Master Thesis
submitted within the UNIGIS MSc. programme at the Centre for Geoinformatics (Z_GIS)
University of Salzburg, Austria
under the provisions of UNIGIS joint study programme with Panjab University, Chandigarh, India
Avalanche Hazard Zonation for Manali - Dhundi Area of Himachal Pradesh (India)
By
Dhirendra Chauhan
Student of
Masters of Geographic Information Science and Systems (MSc.GIS) GIS_UP40562
A thesis submitted in partial fulfilment of the requirements of the degree of
Master of Science (Geographical Information Science & Systems) – MSc (GIS)
Advisor (s):
Dr. S. Shahnawaz
Director (S and SE Asia) UNIGIS Internationals Interfaculty Department of Geoinformatics
University Salzburg. AUSTRIA
UNIGIS@PU_Chandigarh (INDIA)
Date 30-11-2012
Science Pledge
By my signature below, I certify that my master thesis is entirely the result of my own work. I have cited all sources of information and data I have used in my master thesis and indicated their origin.
UNIGIS@PU_Chandigarh (INDIA)
Date: 30-11-2012
Dhirendra Chauhan
Place and Date Signature
Acknowledgements
I thought the time to write this section never would have arrived. Now I have reached the end and it is time to express my deep hearted gratitude to all the people who have encouraged me to finish this work. I am thankful to Almighty God for blessing and grace showered on me to complete this work successfully.
Words are insufficient to express my deep sense of thankfulness to my Project Supervisor Mr. S.K.Dewali, Scientist, ‘D’, Snow & Avalanche Study Establishment, Chandigarh for suggesting this particular research topic and guiding me throughout this work and friendly concern, which led to its fruitful completion.
I wish to extend my sincere gratitude to Dr. Ashwagosha Ganju, Director, Dr Snehmani, Scientist`E’ Group Head (RSA) Snow & Avalanche Study Establishment (SASE), Chandigarh for providing such a nice infrastructure and environment to carry-out this research work.
I would like to express my deep sense of gratitude to Dr. S. Shahnawaz, Director South and South East Asia UNIGIS Internationals, for providing me this opportunity and guiding me throughout the course.
I would like to express my sincere appreciation to all my friends for their continuous support and valuable advice provided at all stages of this task.
Finally I express my gratitude to my parents who stood by me at all times rendering me motivation and more support. Their constant blessing and support could enable me to achieve this success.
Abstract
Avalanches are one of the most destructive phenomena of nature in snow bound areas, and therefore, Avalanche Hazard Zonation is necessary for planning future development activities. Snow avalanche hazard demarcation has the potential to reduce this risk by modeling, mapping and visualizing hazardous terrain using Geographic Information Systems (GIS). My study area is in and around Manali - Dhundi in Himachal Pradesh (India). Study area is located between the latitude of 32°4’30”N to 32°26’30”N and longitude of 77°0’0”E to 77°22’0”E about 40 km north of Kullu town, which is an ideal location for studying well documented avalanche paths that impact the stretch of Manali- Leh Highway. This study presents a method for avalanche hazard zonation based on terrain and ground cover analysis. Modeling of terrain in a GIS is typically done by utilizing a digital elevation model (DEM) of study area, which displays its elevation values. A high resolution digital elevation model (10m), generated from IRS PAN Ortho-Stereo imagery (IRS-P6 Satellite) has been used for extracting the desired terrain parameters such as Slope, Aspect, Curvature. AWiFS (Advanced Wide Field Sensor) multi spectral imagery at resolution of 56m was used for characterization of ground cover. After extracting the terrain parameters and classification of ground cover for the area an MCE (Multi-Criteria Evaluation) and AHP (Analytic Hierarchy Process) methods are applied to the whole terrain, ground cover, elevation factors using GIS and two demarcated avalanche hazard zone maps are produced. These methods are then compared to each other. Registered Avalanche site (by Snow and Avalanche Study Establishment, Chandigarh) data are used for the validation purpose.
This study aims at the demarcation of avalanche hazard zone based on the terrain parameters and ground cover conditions of Manali - Dhundi area using GIS. GIS has been used in combination with DEM and land cover data for the mapping avalanche
hazard in mountainous regions across the world. The results are useful in supplementing traditional field based methods of avalanche hazard mapping as well as providing a tool for risk assessment.
Key Words: Snow Avalanches, GIS, Remote Sensing, DEM, AWiFS, Terrain Parameters, MCE, AHP, Hazard Zone, Registered Avalanche Sites.
Table of Contents Science Pledge
Acknowledgments……….………
Abstract………..……….
Table of Contents………..
List of Tables ……….
List of Figures and Maps……….
Chapter 1 Introduction...10-26 1.1 Background………...10-13 1.1.1 Snow Avalanches………....13-15 1.1.1.1 Types of Avalanches……….………...15-16 1.1.1.2 Main factors for Avalanche Initiation……….16-17 1.1.3 Avalanche trigger site factors……….17 1.2 Objectives of current study………17-18 1.3 Location and general description of study area………...18 1.4 Literature Review……….18-26 Chapter 2 Methodology……….……… 27-32 2.1 Methodology……….27-28 2.2 Software Used………28 2.3 Data used ……….……29-32 2.3.1 Digital Elevation Model……….………..29 2.3.2 AWiFS Imagery……….………..29-31 2.3.3 Registered Avalanche Sites………..………...32
1 2 3 4-5 6 7-8
Chapter 3 Processing of Data and Result………..………..…33-60 3.1 Terrain features………..33-39 3.1.1 Altitude Map………/………33-34 3.1.2 Slope map………...34-36 3.1.3 Aspect map……….36-37 3.1.4 Curvature map………37-39 3.2 Ground cover map…………..………..40-48 3.2.1 Preparation of NDSI model………..40-43 3.2.2 Preparation of NDVI model……….……….44-46 3.3 Multiple-criteria analysis……….……….48-53 3.3.1 Assigning weightages to each terrain and ground cover
parameters………..……….48-53 3.4 AHP Analysis………53-54 3.4.1 AHP weightage scheme……….54 3.4.2 calculation of weightage %...54 3.4.3 Avalanche Hazard Risk zones………..54 3.5 Result and discussion……….54-60 3.5.1 Classification of avalanche hazard zones (MCE Method)……….55-56 3.5.2 Registered Avalanche Sites………..56-57 3.5.3 Classification of avalanche hazard zones (AHP Method)……….57-60 3.6 Comparison between MCE and AHP method………..60 Chapter 4 Conclusion………..60-61 References………..63-66
List of Tables
Table No 1 Avalanche trigger site terrain parameters……….………..17
Table No 2 Parameters of AWiFS camera……….30
Table No 3 Registered Avalanche Sites………..32
Table No 4 Reclassification of Aspect...36
Table No 5 Reclassification Curvature……….39
Table No 6 Slope weightage scheme...50
Table No 7 Aspect weightage scheme...51
Table No 8 Curvature weightage scheme...52
Table No 9 Ground cover weightage scheme...52
Table No 10 Weightage scheme for altitude...52-53 Table No 11 Combined weightage scheme to causitive parameters...53
Table No 12 Comparison scale ...53-54 Table No 13 Weightage scheme during AHP...54
Table No 14 Calculated weights during AHP process...54
List of Figures
Figure No1 Avalanche Triangle ... ...16-17
Figure No 2 Shows risk prediction on the basis of Slope...25
Figure No 3 Shows risk prediction on the basis of Curvature...26
Figure No 4 Shows risk prediction on the basis of Ground Cover...26
Figure No 5 Flow Chart…….……….28
Figure No 6 AWIFS imagery...31
Figure No 7 NDSI Model……….…...41
Figure No 8 NDSI Analysis……….42
Figure No 9 NDSI Mask……….43
Figure No 10 NDVI Model………..……….………..44
Figure No 11 NDVI Analysis………..………45
Figure No 12 NDVI Mask………..………46
Figure No 13 shows the cardinal directions...51
List of Maps
Map No 1 Altitude Map……...34
Map No 2 Slope Analysis………...35
Map No 3 Aspect Analysis...37
Map No 4 Curvature Analysis...38
Map No 5 Classification of Curvature...39
Map No 6 Classification of ground cover...48
Map No 7 Avalanche Hazard Zones using MCE………..………56
Map No 8 Registered Avalanche Sites and Risk Zones………..57
Map No 9 Avalanche Hazard Zones using AHP………...59
Chapter-1 1. Introduction
1.1 Background
Mountains are distributed across all the world’s continents. Mountains include vast diversity of environment: from the wettest to the driest, from hot to cold, and from sea level to 8848 meters at the top of Mount Everest which is also known as Sagarmatha and Chomolungma (Heywood, D .Ian., et al 1992). For centuries, people have been altering the earth’s surface to produce food, gain material and generate energy through various activities. People are moving to mountains, constructing new cottages, hotels, homes etc.
