You are on page 1of 112

Journal of

South Asian Disaster Studies


Journal of the SAARC Disaster Management Centre

Volume 3 l Number 1 l June 2010

SAARC Disaster Management Centre, New Delhi.


All rights reserved. No part of this publication may be reproduced or transmitted in any form or by
any means, electronic or mechanical, including photocopying, recording or by any information storage and retrieval system without permission from SAARC Disaster Management Centre, New Delhi.
ISSN: 0974 6463
Journal of South Asian Disaster Studies Journal is published two times a year and is distributed by
KW Publishers Pvt Ltd.
4676/21,First Floor, Ansari Road, Daryaganj, New Delhi 110002
email: knowledgeworld@vsnl.net / online.kw@gmail.com
www.kwpub.com
Typeset by Black Innovation, New Delhi
Printed by Glorious Printers, A-13, Jhil Mil, Delhi 110095 and Published at SAARC Disaster Management Centre, New Delhi, IIPA Campus, I.P. Easte, Mahatma Gandhi Marg, New Delhi 110002.
Editor: O.P. Mishra and Mriganka Ghatak

ii u Journal of South Asia Disaster Studies

Contents
Editors Note
Notes on Contributors

v
vii

1.


GIS-based Assessment Of Coastal Vulnerability to Severe Cyclones


Along Andhra Pradesh Coast, India
B.S. Prakasa Rao, N. Bhaskar Rao, N. Srinivas, G. Srinivas,
K. Mrutyunjaya Reddy and M. Satyakumar

2.

Glacier Lake Outburst Flood (GLOF) and its Estimation at


Nepals Higher Himalaya
Binod Shakya

21

3.


Usefulness of some Disaster Anticipatory Measures of Indian Railways


in Disaster Management Plan of Mega cities of India with Special
Reference to Ahmedabad city.
J.G. Macwan

31

4.



Algorithm for Information Retrieval of Earthquake Occurrence from


47
Foreshock Analysis using Random Forest implementation in
Earthquake Database Creation and Analysis: A Machine Learning Approach
Pushan Kumar Dutta, Mrinal Kanti Naskar, Om Prakash
Mishra, Kajal Mukherjee

5.

A Concept of a Low-cost Construction Method for Deep Underground Dams 59


Yoshito Kobayashi and Vijay Kumar Jatiya

6.

Assessment of Drought Vulnerability in North Western


Region of Bangladesh
Shams Al-Amin, Umme Salma Rima and Md. Monsurul Huda

81

Vol 3 No 1 June 2010 u

iii

BINOD SHAKYA
7.

Quantifying Disaster Risk and Vulnerability at State


Level: A Case of Gujarat, India
Shital Hardik Shukla

iv u Journal of South Asia Disaster Studies

91

Editors Note
This issue of the Journal of South Asia Disaster Studies is a reflection of the wide spectrum of disaster
management research and practices going on in the South Asian region. The contributors of the research
documents belong to various levels of disaster management initiatives and national organisations, research and academic institutions and research students in the field of disaster management from and
outside SAARC member states.
In the past decades, Geographical Information System (GIS) has emerged as a vital tool for collecting,
integrating, and finally assessing the datasets related to disasters. The first paper of this volume by Rao
et al. is based on application of GIS for impact assessment of Severe Cyclonic Storms (SCS) on the coastal
regions of the southern state of Andhra Pradesh in India. It uses various parameters like population,
resources, distance from the coast, landfall location of cyclones, cyclone track, slope of the ground and
storm surges and the study analysed 100 years data from Andhra Pradesh.
The glacial lakes nestled in the snow clad Himalayas are growing in size due to melting of glaciers as
a direct consequence of global warming. The communities residing near these lakes are exposed to severe
danger due to silent specter of Glacial Lake Outburst Floods (GLOF). The paper of Shakya proposes a statistical model based on lake parameters taking Lower Barun, Imja, Thulagi, and Tsho Rolpa glacial lakes
for which, estimates of GLOF risk in the downstream parts are made by the author.
The Indian Railways has one of the biggest global railway networks, and in spite of having a comprehensive disaster anticipatory mechanism, faces major accidents. Macwans paper addresses this vital issue,
and provides an overview of the strength and shortcomings of Indian Railways in handling disasters,
preparedness at managerial level, and Standard Operating Procedure (SOP) with reference to Ahmedabad
city in Gujarat. It also discusses the measures undertaken by the Ahmedabad Municipal Corporation in
addressing disaster related issues, constraints, etc. In his paper, the author advocates the need of including the related points of Disaster Management Act 2005 of India in the Disaster Management Plan of
Ahmedabad city for effective disaster mitigation practices.
Large number of foreshocks prior to the main shock of an earthquake makes the analysis of earthquake catalogues a difficult task. Dutta et al. attempt a machine learning approach to extract attribute
value pairs from the earthquake catalogue, which enables study of earthquake occurrences on proper
compilation of earthquake data before and after the mainstream earthquake. The results outlined in the
research paper has estimated accuracy of the refined catalogue of 98. 81 per cent for correctly classified
foreshocks.

Vol 3 No 1 June 2010 u

There has been high degree of interest in construction of underground dams for sustainable surface
water management as these have certain advantages over the surface dams. However, there have been
some concerns regarding requirements of high-end technologies, involvement of heavy machinery and
steep construction costs in their construction. The paper of Kobayashi and Jatiya proposes an answer to
these concerns through Transverse Excavating and Infilling Method (TEIM) as a low cost and widely applicable construction method that can be applied for constructing deep underground dams.
In many parts of South Asia, the increasing atmospheric dryness conditions are verging on a state
of irreversible condition leading to potential desertification and degradation of already strained land resources. Al-Amin et al. in their paper have proposed a GIS based dryness index map for the northwestern
part of Bangladesh, commonly known as the Barind tract. The paper uses Aridity Index (AI) as a numerical
indicator for dryness, and the maps are very useful for analyzing the trend of precipitation in the selected
study area.
One of the emerging prerequisites for mainstreaming Disaster Risk Reduction (DRR) into development process is quantification of vulnerabilities of a region to natural disasters. The paper of this issue by
Shukla attempts to quantify vulnerability indicators up to district level for the western state of Gujarat. It
also discusses a regional model for assessing the pattern of vulnerability within the state, and identifying
the drivers of vulnerability in the state for effective identification of the vulnerable district and prioritisation of interventions in planning process.
This issue of the Journal of South Asia Disaster Studies exemplifies the emerging trend of highly informative research, and policy analysis practices that is a reflection of the involvement of policy planners,
academicians, researchers, scientists of the region in the larger interest of making the South Asian region
disaster resilient, and shaping up a region with high degree of DRR standards. The volume is placed before
the disaster management stakeholders at various levels in the region for inviting critical appraisal of the
articles featured herein, which will undoubtedly contribute to the improvement of the journal.

O.P.Mishra
Mriganka Ghatak

vi u Journal of South Asia Disaster Studies

Notes on Contributors
l Binod Shakya, Central Department of Hydrology and Meteorology, Tribhuvan University, Email:
drbinod@lifelinegroup.info
l B.S. Prakasa Rao, Department of Geo-Engineering, College of Engineering, Andhra University, Visakhapatnam-530003. Email: bosukonda@gmail.com
l G. Srinivas, National Remote Sensing Center, Balanagar, Hyderabad-500625. Email: srinivasrao_g@nrsc.gov.in
l J.G. Macwan, Doctoral student (Disaster Management),

Singhania University, Rajasthan, India.

Email: jmacwan45@yahoo.com
l Kajal Mukherjee, Geophysicist,Geological Survey of India, Kolkata, India.
l K. Mrutyunjaya Reddy, Andhra Pradesh State Remote Sensing Applications Centre, Hyderabad-500038.
Email: kmruthyu@yahoo.com
l Mrinal Kanti Naskar, Electronics and Communication Dept., Jadavpur University, Kolkata,India.
l Md. Monsurul Huda, Junior Engineer, Institute of Water Modeling, Dhaka, Bangladesh. Email: hudamonsur@gmail.com
l M. Satyakumar, Indian Meteorological Department, Hyderabad - 500016. Email: msatyak@yahoo.com
l N. Bhaskar Rao, Department of Geo-Engineering, College of Engineering, Andhra University, Visakhapatnam-530003. Email: bhaskar.infotech@gmail.com
l N. Srinivas, Department of Geo-Engineering, College of Engineering, Andhra University, Visakhapatnam-530003. Email: srinu_neeli@yahoo.co.in
l Om Prakash Mishra, Geological Disaster Division, SAARC Disaster Management Centre (SDMC),
New Delhi, India

Vol 3 No 1 June 2010 u

vii

l Pushan Kumar Dutta, Advanced Digital Embedded System Lab, Jadavpur University, Kolkata, India
l Shital Hardik Shukla, Assistant Professor, Sardar Patel Institute of Economic and Social Research,
Opp. Udagam School, Drive In Road, ThaltejTekra, Ahmedabad-380054, Gujarat, India.
l Shams Al-Amin, Lecturer, Department of Civil Engineering, Ahsanullah University of Science and
Technology, Dhaka, Bangladesh. Email: iamtanim@gmail.com
l Umme Salma Rima, Junior Engineer, Institute of Water Modeling, Dhaka, Bangladesh. Email: rimace045_03@yahoo. com
l Yoshito Kobayashi, Searcher, Motive Machinery Group, Center of Search Affairs,Industrial Property
Cooperation Center Fukagawa-gatheria-west 3, Kiba 1-2-15, Koto-ku, Tokyo 135-0042, Japan. Email:
kobayashi-yoshito@kamakuranet.ne.jp

viii u Journal of South Asia Disaster Studies

GIS-Based Assessment of Coastal


Vulnerability to Severe Cyclones
along Andhra Pradesh Coast, India
B.S. Prakasa Rao, N. Bhaskar Rao, N. Srinivas, G. Srinivas,
K. Mrutyunjaya Reddy, and M. Satyakumar

Abstract
Assessment and mitigation of flood hazard is an important component of sustainable development
in coastal areas which are rich in various resources. About 25 per cent of Indias population lives
within 50 km of the coastline. The east coast of India is more prone to cyclones arising in Bay of
Bengal as about 80 per cent of total cyclones, generated in the east coast of India, strike here. Being
one of the coastal states, Andhra Pradesh (AP) has a long coastline stretching about 1,030km out
of 7,500 km of total coastline of India. The severe cyclonic storms during 1831-1970, 20 of them
were landfall on the AP coast and accounted for 12,000 deaths. The damages and economic
losses due to cyclones and floods are relatively higher in AP, which was estimated at Rs.18,469
core in various spheres. Despite the frequent occurrence of these cyclones, adequate measures to be
taken to mitigate their impact need adequate study. Remote Sensing (RS) technology has proved
its capability in providing vital information in a disaster situation; it could be used in disaster
analysis, hazard zonation, and prior risk assessment. The study adopts RS andGIS (Geographic
Information System) to assess the impact of severe cyclonic storms that landfall on the coastal
regions of AP. Hence, an attempt is made to study each taluk along the coast and100 km inland
(86 taluks) for assessment of vulnerability keeping in view the population, resources, distance
from the coast, location of land falling of cyclone and its track, ground slope, and storm surge.
The landfall of severe cyclones over 100 years of data is considered in the study. Vulnerability of
the coastal taluks is arrived by assigning weight factors to each taluk based on various thematic
informations depending upon their capability in the mitigation of storm and economic value. 44
per cent of the study area is categorised as high and very high level of vulnerability, and this area

Vol 3 No 1 June 2010 u

B.S. PRAKASA RAO , N. BHASKAR RAO , N. SRINIVAS , G. SRINIVAS,


K. MRUTYUNJAYA REDDY AND M. SATYAKUMAR
needs special attention during cyclone period to mitigate the human and related loss.
Key words: Coastal Vulnerability, Severe Cyclone, Estuarine region, Landuse/Landcover,
Weight factor.

Introduction
Tropical cyclones are very destructive natural hazards with an average wind velocity of 17
m/s. The east coast of India is more prone to cyclones arising in Bay of Bengal as about
80 per cent of cyclones generated in the sea could landfall on the east coast. Vulnerability
to disaster can be defined as a gap in community coping capacity, or inherent protection,
against the disaster characteristics. Study of historical data on floods and cyclones and
its integration in GIS will throw light on the vulnerability of coastal area. The number of
peer-reviewed papers in the field of science, impacts and vulnerability along the Indian
coasts has been low. Considering the magnitude of these problems and also the longterm interests, research on impacts and vulnerability to the coasts and the necessary
adaptation measures is needed (Unnikrishnan 2011:1273). There are numerous reasons
why we must live and work in regions subject to cyclones. Hence, proper measures should
be adopted to reduce the impact of cyclones. Vulnerability assessment is essentially a
systematic integration of factorstopography, slope, morphological features, engineering
characteristics of structures, and socioeconomic aspects and infrastructure facilities.
Universities need to take up local issues on impacts and adaptation. GIS is capable of
assembling, storing, manipulating, and displaying geographically referenced information
(Burrough 1986).
Coastal vulnerability is studied by Jayanti (1998:115) as a function of cyclone
frequency, coastal topography, bathometry, and population density.Vulnerability was
measured and a disaster risk index for the Indian coast was derived at. The Multi Purpose
Cyclone Shelter Project (MCSP) in Bangladesh (MCSP 1992) has delineated risk zones
by modeling the storm surge with GIS. Damen and Van Western (2003) have studied
vulnerability and risk analysis for Bangladesh using population density and storm
height. Prakasa Rao (2005: 3611)developed a methodology for the estimation of cycloneinduced flood through remote sensing and GIS for two delta regions in AP.The capability
of applying GIS in various aspects of risk assessment has been demonstrated by many
researchers. Specifically, Miller and Onwuteaka (1999: 460) evaluated the vulnerability
of the landscape to oil spills in the East. Granger et al. ( 1999) equally utilised GIS to
synthesise and model the spatial relationship between vulnerability and hazard in

2 u Journal of South Asia Disaster Studies

GIS-BASED ASSESSMENT OF COASTAL VULNERABILITY TO SEVERE


CYCLONES ALONG ANDHRA PRADESH COAST, INDIA
order to study the risk faced by Cairns in Australia to multi-hazard phenomenon.
Thumerer et al. (2000: 265) developed a GIS-based risk assessment model by combining
oceanographic and climatic data with data on sea defenses, elevation values and patterns
of land use to assess the implication of sea level rise along the English east coast using
the Arc-Info GIS package. These include (Jorinet al. 2001: 153) for land suitability
assessment, and (Pramojanee et al. 1997: 17) for flood vulnerability mapping. Krishna
(2005:1339) discussed science for coastal hazard preparedness to tsunamis, seismicity,
storm surges and coastal pollution.Van Westen (2008) used simple data sets from
Colombia (South America) to demonstrate on a national scale the meaning of
hazard, vulnerability and risk. Similar procedures were used by Damen and Van
Westen(2008b) to model cyclone hazard zonation in the South of Chittagong,
Bangladesh;Van Westen and Tertien (2008) to demonstrate the potentials of GIS in
hazard zonation of landslides triggered by earthquakes in Manizales, Colombia, and
Van Westen (2008) to demonstrate the use of quantitatively defined weight values
in the making of hazard maps. The overall aim of Udoh and Ekanem (2011: 205) was
to examine the ways GIS can be used to effectively appraise the degree of threat
posed by oil spill Coastal areas of AkwaIbom State (1989), Nigeria.The present paper
describes in detail the methodology for developing vulnerability degreeby making use of
GIS to map, analyse, assess, and present these indices in the form of national vulnerability
profiles taking taluk as a spatial unit of analysis.
An important component of the coastal population is the seafaring community, of
which fishermen form a sizable part although cyclonic damages are severe in the delta
regions of the Rivers, Vamsadara, Godavari and Krishna, mitigation measures against
floods are said to be meager. Hence, an attempt is made to assess the vulnerability of the
coastal area of AP.The total number of villages perched along the coastline is 2,482500
are within 5 km swathe extending inside from the coast, 601 are within 5-10 km and
1,381 villages are between 10-20 km. Coastal AP covers an area of 98,926 sqkm with a
total population of 3,15,70,722 as per 2001census. The coastal area of the state, with a
significant component of its economy, is related to the sea in some way. This includes
fishing, shipping, ports and harbours, and tourism and allied industries in the coastal
region. The last few years have also seen new investment being made in the Indian
coastal waters (on the continental shelf and slope) for oil and gas exploration. These could
eventually become a critical component of the economy. With these new developments
also come new threats, while these offshore structures are vulnerable to severe cyclones,
storm surges, and tsunamis.

