Reply on RC2 (original) (raw)
Related papers
Remote Sensing, 2019
Although snow avalanches are among the most destructive natural disasters, and result in losses of life and economic damages in mountainous regions, far too little attention has been paid to the prediction of the snow avalanche hazard using advanced machine learning (ML) models. In this study, the applicability and efficiency of four ML models: support vector machine (SVM), random forest (RF), naïve Bayes (NB) and generalized additive model (GAM), for snow avalanche hazard mapping, were evaluated. Fourteen geomorphometric, topographic and hydrologic factors were selected as predictor variables in the modeling. This study was conducted in the Darvan and Zarrinehroud watersheds of Iran. The goodness-of-fit and predictive performance of the models was evaluated using two statistical measures: the area under the receiver operating characteristic curve (AUROC) and the true skill statistic (TSS). Finally, an ensemble model was developed based upon the results of the individual models. Results show that, among individual models, RF was best, performing well in both the Darvan (AUROC = 0.964, TSS = 0.862) and Zarrinehroud (AUROC = 0.956, TSS = 0.881) watersheds. The accuracy of the ensemble model was slightly better than all individual models for generating the snow avalanche hazard map, as validation analyses showed an AUROC = 0.966 and a TSS = 0.865 in the Darvan watershed, and an AUROC value of 0.958 and a TSS value of 0.877 for the Zarrinehroud watershed. The results indicate that slope length, lithology and relative slope position (RSP) are the most important factors controlling snow avalanche distribution. The methodology developed in this study can improve risk-based decision making, increases the credibility and reliability of snow avalanche hazard predictions and can provide critical information for hazard managers.
Toward the development of deep learning analyses for snow avalanche releases in mountain regions
Geocarto International, 2021
Snow avalanches impose a considerable threat to infrastructure and human safety in snow bound mountain areas. Nevertheless, the spatial prediction of snow avalanches has received little research attention in many vulnerable parts of the world, particularly in developing countries. The present study investigates the applicability of a stand-alone convolutional neural network (CNN) model, as a deep-2 learning approach, along with two metaheuristic algorithms including grey wolf optimization (CNN-GWO) and imperialist competitive algorithm (CNN-ICA) in snow avalanche modeling in the Darvan watershed, Iran. The analysis was based on thirteen potential drivers of avalanche occurrence and an inventory map of previously documented avalanche occurrences. The efficiency of models' performance was evaluated by Area Under the Receiver Operating Characteristic curve (AUC) and the Root Mean Square Error (RMSE). The CNN-ICA model yielded the highest accuracy in both training (AUC= 0.982, RMSE=0.067) and validation (AUC= 0.972, RMSE=0.125) steps, followed by the CNN-GWO model (AUC of 0.975 for training, RMSE of 0.18 for training, AUC of 0.968 for validation, RMSE of 0.157 for validation). However, the standalone CNN model showed lower goodness-of-fit (AUC= 0.864, RMSE=0.22) and predictive performance (AUC= 0.811, RMSE=0.330). The approach utilized in this study is broadly applicable for identifying areas where avalanche hazard is likely to be high and where mitigation measures or corresponding land use planning should be prioritized.
What Do Field Observations Tell Us About Avalanche Danger?
The avalanche forecast regions in Canada range from 100 to 30,000 km 2 , far larger than the 10 km 2 covered in a typical backcountry day. This difference in scale could cause the danger a recreationist is exposed to, the local avalanche danger, to differ from the regional bulletin. This study examines the relationship between field observations (instability, snowpack, and weather factors), which do not require digging a snow profile, and the local avalanche danger. The results were grouped for analysis by the dominant avalanche character of the day: Loose Dry, Wet (loose and slab), Wind Slab, Storm Slab, Persistent Slab, and Deep Slab. Throughout the past 6 winters we have created a unique dataset of 28 field observations from 425 field days. Univariate and multivariate cross-validated classification trees were built to examine the predictive capability of the observations for the local danger. Storm, Persistent, and Wind Slab avalanche characters had the most field observations correlate significantly with the local danger, and Wet (loose and slab) had the least. Observations of Slab Avalanche Activity, New Snowfall, and Tree Bombing were applicable for the most avalanche characters. Univariate and multivariate classification trees can be useful to recreationists in interpreting critical observations and the combinations of these observations that indicate elevated danger.
