Nat. Hazards Earth Syst. Sci., 17, 887–904, 2017 https://doi.org/10.5194/nhess-17-887-2017 © Author(s) 2017. This work is distributed under the Creative Commons Attribution 3.0 License. A 240-year history of avalanche risk in the Vosges Mountains based on non-conventional (re)sources Florie Giacona1, Nicolas Eckert2, and Brice Martin1 1Centre de recherche sur les Économies les Sociétés les Arts et les Techniques, Université de Haute Alsace, Mulhouse, 68093, France 2UR ETNA, Irstea Grenoble/Université Grenoble Alpes, Saint Martin d’Hères, 38402, France Correspondence to: Florie Giacona (fl[email protected]) Received: 18 December 2016 – Discussion started: 19 December 2016 Revised: 27 April 2017 – Accepted: 1 May 2017 – Published: 16 June 2017 Abstract. Despite the strong societal impact of moun- 1 Introduction tain risks, their systematic documentation remains poor. Therefore, snow avalanche chronologies exceeding several Databases summing up observations of past events play a decades are exceptional, especially in medium-high moun- crucial role in evaluating the expected frequency and severity tain ranges. This article implements a combination of his- of natural hazards. However, they are typically sparse in their torical and geographical methods leading to the reconstruc- coverage. Their comprehensiveness is hampered by multiple tion, at the scale of the entire Vosges Mountains (north-east factors, including (i) a close link to vulnerability (especially of France), of more than 700 avalanches that have occurred events with human casualties), which results in an underrep- since the late eighteenth century on 128 paths. The clearly resentation of geophysically significant events that did not episodic nature of the derived geo-chronology can be ex- cause damage or fatalities; (ii) a relatively short temporal plained by three interrelated factors that have changed to- coverage; and (iii) incomplete spatial coverage that excludes gether over time: the body and reliability of sources, social large regions. Such factors are particularly prevalent in devel- practices conditioning the vulnerability and the natural haz- oping countries, where there is often an absence of relevant ard itself. Finally, the geo-chronology reflects the changes in records (e.g. von Kotze and Holloway, 1996). In view of the the meaning of the hazard in social space. Specifically, the considerable socio-economic impact of hydro-climatic phe- event which could be retrieved from the historical sources is nomena, these limitations of existing records are paradoxical an aspect of the interaction between society and its environ- (IPCC, 2012). ment. These results confirm the role of the historian in con- Avalanche risk is a characteristic illustration of this situa- textualising and evaluating such data. It transforms these data tion. Snow avalanches strongly impact permanent and tem- into information that is relevant for mitigating risk and under- porary populations in mountain areas in winter (e.g. Mc- standing its change over time. The work also demonstrates Clung and Schearer, 2006; Schweizer et al., 2003), caus- the usefulness of constructing an original database from a ing death and destruction (buildings, tourism infrastructure, diverse suite of historical data and field investigations. This power lines, forest stands), road cut-offs and, more fre- approach could be applied to other risk phenomena in the quently, losses related to an indirect vulnerability of moun- frequent situation in which archival data are sparse. tain societies (e.g. the negative impact of such destructive events on tourism). Databases concerning the hazard and/or the risk have been developed in many countries, but they are heterogeneous in terms of information content (Laternser and Schneebeli, 2002; Borrel and Brunet, 2006; Bourova et al., 2016). In addition, avalanche chronologies exceeding a few decades remain an exception (Corona et al., 2013; Schläppy et al., 2014), and those not biased towards the events causing Published by Copernicus Publications on behalf of the European Geosciences Union. 888 Florie Giacona et al.: A 240-year history of avalanche risk in the Vosges Mountains damage or those triggered artificially even more so (Schnee- The purpose of this article is to demonstrate the value of such beli et al., 1997; Eckert et al., 2013; Podolskiy et al., 2014). an approach for the (re)construction of a geo-chronology of Avalanche chronologies are even completely lacking in some avalanches and avalanche damage of sufficient duration and mountain areas, especially in medium-high mountain ranges, temporal resolution to understand the different characteris- which are often neglected in favour of high mountains. This tics of the phenomenon in the context of global change. is the situation for the Vosges Mountains, a medium-high For this purpose, the Vosges Mountains is an ideal case range located in north-eastern France, whose significance in study. Indeed, they combine topographical and snow features terms of avalanche risk is overshadowed by the Alps. In- that are favourable to avalanche activity with long-lasting