and spending more time there (Richnavasky and Biskupic et al 2011). Mountain terrain greatly influences the climate and vegetation of locations across the earth’s surface. It also influences a variety of human activities, especially those occupations that are directly related to the land. Planners, research scholars, resource managers, environmentalists working in the fields of hydrology, soils, geomorphology, troops and tourists often encounter the need of topographic data. Using DEM (digital elevation model) and DTM (digital terrain model), the analysts incorporate topographic relationships in tasks such as surface characterization, site visibility analysis, wildlife habitat mapping, hazards prediction, satellite imagery classification and modeling human and environmental interactions.(Schneider and Robbins, 2009). Despite the recognized diversity of mountain environments and regions at all scales, it is possible to recognize a common set of issues and problems for the managers who wish to apply GIS to management and research within these regions. GIS have been used for many objectives in mountain areas around the world. To all the goals and purposes, these applications reflect environmental, cultural and economic issues that have come in front of people living in mountainous regions in recent years. The issues faced by people include; forestry, mining ecology,
tourism, hazard mapping, visual impact assessment, and climate change modeling. GIS acts as a tool to assist in resource inventory and the integration of data and as a mechanism of analysis, modeling and forecasting to support decision making (McKendry and Eastman, 1992).
Earth’s land moves at least few centimeters a year. But there will be times when it can move in meters, even kilometers at a time. This can happen when there is a landslide, avalanches or mudslides. As the land slowly moves by a few millimeters a day, it causes tension cracks in the earth and in the bed rock. As the tension increases a significant change in moisture occurs and results in forming things like heavy rainfall or the fast snow melting, the land begins to move more and more. Melting of ice, snow in northern hemisphere can cause ice dams that block rivers and force water to burst shorelines.
(Source [Online]; http://www.naturaldisasters.ewebsite.com/articles/mudslides_-
avalanches-and-landslides.html).
Water in its frozen state accounts for more than 80% of the total fresh water on earth and is the largest contributor to rivers and ground water over major portions of the middle and high latitudes. Snow is an important, though highly variable, earth surface cover. Its presence affects physical, chemical and biological processes and has important economic and societal impacts. The high albedo of snow coupled with its large aerial extent makes it a strong influence on the Earth’s radiation budget. Runoff from snowmelt is an important water resource in many regions (Negi, H.S., et al. 2011) of the world and heavy late season snowfalls can cause disastrous flooding. Disasters are as old as human history but the increase and damage caused by them in recent past have become a cause of national and international concern. Over the past decade there is an increase in natural and man-made disasters. During the years 1994-98, disaster occurrence average was 428 per year while it has gone very high for the years 1999-2003 as 707 events per year. The statistics show that there is an increase of about 60% over previous year. The scenario for India is not different than above mentioned stats. In India, 59% of
land mass is susceptible to seismic hazard, 5% of total geographical area is prone to floods, 8% of total land mass is prone to cyclones, 70% of total cultivable area is vulnerable to draught. Apart from this the mountainous regions are vulnerable to avalanches (snow bound areas)/landslides/hailstorms/cloudbursts. These hazards (Hazard is defined as dangerous condition or event that threaten or have potential for causing injury to life or damage to property or environment. The word hazard is derived from French word ‘HASARD’ and Arabic word ‘AZ-ZAHR’, which means ‘Chance’ and
‘luck’. Hazards are divided into two categories, (Natural and Man-made hazards) which are very frequent and cause of huge damage to life and property (Dey, Balaka. and Singh, R. B., 2006).
The Himalaya, the longest chain of mountains in the world, shows complex variation in snow and meteorological conditions. Glaciers and snowfields normally exist in remote and inaccessible areas and the data collection on regular basis becomes quite difficult and dangerous. Thus, logistically and from manpower point of view, it is very difficult to collect snow data and the point data collected is not representative of the whole area (Joshi and Ganju, 2010).
The satellite remote sensing has a great potential in the study of dynamically changing environments related to the high altitude cold regions mainly because of their repetitive coverage capability, high resolution and synoptic view. As snow cover is very dynamic in nature, the frequent repetitive coverage is an advantage. The advent of multi-spectral, multi-temporal and multi-sensor satellite remote sensing has opened up the possibility of data acquisition in such terrains at regular intervals (Jain, Sanjay. K.,(???)).
Snow is highly reflective in the visible region of the electromagnetic spectrum, making it possible to easily distinguish on an optical image. However, cloud-cover and mountain shadow presents a number of obstacles in snow cover analysis. This problem, to a large extent, can be overcome using all-weather, day and night microwave data.
Snow is an important natural resource of energy, as snow melts, water is used for irrigation, electricity production, domestic purposes (Flerchinger and Cooley 2002, Marks and Winstral, 2001, and Archer 2003) along with this it is also a hazard in the form of avalanches, which in turn, strongly affects the safety of human lives and many other development activities contributing to national economy. Himachal Pradesh, J&K (Jammu and Kashmir), Uttaranchal and certain parts of North Eastern States are affected by avalanche hazard in India (Singh and Ganju 2006).