Vol 3 No 1 June 2010 u

B.S. PRAKASA RAO , N. BHASKAR RAO , N. SRINIVAS , G. SRINIVAS,


K. MRUTYUNJAYA REDDY AND M. SATYAKUMAR

Historic background of cyclonic eruptions in the state


From 1977-2008, there were 29 severe cyclones landfall on AP coast. Table 1 represents
a brief history of severe cyclones that have occurred from 1977 to 2008 and the losses
occurred in various spheres (damage statistics before 1977 are not available):
Table 1: Brief History of Severe Cyclones 1977 to 2008
No of severe cyclones/Cyclonic storms

18

Population Affected (in Lakh)

218.30

Human Loss

12,000

Livestock loss (in lakh)

6.63

Houses damaged (in lakh)

35.71

Damaged Cropped area (in lakh hectare)

59.65

Estimated loss (in Crore)

18,468.57

Factors that have combined to compound natural disaster vulnerability along the AP coast are:
Over 50 per cent of cyclones in the region turn into severe storms, which are often
accompanied by storm surges
The low-lying lands that prevail along Godavari, Krishna, and Pennar deltas and
Pulicat lake are subject to widespread flooding and deep inland sea water incursion
that even a few metres high storm surge can create
High concentration of population, infrastructure, and harbour-related economic
activities along the shore line
Experiences with the cyclones in AP stressed the need for modernisation and
strengthening of existing infrastructure. Therefore, it is proposed to construct new
cyclone shelters as well as missing road links/bridges under Component B of the National
Cyclone Risk Mitigation project (NCRMP, 2009: 4).
The following severe cyclonic storms that landfalls AP coast are considered for the
study (Table 2). These tracks are digitized and created shape files and these are overlaid
on the study area having taluk boundaries to identify the vulnerability of each taluk with
reference to the track.

4 u Journal of South Asia Disaster Studies

GIS-BASED ASSESSMENT OF COASTAL VULNERABILITY TO SEVERE


CYCLONES ALONG ANDHRA PRADESH COAST, INDIA
Table2: Severe Cyclonic Storms 1906 to 2003 during May, October and November
S.No

Year

Month

Period

Nearest location of landfall

1906

October

7th to 9th

Bheemunipatnam

1907

October

1st to 2nd

Chillakur

1914

May

13th to 16th

Ichchapuram

1925

May

13th to 16th

Koduru

1932

November

22nd to 27th

Sullurpeta

1933

November

14th to 18th

Kothapatnam

1938

November

21st to 28th

Nagayalanka

1940

May

17th to 8th

Indukurpet

1966

November

18th to 22nd

Santhabommali

10

1966

November

1st to 8th

Vakada

11

1968

November

2nd to 5th

Chillakur

12

1969

November

6th to 8th

Allavaram

13

1972

November

22nd to 23rd

Sullurpeta

14

1976

November

4th to 5th

Machilipatnam

15

1976

November

16th to 17th

Kavali

16

1977

November

15th to 20th

Chirala

17

1977

October

29th to 31st

Kavali

18

1979

May

8th to 13th

Vidavalur

19

1979

November

21st to 25th

Tada

20

1979

November

23rd to 29th

Nizampatnam

21

1982

October

17th to 21st

Tada

22

1982

October

15th to 16th

Katrenikonda

23

1983

October

2nd to 6th

Bheeminipatnam

24

1984

November

12th to 13th

Vakadu

25

1985

October

10th to 11th

Rambilli

26

1989

November

6th to 8th

Kavali

27

1990

May

6th to 11th

Nizampatnam

28

1998

November

14th to 16th

Visakhapatnam

29

2003

December

14th to 16th

Kruthivennu

Vol 3 No 1 June 2010 u

B.S. PRAKASA RAO , N. BHASKAR RAO , N. SRINIVAS , G. SRINIVAS,


K. MRUTYUNJAYA REDDY AND M. SATYAKUMAR

Study area
The study area covers 86 taluks of AP spread along the coast and about 100-140 km inland.Geographically, it is located between 780 46 to 840 45 East longitude and 130 01
to 190 16 North latitude (Figure 1).The study area is extracted from the digital database
oftaluk map of India.
The coastal districts are primarily agricultural-based. Guntur, Krishna, East Godavari,
West Godavari, and Nellore have highly productive network of irrigation canals of
the Krishna and Godavari rivers. These four districts, along with Srikakulam, also face
the brunt of floods during the monsoons now and then. The coastline ranging from
Srikakulam in the north and Nellore in the south is affected by at least one severe/very
severe cyclone almost every year.

Data and Methodology


The cyclonic tracks of 1906-2003, the huge occurrences in the recent years, were selected
for the study during the months of May, October and November. About 100 years of severe cyclonic storms are considered in the present study. The satellite data is:
Satellite

Path

Row

Date of pass

Resoucesat-(AWiFS)

105

056

January 24, 2009

099

060

March 07, 2009

Resourcesat-AWiFS sensor data are used to prepare the mosaic of the study area after
geo-rectifying the data. This data is selected to show the present status of land use and
land cover and other geomorphologic features, which are important for vulnerability
assessment.
In order to create the vulnerability map, attribute tables are created for the
following input mapssevere cyclonic tracks, population data of all the taluks and
creation of density map, various taluks and their distance from the coast by creating
buffer zones, location of taluks with respective to cyclone tracks,extraction of coast from
image data preparation of slope map from ASTER DEM, and preparation of land use land
cover from AWiFS data. Weight factors are assigned to each taluk and their location with
respective to various above factors. A vulnerability map is prepared by integrated weight
factors of the above themes.

6 u Journal of South Asia Disaster Studies

GIS-BASED ASSESSMENT OF COASTAL VULNERABILITY TO SEVERE


CYCLONES ALONG ANDHRA PRADESH COAST, INDIA

Figure 1: Location of the study area

Vol 3 No 1 June 2010 u

B.S. PRAKASA RAO , N. BHASKAR RAO , N. SRINIVAS , G. SRINIVAS,


K. MRUTYUNJAYA REDDY AND M. SATYAKUMAR

Direction of the cyclone track


The areas in the direction of cyclone track/nearby will have maximum effect due to flood
inundation and strong winds. Hence, this parameter is important in the determination
of vulnerability risk. Considering the direction of the movement of the cyclonic storm,
buffers have been created on taluk map up to a distance of 40 km on the both sides of the
main track, assuming the approximate area of maximum influence of the cyclone would
be up to 40 kmrange on both sides of the track, while the risk would be less with increase
in distance from the track. While assigning the weight factors, 10 is assigned to taluks in
the direction of the track and up to 20 km on either side. Factor 5 is assigned to 20-40 km
distance from the track. The remaining taluks are assigned with factor 1. The landfall of
cyclones is shown over the study area as shown in Figure 2.

Land use/Land cover


In order to study the land use and land cover, the Resourcesat data is taken to build classification map of the study area.The area is covered in two scenes of AWiFS sensor data and
the scene informationis given in the above table. It is 10-bit data with spatial resolution
of 55 m. The scenesare geo-referencedusing reference maps of 1:250,000 toposheetsin
ERDAS software. These scenes are used to prepare mosaic and the study area polygon is
taken to extract the image by preparation of subset (Figure 3). The study area is demarcated with a boundary line in white colour from the total state image of the state.
Land use land cover, one of the important inputs for vulnerability map that has to
be given priority during cyclones, as it reflects the economy of the region. The image is
classified into 7, land use land cover classes using Maximum Likelihood Algorithm in
ERDAS software. The land use classes are agriculture land, fallow land, semi-ever green
forest, deciduous forest, evergreen forest, water bodies, and others. Others class include
settlements, road network, etc.
Classification result shows that about 80 per cent of the study area is occupied with
agricultural land,which is mostly spread along the coast and confined in river deltas and
flood plains. Accordingly, maximum weight is assigned to this class and the other classes
are given low priority. Various land use classes and their areal extent are given in the
Table3.

8 u Journal of South Asia Disaster Studies

GIS-BASED ASSESSMENT OF COASTAL VULNERABILITY TO SEVERE


CYCLONES ALONG ANDHRA PRADESH COAST, INDIA

Figure 2: Landfall of sever cyclonic storms over the study area


Table 3: Land use classes and their areal extent in the study area.
S.No.

Classification

Area occupied in sqkm

% of Total area

Agricultural Land

81,014.00

81.89

Fallow Land

4,371.62

4.42

Semi Green Forest

817.57

0.82

Deciduous Forest

5,546.12

5.60

Ever Green Forest

390.20

0.40

Water Bodies

4,747.49

4.80

Others

2,040.00

2.07

Agriculture land and Wetlands (lakes, etc.) are shown in yellow and blue respectively.
The former is an important component of ecosystem and the latter form the economy of
the region. Generally the coastal areas are highly prone to flood/inundation during sever
cyclones due to heavy rain and storm surge aroused from strong winds. The areas, with

Vol 3 No 1 June 2010 u

B.S. PRAKASA RAO , N. BHASKAR RAO , N. SRINIVAS , G. SRINIVAS,


K. MRUTYUNJAYA REDDY AND M. SATYAKUMAR
intensive agriculture, nearer to the coast and adjacent to the delta have been assigned
weight factor 10 indicating high priority zone where as the areas of barren land, away
from the coast are assigned weight factor 1. Weights are tabulated to all the taluks for the
land use/land cover classes. The satellite image and corresponding classification maps are
shown Figure 3.

Ground slope
DEM of land surface provides significant information for many research activities and
important data as the input of image processing and image analysis, such as image correction due to height of land surface (Ortho rectification), contour mapping, 3D images
generation, disaster management (determination of vulnerable area due to landslide,
flood and tsunami disaster), monitoring land subsidence phenomenon and many others. ASTER (Advance Space borne Thermal Emission and Reflection Radiometer) DEM is
downloaded from the web site. ASTER on board of Terra spacecraft is multi-spectral optical sensor that is launched on December 1999. ASTER sensor has 14 spectral bands that
range from visible to thermal infrared band. There is also stereo coverage in band from 3n
(nadir looking) 3b (backward looking). Therefore, the capabilities of ASTER stereo image
that provides DEM with high spatial resolution (30 m) is very important for remote sensing and GIS (Geographic Information System) users to enhance the accuracy of desired
height information.

(a) AWiFS image of AP state

(b) Land use/land cover


Figure 3

10 u Journal of South Asia Disaster Studies

GIS-BASED ASSESSMENT OF COASTAL VULNERABILITY TO SEVERE


CYCLONES ALONG ANDHRA PRADESH COAST, INDIA
A ground slope map is prepared using the DEM data in Arc GIS. Slope is derived in
degrees and is broadly divided into 4 categories, which ranges from 0 to 5, 5 to 10, 10
to 20, and > 20 for the study area. The range 0 to 5 is considered as plain area, 5 to 10
as moderate slope, 10 to 20 moderately steep, and > 20 is taken as steep slope. The areas
near the coastal plain come under 0 to 5 are given high weight value whereas lower values
are assigned for other categories of slope. The weight factor 10 is assigned to slope 0 to 5
range, factor 5 to moderate slope and 1to moderately steep and steep slopes. The slope
map is shown in (Figure 4a).

Population Density
The population density is high along the coast due to agricultural resources industrial
establishments, which generally depend on water resources of the area. Hence, majority of cities are located along the coast with dense population, ports, and shipyards. The
population density is very important parameter in vulnerability mapping of any area.
Demographic details are obtained from the Department of Statics and Population Studies, Government of India censes 2001. Population density map is prepared taking taluk
as a unit. Collected population data is divided with the area of the taluk and tabulated as
density per square kilometer and spatial distribution of density map is prepared. Population density is shown in 4 classes; namely very high, high, medium and low. Different
colours are assigned to each density class as shown in Figure 4b.Thus, any single taluk
with population in a definite range has single colour. Also more than one taluk may have
a single colour as they may have population with the same range.
In the present analysis, the taluks with more population are located nearer to the coast
or cyclone path have been given higher weights, as there would be heavy human loss if
a cyclone landfall on these areas.The areas with low population density and away from
the track or coastline are given lower weights. All the four categories are assigned by the
factors 2, 6, 8 and 10 according to density ranges in increasing order. Visakhapatnam
and Vijayawada taluks showed high density, and Vizianagaram and Sulurpet showed the
medium level of density followed by remaining taluks.

Distance from the coast


The coastal area and the delta region commonly are very fertile with adequate water resources, and hence are the centers of immense human activity. Therefore, the assessment
of their location with respective to the coast is essential in deriving the vulnerability of
an area for flood hazard.

Vol 3 No 1 June 2010 u

11

B.S. PRAKASA RAO , N. BHASKAR RAO , N. SRINIVAS , G. SRINIVAS,


K. MRUTYUNJAYA REDDY AND M. SATYAKUMAR
Buffer zones have been created for every 20km distance from the coast. Likewise, 4
buffers have been created in Arc GIS and weights are assigned to each taluk. Considering
the distance of the taluk from the coast and main river course, weight factors have been
assigned. The weight factor 10 is assigned to first 20 km and 5 assigned to 20-40km
distance and 1 is assigned to 40-60 and beyond. The nearest/along the coastline, the value
of weight is maximum while the factor decreases with increasing distance from the coast.
There are 31 taluks in the first buffer zone and 20 in the second level of intensity of the
study area (Figure 4c), which shows that 51 taluks are under flood threat.

Storm surge
The storm surge is the sudden rise of seawater due to tropical cyclone and is greatly amplified when the coastal water is shallow, in the estuarine region and the shape of the coast
is like a tunnel. Because of the dense populations along the coast, storm surge associated
the cyclones is the greatest potential threat to life and property. It is, sometimes a large
dense of water from 80 to 150 km wide that sweeps across the coast line where the cyclone makes its landfall caused by the high winds pushing the ocean surface ahead of the
storm. (Krishna Rao, and Ramana Murty, 2003: 2). The level of the surge in a particular
area is determined by the intensity of the storm and the slope of the continental shelf.
A shallow slope off the coast will allow a greater surge to inundate coastal communities.
The phenomenon is so sudden that there is hardly any time for the victim to react
and unless and until prior evacuation is carried out, the loss of life in the effected region
will be nearly complete. The storm surge map of the AP coast is taken from the APCHMP
report and scanned. The scanned map is geo-referenced and digitised the boundary of the
storm surge and overlay on the taluk boundary map. The weight factor 10 is assigned to
taluks covered by the storm surge boundary and 5 is to half or partly covered taluks and
1 to the uncovered. The taluks existing along the coast from Nellore to Kakinada have
shown the effect of storm surge. However, its affect is more in the case of Machilipatnam
and Avanigedda taluks(Figure 4d).

12 u Journal of South Asia Disaster Studies

GIS-BASED ASSESSMENT OF COASTAL VULNERABILITY TO SEVERE


CYCLONES ALONG ANDHRA PRADESH COAST, INDIA

(a)

(b)

(c)
(d)
Figure 4.a. Ground slopeb. Population densityc. Buffer zones from the coast,
d. Storm surge

Vol 3 No 1 June 2010 u

13

B.S. PRAKASA RAO , N. BHASKAR RAO , N. SRINIVAS , G. SRINIVAS,


K. MRUTYUNJAYA REDDY AND M. SATYAKUMAR

Table 3: Weight factors based on location and theme.


