M/F PHD Student Sciences Environnementales

CNRS; France; Sur place; EUR 20 000 - 40 000; Plein temps

Résumé du poste

An environmental research organization is seeking a First Stage Researcher to investigate the influence of snow cover and melting dynamics on landslide occurrences in mountainous areas of France. The research will involve advanced statistical analysis, meteorological modeling, and machine learning to develop predictive models. Candidates should have a Master's degree in a related field and experience with relevant analytical techniques. The position offers a full-time contract commencing on January 1, 2026, aimed at filling gaps in understanding landslide mechanisms and improving predictive tools.

Qualifications

• Experience with statistical analysis and machine learning techniques.
• Familiarity with meteorological modeling and snow dynamics.

Responsabilités

• Conduct statistical analyses and modeling related to snowpack melt and landslides.
• Integrate various data sources to strengthen predictive models.
• Develop innovative tools for predicting landslides.

Connaissances

Statistical analysis

Meteorological modeling

Machine learning

Knowledge of snow dynamics

Formation

Master's degree in environmental science or related field

Description du poste

Organisation/Company CNRS Department Centre de Recherche et d'Enseignement des Géosciences de l'Environnement Research Field Environmental science Biological sciences Geosciences Researcher Profile First Stage Researcher (R1) Country France Application Deadline 1 Jan 2026 - 23:59 (UTC) Type of Contract Temporary Job Status Full-time Hours Per Week 35 Offer Starting Date 1 Jan 2026 Is the job funded through the EU Research Framework Programme? Not funded by a EU programme Is the Job related to staff position within a Research Infrastructure? No

Offer Description

Landslides are one of the major natural hazards in mountainous areas, regularly affecting the French Alps. They cause significant economic costs, disrupt communication networks, and pose direct threats to populations and infrastructure. While heavy rainfall is well recognized as a triggering factor, recent studies have revealed the importance of snow cover and, in particular, rapid melting dynamics in the genesis of gravitational crises. These crises result in hundreds of landslides in a matter of days, as in January 2018, when more than 150 events were recorded in 48 hours. Conventional forecasting models, based solely on liquid precipitation, show their limitations here, as snow acts as a temporary reservoir and releases water to the ground according to specific dynamics. In addition, the frequency and/or seasonality of this type of event is likely to be affected by climate change and the increase in rain-on-snow events and winter thaws, posing new challenges in terms of understanding the processes and anticipating crises.

Current landslide prediction models, based mainly on rainfall thresholds, become ineffective in the presence of snow cover. Snow acts as a temporary reservoir, storing precipitation before releasing it back to the ground in the form of meltwater, at varying rates and intensities. Understanding the exact role of this process in triggering landslides is therefore a major scientific challenge. Recent work shows that the dynamics of the snowpack—in particular rapid melting—play a decisive role. Several geographical and snow meteorological factors interact: altitude, orientation, slope, snow water equivalent (SWE), liquid water content (LWC), rainfall, air and soil temperatures, water saturation, but also radiative fluxes (LWR, SWR) and the degreeday indicator, which reflects the thermal energy available for melting. The central scientific question is therefore: which snow meteorological and geographical parameters control the occurrence of landslides triggered by rapid snowpack melt, and how can they be integrated into predictive models capable of anticipating crises at the regional level over different time horizons?

From a scientific standpoint, the thesis will help fill a significant gap in our understanding of the interactions between snow cover, soil hydrology, and gravitational instability. It will help clarify the respective roles of meteorological, snow and geomorphological parameters in triggering landslides, better characterize the dynamics of winter and spring crises, and quantify their evolution in response to changes in climate and snow cover. Operationally, the results will provide innovative tools and indicators for predicting landslides linked to snowmelt. They will feed into future national monitoring systems (SPIRAL, VIGIMONT) and contribute to the development of risk prevention plans that are better suited to the Alpine context and incorporate non-stationary dynamics. These results will provide useful information for insurers and local authorities to anticipate the consequences of climate change and strengthen the resilience of mountain areas.

« Influence of snow cover and its melting dynamics on the triggering mechanisms and prediction of landslides

The thesis will be based on a multidisciplinary approach combining statistical analysis, level-meteorological modeling, and machine learning. Work carried out during the Master's internship has already identified strong trends and tested statistical and machine learning approaches. The thesis will aim to consolidate and update these results while expanding the methods and data sources and translating the forecasts into future projections. The main objectives are:

• Update and deepen existing statistical analyses based on the BD-RTM, integrating recent years and refining critical periods (early and spring melt).
• Identify and quantify control factors (altitude, slope, orientation, snow water equivalent, air and soil temperature, liquid precipitation, soil water saturation, degreeday, LWR, SWR).
• Diversify data: integrate other event databases (e.g., BDMvt, regional databases), satellite data (snow, soil moisture, surface temperature), and field measurements to strengthen the robustness of the models.
• Exploiting these event databases and comparing them with snow and meteorological data from the SAFRAN‑CROCUS S2M model in order to characterize the spatio‑temporal contexts that favor crises.
• Development of advanced predictive models (multivariate approaches, machine learning) combining event data, snow and weather data, and remote sensing data to characterize the probability of landslide occurrence.
• Multi‑scale validation: application to different mountain ranges (Vercors, Chartreuse, Belledonne) and comparison of results to assess the spatial and temporal transferability of the models.
• Projection of the relationships obtained in future climate scenarios using ADAMONT projections to assess how gravitational crisis episodes will evolve over different time horizons and/or depending on the level of warming.

« Influence of snow cover and its melting dynamics on the triggering mechanisms and prediction of landslides