Snow Avalanches Research Articles

Coupling Different Machine Learning and Meta-Heuristic Optimization Techniques to Generate the Snow Avalanche Susceptibility Map in the French Alps

The focus of this study is to introduce a hybrid predictive framework encompassing different meta-heuristic optimization and machine learning techniques to identify the regions susceptible to snow avalanches. To accomplish this aim, the present research sought to acquire the best-performed model among nine different hybrid scenarios encompassing three different meta-heuristics, namely particle swarm optimization (PSO), gravitational search algorithm (GSA), and Cuckoo Search (CS), and three different ML approaches, i.e., support vector classification (SVC), stochastic gradient boosting (SGB), and k-nearest neighbors (KNN), pertaining to different predictive families. According to diligent analysis performed with regard to the blinded testing set, the PSO-SGB illustrated the most satisfactory predictive performance with an accuracy of 0.815, while the precision and recall were found to be 0.824 and 0.821, respectively. The F1-score of the predictions was found to be 0.821, and the area under the receiver operating curve (AUC) was obtained to be 0.9. Despite attaining similar predictive success via the CS-SGB model, the time-efficiency analysis underscored the PSO-SGB, as the corresponding process consumed considerably less computational time compared to its counterpart. The SHapley Additive exPlanations (SHAP) implementation further informed that slope, elevation, and wind speed are the most contributing attributes to detecting snow avalanche susceptibility in the French Alps.

Remote Sensing Techniques for Assessing Snow Avalanche Formation Factors and Building Hazard Monitoring Systems

Snow avalanches, one of the most severe natural hazards in mountainous regions, pose significant risks to human lives, infrastructure, and ecosystems. As climate change accelerates shifts in snowfall and temperature patterns, it is increasingly important to improve our ability to monitor and predict avalanches. This review explores the use of remote sensing technologies in understanding key geomorphological, geobotanical, and meteorological factors that contribute to avalanche formation. The primary objective is to assess how remote sensing can enhance avalanche risk assessment and monitoring systems. A systematic literature review was conducted, focusing on studies published between 2010 and 2025. The analysis involved screening relevant studies on remote sensing, avalanche dynamics, and data processing techniques. Key data sources included satellite platforms such as Sentinel-1, Sentinel-2, TerraSAR-X, and Landsat-8, combined with machine learning, data fusion, and change detection algorithms to process and interpret the data. The review found that remote sensing significantly improves avalanche monitoring by providing continuous, large-scale coverage of snowpack stability and terrain features. Optical and radar imagery enable the detection of crucial parameters like snow cover, slope, and vegetation that influence avalanche risks. However, challenges such as limitations in spatial and temporal resolution and real-time monitoring were identified. Emerging technologies, including microsatellites and hyperspectral imaging, offer potential solutions to these issues. The practical implications of these findings underscore the importance of integrating remote sensing data with ground-based observations for more robust avalanche forecasting. Enhanced real-time monitoring and data fusion techniques will improve disaster management, allowing for quicker response times and more effective policymaking to mitigate risks in avalanche-prone regions.

Evolutionary characteristics and movement process of the January 2023 Duoxiongla snow avalanche, Tibet

Evolutionary characteristics and movement process of the January 2023 Duoxiongla snow avalanche, Tibet

Areas of maximum intensity of avalanche formation in the Greater Caucasus associated with large anomalies of temperature and humidity regime