DISTRICT

Srikakulam

TALUK

Weight factors
Direction
of the
track

Land
Use/
Land
Cover

Slope

Population
density

Distance
From
the coast

Storm
surge

Result

Sompeta

10

10

High

Pathapatnam

10

Medium

Tekkali

10

10

10

10

V. High

Palakonda

Medium

Srikakulam

10

10

10

10

10

V. High

Narsannapeta

10

10

10

High

Ichchapuram

10

10

10

High

Chipurupalle

10

10

10

10

High

Vizianagaram Parvathipuram

Low

Salur

Low

Bobbili

Medium

Gajapatinaga-

Medium

Vizianagaram

10

High

Srungavarapu-

Medium

ram

kota
Puspatirega

10

10

10

10

V. High

Visakhapat-

Paderu

Low

nam

Anantagiri

Low

Chodavaram

Medium

Chintapalle

Low

Bhimunipat-

10

10

10

10

V. High

10

10

10

10

V. High

Anakapalli

10

10

High

Narsipatnam

10

10

High

Yellamanchili

10

10

10

V. High

nam
Vishakhapatnam

14 u Journal of South Asia Disaster Studies

GIS-BASED ASSESSMENT OF COASTAL VULNERABILITY TO SEVERE


CYCLONES ALONG ANDHRA PRADESH COAST, INDIA
East Godavari

Ellavaram

Low

RampaChoda-

Low

Tuni

10

10

10

10

10

V. High

Prathipadu

10

High

Rajahmundry

10

10

Medium

Peddapuram

10

10

High

Pithapuram

10

10

10

10

10

10

V. High

Kakinada

10

10

10

10

10

10

V. High

Kottapeta

10

10

High

Ramachandrapuram

10

10

10

High

Mummidivaram

10

10

10

10

10

V. High

Amalapuram

10

10

10

10

10

10

V. High

varam

Razole

10

10

10

10

10

V. High

Yanam

Yanam

10

10

10

10

10

V. High

West Godavari

Polavaram

Low

Chintalapudi

Low

Kovvur

Medium

Eluru

Medium

Tadepallegu-

10

High

Tanuku

10

10

High

Bhimavaram

10

10

10

10

V. High

Narsapur

10

10

10

10

10

V. High

Vijayawada

10

10

High

Nuzvid

Low

Kaikalur

10

10

10

High

Gannavaram

10

High

Gudivada

10

10

10

V. High

Machilipat-

10

10

10

10

10

V. High

10

10

10

10

10

V. High

dem

Krishna

nam
Avanigadda

Vol 3 No 1 June 2010 u

15

B.S. PRAKASA RAO , N. BHASKAR RAO , N. SRINIVAS , G. SRINIVAS,


K. MRUTYUNJAYA REDDY AND M. SATYAKUMAR

Guntur

Prakasam

Nellore

Cuddapah

Chittoor

Sattenapalle

10

Medium

Guntur

10

10

10

High

Tenali

10

10

10

V. High

Narasaraopet

Medium

Vinukonda

Medium

Repalle

10

10

10

10

10

V. High

Bapatla

10

10

10

10

10

V. High

Markapur

Low

Addanki

Medium

Chirala

10

10

10

10

10

V. High

Darsi

Medium

Giddalur

Low

Podili

Low

Ongole

10

10

10

10

10

V. High

Kanigiri

Low

Kandukur

10

10

10

10

10

V. High

Udayagiri

10

Medium

Kavali

10

10

10

10

10

V. High

Atmakur

10

High

Kovvur

10

10

10

10

10

V. High

Nellore

10

10

10

10

10

10

V. High

Rapur

10

High

Gudur

10

10

10

10

10

V. High

Venkatagiri

10

High

Sulurpet

10

10

10

10

10

10

V. High

Badvel

10

Medium

Sidhout

Low

Rajampet

Medium

Sri Kalahasti

10

High

Chandragiri

Low

Satyavedu

10

Medium

Puttur

Low

Chittoor

Low

16 u Journal of South Asia Disaster Studies

GIS-BASED ASSESSMENT OF COASTAL VULNERABILITY TO SEVERE


CYCLONES ALONG ANDHRA PRADESH COAST, INDIA

Vulnerability map
As discussed above, the weights are tabulated to all the taluks for all vulnerability parameters as shown in Table 3. To all classes in the above maps, different weight values are
assigned in their attribute table. Finally, all factor maps are analysed in Arc GIS software
to obtain a hazard map which is then classified into four classes: low, moderate, high,
and very high vulnerability hazard. The vulnerability map is shown in Figure 5.
There are 20 taluks, which spread along the coast come under very high category of
vulnerability to cyclone-induced floods. Under high vulnerability category, 9 taluks are
along the coast and 9 taluks are 20 km away from the coast which are spread in Guntur,
Krishna, East Godavari and Northern districts. 44 per cent of the taluks are categorised
as very high and high and the remaining come under medium and low category of
vulnerability. On the levels of extent of damage arrived at form ground truth, for the

Figure5: Vulnerable intensity of the study area

Vol 3 No 1 June 2010 u

17

B.S. PRAKASA RAO , N. BHASKAR RAO , N. SRINIVAS , G. SRINIVAS,


K. MRUTYUNJAYA REDDY AND M. SATYAKUMAR
study area. This ground data are comparable with the levels of vulnerability arrived from
this integrated study. The remaining taluks of the study area come under medium and low
vulnerability. Special attention be given to these taluks in providing better infrastructure,
shelter belts along the coast, cyclone shelters, better drainage facility to mitigate the losses
due to floods. The hazard storm surge could be controlled by mangrove afforestation and
shelterbelts in the case of taluks right on the coast with plains.

Conclusion
The study of the coastal AP for flood vulnerability has high importance to manage the
flood disaster in the state. An ideal flood disaster management system need to support
the activities related to preparedness, damage assessment and rehabilitation based on the
vulnerability map. The space inputs could be used in taking preventive measure through
vulnerability analysis, hazard zonation, and prior risk assessment at regional and local
levels. Hence, an attempt is made to study the coast vulnerability of AP using spatial out
puts and the secondary data.
The area immediately close to the coast is relatively less populated, less accessible due
to poor communication. The people living in the belt of 20 km from the coast generally
comprise fisher folk and weaker sections and majority of them live in thatched houses,
which are vulnerable to the wind pressures of cyclonic gale, which exceed 100 km per
hour. Out of 86 taluks of the study area 38 are under threat to cyclone-induced floods.
The risk index estimated by Jaiyanti (1998: 115) has indicated C (11- 15 of 1 to 25 scale)
category to the above coastal area. However, it is to be rated as B due to its economic
importance and population density in the present context.
The major cities Srikakulam, Visakhapatnam, Machilipatnam, Kakinadaand Nellore
are densely populated with major industries which are within the above zone besides
major deltas of Krishna, Godavari and Pennar rivers. Hence, the study results are to be
adopted to improve flood management strategies, necessary precautions to reduce the
population losses, crop damages, housing problems and infrastructure facilities in the
coastal areas, based on their vulnerability.

Acknowledgement
The first author acknowledges DST for awarding 3D flood simulation project. He also
extends his thanks to AICTE for awarding Emeritus Fellowship.

18 u Journal of South Asia Disaster Studies

GIS-BASED ASSESSMENT OF COASTAL VULNERABILITY TO SEVERE


CYCLONES ALONG ANDHRA PRADESH COAST, INDIA

References
AkwaIbom State, AkwaIbom State- Physical Background, Soils andLand

use

and

Ecological

Problems, Technical Report of the Task Force on Soils and Land use Inventory, AkwaIbom State, 1989.
Burrough P.A., Principles of Geographic Information Systems for Land resources Assessment,(Oxford,
the UK: Clarendom Press, 1986).
Damen M.C.J. and Van Westen C.J., Modelling Cyclone Hazard in Bangladesh. ILWISApplications
03, 2008b.ITC, http://www.itc.nl/ilwis/applications /application03.asp.
Damen M.C.J. and Van Westen C.J., Modeling Cyclone Flood Hazards and Population at Risk in
Bangladesh. Application guide chapter-3, 2003.
Goswami P., and Ramesh K.V., Extreme Rainfall Events: Vulnerability Analysis for Disaster
Management and Observation System Design. Current science, vol. 94, no. 8, April 25, 2008, pp.
1037-1044.
Granger K., Jones T., Leibaand M., Scott G., Community Risk in Cairns: A Multi-Hazard Risk
Assessment. Australian Geological Survey Organization, Commonwealth Australia, 1999.
Jayanti N., Cyclone hazard, Coastal Vulnerability and Disaster Risk Assessment along the Indian
Coasts,Vayumandal, vol.28, no.3-4, July-December1998, pp. 115-119.
Jorin F., Theriaultand F.M., Musy A., Using GIS and Outranking Multicriteria Analysis for Landuse Suitability Assessment, 2001, Int. J. Geogr. Info. Sys, 15(2):153 - 174.
Krishna K.S., Science Plan for Coastal Hazard Preparedness,Current science, vol. 89, no. 8, October
25, 2005, pp. 1339-1347.
Krishna Rao A.V. and RamanaMurty P.V.,Andhra Pradesh Cyclone Hazard Mitigation Project
Integrated Coastal Zone Management in Andhra Pradesh, Storm Surge Preparedness Plan, July
2003.
Miller J.B., and Onwuuteaka J., Oil Spill Emergency Response GIS: Using GIS to Model Environmental
Vulnerability in Coastal Oil Fields, 1999, East Central Nigeria, in Proceedings of ESRI Users
Conference ,http:gis.esri.com/library/userconf/proc99/proceed/papers/pap460/p460.htm.
Multipurpose Cyclone Shelter Programme (MCSP) - Draft final report, vol. IV, Planning and
development issues, UNDP/World Bank/GOB project BGD/91/025,1992.
Prakasa Rao B.S., Ammineduand E., Murty K.S.R., Estimation of Flood Vulnerability Index for
Delta Areas through RS&GIS. International Conference on Geosciences and Remote Sensing,
Coexseoul, Korea, IEEE proceedings vol. 5, July 2005, pp. 3611-3614.
Thumerer A., Jones P. and Brown D., A GIS based Coastal Management Style for Climate Chang
Associated Risk on the East Coast of England, 2000, Geogr. Info. Sci., 14(3): 265-281.
Triskti Bambang and Carolita Ita, Comparison Result of DEM Generated from ASTER Stereo Data
and SRTM. Research officer of the Indonesian National Institute of Aeronautics and Space
(LAPAN), GIS development.net

Vol 3 No 1 June 2010 u

19

B.S. PRAKASA RAO , N. BHASKAR RAO , N. SRINIVAS , G. SRINIVAS,


K. MRUTYUNJAYA REDDY AND M. SATYAKUMAR
Udoh J.C. and Ekanem E.M., GIS-based Risk Assessment of Oil Spill in the Coastal Areas of
AkwaIbom State, Nigeria, 2011.
Unnikrishnan A.S., Assessment of impacts and vulnerability to Indias coastline due to climate
change,Current science, vol.100, no.9, May 10, 2011, p.1273.
Van Westen C.J. and Terlien M.T.J.,Seismic Landslide Hazard Zonation, 2008,ILWIS Applications
07, ITC, http://www.itc.nl/ilwis/applications /application07.asp.
Van Westen C.J., Hazard, Vulnerability and Risk Analysis, 2008, ILWIS Applications 1, ITC,
http://www.itc.nl/ilwis/applications /application1.asp.

20 u Journal of South Asia Disaster Studies

Glacier Lake Outburst Flood (GLOF)


and its Estimation at Nepals
Higher Himalaya
Binod Shakya

Abstract
Changing climate and global warming are the key issues today due to which the world community isfacing several risks. According to D. Douglos, the magnitude of warming was rapid in
the 19thcentury than in the 17th and 18th century in Nepal. In addition, the research shows
that annual warming at Himalayan region of Nepal, between 1977 and 1994, was found to be
0.060C/year. These changes could have large effects on Himalayan glaciers by shrinkage of glaciated areas. Also, there will be substantial increase in the aerial extent of Glacier Lake which may
cause catastrophic Glacial Lake Outburst Floods (GLOF). During GLOF, the rocks and debris are
also released into the valleys downstream with such intensity that the valley floor is badly eroded,
and becomes extremely unstable. The extent of GLOF damage depends on the speed of the release
of the impounded glacier waters. The method presented in paper will enable researchers and planners to estimate GLOF at downstream of lake.
Key words: GLOF, IPCC,Zangzhangbo

Background
Intergovernmental Pannel for Climate Change (IPCC 2001 b) indicates that warming in
the Asian Region is projected to be 30C by 2050. According to Douglos (1995), the magnitude of warming was more rapid in the 19thCentury than in the 17th and 18th century
at higher elevation in Nepal. In addition, the annual warming of the Himalayan Region
of Nepal between 1977 and 1994 was found to be 0.060C/year (Shrestha et al 1999). These

Vol 3 No 1 June 2010 u

21

BINOD SHAKYA
changes could have large effects on Himalayan Glaciers by shrinking of glaciated areas.
Also, there will be substantial increases in the aerial extent of glacier lakes, which may
cause catastrophic Glacier Lake Outburst Floods (GLOFs). Rapid melting of snow enlarge
the existing Glacier Lake or create a new Glacier Lake and a small disturbance can break
the Glacier Lake dam resulting in release of huge amount of water and generates a devastating flood (J.E. Costa, 1988).
In Nepal, Climate change and GLOF studies have been carried out by different institutions including government organisations, international non governmental organisations, academic institutes and the privet sector. CDHM 1999 report showed the growth of
glacier lakes and the formation of new glacier lake in the Nepals Eastern Himalayas with
liable threats of GLOF. Between 1960 and 2000, moraine dammed lakes increased from
33 to 89 in numbers in Khumbhu Region only. Further, it was reported that total moraine
dammed lakes reached 7.254 km2 from 2.291 km2 (Bajracharya et al 2007). In 2001, B.
Shakya identified the most potentially dangerous glacier lakes using criteria based on, size
of a lake and topographic features around the lakeLower Barun glacier Lake, Imja glacier
lake, Thulagi glacier lake, and TshoRolpa glacier lake. Dig Tsho Glacier Lake was considered
a potentially dangerous glacier lake and it burst in August 1985. ImjaGlacierLake is a
similar type of potentially dangerous lake with rapidly expanding and larger lake. The
estimated volume of water stored in the lake was around 28 million m3 in 1992 and 35.8
million m3 in 2002 (Bajracharya et al 2007). The study carried out by B. Shakya 2001 over
9 EasternNepal Himalayan Basin shows active formation of lakes over six basins out of
nine.
From various researches, it was noted that Eastern Nepal and the adjoining Tibet Region (China) have most active glaciers and there are numerous occurrences of glacier
lakes over there. The detailed studies of GLOF of the Himalayan region were not done
properly. But, after GLOF of Zhangzamgbo1981 and Dig Tsho 1985, the scientific attention of the phenomena was realized (WECS Report, 1988). The records show that most
of the cases GLOF have occurred on all major tributaries of the Saptakoshi catchment.
WWF Nepal and Tribhuvan University, Department of Hydrology and Meteorology
have jointly carried out vulnerability assessment from climate change to GLOF in Khumbu Region (sub basin of Saptakoshi catchment) in 2007. Though the environment looks
calmduring assessment, a high threat of GLOF to new settlements, hotels and lodges
in the future was anticipated especially from LakeImja. Some settlements were highly
exposed to GLOF, but such settlements are still growing despite the threat of GLOF (B.
Shakya and K.B. Thapa, 2007).

