during the Holocene (Deline et al., 2015). According to Paranunzio et al. (2016), the 1997 slope failure was associated with a precipitation anomaly in the 3 months preceding the event, low temperatures in the 30 days preceding the event, and a further marked drop in temperature in the 3 days preceding the event.

The Whymper Ice Avalanche, 24–25 January 1997

11In the night between 24 and 25 January 1997, about 25,000 m³ of ice detached from the Whymper Serac at an elevation of about 4000 m, at the top of the Grandes Jorasses Glacier (Mont Blanc Massif, Western Alps; Figure 1, case 5). The ice avalanche initially travelled southwest along the upper part of the Grandes Jorasses Glacier, then it divided into 2 branches. The largest one reached the Planpincieux Glacier (west of the Grandes Jorasses Glacier), travelled all the way down to its hanging front, and was finally channelled into the underlying rock incision, stopping at an altitude of 2000–2100 m, without causing damage (Miletto, 1997). The Whymper Serac had been monitored since March 1996 for the threat posed by recurrent ice failures to the valley bottom (Val Ferret), and the detachment was predicted with a few days of uncertainty (Chiarle et al., 2022).

The Rocciamelone Rock Slide, 26 December 2006

12On 26 December 2006, a rock slide affected a 250 m long stretch of the southern ridge of the Rocciamelone Peak (3538 m a.s.l., Susa Valley, Western Alps; Figure 1, case 6), at an elevation of approximately 3200 m (Turconi et al., 2010). A few tens of thousands of m³ moved downslope for a few tens of meters along the western flank of the ridge. The main movement was followed by minor rockfalls over the next few days. Fractures parallel to the ridge were first detected in September 2004, and their movement accelerated in autumn 2006. The slope instability was reactivated in June 2007 and continued throughout the summer, with about 200 small rock slides and falls documented. In the following years the slope continued to deform, but at a slower rate (Paro et al., 2016). The bedrock affected by rock sliding is made of calcschists and silicate marbles. Factors predisposing to slope instability are the geo-structural setting of the mountain and a deep-seated slope deformation located below the detachment zone of the rock slide. The high air temperatures that occurred in the early 2000s (particularly in summer 2003) may have prepared the slope for failure (Paro et al., 2016). As regards the triggering factors, Paranunzio et al. (2019) found anomalously high temperatures in the months before the failure, and a marked temperature rise in the days immediately preceding the event.

The Crammont Rock Avalanche, 24 December 2008

13On 24 December 2008 at 17:21, a small rock fall detached from the north face of the Crammont (Aosta Valley, Western Alps; Figure 1, case 7). The collapse was preceded by several mixed rock-snow avalanches during the early afternoon, and small rock falls detached from several points around a rock spur. A larger rock fall detached from the base of a spur in the middle of the north face (Deline et al., 2011; Figure 4). Two minutes later, the whole spur collapsed from an elevation of about 2650 m and released a rock avalanche and a dense black dust cloud. The rock avalanche travelled along the Vallon d’Arp down to a snow-covered plateau, where it deposited the bulk of the failed rock mass. A small part of the mass continued downstream and reached the Dora River at 1090 m a.s.l., without damming it. The failed mass entrained snow along the path, and was channelised by moraine ridges. Using the photographic sequence of Figure 4, the mean velocity of the rock avalanche front during the first minute was estimated at 14.5 m s⁻¹. A large amount (possibly as much as 50%) of the accumulation was snow. The rock deposits show sharp limits, have longitudinal ridges, contain mainly coarse debris, and are inversely graded, openwork and clast-supported with an abundant matrix. Their thickness is very variable: from a few cm to 1–2 m. The collapsed ridge summit, which seems to be double-crested in some sections, presented open, vertical cracks. According to Paranunzio et al. (2019), air temperatures had been very low in the 3 months preceding the event, and strongly increased in the 7 days preceding the failure.

Sass Maor Rock Avalanche, 19 December 2011

14On 19 December 2011, a large rock mass detached at the foot of the eastern flank of Sass Maor (2812 m a.s.l., Pale di San Martino, Eastern Alps; Figure 1, case 8), at an elevation of about 2000 m (Bonaccorso, 2011). The scar has an area of about 15,000–16,000 m² and the collapsed rock volume has been estimated at 150,000–200,000 m³, with blocks as large as 80–100 m³. The rock mass split into two branches, collected into two debris channels, and stopped at about 1620 m a.s.l. No one was injured, but three of Sass Maor’s most popular climbing routes were damaged. The failure was attributed to a brittle bedrock and to rock weathering. The event was associated with a long-term warm temperature anomaly, and a sudden temperature drop in the days preceding the failure (R. Paranunzio, pers. com.).

The Val Senales Debris Flow, 27 December 2011

15On the afternoon of 27 December 2011, an unusually high discharge was noticed in the stream flowing through the Maso Corto hamlet in the Senales Valley (Eastern Alps; Figure 1, case 9). In the evening, a flash flood of about 10,000 m³ hit the hamlet, flooding the main road (Bressan, 2011). A campsite and a hotel were evacuated and the nearby roads were closed. A survey carried out the following day allowed to assess that the debris flow was triggered by an outburst flood from a small depression. The latter had filled up with water due to infiltrations in the rock dam of a lake located just above, a few hundred metres below the front of the Giogo Alto Glacier. The outlet of the depression was blocked by a dam of ice and snow, which suddenly broke.

