Mapping snow avalanche debris by object-based classification in mountainous regions from Sentinel-1 images and causative indices

CATENA ◽  
2021 ◽  
Vol 206 ◽  
pp. 105559
Author(s):  
Yang Liu ◽  
Xi Chen ◽  
Yubao Qiu ◽  
Jiansheng Hao ◽  
Jinming Yang ◽  
...  
Keyword(s):  
Snow Avalanche ◽  
Mountainous Regions ◽  
Object Based

scholarly journals Spatial Modeling of Snow Avalanche Using Machine Learning Models and Geo-Environmental Factors: Comparison of Effectiveness in Two Mountain Regions

2019 ◽  
Vol 11 (24) ◽  
pp. 2995 ◽  
Author(s):  
Omid Rahmati ◽  
Omid Ghorbanzadeh ◽  
Teimur Teimurian ◽  
Farnoush Mohammadi ◽  
John P. Tiefenbacher ◽  
...  
Keyword(s):  
Machine Learning ◽  
Goodness Of Fit ◽  
Snow Avalanche ◽  
Slope Position ◽  
Support Vector ◽  
Hazard Mapping ◽  
Ensemble Model ◽  
Mountainous Regions ◽  
Avalanche Hazard ◽  
Statistical Measures

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.

scholarly journals Comparing the Performance of Fuzzy Operators in the Object-based Image Analysis and Support Vector Machine Kernel Functions for the Snow Cover Estimation in Alvand Mountain

2021 ◽  
Author(s):  
Mostafa Karampour ◽  
Amirhossein Halabian ◽  
Akbar Hosseini ◽  
Mostafa Mosapoor
Keyword(s):  
Support Vector Machine ◽  
Snow Cover ◽  
Object Oriented ◽  
Remote Sensing Data ◽  
Kernel Functions ◽  
Support Vector ◽  
Fuzzy Operators ◽  
Mountainous Regions ◽  
Object Based ◽  
Wide Range

Abstract Snow cover is a significant driver in many ecological, climatic, and hydrological processes regarding the mountainous regions and high-latitude areas. Researchers believe that remote-sensing data can provide better estimates of snow cover ranges in comparison with the traditional surveying methods. Therefore, the present study was conducted using Sentinel-2B satellite images to compare the performance of the support vector machine (SVM) kernel functions and object-oriented fuzzy operators in estimating the amount of snow cover in the Alvand Mountain. In this research, the data consists of four Sentinel-2B satellite bands at 10 m spatial resolution (B2, 3, 4 & 8), launched on March 6, 2020. In this research study, the linear, polynomial, radial, and sigmoid SVM kernel functions, as well as the object-oriented fuzzy operators (AND, OR, MGE, MAR, MGWE, and ALP) have been employed. The results indicated that among these algorithms, AND algorithm, which represents the logical commonality, included the lowest return fuzzy value of 98%; therefore, this algorithm seems to provide the overall highest accuracy. Based on these findings, in the digital image classification, the object-oriented processing method can make it possible to achieve the highest accuracy compared to the SVM kernel functions. The reason is that a wide range of information, such as texture, shape, position, content, and bandwidth is associated with the objects in this classification method.

scholarly journals Three-dimensional and real-scale modeling of flow regimes in dense snow avalanches

Landslides ◽  
2021 ◽  
Author(s):  
Xingyue Li ◽  
Betty Sovilla ◽  
Chenfanfu Jiang ◽  
Johan Gaume
Keyword(s):  
Three Dimensional ◽  
Economic Loss ◽  
Density Variation ◽  
Material Point ◽  
Snow Avalanche ◽  
Snow Avalanches ◽  
Computationally Efficient ◽  
Mountainous Regions ◽  
Depth Direction ◽  
Wide Range