22 u Journal of South Asia Disaster Studies

GLACIER LAKE OUTBURST FLOOD (GLOF) AND ITS ESTIMATION AT NEPALS HIGHER HIMALAYA
During GLOF, the rocks and debris are also released into the valleys downstream with
such intensity that the valley floor is badly eroded, and becomes extremely unstable.
The extent of GLOF damage depends on the speed of the release of the impounded glacier waters. If it is gradual, the effect may not be as devastating. The GLOF impact is
generally devastating at nearby settlements and infrastructures of downstream of lake
that are situated near or across the river. In addition to this GLOF occur randomly and
give insufficient warning, they may cause much damage to human settlements of upper
mountain areas. Dig Cho GLOF of August 1985 destroyed nearly completed Namche
small hydroelectric project. Even small GLOF of Chuffing on July 12, 1991 destroyed six
houses and riverbanks. The GLOF of Zhagzanbo at Poque on 1981 July (Bhote - Sunkoshi)
destroyed the diversion wire at the Sunkoshi hydroelectric project in Nepal. Two bridges
were also swept away. Therefore,to reduce the vulnerability from GLOF in Nepal, management of settlement, proper designing of major infrastructures such as bridge, dam, roads,
etc. are necessary. To approximately design Lake burst flood and anticipated Lake burst
flood level for settlement, it is first necessary to estimate the volume of flow at outlet of
lake and at the downstream of the lake. This paper aims to present some formulae to
estimate GLOF at outlet of lake and at the downstream of lake. The research is based on
Scientific Research of Central Asian Hydrometeorological Institute (SANIGMI) conducted
between1998 and 2001. The developed formulas are based on lake volume and statistical
modeling of past GLOFs data from Nepal and other lake burst data. The paper also aims
to present implementation of method on Imja Lake study by WWF project.

GLOF in Nepal
The detail studies of GLOF of the Himalayan region were carried out properly before
1980s, nevertheless, after GLOF of 1981 (Zhangzamgbo) and 1985 (Dig Tsho) the scientific attention of the phenomena was realized (WECS Report, 1988). Most of the GLOF
cases were identified at the all major tributaries of the Saptakoshi catchment. Some of the
recorded events are:
GLOF on Bhotekoshi from Dig Tsho (August 04,1985)
TinkoGLOF of 1982 in Yairuzangbo River of Poique catchment (in Tibet, China and
Arun catchment in Nepal)
GLOF on Bhote-Sunkoshi in Nepal from Zhanzangbo lake 1964, 1981
GLOF from Phuchan glacier lakes on the Tamur river in 1980 and Nare glacier lakes
(South slopes of Mt. Ama Dablam in Nepal) in 1977

Vol 3 No 1 June 2010 u

23

BINOD SHAKYA
GLOF of Ayaco at Northern slope West of Mt. Everest in Pumpque catchment (Tibet,
China) 1968, 1969 and 1970
GLOF of ChubungJuly 12, 1991 Rolwaling valley
GLOF of Zhagzanbo lake at Poqu in July 1981 (Bhote-Sunkoshi)
The location of glacier lakes bursts are presented in Figure 1.1.

Figure 1.1: GLOF in Nepal

Dam Break Models


The GLOF modelling is in the initial stage in Nepal. S. Bajracharya et al 2007 attempted
to predict GLOF attenuation from Imja Lake using Hydrodynamic model. Nevertheless,
many types of Dam Break Flood models have been developed worldwideranging from
simple computation based on historical dam failure data to complicated hydrodynamic
models. It is very complicated and complex to collect detail field survey of Himalayan

24 u Journal of South Asia Disaster Studies

GLACIER LAKE OUTBURST FLOOD (GLOF) AND ITS ESTIMATION AT NEPALS HIGHER HIMALAYA
Glacier Lake and river cross-sections. In such case, the reliable estimation of the peak flow
and flow attenuation at downstream from a dam is statistical modeling from historical
dam failure data. Kirkpatrick in 1977 shows relation between flow volume and height of
dam considering 21 dam failure data and statistical modeling. Similar type of equation
was developed by U.S Soil Conservation Service in 1981 produce relations using 31 dam
fail cases. Hagen in 1982 and Syldler 1984 developed equation between Flood at outlet
and lake volume. Similar method was adopted by Clague and Mathews (1973) glacier lake
outburst (cited J.J. Clague and S.G. Evans, 1994).

Method
For statistical model the basic data needed are historical flood data at lake outlet, flood
attenuation data downstream of lake and other data such as lave volume or dam height
before it burst. Only few data on flood downstream attenuation, flood at outlet of lake
and lake volume prior to burst is available in Nepal. However, some data on surface area
of lake corresponding to lake volume is available for regression analysis.
Therefore, three regression equations are combined to develop GLOF statistical model as:
VL=f(SA)
Qo=f(VL)
QD/Qo=f(D)
where, VL is lake volume, SA is lake surface area, Qo is GLOF at outlet of lake and QD is
GLOF at any downstream distance D.
In other to develop GLOF attenuation formulae, various GLOF data, lake volume and
surface area of lake, and downstream distances are collected from different literatures. Beside
these, primary data from field survey of Phuchan and Dudh Koshi GLOF were also collected.

Key research findings


Based on Lower Barun, Imja, Thulagi and TshoRolpa glacier lakes, and some lakes from Arun
and Kumbhu basins, the best-fit regression equation for glacier lake surface area and volume relation developed is:
VL = 0.0408 x SA1.429 (1)
where, SA is surface area in km2 and VL is total volume of glacier lake in km3. The
square of correlation coefficient is 0.88. The developed regression equation is presented

Vol 3 No 1 June 2010 u

25

BINOD SHAKYA
in Figure 1.2. Similarly based on TshoRolpa and Dig Tshobursts in Nepal and 10 more
GLOF data outside Nepal, regression equation is developed between lake volume and
GLOF at outlet. The developed equation is:
Q0 = 95. 97VL0.5981 (2)
Where, VL is total volume of lake in 106 m3 and Qo is GLOF at outlet of lake in cumecs.
The square of correlation coefficient for the relationships equals to 0.90. The developed
regression equation is presented in figure 1.3. The historical GLOFs with volume of lake
is presented in Table 1.1.

Figure 1.2. Relation between volume and surface


area of glacier lakes

Figure 1.3. Relation between total


volume and discharge at outlet of the lake

Table 1.1: GLOF with lake volume


Name of lake

Burst year

VL , 106 m3

Qo, m3/s

DudhKoshi, Nepal

1977

8.80

1,100

Demmevatn, Norway

1937

11.6

1,000

GraenalonIceland

1939

1500

5,000

GjanupsvatnIceland

1951

20.0

370

George, Alaska

1951

1730

10,100

Tulsequah, British Colombia

1958

229

1,556

Summit lake, British Colombia

1967

251

3,260

Ekalugad valley, Baffin island

1967

4.80

200

Hazard lake, Canada

1978

19.6

640

Gruben, Switzerland

1970

0.17

15

Gorner, Switzerland

1944

6.00

200

Dig Tsho, Nepal

1985

25.0

1,600

Source: Clague, J.J and Evans, S.G. (1994)

26 u Journal of South Asia Disaster Studies

GLACIER LAKE OUTBURST FLOOD (GLOF) AND ITS ESTIMATION AT NEPALS HIGHER HIMALAYA
The records on downstream attenuation of GLOF are rarely available. The GLOF attenuation strength for Zangzhangbo GLOF was presented by Yamada, 1998. The best-fit
regression equation (figure 1.4) of GLOF attenuation is constructed with data of Zhangzangbo GLOF and with primary data from field visit at Phuchen burst site (B. Shakya,
2001) and two historical floods from dam failure in the US (Table 1.2). The developed
equation is:

(3)
where, Q0 and QDare GLOF at outlet of lake and at distance D (km) downstream respectively.
Table 1.2: Flood attenuation, %
Lake name

Distance,
km

Flood
attenuation, %

L.D. creek (US)

0.0

100.0

17.0

82.0

22.0

37.0

80.0

11.0

124

6.3

0.0

100.0

16.0

46.0

138

3.9

178

2.9

Teton (US)
Figure 3.1: Downstream attenuation
from lake burst

0.0
Estimation of Imja GLOF:This lake is located D. Tsho (Nepal)
0
0
at 27 5917 North latitude and 86 5531 East
5.4
longitude over eastern Nepal. The total volume
12.0
at present is about 35.8 million m3 in 2002
15.8
(Bajracharya et al 2007). The WWF project on
Phuchen (Nepal) 0.0
Khumbu climate change (2007) adopted the
7.6
above equations and estimated flood assuming
10.5
incase of GLOF in 2008. The result is shown in
Table 1.4.
Source: cited Shakya B 2001

100.0
60.2
26.9
15.3
100.0
67.0
40.0

Vol 3 No 1 June 2010 u

27

BINOD SHAKYA
Table: 1.4: GLOF at outlet and downstream of Imja incase of burst
Distance, km

10

20

30

40

50

GLOF, m3/s

4,757

2,900

2,344

1,678

1,249

949

735

Conclusion
A detailed research on all glacier lake is a tough task due to the harsh topography, poor
economy and lack of scientific research. The possible GLOF mark at downstream lake
and other management from GLOF threat is still not started. However, Nepal Government lowered a TshoRolpa Lake level using siphons and set up GLOF early stations at and
downstream of Lake. To reduce GLOF vulnerability, estimation of downstream flood from
burst is key data to mark the dangerous level at downstream and to design any infrastructure near or across river. In this connection the only easily achievable data is surface area
of the lake e.g. via Satellite Images. The area also can be extracted from different literatures and reports. Therefore, equations 1,2, and 3 presented above can be used to estimate
GLOF at downstream of lake.

References
CDHM T.U., Glacier Lake Assessment of Nepal Himalayas Uning NASDA satellite Image. Research
Report, Tribhuvan University., 1999.
Chinese Academy of Sciences, WECS, NEA: Report On First Expedition To Glacier And Glacier Lakes
In The Pumqu (Arun) And Poiqu (Bote-Sunkoshi ) River Catchments Xizang (Tibet), China, 1988.
Clague J.J. and Evans S.G., Formation and failure of natural dams in the Canadian Cordillera.
Geological Society, Canada, Bull. 464, 1994.
Clarke G.K.C., Glacier outburst floods from Hazard Lake, Yukon Territory, and the problem of
flood magnitude prediction. Journal of Glaciology, 1982, vol. 28, pp. 3- 21.
Costa J.E. and Schuster R.L., The formation and failure of natural dams.Geological Society of America
Bulletin, vol.100, pp. 1054- 1068, 1988.
Douglos D., Climate Change in Nepal Himalaya, Research Paper, Arizona State University,
Publication, ASU, 1995.
IPCC (2001b) and Climate Change (2001):Impacts, Adaptation and Vulnerability, Contribution of
working group II to the Third Assessment Report of the Intergovernmental Panel on Climate
Change, (Cambridge: Cambridge University Press).
Ives J.D., Glacier Lake Outburst Floods and Risk Engineering in the Himalaya. ICIMOD, occasional
paper no.5, 1986).

28 u Journal of South Asia Disaster Studies

GLACIER LAKE OUTBURST FLOOD (GLOF) AND ITS ESTIMATION AT NEPALS HIGHER HIMALAYA
Shakya B. and Thapa K.B., Integrated Study in Hydrology and Meteorology of Khumbu Region with
Climate Change Perspectives, Research Report, WWF publ., 2007.
Shakya B., Estimation of Main Hydrological Characteristics for Mountain Rivers of Nepal.Research
Paper, publ. Central Asian Research Hydrometeorological Institute, publno. 556.535.3, 2001.
Shrestha A.B., Wake C.P., Mayewski P.A. and Dibb J.E., Maximum Temperature Trends in the
Himalaya and its Vicinity: An Analysis based on Temperature Records from Nepal for the period
1971-1994. In Journal of Climate, publ. no 12: 2775-2767, 1999.
Yamada T., Glacier lake and its outburst flood in the Nepal Himalayas. Monograph no 1, Japanese
Society of Snow and Ice, 1998.

Vol 3 No 1 June 2010 u

29

Usefulness of some Disaster Anticipatory


Measures of the Indian Railways in Disaster
Management Plan of Mega cities of India with
Special Reference to Ahmedabad city
J.G. Macwan

Abstract
This paper focuses on the present disaster anticipatory systems of the Indian railways and usefulness of some measures in the disaster management plan of mega cities of India. Indian Railways
is the biggest network of Asia, and possesses efficient Disaster Management (DM) system. Hence,
some disaster anticipatory measures become very useful and effective in DM plan of a Mega civic
body in India. In this paper, these important useful points of DM plan of Indian Railways are
shown in reference to Ahmedabad city DM plan.
key words: Disaster, anticipatory measures Preventive actions, DM plan, Efficient city
DM plan

Introduction
The Indian subcontinent can be primarily divided into three geophysical regions. The
topographic and climatic characteristics of each region make them susceptible to different type of disasters54 per cent of land is vulnerable to Earthquakes, 40 million hectares of land is vulnerable to floods; hilly areas are at risk from landslides and avalanches.
Moreover, vulnerability to disasters/emergencies of terrorist attack, Chemical, Biological,
Radiological and Nuclear (CBRN) origin also exists.
The Indian Railways came into existence with the running of the first train from Kurla
to Thane in 1853. Ever since then, handling train accidents has been a priority area for
the railways. Thus, in the Indian Railways, disasters may be in the form of natural calami-

Vol 3 No 1 June 2010 u

31

J.G. MACWAN
ties like cyclones, heavy rains, earthquake, landslides, or manmade calamities accidents
like fire, collision or derailment. The Indian Railways has its own DM system and plan.
Ahmedabad is a mega city, commercial capital of Gujarat, the former capital of Gujarat, a major centre for industries as well as trade and commerce, was founded in 1411
AD as a walled city on the eastern bank of River Sabarmati. In the past, due to its textile
industries, Ahmedabad was famous as the Manchester of India. Presently, Ahmedabad is
the 7th largest city of Gujarat and also its commercial capital. Ahmedabad is vulnerable
to multiple natural hazards like earthquake, flood, cyclone, heavy rain due to its geographical position. Moreover, the city is vulnerable to manmade technical and biological
hazards like terrorist attacks, communal riots/civil disorders, fire, Air/rail, road/industrial
accidents and epidemic disasters. In the past, Ahmedabad has faced destruction in 2001
and 2002 due to earthquake and communal riots. To handle disasters, Ahmedabad Municipal Corporation has prepared DM plan. Some important points of DM plans of the
Indian Railways became useful in city DM plan.