1.1.1 Snow Avalanches
Snow avalanche is a mass of snow, often mixed with ice and debris which travels down mountain slopes, destroying everything coming in its path. The word avalanche is derived from the French word avalanche which means descent. It generally occurs because of structural failure of deposited snow mass on the slope (Johnston., 2011). The term ‘Avalanche Hazard’ generally refers to the exposure of people and property to the destructive effects of avalanche. The avalanche causes irreparable losses of life and property, imposes severe constraints on development activities in the hill regions and strongly affects power generation, irrigation, recreation and many other activities which contribute to national economy (Marks et al. 1998). The variations in the local climate, environment and altitude as well as fast snow cover build up and rapid changes in snow characteristics with passage of winter are the major contributing factors to make snow avalanches as one of the threatening problems (Jamieson, Bruce., et al 2002) in the North Western Himalaya. Snow avalanche is one of the natural hazards in India’s mountain areas during winter. This causes about 100 of deaths in India every year as well as damage to mountain villages, settlements, infrastructures and forests (Raho, N. et al 1987). According to a study of several hundred avalanches, 90% of avalanches with (fatal) accidents were directly or indirectly triggered by the victims themselves, only 6%
are of natural causes while only 4% are of known causes (Mc-Cammon., 2000). Snow
avalanches are complex natural phenomena even experts in the same field do not fully understand all their causes. No one can predict avalanche occurrences with certainty.
Sometimes avalanches are the causes of deaths in snow bound areas. The more time you spend skiing, snowboarding, snowshoeing, snowmobiling and enjoying other winter activities, more are the chances of being caught in avalanche (Young, Lance.).
Avalanches cannot be considered as a problem limited mainly to local inhabitants of mountainous area or to their property and infrastructures. They also affect tourists /sportspersons, research scholars and tourists coming from other regions. To overcome this hazard, we have a computer based technique which will help us to minimize the loss (human as well as structural) due to this hazard and will act as a risk assessment tool for avalanche prone areas (McCollister, Chris., and Birkeland, Karl, 2005). The computer based technique is known as Geographic Information System (GIS).
A Geographic Information System is a powerful tool with the capability to capture, store, manipulate, analyze and display spatial information and related attributes of diverse nature (Bernhardsen, Tor., 1992). It is quick, efficient and powerful in “number- crunching” the avalanche problems. GIS in conjunction with Remote Sensing has been discovered as a powerful tool for solving avalanche problems.
Remote sensing is the science for making deductions/influences about object/phenomena by making measurements, without coming into physical contact with them. Remote sensing is a powerful tool, which provides the ability to quantitatively examine the physical properties of snow in remote or inaccessible areas where measurements may be expensive and dangerous (Nolin, Anne. W., 2010). The term remote sensing has been restricted to detect/identify various earth objects by measuring electromagnetic radiation (EM), which is reflected or emitted from the surface. Remote sensing is essentially uses devices called sensors, which are capable of detecting and/or measuring the reflected or emitted radiations. Every object reflects/scatters a part of incident EM radiation depending upon its physical properties or emits radiation depending
upon its temperature and emissivity (Allen and Walsh. 1996).Thus each object has its characteristic reflectance/emmittance pattern, which is called ‘spectral signature’. As we identify a person based on his/her ‘signature’ or ‘finger prints’ spectral signature allows us to identify particular object.
A remote sensing system consists of data acquisition, data processing and analysis. It has the following elements (Walsh, Michel. J., 2003).
Electromagnetic Radiation-Source of Energy
Propagation of radiation through earth’s atmosphere
Interaction of radiation with matter
Active and passive sensors
Platforms, airborne or space-borne
Data products generation
Data analysis, integration of remote sensing data with collateral data and resource management
Hence, Remote Sensing in association with GIS is envisaged to be an ideal tool for assessing, registering, mapping and classification of avalanche hazard prone areas.
Past study has proved that GIS has a capacity to model, analyze, predict map and visualize (Hunter, Laraine. 1998) avalanche terrain to build and improve upon avalanche terrain recognition and education for people living in the area. (McCollister et al).
1.1.1.1 Types of Avalanches
Avalanches vary, though they may seem to be the same. Some avalanches are very dangerous and can cause loss of a person’s life while others only have mild effects on people. There are many types of avalanche, each varies according to cause of occurrence. There are main three types of avalanches, which are mentioned below:-
Powder Avalanche. This type of avalanche often starts from a single point and accumulates snow as it moves down the slope forming a snowball effect. This type
is most common following heavy snowfall or more and often on a smooth surface as after rain or frost.
Slab Avalanche. This is the most common type of winter avalanche, and is formed due to fresh snow fall. Slab avalanches generally occur when a packed portion of snow becomes loose. The slab is difficult to see and avoid, one cannot determine when the snow will move. Most of the time, a person can still travel on it before it falls. This type of avalanche is the primary cause of the most number of casualties among travelers.
Wet Avalanche. This type of avalanches occur after a warm spell or during the spring thaw. Snow becomes heavier as it begins to turn into water. Wet avalanches occur frequently and are generally small and easier to predict than other types of avalanches.
1.1.1.2 Main Factors for Avalanche Initiation
Avalanche hazard in a particular area highly depends on the weather, terrain and snowy conditions of that area (Perla, R.I. 1970). Avalanche occurs as a result of interaction of the snow cover, weather (Meteorological parameters) and terrain. These three factors are known as avalanche triangle (Fredston, J. and Fesler, D., 1994).
Avalanches are also formed due to structural weakness within the snow cove (Buission, Laurent. and Claude, Charlier., 1993). Each factor may work in isolation or together in defining the degree of suitability, because each may enhance or detract the suitability of starting an avalanche at a particular location.
Terrain parameters
Snow pack Weather
Figure No 1 Avalanche Triangle 1.1.1.3 Avalanche trigger site Terrain parameters
Terrain parameters are one of the most important factors responsible for the triggering of avalanches (Mclung, David. and Schaerer, Peter., 2006). These causative terrain parameters are; Slope, Aspect, Ground Curvature, Altitude and Ground Cover. A risk scale is developed to assess the risk on the bases of these parameters. (Stanthan and McMohan., September 2004). It includes the following sub-factors.
SLOPE > 55° too steep for thick snowfall accumulation 25°-45°avalanche formation zone
<25°run-out or transition zone ASPECT NE Quadrant (0°-90°)
SE Quadrant (90°-180°) SW Quadrant (180°-270°) NW Quadrant (270°-360°) TERRAIN SHAPE Convex - Potential release area
Planar - Least potential release area Concave - Accumulation zone GROUND-COVER Snow Area- Potential release area
Vegetation- Least avalanche prone area
Altitude 3500mtrs-4500mtrs - Maximum formation of avalanches Table No 1 Avalanche trigger site terrain parameters
(Source:- Jamieson Bruce and Feldsetzer Torsten 1984-96, Simea Ioana (???), Gleason 1994)
1.2 Objectives of study
The aim of study is to demarcate the avalanche hazard zone areas around Manali-Dhundi region using GIS. Avalanche hazard zonation mapping has gained high importance in land use planning in the mountainous regions (Barbolini et al., 2000; Gruber and Margreth, 2001; Pudasaini and Hutter, 2007). The outcome of current study will act as a
risk assessment tool and plays major role in future development which will imperatively improve the economy of our study area. The objectives of study are:-
Identify and classify the terrain features using DEM (In ArcGIS9.1), classification of ground cover parameters from AWiFS (Advanced Wide Field Sensor) satellite imagery by using NDSI and NDVI models and generation of elevation map from DEM.