AbstractSnow avalanches cause fatalities and economic loss worldwide and are one of the most dangerous gravitational hazards in mountainous regions. Various flow behaviors have been reported in snow avalanches, making them challenging to be thoroughly understood and mitigated. Existing popular numerical approaches for modeling snow avalanches predominantly adopt depth-averaged models, which are computationally efficient but fail to capture important features along the flow depth direction such as densification and granulation. This study applies a three-dimensional (3D) material point method (MPM) to explore snow avalanches in different regimes on a complex real terrain. Flow features of the snow avalanches from release to deposition are comprehensively characterized for identification of the different regimes. In particular, brittle and ductile fractures are identified in the different modeled avalanches shortly after their release. During the flow, the analysis of local snow density variation reveals that snow granulation requires an appropriate combination of snow fracture and compaction. In contrast, cohesionless granular flows and plug flows are mainly governed by expansion and compaction hardening, respectively. Distinct textures of avalanche deposits are characterized, including a smooth surface, rough surfaces with snow granules, as well as a surface showing compacting shear planes often reported in wet snow avalanche deposits. Finally, the MPM modeling is verified with a real snow avalanche that occurred at Vallée de la Sionne, Switzerland. The MPM framework has been proven as a promising numerical tool for exploring complex behavior of a wide range of snow avalanches in different regimes to better understand avalanche dynamics. In the future, this framework can be extended to study other types of gravitational mass movements such as rock/glacier avalanches and debris flows with implementation of modified constitutive laws.

scholarly journals Automated snow avalanche release area delineation in data sparse, remote, and forested regions

2021 ◽  
Author(s):  
John Sykes ◽  
Pascal Haegeli ◽  
Yves Bühler
Keyword(s):  
High Resolution ◽  
Spatial Data ◽  
Large Scale ◽  
Snow Avalanche ◽  
Validation Dataset ◽  
Data Sets ◽  
Validation Data ◽  
Mountainous Regions ◽  
Percentage Points ◽  
Release Area

Abstract. Potential avalanche release area (PRA) modelling is critical for generating automated avalanche terrain maps which provide low-cost large scale spatial representations of snow avalanche hazard for both infrastructure planning and recreational applications. Current methods are not applicable in mountainous terrain where high-resolution elevation models are unavailable and do not include an efficient method to account for avalanche release in forested terrain. This research focuses on expanding an existing PRA model to better incorporate forested terrain using satellite imagery and presents a novel approach for validating the model using local expertise, thereby broadening its application to numerous mountain ranges worldwide. The study area of this research is a remote portion of the Columbia Mountains in southeastern British Columbia, Canada which has no pre-existing high-resolution spatial data sets. Our research documents an open source workflow to generate high-resolution DEM and forest land cover data sets using optical satellite data processing. We validate the PRA model by collecting a polygon dataset of observed potential release areas from local guides, using a method which accounts for the uncertainty of human recollection and variability of avalanche release. The validation dataset allows us to perform a quantitative analysis of the PRA model accuracy and optimize the PRA model input parameters to the snowpack and terrain characteristics of our study area. Compared to the original PRA model our implementation of forested terrain and local optimization improved the percentage of validation polygons accurately modelled by 11.7 percentage points and reduced the number of validation polygons that were underestimated by 14.8 percentage points. Our methods demonstrate substantial improvement in the performance of the PRA model in forested terrain and provide means to generate the requisite input datasets and validation data to apply and evaluate the PRA model in vastly more mountainous regions worldwide than was previously possible.

scholarly journals Methods of snow avalanche nourishment assessment (on the example of three Tian Shan glaciers)

2019 ◽  
Vol 59 (4) ◽  
pp. 460-474
Author(s):  
A. S. Turchaninova ◽  
A. V. Lazarev ◽  
E. S. Marchenko ◽  
Yu. G. Seliverstov ◽  
S. A. Sokratov ◽  
...  
Keyword(s):  
Data Analysis ◽  
Snow Accumulation ◽  
Tien Shan ◽  
Data Availability ◽  
Snow Avalanche ◽  
Weather Data ◽  
Snow Avalanches ◽  
Mountain Glaciers ◽  
Mountainous Regions ◽  
Numerical Assessment