The purpose of this paper is to substantiate the geographical boundaries of areas which are distinguished in the Greater Caucasus by the intensity of avalanche formation at the large-scale hydrometeorological anomalies and can be classified as “epicentres” of the avalanche activity. For the period of snow avalanche observations in the Caucasus in the 20th century, the Nature did twice demonstrate “experiments” in the winters of 1975/76 and 1986/87, during which especially intensive releases of avalanches occurred in the same areas of the southern macro-slope of the Central Caucasus. This work is an effort to answer the question ‒ why the main impacts of the snow disasters occurred within the boundaries of the same territories. The results of the analysis show that the highest intensity of avalanche activity is observed in those areas where three factors of avalanche formation are simultaneously realized: a big number of large avalanche catchments, including ancient glacial cirque zones (corries); high snowiness; types of snow favourable for the mass releases of especially strong avalanches. In the Caucasus, these criteria are met by the usually identified watershed zone of the Main Caucasus Range and, as the analysis showed, the territory of mountainous Georgia in the Inguri River basin (with tributaries of the Nenskra, Nakra, Mestia-Chala, Dolra, Mulkhura rivers), as well as the upper reaches of the river Kodori (Abkhazia). According to the USSR Glacier Inventory (1986), up to 80% of the corrie relief forms of the southern macro-slope of the Caucasus are concentrated in the basins of the Inguri and Kodori rivers, many of which are centres of origin of especially strong avalanches. The high snowiness of this territory, as compared to neighbouring areas, is conditioned by the influence of the orography and the climate effect, that is seen on the map of the solid precipitation distribution in the Atlas of Snow and Ice Resources of the World (1997). Only in this sector of the Greater Caucasus is there a kind of “peninsula” of high snowfall extending for tens of kilometers towards the moisture-bearing airflows from the Black Sea. Temperature conditions in the zone of increased snow accumulation are the factor contributing to the avalanche formation, which is manifested in its structural, textural and strength properties.

AVA-YOLO: image-based multiscale feature fusion enhanced perception model for snow avalanche detection

Abstract The global climate change has led to frequent occurrences of snow avalanche disasters. However, the significant variations in scale and shape during the avalanche process, and complex background imagery pose significant challenges to automated detection efforts. There is an urgent need to combine advanced deep learning technology to research automatic detection and recognition of avalanches in the field. In this paper, a novel deep learning model based on YOLOv8 improved multi-scale detection called AVA-YOLO is proposed to solve this problem. In AVA-YOLO, a key component, AKA (AKConv Combined Attention) module was designed and developed. This module combines the deformable convolutional properties of AKConv with the state-of-the-art self-attention module Exponential Moving Average, aiming to better perceive the feature map information of different shaped avalanches and to enhance the global relevance, thus improving the utilization of the information. Secondly, a new multi-scale sensing network structure was designed by increasing the number of detection heads to four and introducing the AKA module into the key positions of the network, while the association between model layers was newly designed to enhance the fusion of shallow and deep information to improve the detection accuracy. Experimental results demonstrated the effectiveness of AVA-YOLO, achieving 95.7% mAP50 and 75.6% mAP50:95 detection accuracies, as well as an F1 score of 0.92. Finally, a number of experiments were conducted to demonstrate the superior performance of the proposed model in comparison to other versions of YOLO, which will further exploit the potential of webcams as an underutilized technical capability in snow avalanche intelligence and portable monitoring.

Deriving seasonal and annual surface mass balance for debris-covered glaciers from flow-corrected satellite stereo DEM time series

Abstract Glaciers in High-Mountain Asia are experiencing varying rates and patterns of mass loss due to a complex interplay between glacier surface processes, local conditions and climate forcing. Spatially distributed surface mass balance (SMB) estimates can provide valuable insight into these drivers, but observations are currently limited in both space and time. We used very-high-resolution optical stereo images acquired by commercial satellites to prepare time series of digital elevation models (DEMs), and derived contemporaneous surface velocity and elevation change products for six debris-covered glaciers in Nepal. We developed new methods to produce flow-corrected Lagrangian SMB maps to isolate local surface ablation signals with enough detail to study individual ice cliffs. Our results show reduced ablation under thick debris cover and enhanced ablation over ice cliffs. Ablating ice cliffs were responsible for $10\!-\!38\%$ of the total ablation over debris-covered areas, even though they covered $\leq \!11\%$ of the total area. Seasonal SMB products reveal the timing and patterns of summer accumulation and ablation, underscoring the importance of snow avalanches for low-elevation debris-covered glaciers in the region. Our approach can be applied to other glaciers with repeat high-resolution DEM coverage and extended for regional analyses of SMB on seasonal to interannual timescales.