Indian Railways:
Railways were first introduced to India in 1853. By 1947, the year of Indias independence, there were forty-two rail systems. In 1951, the systems were nationalised as one
unit, becoming one of the largest networks in the world. Indian Railways operate both
long distance and suburban rail systems on a multi-gauge network of broad, meter, and
narrow gauges. It also owns locomotive and coach production facilities.
1.1. Organisation
The Indian Railways is Asias largest and the worlds 2nd largest Railway system, next to
Russia. The Indian Railways is the nations largest undertakings employing more than 16
lakh people. The Indian Railways is owned and managed by the Central Government being the principal mode of inland transport.
The responsibility for the administration and management of the Railways vests with
the Railway Board under the overall supervision of the Minister for Railways. The Board
is empowered to function as the Ministry of the Government of India, and exercise all
powers related to the operation of the Railway organisation. The Railway Board consists of six members and chairman who has special position as principal secretary to the
Government of India. The Indian Railways is divided into zones, which are further subdivided into divisions. The number of zones in the Indian Railways increased from six to
eight in 1951, nine in 1952, and finally 16 in 2003. Each zonal railway is made up of a

32 u Journal of South Asia Disaster Studies

USEFULNESS OF SOME DISASTER ANTICIPATORY MEASURES OF INDIAN RAILWAYS IN DISASTER


MANAGEMENT PLAN OF MEGA CITIES OF INDIA -WITH SPECIAL REFERENCE TO AHMEDABAD CITY
certain number of divisions, each having a divisional headquarter. At present, there are
sixty-seven divisions.
Each of the sixteen zones, as well as the Kolkata Metro, is headed by a General Manager (GM) who reports directly to the Railway Board. The zones are further divided into
divisions under the control of Divisional Railway Managers (DRMs). The divisional officers of engineering, mechanical, electrical, signal and telecommunication, accounts,
personnel, operating, commercial and safety branches report to the respective Divisional
Manager, and are in charge of operation and maintenance of assets. Further down the
hierarchy tree are the Station Masters who control individual stations and the train movement through the track territory under their stations administration.
1.2 Technical and operating system
The total length of track used by the Indian Railways was about 1,11,600 km while the total
route length of the network was 64,061 km on March 31, 2010. About 31 per cent of the
route-kilometer and 46 per cent of the total track kilometer was electrified on March 31, 2010.
The Indian Railways operates about 9,000 passenger trains and transports 20 million
passengers daily across 28 states and 2 union territories. Sikkim and Meghalaya are the
only states not yet connected by rail.
1.3 Types of Disasters
Disaster in the railway context was traditionally a serious train accident, caused by human/equipment failure, which may affect normal movement of train services with loss
of human life, or property, or both. This is now extended to include natural and other
manmade disasters.
Different types of disasters along with a few examples are:
(i) Natural Disaster
Earthquakes, floods, cyclones, landslides, tsunami, etc.
(ii) Train Accident related Disaster
Collisions (with a huge number of casualties), train marooned (flash floods), derailments
at a bridge over a river, and coaches falling down; train washed away in cyclone, derailment of a train carrying explosives or highly inflammable material, tunnel collapse on a
train, fire or explosion in trains, and other miscellaneous cases, etc.
(iii) Manmade Disasters
Acts of Terrorism and Sabotage, that is, causing deliberate loss of life and/or damage to

Vol 3 No 1 June 2010 u

33

J.G. MACWAN
property. It includes setting fire to a train, railway installations, etc., bomb blast at railway
station/inside train, chemical (terrorism) disaster, biological and nuclear disaster.
1.4 Nodal department for Policy Formulation on DM in the Indian Railways
There are control offices in the Railway headquarter and the divisional offices which
monitor and control train movements round the clock.
The preparation of the DM plans of the Indian Railways and on the Zonal Railways in
co-ordination with the different departments of the railways, other Central/State Govt.
agencies, NGOs, Private agencies, etc. has to be done by the Safety department in the
Railway Board, on the Zonal Railways and Divisions. Safety Director in the Railway Ministry, Chief Safety Officer in each zonal headquarter, and Sr. Divisional Safety Officer in
the Divisional headquarters are full-time emergency officers.
Besides the above, organisation which functions under the control of Ministry
of Railways, there exists a Commission of Railway Safety headed by Chief Commissioner of Railway Safety with Commissioners of Railway Safety reporting to him. To
maintain independence, the safety Commissioner functions under the control of
Ministry of Civil Aviation and enquires into cases of all serious accidents. Its permission is also needed for opening of new lines, introduction of new rolling stock, etc.
The Commissioner has quasi-judicial powers, and the work is manned by officers
from the Indian Railways.
The Hospital DM plans and the Security arrangements (drills, etc.) are prepared and
coordinated by the Medical and the Security department respectively. The management
of floods, cyclones, earthquakes, landslides, etc., and the preventive actions required for
mitigation are coordinated by the the Civil Engineering Department. The Rescue and
Restoration of DM Plans including preparing plans and procurement of specialised equipment and rescue-centric training of personnel has to be coordinated by the Mechanical
Department
1.5 Disaster Anticipatory Measures of Indian Railways
The Indian Railways have an elabourate scheme for collecting information and pre-warning of the staff located in areas likely to be affected, for taking precautionary arrangements to prevent any mishap to moving trains or passengers by giving relief to passengers
and staff and restoration of communication expeditiously if any stretch of line is affected.
These form part of mandatory instructions to the staff on action to be taken during train
accidents; they also form part of General and subsidiary rule of working of Railways.

34 u Journal of South Asia Disaster Studies

USEFULNESS OF SOME DISASTER ANTICIPATORY MEASURES OF INDIAN RAILWAYS IN DISASTER


MANAGEMENT PLAN OF MEGA CITIES OF INDIA -WITH SPECIAL REFERENCE TO AHMEDABAD CITY
1.6 Strengths of the Indian Railways to handle a Disaster
In handling disasters, the Indian Railways is in a unique position as it has a number of
strengths not available with many other government departments. These include:
Own communication and control network with each Station, Division, and Zonal
headquarter
Own force of RPF/RPSF, fire brigade, territorial army units and civil defense
Railways own medical infrastructure
Full time safety officer and safety department available at division level, Zonal level
and Ministry level to see DM.
Railway organisation is distributed in sections of each 8 km area and section in-charge,
essential staffs with army of gangmen are bound to stay in railway quarters available
in each section 24 hours
All area/divisional officers are bound to stay in railway bungalows provided in each concerned area/division so that they can reach at the time and place of disaster immediately.
Medical and rescue train with essential materials is always ready in each division or
earmarked area to reach assistance and handle disaster.
Essential staffs of each department of Railways are instructed to keep them ready 24
hours so that at the time of occurrence of disaster and when warning siren starts,
they can reach he relief and rescue train placed in the yard with in 30 to 60 minutes
depends on disasterserious disaster-30 minutes, Average disaster-60 minutes. (Intermittent voice of siren means serious disaster and continuous voice of siren means
average disaster.)
Inspection schedule of Railway infrastructure is very powerful as a preventive action.
From the section in- charge to General Manger, all are bound to complete their inspection schedule in fix time frame
1.7 Railways shortcomings to handle Disaster
There are, however, a few inadequacies in the Railways own resources, which are very
essential for handling a specific type of disaster:
Absence of tunnel rescue equipment
Non-availability of trained divers/swimmers for extrication of passengers and/or casualties (dead bodies and drowning/drowned passengers) from rolling stock falling down
in sea/river/lake, etc.
Non-availability of cranes operated from a ship/barge for lifting of the coaches/bogies
from a water body

Vol 3 No 1 June 2010 u

35

J.G. MACWAN
Limited resources to handle a terrorist attack on a train and/or a station, other railway
premises, etc.
1.8 Managerial Preparedness
An emergency organisation is formed to suit each situation. When a disaster causes damages resulting in interruption to traffic, managers and staff are drawn from among the
regular serving personnel, and emergency organisation is formed for supervision till normal situation restores. For attending this activities, suitably equipped relief trains, medical
vans and emergency building /track materials including emergency bridge components
are kept in readiness all the time. Railways carries out potential training programmes just
like periodical training, refresher training, promotional course etc., for staff in executive
and supervisory levels and from field staff on DM aspects, either as part of their regular
training programmes or through special course.
Moreover, in order to expedite work, some unorthodox management methods are
adopted; for example, fixing agencies to carry out work, purchase of urgently required
materials, engaging of labour, revising priority for movement of trains working across
ones jurisdiction in respect of territorial as well as sectional boundaries. Any senior railway officer who happens to travel in a train, which meets with a disaster or faces disaster,
is expected to inform all concerned on the nature of the trouble and take charge of the
situation till the officer in charge of the area arrives at site.
1.9 Action during Disasters
(i) Action on Division/Zones on Orange/Red Alert
On the issue of an Orange Alert (or of a higher level), the Responders have to be activated
as required for relief, etc.:
Mobilisation of gangmen
Hospitals to mobilise doctors and para-medical staff
deployment of railway protection force, state railway police and activate civil defense,
territorial army units, and scouts and guides
Operation and manning of the disaster control room and Coordination amongst various stake holders through advance warnings
(ii) Monitoring/Reporting of Effects of Disaster
The Safety Department. in the Board would be given information regarding Orange/Red
Alerts. On the declaration of an incident as a Disaster by a State Government or District

36 u Journal of South Asia Disaster Studies

USEFULNESS OF SOME DISASTER ANTICIPATORY MEASURES OF INDIAN RAILWAYS IN DISASTER


MANAGEMENT PLAN OF MEGA CITIES OF INDIA -WITH SPECIAL REFERENCE TO AHMEDABAD CITY
Administrator or even by the GM/AGM of the Zonal Railway, the CSO would give timeto-time updates to the Safety Control in Railway Board of the situation. Assistance of
other departments would be made available by the GM to the Safety Department of the
Zonal Railways.
1.10 Standard Operating Procedure (SOP) on Railways
(i) National Disasters
The Civil Engineering Department at the field level and on the divisions gets information through advance warning sent by the respective government departments on the
possibility of floods, cyclones, earthquakes, landslides, etc. Depending on the gravity of
the disaster/crises/calamity expected, the information would be passed on to the Divisional officers through the Emergency Control, which will act as the ICS. Where train
operations have to be suspended or regulated, the operating departments would be suitably advised. After making the train regulation plan, the divisional control would advise
the commercial and security departments for management of the welfare of passengers.
Alerts to the passengers would be issued through the PR Department of the Railways in
the print and electronic media.
The Divisional Railway Managers of the divisions shall ensure coordination among
the departments for ensuring running of train services (including relief special trains)
as also relief arrangements for the passengers and for the Welfare of Railways own staff.
Assistance of other divisions and from the Zonal Railways would be taken through the
Headquarter of the Zonal Railways (i.e. by involving the General Manager). Coordination
with the IOC of Ministry of Home affairs and National DM authority/NDRF would be
through the Emergency Control of each zonal headquarter.
Priority in the rail operating will be shifted towards first giving relief to standard passengers, and then, to the restoration of line to the minimum extent required to enable
trains to pass at slow speed. A foot-by-foot survey is done over the entire length affected
by the disaster at the earliest possible time to assess the extent of damages and estimate
requirement of men and materials communication. Depending on the seriousness of
damages and likely duration of carrying out temporary repairs, a senior Executive from
the Railway headquarter would be sent to the site to take charge of the work. All officials
of different disciplines would be answerable to this executive and progress of work is
monitored at divisional, zonal, and Central (Railway Ministry) level on day-to-day basis.
In some cases, it is done on twice a day basis.
As a large number of men would be working at different locations on a continuous

Vol 3 No 1 June 2010 u

37

J.G. MACWAN
basis, arrangements are made for supply of free cooked food to them in proper time, and
temporary staying arrangements closer to the site are provided. A proper medical and
health team would be posted to look after the hygiene at site and render medical treatment to staff at site itself including inoculation against cholera and typhoid. Engaging a
lot of labour for short duration directly by the Railways paying them and discharge them
on completion of work is not practicable by following laid down procedure of Railways;
same applies to procurement of some construction materials. These are, therefore, done
by engaging contractors. Time being precious, the executive in charge is given freedom
to fix agencies on the site by directly negotiation with experienced contractors who are
available close by and who have the necessary resources readily available with them. The
negotiation are carried out at site itself by committee comprising of engineer concerned,
the finance and accounts officer for the section and another officer. There have been occasions during such emergencies when contractors have been fixed on 24 hour basis for
assessment of damages.
The labour of the track work, bridgework, signal, and telecom and electrical traction
works call for special skill. Hence, such staff is collected from other different sections or
divisions or by diversion from construction work to the said site of work temporarily.
Railways normally hold emergency stock of stones and temporary bridge materials in
different areas.
(ii)Manmade Disasters
Different forms of terrorism fall under the ambit of manmade disasters. A major role
has to be played by the Security Department of the Railways who will coordinate with
the State Governments, and when required, coordinate with the para-military and other
forces. The Security Control of the division will act as the ICS. The headquarter of Security
Control will coordinate with the IOC of MHA.
1.11 Initiative actions after DM plan 2005 of Government of India
Preparation of DM Plans on Zonal Railways with periodical review
Safety Department: Nodal Department for compilation/updating of DM Plans
ISO certification of DM plans
Creation of national disaster response and mitigation funds
Modernisation of relief/rescue during disasters
Crowd control and management of rush at railway stations during festivals
Action plan to handle terrorist attacks on a freight train carrying inflammables

38 u Journal of South Asia Disaster Studies

USEFULNESS OF SOME DISASTER ANTICIPATORY MEASURES OF INDIAN RAILWAYS IN DISASTER


MANAGEMENT PLAN OF MEGA CITIES OF INDIA -WITH SPECIAL REFERENCE TO AHMEDABAD CITY
Strengthening works of Railway infrastructure to reduce destruction due to earthquakes, floods, etc.
Creation of network of mobile medical infrastructure to handle disaster on Railways
Preparation of hospital DM Plan to handle disaster on Railways
Constitution of specialised trained team of Railway Police force in each divisional
headquarter in crowd control and functional support in case of disasters
An integrated security scheme has been sanctioned for installation at 195 stations
of the Indian Railways to prevent explosion in trains and railway premises; the system envisages multi-layered surveillance of vehicles, luggage and passengers in station
premises
Crisis management plans are prepared to handle terrorist acts and hijacking of trains
Modernisation of communication on the Railways through Satellite
The Railways have their own ICS as they have to deal with crises like situations and
mini-disasters in the day-to-day operational working, and especially, with handling of
train accidents. With the setting up of the Rescue Centric Training Institute at Bengaluru, the ICS structure will get streamlined
Coordinating Integrated Command System of Railways with Integrated Operations
Centre of Ministry of Home affairs
Preparation of different DM plans for various disasters like earthquakes, floods, cyclones, landslides and avalanches, biological disasters, chemical Disasters, chemical
(terrorism) disasters, nuclear and radiological emergency
Formation of detailed training programmes for senior officers and subordinate staffs
regarding DM

Ahmedabad
Ahmedabad, the former capital of Gujarat and a major centre for industries as well as
trade and commerce, was famous as the Manchester of India. Presently, it is the 7th largest city of Gujarat, and also its commercial capital.
2.1 Demographic Characteristic
Administrative Body-Municipal Corporation
Area: 458.14 sq km
No of Wards: 57
Polulation: 66,00,000 (as per 2001 census and with new city limit)
No of Households: 10,55, 851

Vol 3 No 1 June 2010 u

39

J.G. MACWAN
2.2 Organisation setup
Administration of Ahmedabad Municipal Corporation is being done by two wings(1)
Elected wing, and (2) Executive wing. The head of the elected wing is called Mayor who
is a first citizen of Ahmedabad. There are so many other committees like standing committee, transport committee, Legal committee, etc., which are headed by the concerned
chairmen. The Mayor, and all committees are policy makers, While Commissioner, a
Secretary rank IAS officer is Administrative head and appointed by state Government,
is assisted by other Dy. and Assistant Commissioners. Presently, Ahmedabad Municipal
Corporations staff strength is 35,000. For administrative conveniences, municipal area is
divided in 5 zones and 57 wards, which are headed by Dy. Municipal Commissioner and
Ward Officer respectively.
2.3 Potential Hazard Possibility In Ahmedabad
Ahmedabad is vulnerable to multiple natural, manmade, technical, and biological hazards:
Table 1 Possibility and intensity of Disasters in Ahmedabad
Type of disasters