Integrate the above data in a GIS environment and analyze terrain parameters (slope aspect, curvature), ground cover parameters (snow/vegetation/barren land) and elevation map.
Assign weightages to the individual causative parameter for triggering of an avalanche phenomenon, and integrate them.
Demarcation of avalanche prone areas.
Use Avalanche triggering sites data for validation purpose.
1.3 Location and general description of Manali - Dhundi area
In present study, Manali - Dhundi area of Himachal Pradesh (India) Indian Himalaya, has been selected as a study area. The study area is located between latitude 32°4’30”N to 32°26’30”N and longitude 77°0’0”E to 77°22’0”E about 40 km north of Kullu town. Study has been conducted for 14th December 2008. Climatically, this area falls in the lower Himalayan zone (Sharma and Ganju 2000) and characterized by moderate temperature as mean minimum temperature −1.6°C, lowest minimum−12 °C and mean maximum 7.7°C during winter (November to April) and high precipitation as mean
standing snow 165 cm and mean cumulative snowfall 363 cm in the month of January (Singh and Ganju 2006).
1.4 Literature Review
Allen and Walsh, 1996 has studied the patterns of alpine treeline across the area of Glacier National Park, Montana. To classify the ground cover satellite image has been used, digital terrain modeling and GIS measurements of landscape structure provided important tool for the analysis.
Almuhairi, Maitha. Aylan., 2008 has studied the digital image analysis. During this analysis author has applied different- different image classification techniques.
Supervised and unsupervised classification techniques are most frequently used techniques, these techniques are mostly used for image classification. Authors have described the difference between hard image and digital image during the analysis.
Gleason, J.A. 1994 has studied the avalanche hazard on mountainous areas depending on terrain, weather and snow cover parameters. These parameters acts as an avalanche triangle. He observed that mostly avalanches occurs due to interaction between snow cover, terrain parameters and weather of the mountainous region.
Stethem et al 2002 authors have developed an avalanche course in Canada. To develop this course they have defined some objectives. Objectives of study were to prepare uniform national guidelines for risk analysis and avalanche mapping, to inform land managers about recognition and mitigation of avalanche hazards and develop a training course in avalanche mapping for land-use planning. Authors have divided risk zones into three categories named, White zone, Red zone and Blue zone.
McClung. 2001 statistically analyzed the parameters characterizing the release areas of 76 avalanche sites. Slope inclination, wind index, ground surface roughness, vegetation cover, slope orientation, cross slope curvature, elevation, vegetation heights and avalanche size were the terrain parameters considered. For each of these, the most
frequently recurring classes, within past release area, were highlighted; showing their individual contribution to avalanche initiation.
Thome et al 2006 have studied the Indian Remote Sensing (IRS) P6 Advanced Wide Field Sensor (AWiFS). Authors have described the parameters of AWiFS sensor.
AWiFS camera is split into two separate electro-optic modules (AWiFS-A and AWiFS-B).
It has a resolution of 56 meters.
Maxton, Anupam. Andrew., 13 October 2009 estimated the daily Evapotranspiration for Tons Canal Command area using remote sensing. The study was conducted for 20th November 2000. Authors used landsat data to reach the goal of study.
He has concluded that GIS in conjunction with Remote Sensing has been discovered a very useful decision making technique.
Delparte, Donna. M. 2008 has proposed a method for avalanche hazard zonation on the basis of terrain parameters. In this, he has described the analysis of satellite data in conjunction with terrain parameters using GIS.
McCollister, Chris., and Birkeland, Karl, 2005 have observed that if used correctly, GIS can be a useful and powerful tool for avalanche work. GIS can be used to effectively display and interpret the increasing volumes of data available to avalanche forecasters. Authors have strong belief that GIS technology will continue to improve, Digital Elevation Models will get better and Remote Sensing techniques for identifying land-cover will also get better in future.
McKendry and Eastman, 1992 have studied that mountains are spread across all of the world’s continents and these include vast diversity of environment. Despite the recognized diversity of mountain environments and regions at all scales, it is possible to recognize a common set of issues and problems for the managers who wish to apply GIS to management and research within these regions. GIS have been used for many objectives in mountain areas around the world. To all the goals and purposes, these applications reflects environmental, cultural and economic issues that have come in front
of people living in mountainous regions in recent years. The issues those are faced by people; include forestry, mining ecology, tourism, hazard mapping, visual impact assessment, and climate change modeling. GIS acts as a tool to assist in resource inventory and the integration of data and as a mechanism of analysis, modeling and forecasting to support decision making. The authors have recommended GIS as a decision making technique.
Richnavasky and Biskupic et al 2011 have described the avalanche incidence of Magurka 1970. Authors have also studied the loss caused by this avalanche in the study area and suggested that to minimize the loss, for forecasting of avalanches, GIS study is required. After the GIS analysis of avalanche incidence of study area it was concluded that the introduction of GIS technology has opened up new perspectives to mapping and assessing hazards from snow avalanches.
Roshani et al 2008 have measured snow cover area for AlamChal glacier located in the vicinity of Kelardasht city in the north of Iran, using remote sensing data. It is observed that measuring snow cover on the basis of ground observations is more difficult than using satellite data. The authors have used landsat data to classify snow and non snow areas.
Heywood, D. Ian., et al 1992 Mountains are spread across all of world’s continents. They include a vast diversity of environments. The authors provide an overview of current applications, issues and challenges that are faced by people living in mountainous regions and receiving attention by mountain scientists using GIS. Beginning with a consideration of the special characteristics of mountain regions, it examines the extent to which the use of GIS in these regions reflects such characteristics and assesses whether current GIS technology can meet the demands of scientific research and management in mountain. It is also observed that GIS has broad use in decision making, but technologically it has received relatively limited use in mountain environment.
Dey, Balaka., and Singh, R. B., 2006 have studied natural hazard and disaster management . Authors have termed disasters as old as human history. In the chapter it is also mentioned that disasters are increasing year by year. In the article, authors have described major disasters occurred in India since 1970 (Super cyclone in Orissa 1999, Earthquake in Gujarat 2001, Tsunami in 2004 and floods in north-eastern states) and their impact on lives, environment, structural properties has also been mentioned.
Negi, H. S. et al 2009 authors have developed a methodology for the mapping of snow cover in Beas basin, Indian Himalayas using AWiFS (IRS-IP6) Satellite data. In the study, author has also described NDVI & NDSI models for snow and vegetation classification.
Negi, H. S, and Kokhanovsky, A., 2011 have studied the seasonal snow cover for lower Himalayan zones. He described snow as an important natural resource to the people living in the mountainous areas. During the study author has studied ART (Analytical Radiative Theory) and observed that this theory is very useful for the retrieving of the snow properties.
Jamieson, Bruce and Feldsetzer 1984-96 have mentioned the various patterns and trends of avalanche accidents between 1984-96, for this the have utilized the terrain, snow cover, weather data of study area for relevant time period.
Stanthan, Grant. and McMohan, Bruce. 2004 Authors have described an avalanche risk scale based on terrain parameters. In this he has described the importance of individual terrain as well as ground cover factors in the initiation of avalanche.