Forecasting forest avalanches: A review of winter 2011/12
Mountain forests play a crucial role in avalanche mitigation by hindering avalanche for-mation. Nevertheless, due to the complex interactions between ecological conditions, terrain, snow-pack and meteorological parameters the protective effect of forests may be reduced. Therefore, so-called 'forest avalanches' do occur and may be a threat to roads, railways and ski-runs below the for-est. Due to their sporadic occurrence, gaining experience in forecasting forest avalanches is challeng-ing for local avalanche forecasters. We describe a period of widespread high activity of forest ava-lanches which occurred in late February 2012. The weather situation was characterized by a sudden increase in air temperature which led to a high activity of wet and glide snow avalanches in general; however, the high number of avalanche releases in forests was surprising and challenging for safety authorities since they buried roads, railways and ski-runs. We tested an approach for forest avalan...
2022
Climate change impact on avalanches is ambiguous. Fewer, wetter, and smaller avalanches are expected in areas where snow cover is declining, while in higher altitude areas where snowfall prevails, snow avalanches are frequently and spontaneously triggered. In the present paper, we assess 39 years (winters of 1979-1999 to 2002-2020) of avalanche activity related to meteorological and snow drivers in the Krkonoše Mountains, Czechia, Central Europe. The analysis is based on an avalanche occurrence dataset for mostly south, south-easterly oriented 60 avalanche paths and related meteorological and snowpack data. Since 1979, 179 / 531 wet-snow / slab avalanches have been recorded. The aim is to analyze changes in avalanche activity: frequency and magnitude, and detect driving weather variables of wet and slab avalanches with quantification of variable importance. Especially, the number of wet avalanches in February and March has increased in the last three decades, while the number of slab avalanches has decreased with decadal variability. Medium, large, and very large slab avalanches seem to decline with decadal variability since 1961. The results indicate that wet avalanches are influenced by 3-day maximum and minimum air temperature, snow depth, wind speed, wind direction, and rainfall. Slab avalanche activity is determined by snow depth, rainfall, new snow, and wind speed. Air temperature-related variables for slab avalanches were less important than rain and snow-related variables based on the balanced random forest (RF) method. Surprisingly, the RF analysis revealed less significant relationship between new snow sum and slab avalanche activity. This could be because of the wind redistributing snow in storms in low altitude mountains. Our analysis allows the use of the identified wet and slab avalanche driving variables to be included in the avalanche danger levels alerts. Although it cannot replace operational forecasting, machine learning can allow for additional insights for the decision-making process to mitigate avalanche hazard. 1 Introduction Snow avalanches are major natural hazards. As rapidly moving snow masses, snow avalanches pose a serious threat to people, property, and infrastructure. The growth in popularity of winter tourism has led to an increase in numbers of avalanche acci-1
Computer Assisted Avalanche Forecasting: Skier-Triggered Avalanches
2004
Computer assisted avalanche forecasting has become a valuable tool in some forecasting operations in Canada. Avalanche forecasting models are based on statistical, numerical and/or rule-based methods, but usually predict natural rather than skier-triggered avalanches. A nearest neighbour forecasting model for skier-triggered avalanches on post storm weak layers was improved by adding snowpack properties including a stability index to meteorological variables. However, a skier stability index did not improve the forecasting success on storm snow instabilities. The best forecast could be achieved by using meteorological parameters.
Snow avalanche hazard prediction using machine learning methods
2019
Snow avalanches are among the most destructive natural hazards threatening human life, ecosystems, built structures, and landscapes in mountainous regions. The complexity of snow avalanche modelling has been discussed in many studies, but its modelling is not well-documented. Snow avalanche modeling in this study was done using three main categories of data, including avalanche occurrence locations, meteorological factors, and terrain characteristics. Two machine learning models, namely support vector machine (SVM) and multivariate discriminant analysis (MDA), were employed. A ratio of 70 to 30 of data was considered for calibrating and validating the models. Results indicated that both models had an excellent performance in snow avalanche modeling (area under curve, AUC > 90), although hits and misses analysis demonstrated the superior performance of MDA. Sensitivity analysis indicated that the topographic position index, slope, precipitation, and topographic wetness index were the most effective variables for modeling. A snow avalanche map indicated that the high snow avalanche hazard zone was mostly near the streams and was matched with hillsides around the water pathways. Findings of study can be helpful for land use planning, to control snow avalanche paths, and to prevent the probable hazards induced by it, and it can be a good reference for future studies on modeling snow avalanche hazards.