Snow Avalanche Susceptibility Mapping of Transportation Corridors Based on Coupled Certainty Factor and Geodetector Models

Avalanche susceptibility assessment is a core aspect of regional avalanche early warning and risk analysis and is of great significance for disaster prevention and mitigation on proposed highways. Using sky–ground integration investigation, 83 avalanche points within the G219 Wen Quan to Horgos transportation corridor were identified, and the avalanche hazard susceptibility of the transportation corridor was partitioned using the certainty factor (CF) model and the coupled coefficient of the certainty factor–Geodetector (CF-GD) model. The CF model analysis presented nine elements of natural conditions which influence avalanche development; then, by applying the Geodetector for each of the factors, a weighting coefficient was given depending on its importance for avalanche occurrence. The results demonstrate the following: (1) According to the receiver operating characteristic (ROC) curve used to verify the accuracy, the area under the ROC curve (AUC) value for the CF-GD coupled model is 0.889, which is better than the value of 0.836 of the CF model’s evaluation accuracy, and the coupled model improves the accuracy by about 6.34% compared with the single model, indicating that the coupled model is more accurate. The results provide avalanche prevention and control recommendations for the G219 Wen Quan to Horgos transportation corridor. (2) The slope orientation, slope gradient, and mean winter temperature gradient are the main factors for avalanche development in the study area. (3) The results were validated based on the AUC values. The AUCs of the CF-GD coupled model and the CF model were 0.889 and 0.836, respectively. The accuracy of the coupled model was improved by about 6.34% compared to the single model, and the coupled CF-GD model was more accurate. The results provide avalanche control recommendations for the G219 Wen Quan to Horgos transportation corridor.

OpenFOAM-avalanche 2312: depth-integrated models beyond dense-flow avalanches

Abstract. Numerical simulations have become an important tool for the estimation and mitigation of gravitational mass flows, such as avalanches, landslides, pyroclastic flows, and turbidity currents. Depth integration stands as a pivotal concept in rendering numerical models applicable to real-world scenarios, as it provides the required efficiency and a streamlined workflow for geographic information systems. In recent years, a large number of flow models were developed following the idea of depth integration, thereby enlarging the applicability and reliability of this family of process models substantially. It has been previously shown that the finite area method of OpenFOAM® can be utilized to express and solve the basic depth-integrated models representing incompressible dense flows. In this article, previous work (Rauter et al., 2018) is extended beyond the dense-flow regime to account for suspended particle flows, such as turbidity currents and powder snow avalanches. A novel coupling mechanism is introduced to enhance the simulation capabilities for mixed-snow avalanches. Further, we will give an updated description of the revised computational framework, its integration into OpenFOAM, and interfaces to geographic information systems. This work aims to provide practitioners and scientists with an open-source tool that facilitates transparency and reproducibility and that can be easily applied to real-world scenarios. The tool can be used as a baseline for further developments and in particular allows for modular integration of customized process models.

Universal Snow Avalanche Modeling Index Based on SAFI–Flow-R Approach in Poorly-Gauged Regions

Most high-mountain regions worldwide are susceptible to snow avalanches during the winter or all year round. In this study, a Universal Snow Avalanche Modeling Index is developed, suitable for determining avalanche hazard in mountain regions. The first step in the research is the collection of data in the field and their processing in geographic information systems and remote sensing. In the period 2023–2024, avalanches were mapped in the field, and later, avalanches as points in geographic information systems (GIS) were overlapped with the dominant natural conditions in the study area. The second step involves determining the main criteria (snow cover, terrain slope, and land use) and evaluating the values to obtain the Snow Avalanche Formation Index (SAFI). Thresholds obtained through field research and the formation of avalanche inventory were used to develop the SAFI index. The index is applied with the aim of identifying locations susceptible to avalanche formation (source areas). The values used for the calculation include Normalized Difference Snow Index (NDSI > 0.6), terrain slope (20–60°) and land use (pastures, meadows). The third step presents the analysis of SAFI locations with meteorological conditions (winter precipitation and winter air temperature). The fourth step is the modeling of the propagation (simulation) of other parts of the snow avalanche in the Flow-R software 2.0. The results show that 282.9 km2 of the study area (Šar Mountains, Serbia) is susceptible to snow avalanches, with the thickness of the potentially triggered layer being 50 cm. With a 5 m thick snowpack, 299.9 km2 would be susceptible. The validation using the ROC-AUC method confirms a very high predictive power (0.94). The SAFI–Flow-R approach offers snow avalanche modeling for which no avalanche inventory is available, representing an advance for all mountain areas where historical data do not exist. The results of the study can be used for land use planning, zoning vulnerable areas, and adopting adequate environmental protection measures.