Possibility

Earthquake

Medium

Flood

Medium

Heat wave

High

Cold wave

Medium

Cyclone

Medium

Terrorist Attack

High

Civil Disorder

High

Mutinity of state police force

Less

Road Accident

Medium

Railway Accident

Less

Air Accident

Medium

Fire

High

Wall/Building Collapse

High

Industrial Accident

Medium

Epidemic Disease

High

Animal Disease

Very less

40 u Journal of South Asia Disaster Studies

USEFULNESS OF SOME DISASTER ANTICIPATORY MEASURES OF INDIAN RAILWAYS IN DISASTER


MANAGEMENT PLAN OF MEGA CITIES OF INDIA -WITH SPECIAL REFERENCE TO AHMEDABAD CITY
2.4 Strengths of Ahmedabad Municipal Corporation to handle a Disaster
Own Fire brigade staffs who are trained personnel, to handle relief rescue work of disaster like flood, earthquake, cyclone, etc.
Own medical infrastructure with 5 big hospitals and 79 small hospitals/urban health
centre. ( Civil hospital, the biggest hospital of Asia is in Ahmedabad also)
Emergency medical assistance service (1-0-8) established by the Government of Gujarat in public-private partnership basis is available
Own city transport service with about 1,000 buses and other vehicles
Well established ward and zonal offices with about 35,000 staff
Efficient public information system through cable network, mobile vans, Doordarshan, Radio/Akashvani
Big police force is available in city, which is useful at the time of disaster
Assistance of various government offices situated in the city area are available at the
time of disaster, that is, Collectors office, PWD, Meteorological department, Electricity Company/Gujarat Electricity Board, Railway, State transport, Water supply Board,
BSNL, etc.
Services of various NGOs like Civil Defense, Scouts and Guides, NCC, Red cross, Indian Medical Association, Rotary club, Lions club, etc. are available
2.5 Shortcomings of Ahmedabad Municipal Corporation to handle Disaster
Police force and communication systems are major requirements for DM but these two
are not under the control of the Ahmedabad Municipal Corporation; only assistanceship
is available.
2.6 Disaster Anticipatory Measures of Ahmedabad Municipal Corporation
(i) System to handle Natural disaster
The Ahmedabad Municipal Corporation receives the forecast and warning messages from
the meteorological department regularly for natural disasters like cyclones, and heavy
rain, while it receives warning/forecast from the State Irrigation Department. Whenever
an emergency is feared, the action plan is initiated. To handle the situation, Zonal Dy.
Municipal Commissioner starts work with the concerned ward officer, fire brigade and
other staff under the guidance/instruction of the Municipal Commissioner. In the case
of floods/heavy rain, earthquakes, cyclones, etc., help of the police department and collectors office is taken for relief and rescue work. Regarding situation of disaster, the Municipal Commissioner sends report to the GSDMA, and depending on the situation, help

Vol 3 No 1 June 2010 u

41

J.G. MACWAN
from the State Government is provided. In worst situations, on the report of the State
Government, the Central Government provides help.
(ii) System to handle Manmade Disaster
In the case of manmade disasters like terrorist attacks, civil disorders, the function of the
Municipal administration is very limited. Two main organsthe Fire brigade and the
hospital infrastructuresare activated to help the police authority to handle the situation.
In the case of accident-related disasters, all hospitals situated in the city area provide
emergency medical assistance to the victims, and fire brigade provides services in relief
and rescue work.
While in the case of epidemic diseases, the Health depart of the Ahmedabad Municipal Corporation provides services through their hospitals, and public relation department
issues various advertisements in print and electronic media to provide information about
the diseases to the public so that they can get preventive actions. In the worst situation
of epidemic disasters, the Ahmedabad Municipal Corporation contact the State Health
Ministry for assistantship, and depending on the situation, the State Ministry gets help
from the Central Ministry.
2.7. Initiative actions after Disaster Management plan 2005 of Government
of India
Before DM plan 2005 of the Government of India, Gujarat State Disaster Management
Authority (GSDMA) and the United Nation Development Programme (UNDP) had jointly
started Urban Earthquake Vulnerability Reduction Programme (UEVRP) 2002-07, in Mega
cities of India. Under this scheme, ward DM plan and City DM plan were prepared.
2.8 Useful Anticipatory Measures of Indian Railway
Looking at the above facts, it is understood that the anticipatory measures of the Railway
department is more efficient. Some important anticipatory measures, which become useful in the city DM plan, are:

42 u Journal of South Asia Disaster Studies

USEFULNESS OF SOME DISASTER ANTICIPATORY MEASURES OF INDIAN RAILWAYS IN DISASTER


MANAGEMENT PLAN OF MEGA CITIES OF INDIA -WITH SPECIAL REFERENCE TO AHMEDABAD CITY
Important Points

Railway Organisation

Ahmedabad Municipal Corporation

Preparation of DM

3-tier DM plan is prepared (Divi-

2-tier DM plan is prepared (ward level and

Plans

sional, Zonal and Ministry level)

city plan); Zonal plan not prepared

Preparation of

Separate sub-plan for each natu-

separate sub-plan for each disaster is not

sub-plan for each

ral disaster is prepared

prepared

disaster

separate Hospital DM plan is

separate Hospital DM plan is not pre-

prepared

pared

Periodical review of

Periodical updating of DM plan

updating of DM plan is being done partly

DM plan and mock

is being carried out by the safety

Mock drills are not being arranged periodi-

drill arrangement

department

cally

related to DM

Mock drills are being arranged


periodically
Safety Depart-

Full time safety officer and

Full time safety officer/Nodal officer or

ment- Nodal

safety department is available in

department is not in existence but Dy.

Department for

division office, headquarter and

Municipal Commissioner is designated as

Compilation/Up-

Ministry office

the Nodal officer and the Chief fire officer

dating of DM Plans

ISO certification received for

as the liaison officer to look after disaster

the DM plans (Ministry level)

management work. One project officer appointed by GSDMA is looking after UEVRP
project

Availability of Basic

Possess Own police force, com-

Possess Own fire brigade &Hospital infra-

infrastructure

munication system and Hospital

structure (for police force and communica-

infrastructure and fire brigade

tion system depends on state Government)

Availability of

Essential items to handle the

Essential items to handle the disaster are

essential items to

disaster are available at the ear-

available in city but in each ward all essen-

handle the disaster

marked place in each section

tial items are not available


Essential items to handle the terrorist attacks, civil disorders, etc. are not available;
it is with the State Government

Availability of

An army of gangmen spread out

Limited workers/labourers are available

essential technical

all over the Indian Railways

Ward officer/zonal officer and essential

staff, officers and

All essential staffs/officers are

staffs including workers/labourers are not

workers in each

bound to stay in section area and

bound to stay in concerned ward/zone and

area

available in each section 24 hours

hence not available 24 hours

Vol 3 No 1 June 2010 u

43

J.G. MACWAN
Control network

Well-planned Control network

Well planned control room for fire brigade

on each Division linked with

and ambulance services

each Station and for Division to

Control room at Municipal corporation

Zonal headquarters

headquarter for upward and downward

Control room at Zonal head-

information /instruction

quarters to Ministry level

zonal level and ward level control room


are not efficient

Inspection network

Inspection network is well-

Inspection network is very poor

as a preventive

planned from section in-charge

There is no inspection schedule for Higher

action

level to General Manager level.

officers like Municipal

Each officer has a schedule of in-

/Deputy Commissioner.

spection as the preventive action

There is no surprise night inspection

Surprise night inspection is also

schedule

commissioner

a part of duty
Services of civil de-

Civil defense and territorial

Municipal staff/officers are not trained for

fense scout guide,

army wing are available in Rail-

civil defense/territorial army service

NCC, and territo-

way staff

Services of scout guide, NCC are also avail-

rial army wing at

Services of scout guide, NCC

able from Municipal schools and granted

the time of disaster

are also available from Railway

government schools at the time of disaster

schools at the time of disaster


Availability of

Railway department does not

Has its own swimmers and boats

swimmers, boats,

have its own swimmers/boats;

Some quantity of relief /rescue materials

cranes, Gas cut-

hence; depend on State govern-

like cranes, gas cutters, etc. are available

ters at the time of

ment/civic bodies

disaster

Some quantity of relief/rescue


materials like cranes, gas cutters,
ropes, etc. are available.

Duty list of depart-

Detailed duty list of each de-

Duty list of concerned department and

ment and officer to

partment and important officers

officers are not shown in DM plan; only an

handle disaster

are prepared and available in DM

outline is given

plan
Availability of

medium size transport vehicles

Own transport services; hence, buses and

vehicles at the time

are available

medium vehicles are available

of disaster

Depends on emergency, buses


from state transport are taken

44 u Journal of South Asia Disaster Studies

USEFULNESS OF SOME DISASTER ANTICIPATORY MEASURES OF INDIAN RAILWAYS IN DISASTER


MANAGEMENT PLAN OF MEGA CITIES OF INDIA -WITH SPECIAL REFERENCE TO AHMEDABAD CITY
Training pro-

As a preventive action, various

As a preventive action, planning for re-

grammes related

training programmes are being

fresher training for the staff attached with

to DM

run for the staff and officers who

DM is made in DM plan

are attached with DM; for example, refresher training, periodical training, periodical medical
examination
Crisis management

Crisis Management Plan is

There is no crisis management plan to han-

plan to handle

prepared with the coordination

dle terrorist attacks

terrorist attack and

of National Crisis Management

hijacking of train

Committee (NCMC) and Crisis

situation

Management Group (CMG) at


the Railway Board level and at
the zonal level. Quick Reaction
Teams (QRTs) of RPF personnel
should be available round the
clock at bigger stations, which
will be of immense help to tackle
such situations during initial
phases, especially, in cases of terrorist attacks

Creation of Disas-

Railway organisation has created

ter Response and

disaster response and mitigation

There is no specific fund allotted for DM

Mitigation funds

funds

Strengthening of

Strengthening works of the

To reduce destruction due to earthquakes

infrastructure to

Railways infrastructure are carried

and floods, strengthening of important

reduce destruction

out to reduce destruction due to

buildings are essential but this type of pro-

due to earthquakes,

earthquakes, floods, etc.

gramme is not in existence.

floods, etc.

Conclusion
The Indian Railways has well managed DM system. In handling disasters, it is in a unique
position as it has a number of strengths not available with many other departments of

Vol 3 No 1 June 2010 u

45

J.G. MACWAN
Government of IndiaUniformed force of RPF/RPSF, own medical infrastructure, control
network on each division linked with each station, own communication network, an
army of gangmen spread everywhere in the Railways, civil defense and territorial army
wing , well-planned inspection network from bottom to top level, and full-fledged safety
department with full time safety officer.
On the other hand, the Ahmedabad Municipal Corporation or any civic body has very
limited infrastructure to fight against disaster. The DM plan prepared by the Ahmedabad
Municipal Corporation is only meant for natural and some accident-related disasters like
earthquakes, cyclones, floods, industrial/transport-related accident and fire but for other
disasters like terrorist attacks, civil disorders, epidemic diseases, its DM plan is silent because AMC is fully dependent on state and other authorities to handle these disasters.
Moreover, the present DM plan is also not sufficient to handle natural disasters because
detailed duty list of each department and officers at the time of disaster are not prepared.
The city DM plan shows brief actions after occurrence of disaster but preventive actions
are absent. Preventive actions become effective only through well-planned inspection
network; in the Ahmedabad Municipal Corporation, this point is very weak. Overall city
DM plan is prepared but Separate DM plan for each disaster as well as hospital DM plan
to handle disasters is not prepared. Moreover, to handle any disaster, work force and essential staffs are required but as workforce and even nodal officers are not bound to stay
in concerned ward at the time of disaster, they are unable to reach to their duty station.
The only positive aspect of the Ahmedabad Municipal Corporation is its well-trained fire
brigade staff on which it is fully dependable. Hence, important points of the DM plan of
the Indian Railways should be adopted in the city DM plan to get effective results at the
time of disasters.

Reference:
City Disaster Management plan Ahmedabad city.
Text book no. 15 Technological Disasters (Master of Disaster Mitigation): priya Ranjan Trivedi
Institute of Postgraduate Environmental Education & Research, New Delhi.

Internet resources:
Disaster Management plan of the Indian Railways.
Management in India-A status Report by NDMA.
Gujarat State Disaster Management policy by GSDMA.

46 u Journal of South Asia Disaster Studies

Algorithm for Information Retrieval of Earthquake


Occurrence from Foreshock Analysis using Random
Forest Implementation in Earthquake Database
Creation And Analysis: A Machine Learning Approach
Pushan Kumar Dutta, Mrinal Kanti Naskar, Om Prakash Mishra,
and Kajal Mukherjee

Abstract
The task of modelling the distribution of a large number of earthquake events with frequent tremors detected prior to a main shock presents us unique challenges to model a robust classifier tool
for earthquake catalog analysis in different seismic environments. The proposed study evaluates
a machine learning approach to extract attribute-value pairs from earthquake catalog through
event tagging in order to augment databases by representing each precursory foreshock event in a
relational database framework. We have designed using SQLite, a relational database for running a geophysical modeling application after connecting database record of all clusters of foreshock events from (1998-2010) for a complete catalog of seismicity analysis for the Himalayan
basin by Nath et al,2010. In this study, a proposed method of machine learning based database
processing output system (MOPDO) has been made to observe the occurrence of earthquakes
preceded by seismic premonitory patterns as foreshock events occurring prior to mainshocks in
the last 12 years for 5-7 magnitude earthquakes occurring in the Himalayan region. The proposed study has conducted a comprehensive and coherent evaluation through extracting information for a particular set of geographical co-ordinates leading to the spatial discovery of future
earthquake having prior foreshock.This paper proposes a framework for extracting, classifying,
analysing, and presenting semi-structured catalog data sources. The database framework has
been proposed to extract relevant information from different data sources through a classification
based machine learning approach. Among all the decision classifier based models for analysis
of foreshock frequency and subsequent earthquake occurrence, the random forest with bagging
algorithm has an accuracy values of 98.816 per cent correctly classified instances.

Vol 3 No 1 June 2010 u

47

PUSHAN KUMAR DUTTA, MRINAL KANTI NASKAR, OM PRAKASH MISHRA AND KAJAL MUKHERJEE
Keywords: Catalog, Foreshock, Event tagging, Information extraction, Machine learning, Databases

Introduction
Detecting patterns through text retrieval and analysis is a central part of natural language
processing. Words ending -shocks tend to be earthquakes; observable patterns like text
structure and text frequency happen to correlate with particular aspects of occurrence. The
issue of correlating, integrating and presenting related information to issue early warning to impending natural disasters has become a burning issue in the present millennia.
A multitude of search model framework has been used to seek specific information about
past disaster and their occurrence history generating results which are certain event based
tags about the availability of the desired information. This type of query-based search
can be used to develop a system based on machine learning methods that improves the
acquisition of natural disaster data like earthquake. The study of earthquake occurrencesinvolves analytical insight into the statistical properties of earthquake occurrence extracted from larger database and then performing a series of complete rigorous analytical
classification techniques to get a solution to the future occurrence of earthquake in the
region.Classification of different tremors as foreshock-mainshock-aftershock label in a
database framework as done by Valero et al 2009 involves the task of choosing the correct
class label for a given input. Earthquake data is generally comprised of an ever increasing collection of earth science information catalogs for post-processing analysis available
in spreadsheets and database formats. Analysing seismic activity data by studying the
likelihood of pre-seismic patternsfor foreshock occurrence leads to explanations of the
nature of a future earthquake event. The spatial distribution of foreshocks is predicted to
migrate toward the mainshock occurrence with time by the mechanism of a cascade of
seismic triggering leading to a succession of failures in the tectonic environment (Hough
and Jones,1997). Among proposed anomalous properties are proportion forseismic occurrence of normal to large, medium versus small foreshocks occurrence patterns prior
to earthquake occurrence. In this proposed study, we apply a set of rigorous analytic tests
to the earthquake catalogue to search for complex co-relational seismicity patterns for
foreshock-main shock occurrence in a spatio-temporal location for the Himalayan basin.
For the first time, using a set of rigorous text analytic and machine learning tool like weka
(Witten and Frank, 2005), we have tried to interpret the nature of the foreshock patterns
for an earthquake event to analyse by classifying the earthquakes for a certain patterns
of foreshock occurrence. The components of the proposed database framework include

48 u Journal of South Asia Disaster Studies

ALGORITHM FOR INFORMATION RETRIEVAL OF EARTHQUAKE OCCURRENCE FROMFORESHOCK ANALYSIS USING RANDOM
FOREST IMPLEMENTATION IN EARTHQUAKE DATABASE CREATION AND ANALYSIS: A MACHINE LEARNING APPROACH

Query Interface (QI), which is used to accept users queries and search catalogs through
users queries; Information Extraction (IE), which is used to extract and classify the data
sets obtained from QI and convert the extracted and classified event tags in classical form
and named Entity Event Tag Recognition Analyzer (ETRA). This ETRA is a classifier that
is used to determine the relevant information extracted from IE.The system applies a traditional machine learning method (check wekadocs) with traditional decision tree techniques studied by Moore;it differs from other previous IE systems asit does not depend on
sophisticated resources but event tag based techniques for natural language processing.
In particular, it only uses lexical features and avoids the usage of complex syntactic attributes. This is precisely the subject of the present work through enhancement of the acquisition process of natural disaster data. The approach of the present problem has been
divided into sections based on study; the next section describes the most recent related
work organised by idea. In section 3, identification of particular features of foreshock data
that are salient to classification and general design of the database framework for classifying the catalog data has been proposed. In Section 4, we give a brief overview of the
robust natural language processing algorithm for a machine learning oriented database
processing output system MOPDO. Finally, result analysis and directions for future work
through development of a relational database framework is presented in Section 6.