hu- deed, the latter region has undergone systematic observation man occupation at relevant elevations, which should ensure of avalanches since the late nineteenth century and has been the existence of appropriate data sources. After a detailed de- the focus of substantial interest regarding snow avalanche scription of the geographic and socio-historical background, science and risk engineering (Mougin, 1922; Giacona et al., what follows describes the proposed methodology. It con- 2017b). sists of an original combination of (re)sources (Table 1) of Yet researchers and risk managers need long data series different types: administrative, printed, graphical, verbal and to detect changes in risk with time (due to both social and those resulting from spatial analyses. The results provided by environmental factors) and estimate the relevant probability this approach are then presented, interpreted and discussed of event occurrence. That is, long data series allow one to in light of the socio-historical, geographical and biophysical understand the dynamics of the phenomenon (return period), contexts of the region. to define reference scenarios and to contextualise them. His- torical analysis is now widely used to reconstruct chronolo- gies and specify the characteristics of past events, particu- 2 Description of the territorial context larly in terms of flood risk (Stedinger and Cohn, 1986). How- ever, avalanche risk stands out as an exception to this pat- The Vosges Mountains are located in the north-east of tern, with very few studies on this topic by historians, aside France, in the western part of the Rhine Valley. They are gen- from those by Granet-Abisset and Brugnot (2002) in France, erally associated with an imagery and landscape that contrast Bruno (2013) in the former Soviet Union, Laely (1984) in with the idea of massive avalanches: a mountain range on Switzerland and di Stefano (2013) in the eastern United a human scale, described using a lexicon that conveys soft- States and Canada. In addition, these rare studies have fre- ness, accessibility and friendliness (Giacona et al., 2017a). quently been conducted within a framework that combines Their dimensions are small, with a compact shape: 7300 km2, snow avalanches with other types of gravitational hazards, 1423 m a.s.l. at the highest point and 150 km long with a and at the spatial scale of a village or a mountain region width varying between 20 and 60 km. However, the southern (Leone, 2006; Granet-Abisset and Montredon, 2007; Favier part of the Vosges Mountains shows marked signs of Quater- and Granet-Abisset, 2000; Favier and Remacle, 2007). How- nary glaciation: U-shaped valleys, glacial cirques, moraine ever, the study of the evolution of avalanche activity would deposits and slopes greater than 30◦ (Flageollet, 2003). Ori- require specific focus on snow avalanche risk at the scale of ented north–north-east to south–south-west, they form the a mountain area that is coherent in terms of climate and to- first orographic barrier encountered by low-pressure airflows pography. coming from the Atlantic (Fig. 1). The cold and humid cli- The contribution of historians to research on risk goes sig- mate permits the accumulation of a deep snow cover. The nificantly beyond the simple production of large data sets latter persists in the form of snow patches until late spring from a variety of disparate sources. Their expertise is es- and sometimes even until early summer in the glacial cirques sential for the occurrence of a natural phenomenon and its located on the eastern slope. The proportion of snow precip- societal interpretation to be placed in their socio-historical itation to total precipitation is 20, 30 and 60 % at 700, 1000 and geographical contexts, so as to elucidate their mean- and 1350 m a.s.l., respectively (Wahl et al., 2009). The main ing (Cœur and Lang, 2011). More broadly, contextualisation ridge stands perpendicular to the prevailing winds that sweep helps to capture the temporality and dynamics of the vari- its flat and bald apical surface. This loads snow on the east- ous components involved in the risk system: biophysical fac- ern slopes, with impressive cornices forming at the break of tors (climate, hazard), practices and land use and vulnerabil- slope (Fig. 1e). Such commonly produced wind slabs are a ity factors. The latter includes stakeholder perceptions, rep- characteristic feature of the avalanche dynamics in the Vos- resentations of risk and the relation of societies to risk, as ges Mountains (Wahl et al., 2007). well as the capacity for society to deal with damageable phe- In the Vosges Mountains, areas in which avalanche activ- nomena. The historical approach is highly dependent on the ity is significant do not constitute permanent living spaces. existence of sources and, consequently, only possible if the However, human activity has been present in these areas for area has been occupied for a long time. In cases for which the centuries (Kammerer, 2003) and several activities have left a usual sources are deficient, geohistorical tools (Table 1) can durable imprint on the landscape.