Integrating Sequential Backward Selection (SBS) and CatBoost for Snow Avalanche Susceptibility Mapping at Catchment Scale

Integrating Sequential Backward Selection (SBS) and CatBoost for Snow Avalanche Susceptibility Mapping at Catchment Scale

Extreme erosion and bulking in a giant submarine gravity flow.

Sediment gravity flows are ubiquitous agents of transport, erosion, and deposition across Earth's surface, including terrestrial debris flows, snow avalanches, and submarine turbidity currents. Sediment gravity flows typically erode material along their path (bulking), which can dramatically increase their size, speed, and run-out distance. Hence, flow bulking is a first-order control on flow evolution and underpins predictive modeling approaches and geohazard assessments. Quantifying bulking in submarine systems is problematic because of their large-scale and inaccessible nature, complex stratigraphy, and poorly understood source areas. Here, we map the deposits and erosive destruction of a giant submarine gravity flow from source to sink. The small initial failure (~1.5 cubic kilometers) entrained over 100 times its starting volume, catastrophically evolving into a giant flow with a total volume of ~162 cubic kilometers and a run-out distance of ~2000 kilometers. Entrainment of mud was the critical fuel, which promoted run-away flow growth and extreme levels of erosion.

Dendrochronological analysis of conifers in avalanche catchments of the Northwestern Altai (Korgon River basin)

The purpose of the study was to date the powerful snow avalanches occurred in the avalanche catchment area in the river Korgon basin (Northwestern Altai) by the dendrochronological method. The paper presents the results of the analysis of a number of dendrochronological indicators of avalanches: the age of trees, the relation between widths of annual rings from opposite sides of the trunk, the presence of reactive (compressional) wood and traumatic resin canals, the dates of death and formation of wounds in trees, the presence of clearing effects (a sharp increase in growth) in three avalanche catchments, as well as a complex dendrochronological index of the avalanche activity. The dates of releases of the powerful avalanches were established down to 1570. It has been found that against the background of increasing amount of winter precipitation in these catchments, there are different trends in occurrence of the avalanches, which is determined by their morphological properties. In the avalanche areas with steeper slopes, due to repeated unloading of the avalanche centers in winter, the probability of releasing of powerful avalanches decreases. For the same reason, the powerful avalanches do not release during the years of maximum snow accumulation. The trends in the avalanche activity in the region under consideration had been obtained for the first time. The dendrochronological index of the avalanche activity, which is the ratio of the number of the tree growth failures to the number of the examined trees, is a good indicator of a release of a powerful avalanche even at a value of 0.85, provided that the number of examined trees in the lower part of the transit zone and in the zone of accumulation is 25–35 units. The results of this study may be used to predict the territorial differentiation of changes in the avalanche activity due to climate change. In connection with the recreational development of the Altai territory, they may also be of practical importance.

Safety improvement in open-pit mines with challenging mining conditions through upgrading avalanche prevention measures

One of the most common problems mining companies face in developing open-cut deposits in complicated geological and climatic conditions is avalanches during the spring-winter work period. This phenomenon is accompanied by a rapid displacement of snow masses, which can lead to catastrophic consequences. In this regard, the article aims to develop a modernized set of measures for avalanche protection. Prevention of avalanches of catastrophic scale, leading to blockages, road closures, and accidents, and the introduction of measures to reduce accidents and injuries in quarries and dumps, is a time-consuming task that requires serious approaches to its solution. The main research method is analyzing the applied anti-avalanche measures for active impact on snow cover and considering the possibility of using safer methods for active impact on avalanches at a particular mining enterprise. This approach allows for the evaluation of the advantages and disadvantages of the avalanche measures used and for choosing a set of measures to ensure more effective avalanche protection. The authors analyzed the existing shortcomings in the regulatory documents related to the provision of avalanche protection and anti-avalanche measures. It was found that they are not able to effectively ensure the development of internal local documents of enterprises in the field of anti-avalanche measures at the present stage. The resulting study showed that a comprehensive solution to the problem is needed through the introduction of several anti-avalanche devices to reduce the likelihood of avalanches of catastrophic scale. For the first time, the authors attempted to justify the possibility of using a modernized set of measures for the preliminary descent of snow avalanches to solve the problem of avalanche protection in the conditions of a real mountain object. The results obtained can be used to reduce the level of occupational injuries by reducing or eliminating accidents associated with unforeseen avalanches during the operation of an existing quarry.