Related work
In the present era,quite a few procedures have been made as learning algorithms for
natural language based processing usually with limited success. In particular, representations based on phrases(Dumais et al 1998; Fuernkranz et al 2000), named entities (Kumaran and Allan, 2004), and the term clustering (Lewisand Croft, 1990; Bekkerman,
2003) has been explored. Content-based recommender systems (Burke, 2009) are classifier systems derived from machine learning research has been used in catalog analysis and
text based search queries.Queries that return large numbers of rows are highly inefficient,
and each retrieved entity must be allocated the proper address in the heap. Typically,
good classifiers cannot be learned until the user has rated many items. In this analysis, we
have applied a number of classification techniques for study (Dimov, 2007) using weka
tool to the same dataset for text-based study.

Foreshock Data Representation and Modeling


Sometimes, all the energy to be released prior to an earthquake gets released in trains of
smaller tremors before large earthquakes by continuing collapse and slippage along the

Vol 3 No 1 June 2010 u

49

PUSHAN KUMAR DUTTA, MRINAL KANTI NASKAR, OM PRAKASH MISHRA AND KAJAL MUKHERJEE
fracture as the fault shift is preceded by the small structural failures we detect as foreshocks. An elaborate study for the global rates of foreshock occurrence involving shallow
M S 6 and M S 7 mainshocks and M S 5 foreshocks had been done by Reasenberg,2001 to
reveal the process of seismogenesis through differences in foreshock rate found among
subsets of earthquakes defined by their focal mechanism and tectonic region (Abercrombie and Mor, 1996) for a region. In the proposed study based on previously recorded
data, input dataset foreshock.arff has been created with four attributes measuring the
spatial coordinates for the earthquake events and its magnitude, and a nominal attribute
for the rate of foreshock occurrence detected for events in the range from 75to 960E
longitude and from 260 to 320N latitude. By stacking many foreshock sequences, a welldefined acceleration of the seismicity preceding mainshocks emerges (Helmstetter, 2003;
Helmstetter and Sornette, 2002). The objectives of this work done are three-fold. First,
we perform an analysis of a 12-year archive of earthquake data in order to extract nature
of foreshock patterns prior to an earthquake event by studying the frequency of tremors
that are not confirmed by analysts as mainshock occurrence. These are foreshocks whose
frequency measures the likelihood of occurrence of catastrophic phenomena.
Figure 1: Pairs of feature sets and labels fed into the machine-learning algorithm to

generate a model during training. During prediction, the same feature extractor is used to
convert unseen inputs to feature sets

Second, we propose the application of supervised Machine Learning (ML) to train

50 u Journal of South Asia Disaster Studies

ALGORITHM FOR INFORMATION RETRIEVAL OF EARTHQUAKE OCCURRENCE FROMFORESHOCK ANALYSIS USING RANDOM
FOREST IMPLEMENTATION IN EARTHQUAKE DATABASE CREATION AND ANALYSIS: A MACHINE LEARNING APPROACH

predictive models on data from tagged event libraries within the dataset. We show that
these models are capable of identifying false events produced by the automatic system,
which can be automatically screened out without manual analyst review. We give several
examples of machine learning methods applied to this problem, ranging from simple
methods yielding highly interpretable models, to more elaborate methods whose models
may be more difficult to interpret.

Proposed Information Extraction (IE) System


Pre-processing Data
The SQLite is a C library that provides a lightweight disk-based database that doesnt require a separate server process and allows accessing the database (check reference). SQLite
provides a compact relational database framework working with a single executable file
supporting cross-platform, single database file system having variable length records. In
the proposed work, SQLiteis implemented with python involving a series of procedural
tasks like querying the database for frequency of foreshock analysis and finally create the
output module for information retrieval.There are two tasks that are essential for building an information retrieval system (Bouckaert, 2002; Hobbs, 1992) first by detecting the
set of information units that will be extracted; and then to find the information patterns
for the system. A search and information retrieval system involves automatic discovery
of the extraction patterns (Muslea, 1999; Peng, 1999; Stevenson and Greenwood, 2006;
Turno, 2003). In particular, modern IE approaches are supported on machine leaning
(ML) techniques (Ireson et al 2005) involving Text Categorisation (TC) and Information
Extraction (IE), respectively. The objective of TC (Jackson and Moulinier, 2007) is to automatically determine the category (or topic) of a document within a previously defined
set of possible categories; whereas, on the other hand, the goal of an IE system (Moens,
2006) is to identify and extract facts from natural language.
In this paper, we present the architecture and evaluation of MODPO, a machine learning oriented database processing output system that automatically extracts information
from earthquake catalog and offline data for offline information retrieval system. This
system is entirely based on a machine learning approach and its architecture consists of
a set of components that, firstly, identify the texts related to natural disasters based on
the approach by Bird et al and subsequently, extract the relevant data for populating a relational database. The evaluation shows its effectiveness for detecting the relevant documents about natural disasters (reaching an F-measure of 98 per cent), and for extracting

Vol 3 No 1 June 2010 u

51

PUSHAN KUMAR DUTTA, MRINAL KANTI NASKAR, OM PRAKASH MISHRA AND KAJAL MUKHERJEE
relevant facts to be inserted into a given database (reaching an F-measure of 76 per cent).
It is then a decision learner to train the metadata has been applied in this study. A step by
step analysis for information search and retrieval has been proposed.
Proposed robust natural language processing based algorithm
1. First step in creating a classifier is deciding what features of the input are relevant, and
how to encode those features involving named entity recognition of the word -shock.
2. Create SQLite3 database with python for creating relational database.
3. Creating a classifier to find features of the input are relevant, and to encode those features and corresponding class labels for applying the proposed processing algorithms.
4. Working with large corpora, constructing a single list that contains the features of every instance (eg. foreshocks in Himalaya basin when main shock occurred in latitude
and longitude) that can use up a large amount of memory; for a catalog description;we
begin by constructing a list of the 50 most frequent words in the overall corpus. We
can then define a feature extractor that simply checks whether each of these words is
present in a given document.
5. Defined our feature extractor, we can use it to train a new decision tree classifier
through a list of examples and corresponding class labels.
6. Capture the dependencies between related classification tasks using joint classifier
models by choosing an appropriate labeling for a collection of related inputs. In the
case of event tagging, a variety of sequence classifier models can be used to jointly
choose event-based tags for all the similar records in a given sequence.
7. Subdivide the errors through modeling the linguistic data found in corpora can help
us to understand linguistic patterns, and can be used to make predictions about new
language data for classification of -fore, -main and after.
8. Applying supervised classifiers use labeled training corpora to build models that predict
the label of an input based on specific features of that input. Look for exact match, overlapping, and mutually disjoint for set of tokens saved in the array and Attr (DB). Extract
set of tokens by matching with Sub_Attr (DB) and (ii) extract G_Sub (WP) by matching
with each G_Sub (DB). Identify the index number of Attr (DB) that is matched and
group the extracted attributes and sub-attributes based on the index number.
9. Identify the index number of Attr (DB) that is matched through information extraction for theAttr (WP), Sub_Attr (WP), and value of Sub_Attr (WP) and later identifies
the index of Attr (DB) that is matched.
10. Based on the index number we apply the measured foreshock frequency as high, low,

52 u Journal of South Asia Disaster Studies

ALGORITHM FOR INFORMATION RETRIEVAL OF EARTHQUAKE OCCURRENCE FROMFORESHOCK ANALYSIS USING RANDOM
FOREST IMPLEMENTATION IN EARTHQUAKE DATABASE CREATION AND ANALYSIS: A MACHINE LEARNING APPROACH

and medium.
11. Train the set of meta-learners using weka model using 10 fold cross validation to
evaluate the best classifier model the decision based algorithm to find degree of similarity between instances in the test set and those in the development set.
12. Apply for training data and checked that the random forest with bagging decision
learner has classified the unknown data set as in Figure 2 and serves as the best classifier model as evaluated from comparative results as in Table 1.

Experimental results
In the experiments performed in this work we used the evaluation technique for 10-fold
cross-validation, which consists of randomly dividing the data into 10 equally-sized subgroups and performing different experiments. We separated one group along with their
original labels as the validation set;another group was considered as the test set; from the
remaining data a random selection had been done. The effectiveness of the decision algorithms by studying a series of metrics reveals that random forest with bagging serves as
the best decision tree classifier that will fit in a relational database framework. In Table 1,
we summarise the results for test data attributes and found that found that random forest
using bagging is the best decision based classifier for foreshock analysis.
Random forest
Table 1: Classification algorithms for test data attributes and the identified metrics.
Classifier

MAE

Kappa

CCI %

RMSE

Precision

TP rate

F measure

Random forest with bagging

.179

0.9816

98.506

.2201

.986

Multi classifier

0.4543

0.2083

62.5

0.4967

0.619

0.625

0.612

Jrip

0.4859

0.0022

57.727

0.5776

0.57

0.523

0.512

Classification and regression tree

0.4247

47.17

0.4608

0.222

0.471

0.302

Classification via regression

0.4965

0.1924

62.5

0.4965

0.623

0.625

0.598

Classification via clustering

0.4205

0.1572

57.954

0.6484

0.589

0.58

0.581

Decision table

0.4732

0.1927

63.634

0.4879

0.675

0.636

0.579

Metamulti classifier

0.4732

0.2083

62.5

0.4767

0.619

0.625

0.612

Rep Tree

0.4835

0.0194

55.6818

0.5205

0.525

0.557

0.493

Simple CART

0.5486

-0.2027

45.455

0.5875

0.35

0.455

0.381

Vol 3 No 1 June 2010 u

53

PUSHAN KUMAR DUTTA, MRINAL KANTI NASKAR, OM PRAKASH MISHRA AND KAJAL MUKHERJEE
Random forests (Breiman and Cutler, 2001) blend elements of random subspaces and bagging in a way that is specific to using decision treesas base classifier. At each node in the
tree, a subset of the available features is randomly selected and the best split available
within those features is selected for that node. Internal estimates, which are obtained using
the out-of-bag examples, can give estimates for the performance on unseen data as well as
estimates for the strength of each tree, and the correlation between trees. The classification
of unseen points is done by voting while bagging is used to create the training set of data
items for each individual tree. The number of features randomly chosen (from n total) at
each node is a parameter of this approach. Each tree gives a vote that indicates the trees
decision about the class of the object. The forest chooses the class with the most votes for
the object. The advantage of random forest over the rest for complete evaluation lies in
accuracy among the current data mining algorithms. Random forest, with Meta bagging,
outperforms other classification trees, as it is computationally effective and offer good prediction performance due to restriction towards over-fitting components compared to other
classification tree models. It runs efficiently on large data sets with many features giving
estimates of what features are important. It has no nominal data problem and does not
over-fit for both balanced and unbalanced data sets. In random forests, there is no need for
cross-validation or a test set to get an unbiased estimate of the test error. Since each tree is
constructed using the bootstrap sample, approximately one-third of the cases are left out
of the bootstrap samples and not used in training. These cases are called out-of-bag (oob)
cases. These oob cases are used to get a run-time unbiased estimate of the classification
error as trees are added to the forest.The dataset for foreshock classification may belong to
multiple categories (i.e. the multi-label problem) there may be possible errors in the manually generated labels of the training sets (i.e., categories) of future tremors likelihood, which
can impact the performance of learning algorithms.Random inputs and random features
produce good results in classification- less so in regression. For larger data sets, we can gain
accuracy by combining random features with bagging.
Bagging
The decision tree for identifying the class attribute for the foreshock occurrence is classified using a voting algorithm. If we have an infinite number of independent training
sets, test instance can be classified and a single answer determined for the majority vote
by means of a bias variance decomposition. Bias is the mean square error expected when
averaging over models built from all possible training sets of the fixed size and variance
is the expected error of single model built from particular training data. The above clas-

54 u Journal of South Asia Disaster Studies

ALGORITHM FOR INFORMATION RETRIEVAL OF EARTHQUAKE OCCURRENCE FROMFORESHOCK ANALYSIS USING RANDOM
FOREST IMPLEMENTATION IN EARTHQUAKE DATABASE CREATION AND ANALYSIS: A MACHINE LEARNING APPROACH

sification method generates an ensemble of classifiers, thus, giving excellent results.


Metrics
Correctly Classified Instances
Incorrectly Classified Instances
Kappa statistic
Mean absolute error
Root mean squared error

86

98.8506 %

1.1494 %

0.9819
0.179
0.2201

Relative absolute error

42.0817 %

Root relative squared error

47.7606 %

Total Number of Instances

87

To measure the performance of a classification method we use four metrics: precision,


recall, aggregate precision, and F-measure. The precision of an algorithm is the ratio of
True Positives over the sum of True Positives and False Positives. Recall is the ratio of True
Positives over the sum of True Positives and False Negatives, or the percentage of flows
in an application class that are correctly identified. Aggregate precision is the ratio of the
sum of all True Positives to the sum of all the True Positives and False Positives for all
classes. The two former metrics to evaluate the quality of classification results for each application class, and the latter metric to characterise the overall accuracy of a classifier on
the whole trace set. Finally, F-Measure combines precision and recall into a single metric
for test accuracy by taking their harmonic mean; 2precisionrecall/(precision + recall).
We use this metric to compare and rank per-application performance of machine learning algorithms (Table 1) included in weka.The error levels when applying the classifier
to the training data gives error levels during a 10-fold cross-validation. For our purposes
the most important figures here are the numbers of correctly and incorrectly classified
instances.

Results and futuredirections


This paper presented some ideas for enhancing the acquisition process of natural disaster
based foreshock data. In particular, a system is proposed that automatically populates a
natural disaster database by extracting information from offline catalog data and information retrieval based system. Better results can be achieved by use of new and enhanced
features, and especially by the use of cost-sensitive learning to bias the categoriser towards
lowering the false positive rate at the expense of the false negative rate for a classification

Vol 3 No 1 June 2010 u

55

PUSHAN KUMAR DUTTA, MRINAL KANTI NASKAR, OM PRAKASH MISHRA AND KAJAL MUKHERJEE
tree algorithm reflect actual monitoring system goals. Future work will fall under three
main areasidentification of additional features to help inform categorisation, improved
categorisation algorithms suited for specific task and events, and the use of cost-sensitive
learning to help improve results. The need for a real time dynamic resource description
framework is needed to implement our classification tree. The sharing of virtual servers
updated real time can be used to tag events across different data sets. Machine learning
algorithm can be used to give causal relationships between those features and patterns in
natural language processing can test for statistical inference(Figure 3).
Figure2:Dynamic resource sharing framework for event tagging based on machine

learning approach.

This kind of integrated resource development framework can be helpful in tracking


down many the most essential target for a natural dissipative system.The proposed system has two main limitations, which is should be solved. On one hand, it does not correctly extract information from documents describing more than one disaster, and on the
other hand, it does not allow integrating the data extracted from different documents
related to the same disaster. Our future work will be mainly focused on the solution of
these two inconveniences. In addition, it will consider the collection of a bigger training
set and the construction of a set of binary classifiers, one for each kind of desired data.