Joshi and Ganju, 2010 have studied that temperature and fresh snow are very essential inputs in an avalanche forecasting model. Without these parameters, prediction of avalanche occurrence for a given region would be very difficult. This study is aimed at to estimate temperature and precipitation intensity on various regions of Indian Himalaya by using Barnes objective analysis, which is a convergent weighted averaging interpolation technique that can be used to derive the desired amount of details in the
analysis of a set of randomly spaced data. The developed scheme is tested at various places over Indian Himalaya and compared the result with observations.
Campbell, James. 10 December 2001 has studied different methods of image classification, he had described that unsupervised method is the method where human errors are very less and there is need of prior knowledge of study area.
Buission, Laurent. and Claude, Charlier. 1993, described the two methods for the calculation of snow cover stability. In first method he studied that if the upper layer is slab and lies on weak layer with no cohesion or sliding surface then the stability depends on value of slope angle. While in second method he studied the soil mechanism and studied the equation for stability and observed that if h > h criteria then the upper layer is unstable layer.
Hunter, Laraine. 1998, author has described two most destructive hazard events such as the 1989 New Castle earthquake and 1990 Sydney hailstorm. It is also observed in the study that GIS has proven to be valuable tool in assessing the spatial pattern’s of damage caused by such hazards. Author has also described that how GIS remains an integral part of modeling process and the change in basic GIS programs over the years have allowed us to improve modeling techniques significantly.
Kulkarni 2007, has observed the effect of global warming on the Himalayan Cryosphere, in which he described the change in glacial extent, glacial mass extent and the change in seasonal snow cover for Chenab,Parbati and Baspa basin. He has described the NDSI method, to extract snow and non snow areas parameters from AWiFS imagery.
Nolin, Anne. W., 2010, has described remote sensing as a powerful tool providing the ability to quantitatively examine the physical properties of snow in remote or inaccessible areas where measurements may be expensive or dangerous. Author has highlighted the advanced analysis techniques that have been developed during past decade. The areas of advancement include improved algorithms for mapping snow cover
extent, snow albedo, snow grain size, snow water equivalent, melt detection and snow depth as well as new uses of instruments such as a multi-angular spectro-radiometers, scatterometry and lidar.
Strobl, Josef. 2007, has presented MCE method in which he found that with the help of ranking and weights, information from two or more than two spatial datasets can be combined to produce a single index of evaluation. To combine all these information he recommended Overlay spatial tool.
Rao,N. Mohan. et al 1987, has studied problems faced due to avalanche hazard in Himachal Pradesh and also has described an avalanche incidence of 6th March 1979, when an avalanche triggered from a slope of Guskiar village in Kelong district taking a heavy toll of human lives, forest wealth and properties. Avalanche has a formation zone area of about 34 hectare, triggered from an altitude 4500 m. The average slope avalanche path in the formation zone is 36°. The length of track was nearly 3.5 km with an average slope of 24°.
Singh and Ganju. 2006, have studied the complex mountain ranges of Western Himalayas and its diverse snow climatic zones. Authors have also developed a mountain range specific analog weather forecast model by utilizing surface weather observations of reference stations in each mountain range in North West Himalaya. The developed analog weather forecast model is tested with independent dataset of more than 717 days (542 days for Pir Panjal range in HP) of the past four winters 2003-04 to 2006-07. The results of independent test are reasonably good and suggest that there is some possibility of forecasting weather in operational weather forecasting by applying analog method over different mountain ranges in NW Himalaya.
Young Lance (???), author has described terrain parameters causing avalanches in mountainous areas. He has also studied the complex nature of avalanche and its types.
According to author even the experts in the field of snow avalanches do not fully understand all their causes and he suggested that only the knowledge can help an
individual to avoid being caught in avalanche; it may also help you survive if you are buried.
1.5 Topographic parameter affecting avalanche release
General risk assessment generated on the basis of current study is shown in the following figures indicating which areas are safe and which are not. This type of information is very useful for the travelers, troops, research, skiers and other people during the winters while carrying out their respective activities in this snowbound area.
Figure No.2 Risk prediction on the basis of Slope
Figure No.3 Risk prediction on the basis of Curvature
Figure No.4 Risk prediction on the basis of Ground Cover
Chapter-2 2.1 Methodology
Mountains are distributed across all of world’s continents. These include a vast diversity of environment. Given the dynamic nature of mountain environments, natural hazards, such as avalanches, landslides, fires and floods are major problems of concern for management. GIS can be used in many various ways to increase the understanding of these phenomena and assist in developing adaptation and prevention strategies .Avalanches are one of the natures most destructive phenomena in snow bond areas. So avalanche hazard zonation is must for planning and future development. To accomplish the objective of study, high resolution (10m) Digital Elevation Model generated from IRS PAN Ortho-Stereo Imagery has been used. The DEM generated is used to derive terrain parameters such as slope, aspect, curvature and elevation. AWiFS Imagery in study is used for the classification of snow and vegetation existing in study area. Snow and vegetation maps are generated using NDSI and NDVI algorithm. Two different methods are evaluated to demarcate avalanche risk zonation. Multi-Criteria Evaluation (MCE) developed by Edwin et al 1990 and Analytic Hierarchy Process (AHP) developed by Saaty, 1977 are the methods that integrate all the terrain features and ground cover information into a single index evaluation. The two different risk zonation maps from MCE and AHP are developed. The classified maps are categorized into four risk zones i.e., No risk area, Low risk area, Moderate risk area and High risk area) (Delparte, Donna. M., 2008). Outcome of these two methods are compared to check the accuracy level. Figure 5 shows the methodology flow chart. Following procedure have been adopted for the execution of the project objectives:
1. Derivation of terrain parameters such as, slope, aspect, curvature from DEM 2. Generation of snow cover area (NDSI) and vegetation cover (NDVI) from AWiFS
3. Integration of terrain parameters and ground cover information in a GIS environment (ArcGIS).
4. Use MCE and AHP to integrate the parameters
5. Assign weightages to individual components based on causative parameters 6. Identify and map the avalanche hazard prone areas
Figure No 5 Flow Chart of methodology
2.2 Software used
To accomplish the study GIS and Remote Sensing software have been used.
Software used to generate terrain features and classify AWiFS imagery are:-
ArcGIS (ver. 9.1) is used for data integration
ERDAS Imagine Professional (ver. 9.1) used for ground cover classification.
2.3 Data Used
The study area falls under the SOI toposheet No., 52h/3, High resolution digital elevation model of Manali - Dhundi area and IRS-P6 satellite imagery (Captured by AWiFS Sensor) has been used in study.
2.3.1 Digital Elevation Model of Study Area
DEM provides the digital representation of elevation values for our study area. A high resolution digital elevation model (10m), generated from IRS PAN Ortho-Stereo imagery is used to derive all the terrain parameters, all the terrain parameters are extracted from DEM with the help of Spatial Analyst tools in ArcGIS (9.1).
2.3.2 AWiFS Imagery
We perceive the surrounding world through our five senses. Out of these senses some senses (taste and touch) require contact of our sensing organs with the objects.
Remote Sensing is defined as the science and technology by which the characteristics of objects of our interest can be identified, measuring or analysed without direct contact.