The effect of ambient air temperature on meltwater production and flow dynamics in snow avalanches

The effect of ambient air temperature on meltwater production and flow dynamics in snow avalanches

Estimating Daily Snow Density Through a Spatiotemporal Random Forest Model

AbstractSnow density is of paramount importance in water resource management, snow avalanche warning, and climate change research. However, the lack of competent methods for long‐term and vast‐scale snow density mapping persists due to the intricate spatiotemporal dependencies inherent in snow density, resulting in scarce and inaccurate snow density products. To address this challenge, a spatiotemporal random forest (STRF) model is constructed by leveraging in‐situ measurements, multisource remote sensing, and reanalysis data. It tackles the spatiotemporal dependencies in snow density arising from its inherent heterogeneity and the relations involving snow density and nonlinearly connected meteorological, terrain, vegetation, and snow‐related factors. The effectiveness of the model is substantiated through rigorous validation methods, including random, temporal, and spatial block cross‐validations as well as independent validation, apparently surpassing ERA5‐Land snow density. The estimated snow density is also demonstrated to be able to improve existing snow water equivalent data set using fixed snow density. Utilizing the proposed model, a data set of daily 25‐km snow density from 1980 to 2018 is constructed for stable snow cover areas in China, which holds significant potential for research and applications in the realm of snow hydrology.

Quantifying the resistance of mixed-forests against natural hazards in the Pyrenees

Mountain protection forests can prevent natural hazards by reducing their onset and propagation probabilities. In fact, individual trees act as natural barriers against hydrogeomorphic events. However, assessing the structural strength of trees against these hazards is challenging, especially in a context of climate change due to the intensification of extreme events and changes in forest dynamics. Here, we focus on the mechanical analyses of two of the most common tree species across the Pyrenees (Abies alba Mill. and Fagus sylvatica L.) growing in two different areas (Spain and France), and affected by recurrent snow avalanche and rockfall events. We first performed 53 pulling test on mature trees, where the root-plate stiffness and the modulus of elasticity of the stems were evaluated. To further analyse the impact of forest management and climate on protective forests, we yielded information on tree growth using dendroecology techniques. Then, we assessed structure and neighbourhood characteristics for each target tree to account for the surrounding forest structure. Finally, using linear and structured equation models we tested if the mechanical capacity of the trees is determined either by functional traits (e.g. species, tree growth, diameter and height) or forest structural traits (e.g. tree density, tree structure and slenderness) or both. Our results suggest that the forest neighbourhood influences tree mechanical capacity through two pathways, including both functional and structural traits. The individual stiffness parameter of trees is influenced by their functional traits, while their structural traits are more closely related with changes in the modulus of elasticity. Both species exhibit varying levels of dominance in different locations, which is related to their resilience to the diverse natural hazards they confront. Our findings provide relevant insights to anticipating management strategies for forests that serve as a protective barrier against natural hazards in the context of a changing climate.

The role of local knowledge in snow observation and applied snow avalanche forecasting in Longyearbyen, Svalbard

In this article, we present the results of our study of knowledge in the actor-network of snow avalanche forecasters and observers in Longyearbyen, Svalbard. Our study aimed to understand how observers and forecasters make sense of snow conditions and snow avalanche risks. To do so, we conducted empirical investigations, which included interviews with key observation, forecasting, and decision-making actors in Longyearbyen’s risk governance system and autoethnographic participatory observation by the first author during the 36 months they worked closely with several informants in Longyearbyen. We draw on perspectives from literature on sensemaking, sensework, and tacit knowledge to conceptualize the modus operandi of observers and forecasters in a manner that allows for generalization to contexts beyond snow avalanche forecasting in Longyearbyen. We find that avalanche-related knowledge construction is highly dependent on the experiences of the observers from their non-professional interaction with the local topography and snow as part of their leisurely activities. Further, social relations and communication among observers and forecasters are essential in knowledge construction. Based on the different processes and resources entering the knowledge construction, we describe three dimensions of local knowledge involved in snow avalanche forecasting: situated, social knowledge; embodied knowledge; and synthetic knowledge. We elaborate on these dimensions of knowledge and discuss implications for the development and administration of avalanche warning systems in contexts characterized by a high degree of transience and change in phenomena.