The Tre Amici Rock Avalanche, 16 December 2015

16In the night between 16 and 17 December 2015, a rock fall detached from the northern side of the Tre Amici Peak (3780 m a.s.l., Monte Rosa Massif, Western Alps; Figure 1, case 10), at an elevation of about 3400 m, partially involving the Signal Glacier (Chiarle et al., 2016; Figure 5). The rock fall rapidly evolved into a rock avalanche of about 200,000 m³, which crossed the Northern Locce Glacier and reached the southern shore of the Locce Lake (2200 m a.s.l.). The same slope was affected in September 2010 by a previous rock fall and is still unstable, as testified by the recurrent block falls and debris flows from the same source area of the 2015 rock fall. The morphological configuration acquired by the slope after the failure is now also predisposing the slope to the development of snow avalanches. Late autumn 2015 in the Western Italian Alps had exceptionally warm temperatures and no snow. After a cold interlude in early December, the temperature returned well above the seasonal average in the days immediately preceding the failure.

The Monviso Rock Fall, 26 December 2019

17On 26 December 2019, a large rock fall developed at about 3200 m a.s.l. on the northern slope of Torrione Sucai, a rock spur composed of metabasites and green schists in the Monviso Massif (3841 m a.s.l., Western Alps; Figure 1, case 11). The detachment scar was 150–160 m high and 40–70 m wide, for an estimated volume of collapsed rock of about 60,000–80,000 m³ (ARPA Piemonte, 2021). The rock fall material formed a fan-shaped accumulation, up to 200 m wide, on the debris cone at the foot of the failed rock slope, with blocks up to 500 m³ in size (Figure 6). Despite the lack of reports on similar slope failures in the antecedent years, a large debris cone at the base of the slope and the recurrent appearance of new large rock blocks in aerial images from the 2000s testify an intense instability of the slope. After the main collapse, the slope showed a residual instability through early summer 2020. The rock fall did not cause any damage, but the terminal deposits came to rest 100 m above a very popular trail that leads to the Quintino Sella Hut, along the path to the summit of Monviso. The slope failure was attributed to the high degree of fracturing of the rock, and to permafrost degradation (ARPA Piemonte, 2021). The event was associated with abundant precipitations in the previous 3 months, and with a sudden drop in temperature in the week preceding the failure (R. Paranunzio, pers. com.). The detachment scar is very close to the Superiore del Coolidge Glacier, which collapsed almost entirely in 1989, producing one of the largest documented ice avalanches in the Italian Alps (Chiarle et al., 2022).

The Punta San Matteo Rock-Ice Avalanche, 20 December 2020

18Between 19 and 20 December 2020, a rock-ice avalanche developed from the glacier-covered northern slope of Punta San Matteo (3678 m a.s.l., Ortles-Cevedale Group, Central Alps; Figure 1, case 12). The failure occurred at an altitude of 3650 m, and the avalanche stopped on the Forni Glacier, at 2945 m a.s.l. (Scotti & Cola, 2021; Figure 7). The estimated volume was of 80,000–90,000 m³, of which 90% rock, 10% firn and ice. Early signs of instability started in the morning of 19 December, while the main collapse occurred on 20 December, with minor slope failures documented in the following days. In the last 30 years, the summit of Punta San Matteo has undergone a dramatic deglaciation, which has led to the formation of large bedrock outcrops. Therefore, the detachment zone probably was probably subject to post-glacial decohesion, as well as to marked thermal fluctuations and cryoclastic processes due to the percolation of meltwater into rock fractures.

Results and discussion

19The main characteristics of the case studies analysed in this work are summarised in Table 1. At high elevation, winter mass movements are quite rare, as normally gravitational events concentrate in spring and autumn, and in summer. Nigrelli et al. (2023b) analysed the monthly distribution of 441 slope instabilities that occurred above 1500 m a.s.l. in the Italian Alps in the period 2000–2022. Figure 8 shows that slope failures occurring in the 1500–2500 m elevation range have a maximum frequency in May, and a secondary maximum in October-November. Conversely, slope failures that occur above 2500 m a.s.l. show a clear frequency peak in August. Notably, December, January, February, and September are the only months in which the number of events—however extremely limited—is almost the same for both elevation ranges.