Remote Sensing deals with gathering information about the earth from a distance. This can be achieved from few meters from Earth’s Surface, with the help of aircraft flying hundreds thousands of meters above the surface or by a satellite orbiting hundreds of kilometers above Earth. Remote Sensing satellites are equipped with sensors which look down at earth and captures the images for time and again. The sensors are like eyes in the sky constantly observing the earth. Remote Sensing satellite images provide a synoptic view of any place on Earth’s surface. This allows us to study, map, and monitor the Earth’s surface at local, regional and global scales. Remote Sensing is cost effective and provides a better spatial coverage as compared to ground sampling. Electro-magnetic radiation reflected or emitted from an object is the important source of remote sensing.
Each object has a unique characteristic of reflection or emission if the environment conditions are different. To detect the electro-magnetic radiation reflected or emitted sensors are used. These sensors are mounted on the satellites and different satellite systems have different characteristics for example resolution, number of bands, and have their own importance for different applications (Shrestha et al 2001). In the current study a satellite image (of 14th December 2008) captured by AWiFS (Advanced Wide Field Sensor) sensor has been used to classify ground cover of Manali - Dhundia, Himachal Pradesh (India). Past studies in relevant field concludes that satellite images provide a powerful tool for the identification of ground cover when used in combination with geographical information systems (GIS). It has been proved to be effective for demarcation of avalanche prone areas and land use (Maxton, Anupam. Andrew., 13 October 2009). Advanced Wide Field Sensor (AWiFS) provided by IRS P6 (Indian remote sensing satellite known as ResourceSAT-1) satellite has been used in study to develop snow and vegetation cover. Figure 7 shows AWiFS imagery with study area. The technical specification of AWiFS sensor is given in table No 2.
Sl.
no. Parameter Values
1. Ground sampling distance (m) Across track
Along track
56 (nadir), 70 (off-nadir)
66 for an integration time of 9.96 ms 2. Swath (w/o earth curvature effect)
740km
3. Bands (μm) B2 B3 B4 B5 0.52–0.59 0.62-0.68 0.77–0.86 1.55–1.70
4. Quantization (Bits) 10
5. Signal-to-noise ratio @ saturation
radiance > 512 (for all bands)
6. Band-to-band registration (pixel) ≤±0.25
Table No 2 Parameters of AWiFS camera (source: Thome et al 2006).
Figure No 6 AWiFS Imagery (Date of satellite pass 14 Dec 2008) Manali
Dhundi
2.3.3
Avalanche Sites Data
The locations of registered avalanche sites across Manali - Dhundi area have been provided by Snow and Avalanche Study Establishment, Chandigarh, India. These sites are created after the field survey carried by SASE team across Manali - Dhundi area.
Table No 3 shows the registered avalanche sites along with the area (in hectares) for each site.
Table No 3 Registered avalanche sites Registered
Avalanche Sites
Area (In hectares)
MSP1 14.915 MSP2 12.920 MSP3 222.84 MSP4 24.835 MSP5 18.620 MSP6 6.8102 MSP7 347.80 MSP8 12.267 MSP9 16.800 MSP10 32.292 MSP11 29.735 MSP12 31.982 MSP13 35.015
Chapter-3
3 Processes and Results
Avalanches are one of the most destructive phenomena of nature in snow bound areas, and therefore avalanche hazard zonation is necessary for planning future development activities. Current study presents the method for avalanche hazard zonation in which terrain and ground cover parameters have been integrated in GIS environment (MCE and AHP method are applied to generate risk zones). The study was conducted for 14th December 2008, for Manali Dhundi areas, Himachal Pradesh, India.
3.1 Terrain features
High resolution (10m) Digital Elevation Model is used to generate the terrain features such as altitude, slope, aspect and curvature. GIS software ArcGIS is used to generate these maps. All these causative parameters are combined in a reasonable way and according to multi-criteria and analytical hierarchy process analysis will allow to us to establish the existing avalanche hazard within the study area. For these various thematic data layers corresponding to causative factors namely, altitude, slope, aspect and curvature (using DEM) and kind of ground cover (Using AWiFS Imagery) have been prepared.
3.1.1 Altitude Map
As altitude increases, risk factor also increases because environmental, spectral conditions changes from region to region (Arora and Mathur 2001). Snow fall, wind and temperature varies with increase in elevation. Higher slopes tend to get more snow, which favours stability, but they are more susceptible to wind and are colder, encouraging
faceted crystal. Lower slopes are warmer and stabilize faster after storms, but the shallower snow favours faceted crystal. Lower elevations are prone to rain-on-snow events that overload and weaken the snow pack or create icy problem layers.
Avalanches are usually more likely at higher altitude because there is more snow and wind than at lower altitude, and there are fewer trees, bushes and logs to anchor the snowpack.
Map No1 Altitude Map
3.1.2 Slope Map
A terrain slope is inclined or steepness of a surface and is calculated by difference between the elevation of two neighbouring cells and then, using this distance, which is calculated on the basis of cell resolution, calculating the tangent between them.
tanQ=x2-x1/y2-y1
Map No2. Slope Analysis
In the study area it was observed that the most of the avalanche activities were associated with slope angles Map No 2 slope classification. Through years of field research, a range of slopes extending between 25 -45 degrees has been identified as the most potential for avalanche formation. Below 30 degrees, generally the moderate angles do not promote downhill movement of cohesive snow packs. Above 45 degrees, the terrain is typically too steep to collect any significant snowfall accumulation.
3.1.3 Aspect Map
Aspect is defined as the direction in which surface faces with respect to sun.
Differing aspects receive different intensities and incident angles of sunlight. These variations result in inconsistent energy exchange between the pack and the atmosphere, as well as within the snow pack itself. Differing aspects also create different snow cover environments leading to stress discontinuity in the snowpack, which greatly influences the formation of avalanches. Map No 3 shows aspect thematic layer which displays 6 classes, based on the avalanche occurrence researches. North & North East facing slopes are considered more favourable for the avalanche formation. Aspect map generated in ArcGIS has been reclassified into five categories, values used for the reclassification are mentioned in table no. 4.
Aspect (in Degrees) Orientation
0‐45 N‐NE
45‐135 NE‐E‐SE
135‐225 SE‐S‐SW
225‐315 SW‐W‐NW
315‐370 NW‐N
Table No. 4 Reclassification of Aspect
Map No 3 Aspect Analysis
3.1.4 Curvature map
Avalanches tend to start on recognizable geomorphological features, such as convex slopes, planar slopes or along ridges with cornices. They also initiate at changes in slope profile or at changes in groundcover. Terrain shape influences the shape of accumulated snow and where snow pack weaknesses will occur. For example, on a convex slope, the weight of snow on the steeper part of the curve sets up tensile stresses in the snow pack in the rounded upslope part of the curve.
Map No 4 Curvature Analysis
To recogonise the concave, convex and plan areas, above map has been reclassified into 3 classes on the basis of following criteria applied during the process of reclassification:-
+ ve values indicate the convex terrain shape.
-ve values indicate the concave terrain shape.
0 value indicates the planar terrain shape.
Map No 5 shows the reclassified map displaying negative and positive values that portray concave and convex surfaces respectively. Areas having 0 pixel values are known as plan surface.
