Postdoctoral position (M/F) on the snow and meteorological processes driving the evolution of avalanche activity in the Alps under climate change

Job Description

Missions

The IPCC Special Report on the Ocean and Cryosphere (IPCC, 2019) states that mountain hazards [including avalanches] "are projected to occur in new locations and during different seasons in the future". Recent work carried out at CNRM/CEN and INRAE has quantified the evolution of natural avalanche activity under climate change in an exemplary alpine valley, the Haute-Maurienne (Doussot et al., 2026). Combining a statistical learning model trained on the EPA database (Enquête Permanente sur les Avalanches, the French long-term avalanche survey) with snow and meteorological simulations (SAFRAN/S2M reanalysis, ADAMONT climate projections), this study shows a decrease in the annual number of avalanches of about 9% per decade between 1958 and 2023, and a continuation of this decline over the 21st century (about 5% and 9% per decade under the RCP4.5 and RCP8.5 scenarios respectively), mainly driven by a strong decrease in springtime avalanche activity, while major avalanche cycles (30-year return level) decline at a more moderate rate. These trends are now quantified, but the physical processes explaining them remain to be established.

The objective of this position is to identify and rank the snow and meteorological mechanisms driving the decrease in avalanche activity. Several non-exclusive hypotheses are to be explored: the shortening of the snow cover duration and the decrease in snow depth, which reduce both the time window and the volume available for release; the decrease in the frequency and intensity of snowfall, in particular extreme accumulations, the main triggering factor; the earlier onset of snowpack wetting, which shifts wet-snow activity from spring towards mid-winter before eventually suppressing it; changes in melt-freeze crusts and nocturnal refreezing cycles, which temporarily stabilize the snowpack; and the evolution of persistent weak layers related to changes in snow metamorphism (temperature gradients, rain-on-snow events). The aim is to move from an observed trend to a quantitative attribution of its causes, distinguishing between dry-snow and wet-snow avalanches, and characterizing the dependence of these mechanisms on elevation and aspect.

A second objective is to extend these trends, so far established at the valley scale, to large spatial domains such as the Northern and Southern French Alps, whose contrasting snow-climate regimes (Mediterranean influence, "retour d'est" easterly flow events, elevation gradients) are likely to modulate the amplitude and seasonality of the trends. This extension raises the question of the transferability of the impact model to massifs where avalanche observations are scarcer or less homogeneous. The target spatial domain is the French Alps, starting from the well-documented case of the Haute-Maurienne. The temporal domain covers the period 1958–2100.

- Doussot, F., Viallon-Galinier, L., Eckert, N., Hagenmuller, P., 2026. 1950-2100 climate trends in avalanche activity in Haute-Maurienne, French Alps. EGUsphere [in press]. https://doi.org/10.5194/egusphere-2026-336
- Reuter, B., Hagenmuller, P., Eckert, N., 2025. Trends in avalanche problems in the French Alps between 1958 and 2020. Cold Regions Science and Technology, 238, 104555.
- Mayer, S., Hendrick, M., Michel, A., Richter, B., Schweizer, J., Wernli, H., and van Herwijnen, A.: Impact of climate change on snow avalanche activity in the Swiss Alps, The Cryosphere, 18, 5495–5517, 2024.
- Eckert, N., Corona, C., Giacona, F., 2024. Climate change impacts on snow avalanche activity and related risks. Nature Reviews Earth and Environment, 5, 369–389. - IPCC (2019). IPCC Special Report on the Ocean and Cryosphere in a Changing Climate. https://www.ipcc.ch/srocc/

Activity

To identify the processes driving the decrease in avalanche activity, we propose a modelling and analysis effort structured into the following tasks:

1) Characterise the past and future evolution of the snow and meteorological variables that are candidate explanations for the trends: snow cover duration and snow depth, frequency and intensity of snowfall (including extremes), liquid water content and date of first wetting, occurrence of rain-on-snow events, formation of melt-freeze crusts, and indicators of persistent weak layers, based on the S2M reanalysis and the ADAMONT climate projections.

2) Quantify the contribution of each family of predictors to the simulated trends in avalanche activity, using interpretability methods for machine learning models (variable importance, Shapley values) applied to the model developed by Doussot et al. (2026).

3) Carry out counterfactual sensitivity experiments: re-run the impact model while holding selected groups of variables at their reference climatology (for example, removing the trend in snowpack wetting, or in new snow accumulation) in order to isolate the share of the trend attributable to each process.

4) Disentangle the trends by avalanche type (dry snow / wet snow) and by avalanche problem (new snow, wind slab, persistent weak layer, wet snow, gliding snow), building on existing avalanche problem diagnostics (Reuter et al., 2025).

5) Extend the trend analysis to large spatial domains (Northern and Southern French Alps), addressing the transferability of the impact model between massifs (re-training, domain adaptation, accounting for topographic specificities and for the number of monitored avalanche paths) and analysing the dependence of the identified mechanisms on elevation, aspect and snow-climate regime.

6) Quantify the associated uncertainties (internal variability, GCM/RCM inter-model spread, impact model uncertainty) using robust statistical methods.

7) Summarise the scientific results in a journal article and in a report addressed to avalanche risk managers, detailing the mechanisms of the decline in activity, its regional variability, and their implications for long-term prevention strategies.

Your Profil

Skills

- Experience in numerical simulation of climate and/or snowpack, in particular in mountain regions
- Knowledge of snow physics (metamorphism, wetting, stability)
- Knowledge of statistical tools, in particular interpretability methods for machine learning models and rare event statistics
- Ability to implement a model in a programming language
- Strong interest in cryosphere and climate research
- Ability to synthesise

Your Work Environment

This position is offered within the framework of the agreement between the French Directorate General for Risk Prevention (Direction Générale de la Prévention des Risques, DGPR) and CNRM (Météo-France – CNRS). CNRM is mainly located in Toulouse, but this position will be based in the offices of the CNRM/CEN group in Saint-Martin-d'Hères (Grenoble metropolitan area — not in Toulouse). The work will be supervised by Pascal Hagenmuller and L. Viallon-Galinier (CNRM/CEN). A collaboration with N. Eckert (INRAE / IGE) is also planned.

Constraints and risks

This position is not compatible with fully remote work.

Compensation and benefits

Compensation

Monthly gross salary between €3,042 and €4,216 according to experience

Annual leave and RTT

44 jours

Remote Working practice and compensation

Pratique et indemnisation du TT

Transport

Prise en charge à 75% du coût et forfait mobilité durable jusqu’à 300€

About the offer

About the CNRS

The CNRS is a major player in fundamental research on a global scale. The CNRS is the only French organization active in all scientific fields. Its unique position as a multi-specialist allows it to bring together different disciplines to address the most important challenges of the contemporary world, in connection with the actors of change.

CNRS

The research professions