СТАТИСТИЧЕСКИЙ АНАЛИЗ МНОГОЛЕТНИХ ДАННЫХ О СХОДЕ СНЕЖНЫХ ЛАВИН И СНЕЖНО-МЕТЕОРОЛОГИЧЕСКИХ УСЛОВИЙ В ГОРАХ ИЛЕ АЛАТАУ

Based on long-term observations at snow avalanche stations in the vicinity of Almaty, more than 5 thousand snow avalanches have been recorded since 1966. After statistical analysis of the data, the main characteristics of the avalanches were determined and snow and meteorological conditions were analyzed. The average volume of avalanches is 4 thousand m3, but this value is distorted due to the nonparametric distribution law. The median value of the volume of avalanches is 750-800 m3. 50% of cases within the upper and lower quartiles are 250-2500 m3. Catastrophic avalanches of more than 30 thousand m3 account for 3% of cases, and more than 100 thousand m3 only 0.3% of cases. Typical weather conditions during periods of avalanche danger are: snow height on the slopes 80-90 cm, on the meteorological site 50-60 cm, precipitation 15-19 mm, snow growth 25-30 cm. The dominant factors in avalanche formation were the height of the snow cover on the slopes, the amount of snowfall precipitation, snow cover increase, maximum air temperature. Spearman's correlation coefficients were found to be statistically significant for these parameters at the standard 5% significance level (a-level) with a probability of error (p-level) of less than 0.01. The results of factor analysis showed that avalanches are influenced by three groups of predictors: snow depth, precipitation and snow growth, and temperature conditions. Analysis of catastrophic avalanches in the indicator avalanche collection in the Kotyrbulak gorge showed typical conditions for mass avalanches in Ile Alatau in the spring - snow height at the weather site 50 cm, precipitation amount 53 mm, snow increase 28 cm. The agreement criteria we calculated turned out to be statistically significant by 5% level of significance. The standard errors of some data were within the recommended limits of 10%, which indicates the reliability of the results.

Analysis of Superresolution Methods in the Application to the Subsurface Sounding Radar for Detecting Hidden Objects

Adaptive methods have been developed for generating a radar image of a probed surface area. A block diagram of a subsurface sounding radar is described. The results of modeling and experimental studies are presented. Their use makes it possible to develop a subsurface sensing radar capable of effectively identifying hidden objects in concrete walls, as well as detecting cracks, voids and defects in the walls of buildings and structures, snow debris and avalanches, underground communications and organic bodies in the ground.

Retrieval of snow liquid water content from radiative transfer model, field data and PRISMA satellite data

The amount of liquid water (LWC) present in the snowpack is critical for predicting wet snow avalanches, forecasting meltwater release, and assessing water availability in river basins. However, measuring this variable using traditional in situ methods is challenging. Space imaging spectroscopy is emerging as a promising approach to map the spatial and temporal variations of snow parameters. While some studies suggest the potential of hyperspectral remote sensing to infer liquid water content, field validation is still lacking. In this context, we propose a new spectral index, namely Snow Surficial Water Index (SSWI), which is designed to be sensitive to the percentage of surficial liquid water content in snow. We developed the index using the BioSNICAR radiative transfer model and then we tested it on both field spectral data and satellite PRISMA imagery. Validation was performed using field data collected with a Snow Sensor during four campaigns in alpine environments, one of which simultaneously with PRISMA. Through a k-fold cross-validation analysis, we achieved a coefficient of determination of 0.7 and a Root Mean Square Error equal to 3%, demonstrating the effectiveness of the proposed index in retrieving LWC from field data and mapping LWC from PRISMA data. A spatial analysis at the catchment level reinforced the results, showing an LWC distribution consistent with orography. The proposed method can be easily applied to other space imaging spectroscopy missions.