Table No 5 Reclassification of Curvature
Map No 5. Classification of curvature Pixel Values Surface -151.5555573 - -0.01 Concave
-0.01-0 Plane 0-467 Convex
3.2 Groundcover Map
Image classification is the most important part of digital image analysis. It is very nice to have a "pretty picture" or an image, showing a magnitude of colours illustrating various features of the underlying terrain, but it is quite useless unless it is known what the colours mean. (PCI, 1997). So classification of image is required, aim of image classification is to classify and represent features (Lu, Xiong., (???)) as a unique color, the feature existing in an image in terms of land cover type, those feature actually existing on ground. There is a variety of image classification methods which can be used but there are two main image classification methods (Supervised and Unsupervised image classification method).
Unsupervised classification has been applied to classify AWiFS imagery. This identifies the desired categories of natural groups within multi-spectral data (Campbell, James., 10 December 2001). Hence there are very less chances of error during the execution of this method. To accomplish current study AWiFS imagery has been classified into three categories that is snow, barren land and vegetation. Before the final classification I distinction was made between the snow and non snow area (Roshani et al 2008), vegetation and non vegetation areas with the help of NDSI and NDVI models.
3.2.1 Preparation of NDSI Model
In the Himalayan region, during winter time automated technique for classification is difficult to apply due to mountain shadows and cloud cover. This problem can be partially solved, if systematic procedure based upon Normalized Difference Snow Index (NDSI) is properly developed. The utility of NDSI for snow cover mapping is based upon snow reflectance characteristics. Snow reflectance is high in visible region and low in SWIR region. Advantage of NDSI is delineation and mapping of snow in mountain shadows (Hall et al 1995, Negi, H. S. 2009). Field and satellite observations suggest that NDSI values in shadow and non-shadow region are same (Kulkarni et al 2002b). This is possibly due to reflectance from diffuse radiation in shadow areas. NDSI is defined by following relation and it ranges from −1to+1.
NDSI = (Green− SWIR) / (Green + SWIR)
Figure No 7. NDSI Model
Where Green (band2) and SWIR (band5) are reflectance bands (Kulkarni 2007)
Before measuring the final ground cover, NDSI and NDVI models are generated for the accuracy of ground cover. Using NDSI model a thematic map showing the snow in the study area has been generated. Using this NDSI image a mask showing snowy and non snowy areas has been generated. The analysis has been performed in Erdas Imagine (Remote Sensing Software) software. From NDSI image it is observed that during the month of December 2008 the snow was generally present at high altitudes only while there was no snow in the lower valley at that time.
Figure No 8 shows the snow and non-snow areas in Manali-Dhundi area.
Figure No 8 NDSI Analysis
A mask has been created for above NDSI image, which shows snow and non snow classes in study area. Figure no 9 shows the NDSI mask.
Figure No 9 NDSI Mask
3.2.2 Preparation of NDVI model
Non-snow image has been extracted from the reflectance image having threshold less than 0.4 in the NDSI image. A threshold value for NDSI of 0.4 is defined for the pixels that are approximately 50% or greater covered by snow from the imageries of different sensors (Xiao et al 2001). NDVI has been further used to check the non snow area derived from NDSI image.
NDSI = (IR− R) / (IR + IR)
Figure No.10 NDVI Model
Where IR is reflectance near infrared band while Red is visible band (Almuhairi, Maitha.
Alyan., 2008).
NDVI image (Figure No 11) displays the distribution of vegetation across Manali - Dhundi area. Than vegetation class has been extracted from NDVI image, resulting a mask showing vegetation and non vegetation in study area.
Figure No 11 NDVI Analysis
Like NDSI a mask showing vegetation and non vegetation has been created from NDVI image. Figure No 12 shows NDVI Mask.
Figure No 12 NDVI Mask
The NDSI and NDVI were performed with the help of Erdas Imagery using model maker. The ground cover is classified into three categories, Snow, Barren Land and Vegetation. Map No 6 shows the ground cover map of Manali - Dhundi area derived using AWiFS imagery. Satellite images are normally digital images. In order to extract useful information from these images, image processing techniques are applied to enhance the image to help visual interpretation, and to correct or restore the image if the image has been subjected to geometric distortion, blurring and degradation by other factors. There are many image classification techniques available and the method used depends upon the requirements of specific problem concerned. To classify ground cover in current study unsupervised classification technique has been applied. This method of classification generates the clusters and then the clusters are classified into desired categories. To apply this technique of image classification there is no need of prior knowledge of study area. Accuracy of ground cover image can be cross checked with NDSI and NDVI images.
Avalanches are greatly affected by groundcover type because it affects the stability of the overlying snow pack. The rougher the ground surface, the more snow depth is required before an avalanche will take place. When the snowpack is thick, avalanches are not influenced by surface roughness. The most defining element for avalanche initiation is whether a slope is forested or not, because forest cover reduces the effect of wind exposure to snow pack distribution. Areas above the tree line, rocky outcrops and cut blocks are more exposed to the effect of wind on snow distribution. Forested areas offer considerably more shelter from the wind and therefore do not tend to accumulate dangerous snow packs. In reality, this is not always the case, for example a forested gully below a ridge or convex slope could pose a serious risk to travelers.
Map No 6 Classification of Ground Cover
3.3 Multi-Criteria Analysis
Multi- criteria evaluation is primarily concerned with how to combine the information from several criteria to form a single index of evaluation. This method involves
either qualitative or quantitative weighting, scoring or ranking of criteria in terms of their importance to achieve one or more objectives. Such an informal weighting and scoring method is the most common modeling technique used by the decision makers (Strobl, Josef 2007, Video on you tube by Forestclimsoftdev 18 July 2011). The important parameters required to perform MCE to achieve the objectives are Slope, Aspect, Ground Configuration (Curvature) Ground Cover truth and Elevation. So a ranking method was used to rank the all generated thematic layers (terrain and ground cover parameters) on the basis of their importance in avalanche initiation. Thus, numeric estimation of the risk, i.e. the adopted values by each parameter involved was carried out by using ArcGIS (9.1) software.
3.3.1 Assigning Weightages to each Terrain parameters
Assigning Weight for Slope. Slope is the incline or steepness of a surface. It is normally described by the ratio of the "rise" divided by the "run" between two points on a line. Among all the terrain variables, the slope factor has higher influence on the avalanche initiation. This is easy to understand since an avalanche is only a snow mass moving downhill so that the gravity and the friction will be the forces that will decide if the avalanche will occur or not. While the previous research has proven that the critical slope for the avalanche triggering range from 25 degree to 45 degrees, so these slope categories have been assigned maximum value. Lower slopes contribute to reach the needed stability to cancel the chance of an avalanche event, while higher slopes allow continuous
“purges” i.e. small slides of snow that prevent the formation of the big amounts of snow necessaries to trigger a considerable avalanche. The risk factor for the slope range is given in Table No 6. Basically, it is divided into 4 zones and respective avalanche hazard zones.
Table No 6. Slope Weightage scheme
Assigning Weight for Aspect: Aspect is the direction that a surface faces. There are four cardinal directions (N,E,W,S).The intermediate (Inter-cardinal) directions are north-east (NE), south-east (SE), south-west (SW), and north-west (NW).