The contribution of snow avalanches to the seasonal snow accumulation on a glacier is among the least studied components of the glacier’s mass balance. The methods for the numerical assessment of avalanche accumulation are still under development, which is related to poor avalanche data availability and difficulties in obtaining such data on most of mountain glaciers. We propose a possible methodology for the numerical assessment of snow avalanche contribution to snow accumulation at mountain glaciers based on DEM and weather data analysis using GIS and numerical modeling of snow avalanches. The developed methodology consists of the following steps: terrain analysis; weather data analysis; snow avalanche volume assessment during an analyzed balance year; numerical simulation of snow avalanches using RAMMS; evaluation of snow avalanches contribution into a glacier accumulation. The proposed methodology was tested on three glaciers located in the Inner Tien Shan: Batysh Sook, № 354 and Karabatkak during the 2015/16 balance year. To evaluate snow avalanche contribution to the seasonal accumulation, we reconstructed avalanche release zones that were most probably active during the 2015/16 balance year and corresponding snow fracture height in each of these zones. The numerical simulations of most probable released snow avalanches during the winter period 2015/16 using avalanche dynamics software RAMMS were performed and compared with the field observations and UAV orthophoto image from July 2016. The outlines of avalanches deposits were realistically reproduced by RAMMS according to the results of field observation. The estimated share of snow avalanche contribution to the accumulation on the research glaciers during the 2015/16 balance year turned out to be: Batysh Sook – 7,4±2,5%; № 354 – 2,2±0,7%; Karabatkak – 10,8±3,6% of the total accumulation. The next step would be to test the proposed methodology based on the data and regional dependences from the Inner Tien Shan in other mountainous regions. This methodology is applicable in the regions where DEMs, regular meteorological observations as well as data on the regional avalanche formation factors are available.

scholarly journals Mass wasting susceptibility assessment of snow avalanches using machine learning models

2020 ◽  
Vol 10 (1) ◽  
Author(s):  
Bahram Choubin ◽  
Moslem Borji ◽  
Farzaneh Sajedi Hosseini ◽  
Amirhosein Mosavi ◽  
Adrienn A. Dineva
Keyword(s):  
Machine Learning ◽  
Boosted Regression Trees ◽  
Snow Avalanche ◽  
Superior Performance ◽  
Recursive Feature Elimination ◽  
Support Vector ◽  
Snow Avalanches ◽  
Susceptibility Assessment ◽  
Mass Wasting ◽  
Mountainous Regions

Abstract Snow avalanche is among the most harmful natural hazards with major socioeconomic and environmental destruction in the cold and mountainous regions. The devastating propagation and accumulation of the snow avalanche debris and mass wasting of surface rocks and vegetation particles threaten human life, transportation networks, built environments, ecosystems, and water resources. Susceptibility assessment of snow avalanche hazardous areas is of utmost importance for mitigation and development of land-use policies. This research evaluates the performance of the well-known machine learning methods, i.e., generalized additive model (GAM), multivariate adaptive regression spline (MARS), boosted regression trees (BRT), and support vector machine (SVM), in modeling the mass wasting hazard induced by snow avalanches. The key features are identified by the recursive feature elimination (RFE) method and used for the model calibration. The results indicated a good performance of the modeling process (Accuracy > 0.88, Kappa > 0.76, Precision > 0.84, Recall > 0.86, and AUC > 0.89), which the SVM model highlighted superior performance than others. Sensitivity analysis demonstrated that the topographic position index (TPI) and distance to stream (DTS) were the most important variables which had more contribution in producing the susceptibility maps.