20The small number of mass movements that occur at high elevation during the winter does not allow statistically based conclusions to be drawn. However, some preliminary considerations can be drafted:

1. Some significant differences emerge between the oldest and the most recent events: A) the mass movements that occurred before 1997 started mainly in the 1700–2000 m elevation range, i.e., on slopes without permafrost; since 1997 the starting zones are located in the elevation range 2200–4000 m a.s.l., and many above 3200 m a.s.l., i.e., on slopes where permafrost is expected. B) The aspect of slopes where movements initiated before 2000 range between NE and S (mainly S and SE), while that of slopes which failed after year 2000 is between SW and E (mainly N). It is well known that slopes facing from E to S are the ones most subject to thermal fluctuations, both on a daily and seasonal scale, being the most exposed to solar radiation (Nigrelli et al., 2022). However, several studies (e.g., Nagai et al., 2013) have shown that high-elevation slopes facing N are the most subject to instability due to permafrost degradation. C) Mass movements up to 1997 are associated with abundant precipitations in the months preceding the collapse, and often with a temperature drop in the days immediately before the failure. The triggering of mass movements after 2000 is, however, associated with temperature anomalies only (with the exception of Monviso 2019). All these observations point to a change in the winter instability pattern after 2000, with an increased involvement of rock slopes under permafrost conditions for which no winter failures are known before 1997, and a transition from heavy-precipitations controlled mass movements, to temperature anomalies and fluctuations controlled mass movements.

2. In winter, all types of mass movements can occur, including debris flows, even if rock slope failures prevail.

3. Most of the documented winter mass movements have quite high magnitudes (10^4–10⁶ m³). This could be partially due to the low frequentation of mountain environments during the winter, and therefore to the possibility that smaller events are gone unreported. However, in authors’ opinion, slope failures with a volume of 10³ m³ are unlikely to go unnoticed: this speculation is confirmed by the images of a growing number of webcams installed in alpine areas, which allow a visual monitoring throughout the winter.

4. The mobility of documented mass movements (Height difference/Runout, H/L in Table 1) is mostly in the range of 0.41-0.61, consistent with values found in the literature (Lucas et al., 2014). Only in 3 cases (Monviso, Fiernaz, Whymper) the mobility is significantly lower (0.80-1.41).

5. Most events occurred in the last 25 years. This may partly depend on climate change, partly on the growing attention towards this type of process.

21In summary, based on the case studies presented in this work, we can hypothesise that winter mass movements at high elevation in the Italian Alps depend on the following factors:

1. Predisposing factors: the litho-structural conditions of the bedrock are the main factor predisposing slopes to instability (see e.g., Alleghe, Fiernaz, Crammont, Tre Amici). Schistose rocks prevail. However some failures developed in more rheologically compact rocks (such as limestones/dolomites and gneisses/granites).

2. Preparatory factors: for mass movements that occurred before 2000, heavy precipitations in the weeks/months preceding the collapse played a fundamental role, by building up critical water pressures in the unstable slopes. In one case (Colmandro 1825), land use practices were also key in preparing the failure. For mass movements that occurred after 2000, long-term positive temperature anomalies appear to have played a major role, likely favouring ice, snow, and permafrost melt. Additionally, studies have shown that the permafrost active layer reaches its maximum depth in late autumn (e.g., Magnin et al., 2015): the delay of warm temperatures from summer penetrating at depth might cause the instabilities to manifest after a long interval, coinciding with the Christmas period. This process may also explain the relatively high magnitude of winter slope failures. In some cases, slope deglaciation has also been considered to contribute to the slope instability. Finally, freezing conditions at the slope surface, characteristic of the winter season, can also be considered a preparatory factor for failure, or even a trigger, as they prevent liquid water circulating in the slope from escaping, thus leading to water pressure build-up (Govi et al., 1993).

3. Triggering factors: in some cases, predisposing and preparatory factors may be sufficient to cause slope failure initiation, once equilibrium conditions are exceeded (e.g., Alleghe, Brenva). In several cases, a key role was probably played by a rapid temperature increase in the days preceding the failure, which could have caused snowmelt (Colmandro, Fiernaz, Rocciamelone, Tre Amici), providing the necessary water input for the final trigger. In some cases, marked temperature drops may have been the trigger, by freezing spring waters and causing water pressure build-up in the slopes.

Concluding Remarks

22Mass movements at high elevation in winter are rare events in the Italian Alps. This study has shown that any type of instability process can occur in winter, including debris flows, with landslides, and in particular rock falls/avalanches, prevailing with often quite high magnitudes (10⁴–10⁶ m³). In the recent decades, an apparent increase in the frequency and elevation of occurrence of winter mass movements can be observed: in particular, no slope failure is documented in permafrost areas before 1997. These findings may in part stem from the growing attention dedicated in the recent years to the impacts of climate change and the related risks, even though it seems unlikely that major events, even in the past, could have completely gone unreported. It is more likely that the rapid changes taking place in high mountain, have profoundly altered the equilibrium of the slopes, debutressing and exposing the bedrock to the atmospheric agents, and increasing the amount of liquid water circulating inside the slopes. In this context, and considered the growing frequentation of Alpine areas also during the winter, especially for tourism purposes, it is important to expand our knowledge on winter mass movements.

23Additional information on these types of collapses can come from the growth of case studies, thanks to emerging technologies (i.e., satellite and drone imagery, webcams) and from the comparison with high altitude/latitude areas outside Italy. It is worth mentioning that citizens are also giving a valuable contribution in reporting, documenting and disseminating information on mass movements through social media, and must become increasingly aware and responsible users of mountain environments.