Below is the diagram showing these direction:-
Figure No13. Cardinal directions
The derived aspect from DEM displays six directions. Aspect is important parameter due to two different factors interacting the prevailing wind direction and the Sun’s path along the day. There are other terrain parameters those contribute to whether or not an avalanche triggers. Aspect is helpful to decide which part of mountain areis safer for travelling. The orientation of a slope to the prevailing Sun is an important factor in the triggering of avalanches in study areas. As it is well known, in the North Hemisphere the Sun travels the sky following a path almost always at the south of the observer. Because of this fact,
Slope angle in degrees Weightage
0-10 1 10-25 3 25-45 4 45-55 3 55-90 2
the north face keeps the snow cover colder and compacter than the other faces.
Hence there are maximum chances of avalanche initiation in North and North East facing slopes (Tremper, Bruce., October 2001). Moreover, it must be taken into account that, in the special case of wet snow avalanches usually in the spring; it is in the south faces where the avalanche hazard zones are higher due to the snow fusion and the decrease in friction caused by the pressure rise inside the pores of the snow mantle. The follwing table No.7 shows the assigining of weightages to the different direction angles with respect to sun's path along the day.
Categories Weightage
Flat 1
N-NE 4
NE-E-SE 3
SE-S-SW 1
SW-N-NW 2
NW-N 3
Table No 7 Aspect weightage scheme
Assigning Weight for Terrain Shape factor (Curvature), In the present study the plan curvature has been chosen, based on the understanding that defines clearly the existence of tensile stress forces in the hillside direction (in which the main involved force acts by gravity). In current study, curvature is defined as the difference of slope angle. Positive values are defined as convex surfaces because slope angle decreases and the difference between initial - final increases and results in the positive values. Convexity causes unstable condition in snow cover, due to tensile stress. Hence chances of avalanche initiation are very high, so this category is assigned maximum weightage value. Values equal to 0 shows the plane area and in the case of concave surface slope angle increases, the difference in slope decreases, so these areas are assigned
negative values. Concave surface favour the snow stabilization due to compression. So low weightage be given to this type of surface. The weightages in the map according to the plan curvature are shown in table no. 8.
Curvature Weightage -151.555573 - -0.01 3
-0.01-0 1 0-467 4
Table No 8 Curvature Weightage scheme
Assigning Weight for Ground Cover. Past researches in the relevant field state that the 90% of avalanches affect the terrain with a green cover having average height lower than 2500 m. Chance of an avalanche triggering on forested areas is almost null, so these areas have been ignored in this study by assigning to these areas a lowest weight. The snowy surfaces have given maximum weightage because as there is a fresh snow fall which creates instability in snow pack due to weak bonding in snow layers. To estimate the green cover, an AWiFS image has been used. The ground cover is classified into three classes,
Categories Weightage
Snow 4
Vegetation 1
Barren Land 3
Table No 9 Weightage scheme for ground cover
Assigning Weight for Altitude. The altitude map was reclassified according to potential release area factor. Through back analysis, it has been found out that risk factor enhances with increases in altitude around 4500 meters in study area.
Because higher altitude gets more snow and the environmental conditions changes
as altitude changes. The ranking for different altitude is given in the following table no10.
Altitude (meters) Weightage 1918-2500 2 2500-3500 3 3500-4500 4 4500-4760 3
Table No10 Weightages given to various altitudes
3.3.2 Assigning Combined Weightage to the Avalanche Causing Parameters: Till now we have assigned weightages to the individual categories of causative parameters, now we will assign weightages % to each causative parameters according to its importance in avalanche initiation.
Terrain Parameters Influence in Percentage Reclassified Groundcover 12
Reclassified Altitude 35 Reclassified Curvature 12 Reclassified Aspect 11 Reclassified Slope 30
Table No11 Combined weightage to avalanche causing factors
3.4 AHP Method
AHP stands for Analytic Hierarchy Process which was introduced by SAATY (1977) and is a very popular means to calculate the needed weighting factors with the help of preference Matrix where all the identified relevant criteria are compared against each other with reproducible preference factors. All the reclassified thematic layers are put into AHP tool and are compared according to the scale for comparisons (SATTY and VARGAS, 1991). The comparison is made on the basis of following table no12.
Table No12 Comparison scale according to SATTY and VARGAS, 1991
3.4.1 AHP Weightage Scheme
Terrain and ground cover parameters are compared to each other by using AHP tool. In this method, one parameter is compared with other individually on the basis of their importance in avalanche initiation and values are assigned with respect to comparison table and results in % weightage that a individual parameter has in avalanche occurrence.
Terrain Factors Slope Elevation Aspect Curvature Ground Cover
Slope 1 0.3333 5 5 3
Altitude 3 1 5 5 3
Aspect 0.2 0.2 1 1 1
Curvature 0.2 0.2 1 1 1
Ground Cover 0.3333 0.3333 1 1 1 Table No 13 Weightage scheme applied during AHP process
After performing above required steps for AHP analysis, the weightage % for every individual parameter is to be calculated and a single index evaluation map showing the avalanche risk zones is generated. Shown in table No 14.
3.4.2 Calculated Weights for all the terrain factors from above process Intensity Of
Importance
Description
1 Equal importance
3 Moderate importance of one factor over other
5 Strong or essential importance
7 Very strong importance
9 Extreme importance
2,4,6,8 Intermediate values
Reciprocals Values of inverse comparison
Terrain Factor Calculated Weights Approximate Weightage %age
Slope 0.2894 28
Altitude 0.46 46
Aspect 0.0771 8
Curvature 0.0771 8
Ground Cover 0.096 10
Table No 14 Shows the calculated weights for each terrain factor
3.4.3 Risk Zones generated from AHP. After comparing all parameters to each other on the basis of their importance in avalanche initiation, the total weightege % is calculated given in table No 13 and a raster is also generated which is shown in Map No 8.
3.5 Results
The demarcation of avalanche hazard zones from MCE and AHP method according to the risk level are mapped. The result maps displays the study area into 4 avalanche hazard risk zones, No Risk Areas, Low Risk Areas, Moderate Risk Areas and High Risk Areas.
3.5.1 Classification of Avalanche prone areas resulted by MCE method Shown in Map No 7
The objective of the study has been achieved successfully. An avalanche hazard zonation scheme has been developed by combining terrain and ground cover parameters in GIS environment. This section provides the zonation scheme developed using MCE method. This method involved weighting and ratings of all caustic terrain and ground cover parameters. Then these weighted thematic data layers are combined in ArcGIS (using weighted overlay tool) to provide a single index evaluation information. The risk zones generated from this method are:-
No risk Areas. These areas for avalanche risk zone are at lower altitude in valley, except some area at higher altitude near ridges/spurs, where the slopes are less than 10 degrees.(i.e. the flattest areas).
Low Risk Areas. The areas are showing variation in our study area. These are generally at low slopes at low altitude in lower valley. Some areas are at altitude between 2500-3000, these are concave surface and slopes are south facing.
Moderate Risk Areas. These areas lie at all aspects on the slope angle which is having variation in between 10-25 degrees. The surfaces are mostly convex in nature at an altitude of 2500-4500 meters.
High Risk Areas. These areas are at the altitude between 4500-4700 m at slope angle between 25-45 degrees on convex surfaces.