From avalanche-obstacle interaction processes to physics-based impact pressure calculations – an insight from DEM simulations

2021 ◽  
Author(s):  
Michael Lukas Kyburz ◽  
Betty Sovilla ◽  
Johan Gaume ◽  
Christophe Ancey
Keyword(s):  
Drag Coefficient ◽  
Flow Regime ◽  
Dynamic Pressure ◽  
Flow Regimes ◽  
Impact Pressure ◽  
Snow Avalanche ◽  
Hazard Mapping ◽  
Mountainous Regions ◽  
Avalanche Flow ◽  
The Impact

<p>Calculating snow avalanche impact pressure is an essential task for safe construction and hazard mapping in mountainous regions. Although the avalanche-obstacle interaction crucially depends on the flow regime, practitioners mostly assume that the impact pressure is similar to the dynamic pressure in inviscid fluids, that is, it is proportional to the square velocity weighted by an empirical drag coefficient. Field measurements indicate that the drag coefficients cover more than one order of magnitude. In the absence of a physics-based framework, setting the right drag coefficient requires good working knowledge and experience from practitioners. Indeed, even for trained engineers it may be unclear how the impact pressure depends on the expected flow regime, on obstacle width, or on terrain configuration. To address these questions, we simulate the avalanche impact pressure on obstacles of varying geometry for four distinct avalanche flow regimes using the Discrete Element Method and a cohesive contact model. The results allow us to quantify the influence of the obstacle width and shape on the average impact pressure, as well as the detailed pressure distribution on the obstacle surface. Furthermore, we propose a novel method for estimating the drag coefficient based on simple geometrical considerations and key characteristics of avalanche flow. Our results are validated using experimental data from the Vallée de La Sionne test site, and make a step forward in the derivation of a physics-based framework for computing snow avalanche impact pressures for varied flow regimes depending on obstacle shape and dimensions.</p>

scholarly journals Automated snow avalanche release area delineation – validation of existing algorithms and proposition of a new object-based approach for large scale hazard indication mapping

2018 ◽  
Author(s):  
Yves Bühler ◽  
Daniel von Rickenbach ◽  
Andreas Stoffel ◽  
Stefan Margreth ◽  
Lukas Stoffel ◽  
...  
Keyword(s):  
Return Period ◽  
Large Scale ◽  
Snow Avalanche ◽  
Hazard Maps ◽  
Digital Terrain Models ◽  
Avalanche Hazard ◽  
Digital Terrain ◽  
Terrain Models ◽  
Object Based ◽  
Release Area

Abstract. Snow avalanche hazard is threatening people and infrastructure in all alpine regions with seasonal or permanent snow cover around the globe. Coping with this hazard is a big challenge and during the past centuries, different strategies were developed. Today, in Switzerland, experienced avalanche engineers produce hazard maps with a very high reliability based on avalanche cadastre information, terrain analysis, climatological datasets and numerical modelling of the flow dynamics for selected avalanche tracks that might affect settlements. However, for regions outside the considered settlement areas such area-wide hazard maps are not available mainly because of the too high cost, in Switzerland and in most mountain regions around the world. Therefore, hazard indication maps, even though they are less reliable and less detailed, are often the only spatial planning tool available. To produce meaningful and cost-effective avalanche hazard indication maps over large regions (regional to national scale), automated release area delineation has to be combined with volume estimations and state-of-the-art numerical avalanche simulations. In this paper we validate existing potential release area (PRA) delineation algorithms, published in peer-reviewed journals, that are based on digital terrain models and their derivatives such as slope angle, aspect, roughness and curvature. For validation, we apply avalanche cadastre data from three different ski resorts in the vicinity of Davos, Switzerland, where experienced ski-patrol staff mapped most avalanches in detail since many years. After calculating the best fit input parameters for every tested algorithm, we compare their performance based on the reference datasets. Because all tested algorithms do not provide meaningful delineation between individual potential release areas (PRA), we propose a new algorithm based on object-based image analysis (OBIA). In combination with an automatic procedure to estimate the average release depth (d0), defining the avalanche release volume, this algorithm enables the numerical simulation of thousands of avalanches over large regions applying the well-established avalanche dynamics model RAMMS. We demonstrate this for the region of Davos for two hazard scenarios, frequent (10–30 years return period) and extreme (100–300 years return period). This approach opens the door for large scale avalanche hazard indication mapping in all regions where high quality and resolution digital terrain models and snow data are available.

Sign in / Sign up

Export Citation Format

Share Document