A manual for assessing, mapping and mitigating snow avalanche risk
2018
In summer 2018, the Canadian Avalanche Association will publish a book entitled Planning Methods for Assessing and Mitigating Snow Avalanche Risk. This book describes the methods used to assess, map and mitigate snow avalanche hazard and risk. The book is intended for the consultants, engineers, geoscientists, and their teams who prepare the reports and maps. However, to encourage readers interested in, or starting land-use planning for snow avalanche risk, the book includes hypothetical examples and illustrations in which qualitative, semi-quantitative and quantitative assessment and mapping methods are applied to diverse situations where elements at risk are exposed to snow avalanches. The book does not prescribe which methods are to be used in specific situations or jurisdictions; rather it provides a toolbox of methods for practitioners to select from, adapt and apply. The assessment and mapping chapters may be most relevant to North America and other regions where there are few written records of avalanche runouts, dynamic models are poorly calibrated, yet vegetation damage from extreme runouts are often available. The book does not cover the operational (day-today) management of snow avalanche risk by avalanche forecasters, ski guides, etc. There are 14 chapters: an introduction that frames the methods in the ISO 31000 context, six chapters about characterizing the terrain and avalanches for the situation of interest, four chapters about assessment and mapping methods, and three chapters that overview mitigation methods. The 280-page book has 16 authors with diverse experience in assessing, mapping and mitigating snow avalanche hazard and risk.
On Estimating Avalanche Danger From Simulated Snow Profiles
International Snow Science Workshop Grenoble Chamonix Mont Blanc October 07 11 2013, 2013
Estimating avalanche danger is the primary goal of avalanche warning services. Typically avalanche danger is estimated based on a variety of information such as manual snow profiles, avalanche observations as well as weather data. However, this required information is often not available especially in data sparse areas, which are common in Canada. It has been shown that coupled snow cover and numerical weather prediction models can provide such information on the snow cover. For this study we simulated the snow cover for three elevation bands -alpine, tree-line, below tree-line -at Glacier National Park, B.C., Canada for the winter season 2012-2013 between December and March. Snow cover simulations were performed using the Swiss snow cover model SNOWPACK forced by weather data from the Canadian high-resolution numeric weather prediction model GEM-LAM. Experienced forecasters estimated the regional avalanche danger (Low to Extreme) daily during the same period for the three elevation bands. Multivariate classification trees were used to estimate the avalanche danger from the simulated profiles. Classification trees were built using four parameters derived from the simulated profiles. These four parameters were the new snow amountsmaximum over 24-hours and 3-days -as well as measures for the likelihood of triggering and the expected avalanche size -based on a skier stability index and the depth of a critical layer. A comparison of the avalanche danger estimated from the simulated profiles with the forecasted avalanche danger showed that the avalanche danger was estimated correctly with an accuracy of 77% for the alpine, 76% for tree-line and 70% below tree-line -overall accuracy about 74%. Although the simulated avalanche danger tends to be slightly underestimated, especially for the alpine and treeline, such a model chain can be a valuable tool for avalanche warning services especially for data sparse areas.
Snow avalanche hazard assessment and risk management in northern Quebec, eastern Canada
Natural Hazards, 2015
In the northern environments of Quebec (eastern Canada), snow avalanche hazards have been ignored for a long time because no major incident was recorded before the tragedies of Blanc-Sablon (Lower North Shore of the St. Lawrence River) in 1995 and Kangiqsualujjuaq (Nunavik) in 1999. To enhance risk reduction at these sites, this research on process characteristics describes prone terrain, run-out distance and triggering factors, and prompted efforts (permanent and temporary measures) made to mitigate and prevent future snow avalanche tragedy from short, steep slopes. Considering the high vulnerability of these communities related to the growing population of Nunavik and the lack of knowledge of avalanches on the Lower North Shore, acceptable risk was based on the implementation of a snow avalanche forecasting and warning program over 3 years, the first one in eastern Canada. Community participation and the involvement of the municipal and provincial authorities have enabled the efficient operation of the program and accentuate the sensitivity and resilience of the populations to avalanche hazard and risk, as evidenced by the subsequent identification of avalanche sites by the communities themselves. These case studies demonstrate the importance of adequate and safe land planning, notably in the context of climate change, and particularly for isolated northern communities.