56 u Journal of South Asia Disaster Studies

ALGORITHM FOR INFORMATION RETRIEVAL OF EARTHQUAKE OCCURRENCE FROMFORESHOCK ANALYSIS USING RANDOM
FOREST IMPLEMENTATION IN EARTHQUAKE DATABASE CREATION AND ANALYSIS: A MACHINE LEARNING APPROACH

References
Abercrombie R.E. and Mor J. , Occurrence Patterns of Foreshocks to Large Earthquakes in the Western
United States, Nature 381, 303- 307; DOI:10.1038/381303a0, 1996.
Bekkerman R., Distributional Clustering of Words for Text Categorization. Masters thesis, Technion, 2003
Bird S., Klein E. and Loper E., Natural Language Processing with Python, (London: Oreilly Publishers,
2009). <http://www.nltk.org/code>, Oreilly Publishers,London.(Viewed on June 07, 2011).
Bouckaert R., Low level information extraction. In Proceedings of the Workshop on Text Learning
(TextML-2002), Sydney, Australia.
Breiman L. and Cutler A., Random forests.Machine Learning, 2001.www.stat.berkeley.edu/~breiman/
RandomForests/(Viewed on March 27,2011).
Burke R R., Hybrid Web Recommender Systems,In P. Brusilovsky, A. Kosba,W. Nejdl,eds.:The Adaptive
Web:Methods and Strategies of Web Personalization.vol.4321 of Lecture Notes in Comp Science.
Springer-Verlag Heidelberg, New York, 2007.
Dimov R., Weka: Practical Machine Learning Tools and Techniques.<www.dfki.de/~kipp/seminar_
ws0607/slides/Dimov_WEKA.pdf>(Viewed on April 28,201)
Dumais S., Platt J., Heckerman D. and Sahami M., Inductive learning algorithms and representations
for text categorization. In CIKM1998.
Fuernkranz J., Mitchell T. and Riloff E., A Case Study in using Linguistic Phrases for Text Categorization
on the WWW. Learning for Text Categorization,AAAI Press.
Helmstetter A., Is Earthquake Triggering Driven by Small Earthquakes? Physical Review Letters, vol. 91,
no. 5, DOI: 10.1103/PhysRevLett.91.058501.
Helmstetterand

A.,

Sornette

D.,

Foreshocks

and

Earthquake

Predictability,

2002.<arXiv:physics/0210130v1 [physics.geo-ph] > (Viewed on May 17,2011)


Hobbs J.R., The Generic Information Extraction System. In B. Sundheim, editor. Fourth Message
Understanding Conference (MUC-4), Mc Lean, Virginia Distributed by Morgan KauffmanPublishers,
Inc., San Mateo, California, 1992.
Hough S.E. and Jones L.M., EOS, TransactionsAmerican Geophysical Union 78,N45, 1997, 505-508.
Ireson N., Ciravega F., Califf M.E., Freitag D., Kushmerick N. and Labelli A., Evaluating Machine
Learning for Information Extraction, In Proceedings of the 22nd International Conference on
Machine Learning, Bonn, Germany, 2005.
Jackson P. and Moulinier I., Natural Language Processing for Online Applications: Text Retrieval,
Extraction and Categorization,John Benjamins Publishing Co, second edition, 2007.
Kasper R.T. and Rounds W.C.A., Logical Semantics for Feature Structures,in Proceedings of the 24th
Annual Meeting on Association for Computational Linguistics, pp.257266.

Vol 3 No 1 June 2010 u

57

PUSHAN KUMAR DUTTA, MRINAL KANTI NASKAR, OM PRAKASH MISHRA AND KAJAL MUKHERJEE
Kumaran G. and Allan J., Text classification and named entities for new event detection,SIGIR 2004.
Lewis D. and Croft W., Term clustering of syntactic phrases,SIGIR1990.
Moore A.W., Statistical Data Mining Tutorialsslides, March 07, 2011.<http://www.autonlab.org/
tutorials/>viewed
Muslea I., Extraction Patterns for Information Extractions Tasks: A Survey,In Proceedings of the
AAAI Workshop on Machine Learning for Information Extraction,Orlando, Florida, 1999.
Nath S.K., Thingbaijam K.K.S. and Ghosh S.K., Earthquake Catalogue of South Asia - a Generic MW
Scale Framework, 2010.
Peng F., Models Development in IE Tasks - A survey, CS685 (Intelligent Computer Interface) course
project, Computer Science Department, University of Waterloo, 1999.
Reasenberg P.A., Foreshock Occurrence Rates before Large Earthquakes Worldwide, vol.155, no.2-4,
355-379, DOI: 10.1007/s000240050269.
Stevenson M. and Greenwood M.A., Comparing Information Extraction Pattern Models, In
Proceedings of the Workshop on Information Extraction Beyond The Document, Association for
Computational Linguistics, Sydney, 2006, pp.12-19.
Turno J.,

Information Extraction, Multilinguality and Portability. RevistaIberoamericana de

Inteligencia Artificial, no. 22, 2003, pp. 57-78.


Valero T., Gmez M.M., Pineda L.V., and Penas A., Learning to Select the Correct Answer in Multi-Stream
Question Answering, Information Processing and Management, viewed June 07,2011 <www.
sciencedirect.com/science/ article/pii/S0306457310000270>doi.org/10.1016/j.ipm.2010.03.007
Weka Documentation, http://wekadocs.com/node/13, accessed on 21 March 2010.
Witten I., H. and Frank E., Data Mining: Practical Machine Learning Tools and Techniques. 2nd
Edn., Morgan Kaufmann, San Francisco, 2005,ISBN: 0-12-088407-0, pp: 560.
How is sqlite different<http://www.sqlite.org/different.html>accessed 15th July,2011.

58 u Journal of South Asia Disaster Studies

A Concept of a Low-cost
Construction Method for Deep
Underground Dams
Yoshito Kobayashi and Vijay Kumar Jatiya

Abstract
This paper proposes a concept of a new construction method, TEIM, for deep underground dams
with low cost and without using heavy machines. The TEIM consists of three basic processes and
the finish processthe first process makes a working tunnel of small cross section for manual
work, along the full length of the bottom of the planned underground dam wall; the second process is the combination of excavating with a little depth at the roof and infilling with the filler,
which includes low permeable dam wall materials such as fine clay, with the same depth at the
bottom of the working tunnel, and the third process, where the layers of the fillers at the current
step and at previous step are unified by ramming, repeats the second process to make the dam
wall little by little from its bottom to the top while maintaining the original small cross section
of the working tunnel. The final process backfills the working tunnel with soil.

Introduction
There are huge semi-arid regions in the world such as in sub-Saharan Africa, in western
Asia, in northern China, in northeastern Brazil, etc. These semi-arid regions, where huge
amount of water runs off uselessly in rainy season (Jatiya, 2002), can be transformed into
farmland for year-round agriculture by the application of the appropriate water supply
and drainage system, in consideration of desalinisation in case of necessity.
Well, qanat, and underground dams have been available as the utilisation technologies of groundwater resource. Wells can supply small quantity of water flow through the
wholewall, which have not wide area for seeping of groundwater out of the earth in aqui-

Journal of South Asia Disaster Studies u

59

YOSHITO KOBAYASHI AND VIJAY KUMAR JATIYA


fer. Drawing too much groundwater from the wells may induce to lower the groundwater
level or to run dry the well. Qanatshavethe maximum water flow, which limits the water
flow velocity to prevent the erosion at the water-channel bed.
Underground dams can store water in the rainy season, when much water runs off
uselessly, for using in the dry season in the seasonal river area. Underground dams can
provide the suitable groundwater level, at around 3 to 7 m depth in the broad stretch of
reservoir area of upstream side of the underground dam wall, so that farmers can easily
take groundwater with hand pump from shallow wells and so that the soil surface may
not be salinised with the capillary action of groundwater.
Underground dams have the advantages below by comparison with water sources
such as the ground dam and the well (Ministry of the Environment, Japan and Overseas
Environmental Cooperation Center, Japan, 2004).
(1) No accompanying submerging area
(2) Little evapotranspiration
(3) Hygienic water
(4) No danger of collapsing
(5) No possibility of depletion of groundwater
Several deep underground dams have been constructed in Japan with using heavy
machine (Nishigaki et al 2004) of auger type drilling machine. The application of heavy
excavation machines such as the chainsaw type trench cutting machine, the bucket type
excavation machine as well as auger type drilling machine for construction of deep underground dams may be less feasible in the above-mentioned semi-arid regions from a
cost-benefit standpoint.
Many underground dams of shallow wall less than 10 m deep have been constructed
by open cut excavation method with manual excavation in the world, that are around
500 units in Brazil (Foster and Tuinhof, 2004) and Kenya (Foster and Tuinhof, 2004),
several units in India (Nishigaki et al 2004; Nilsson, 1988), Tanzania (Nilsson, 1988) and
Iran (Salih, 2006; Yazdani et al 2006), and an unknown number of units in Italy (Nilsson,
1988), France (Nilsson, 1988), Bolivia (Nilsson, 1988), Germany (Nilsson, 1988), Burkina Faso (Ministry of the Environment, Japan and Overseas Environmental Cooperation
Center, Japan, 2004; Nishigaki et al 2004), Taiwan (Ting, 2009) , etc.
The open cut excavation method, without excessive expense for ground support work,
might have limited the depth of the underground dam wall up to 10 m.

60 u Journal of South Asia Disaster Studies

A CONCEPT OF A LOW-COST CONSTRUCTION METHOD FOR DEEP UNDERGROUND DAMS

Figure 1: The status of the construction cost of underground dam and the target
cost of the present method

Figure 1 shows the construction costs of water cut off wall for underground dams
per unit area in relation to their wall depths (ASAL Consultant; Nilsson, 1988; Okinawa
General Bureau). The existing underground dams can be divided into two categories of
the shallow and low-cost type without using heavy machines and the deep and high-cost
type with using heavy machines.
Vast seasonal river areas, which have the impermeable layer deeper than 10 m, have
been left uncultivated in the world.
The aim of this paper is to propose a newly invented method, TEIM (Kobayashi, 2008,
2009), which enables to construct deep underground dams without requiring heavy machines and high-level technologies with low cost in semi-arid regions inremote areas with
using the combination of ancient and novel technologies. The target cost area of the
present method is shown in Figure 1.

Processes of the Traverse Excavating and Infilling Method: TEIM


The features of TEIM different from the existing methods for construction of water cut
off wall in underground are:
(1) Non-open cut excavation method
(2) Upward excavation in the main process

Vol 3 No 1 June 2010 u

61

YOSHITO KOBAYASHI AND VIJAY KUMAR JATIYA


(3) Minimisation of ground support work
(4) Utilisation of a part of excavated soil for the dam wall material without carrying
away
(5) Applicable both to manual-excavation and machine-excavation

The TEIM consists of three basic processes and the finish process:
1st Process: A working tunnel of small cross section is excavated along the bottom of
the planned dam wall site.
2nd Process: The combination of the excavating at the roof of the working tunnel with
a little depth and the infilling with the dam wall material at the bottom with the same
depth.
3rd Process: The repetition of the second process to make the dam wall little by little
from the bottom to the top while maintaining the original small cross section of the
working tunnel.
Finish Process: The working tunnel is back filled with soil after completion of dam wall.
Figure 2 show construction with process of TEIM for an underground dam in longitude-sectional view. Figure 2(a) shows the underground dam wall to be constructed
in longitude-sectional view. Vertical shafts are dug at first into impermeable layer at
appropriate intervals at full length of the planned underground dam wall site. Then
approximately horizontal working tunnel, which has the width of around 0.8 m and
the height of around 1.2 to 1.5 m for manual excavation, is excavated through the
bottom points of the vertical shafts along the full length of the bottom of the planned
underground dam wall as shown in Figure 2(b). Vertical shafts are utilised for test
boring to characterise underground geology and hydrology, for carrying away of the
excavated surplus soil, for ventilation, for drainage, etc. Figure 2 (c), (d) show the
second process which comprises the combination of the excavating at the roof of the
working tunnel with a little depth, around 0.1 to 0.2 m, and the infilling with the
filler including the dam wall material at the bottom with the same depth. Figure 2 (e),
(f), (g) show the third process. Figure 2 (g) shows the last stage of the third process,
where the top surface of the filler has reached at the depth equal to the top of the
plannedunderground dam wall. Figure 2 (h), (i) show the finish process, where the
working tunnel is backfilled and a part of vertical shafts are utilised as water supply
wells without backfilling.

62 u Journal of South Asia Disaster Studies

A CONCEPT OF A LOW-COST CONSTRUCTION METHOD FOR DEEP UNDERGROUND DAMS

Figure 2: Construction process of TEIM for an underground dam in longitude e-sectional view

Vol 3 No 1 June 2010 u

63

YOSHITO KOBAYASHI AND VIJAY KUMAR JATIYA

Figure 3: Construction process of TEIM for an underground dam in cross sectional view

Figure 4: Construction process for the multiple-layer structure of underground dam wall

Figure 3 shows the construction processes with TEIM in cross-sectional view in the
case of simple underground dam wall having a single-phase layer, such as compacted fine
clay, with cement when it is affordable. The size and shape of the working tunnel in the
TEIM for manual excavation are chosen similar to those of the qanat tunnel which have

64 u Journal of South Asia Disaster Studies

A CONCEPT OF A LOW-COST CONSTRUCTION METHOD FOR DEEP UNDERGROUND DAMS


been designed after long time experience. The roof excavation depth in one-step in the
second and third processes is chosen for the effective compaction thickness in the filler at
the bottom. The equalisation of depth in one-step between in the roof excavation and in
the bottom infilling in the second and third processes induces to keep the original small
cross-section of the working tunnel up to the final stage of the third process.
The construction work in the underground below the groundwater level needs
additional consideration, such as the application of the inclined working tunnel and
water pumping facility, to drain off water from the working tunnel. Figure 4 shows the
examples of the construction processes for the multi-layer structure of underground
dam wall. Figures 4 (a), (b), and (c) show the parts of the construction processes for
the low permeable layer at the center potion and the high permeable layer at the side
portions to drain water. Figure 4 (b) shows the shuttering plates for infilling separately
with fine clay at the center portion and with pebble, boulder, etc. at the side portions
of the working tunnel. The fine clay at the center portion is compacted with a proper
percentage of water to reduce the void ratio. A careful design examination on the
particle size, void ratio and thickness of the center portion as well as the hydrostatic
pressure gradient in the low permeable layer in consideration of the Kozeny-Carman
Formula (Carrier, 2003), etc. is necessary to meet the requirement of impermeability
of the dam wall.
The layers of larger particles of pebble, etc. at the side portions take part in drains.
Figure 4 (c) shows the completed underground dam wall with buffer layers. In this case,
pebble and boulder, etc. are separated by sieving from the excavated soil at the roof of the
working tunnel and are utilised as for buffer layers material without carrying out from
the working tunnel.
Figure 4 (d), (e) show the construction process of an underground dam with using
impermeable plastic sheet, which has been applied in Brazil (Foster and Tuinhof, 2004),
with TEIM in cross-sectional view. The impermeable plastic sheet is placed between the
sheet protecting layers of clay and sand in the central portion to separate it from sharp
stones in the buffer layers in the working tunnel. In this case, the total cost for the construction should be estimated concerning the increase in expense for the application
of the impermeable plastic sheet and the decrease in expense for the reduction of the
amount of soil to be carried out and of the amount of underground dam wall materials
to be carried in. As impermeable barrier stainless-steel sheet, bricks, cement, etc. are also
available in the TEIM.

Vol 3 No 1 June 2010 u

65

YOSHITO KOBAYASHI AND VIJAY KUMAR JATIYA

The Combination of the TEIM and Other Technologies


(1) Permeability of Dam Wall
In Figure 4, centre layer made of fine clay has to be low permeability for dam wall. The
permeability k of porous media like sand layer depends on the diameter D of the particles
and on the void ratio e of the soil as shown in the following Kozeny-Carman formula
(Kozeny, 1927; Carman, 1938; Carman, 1956; Carrier, 2003) modified by W. D. Carrier
together with Hazen formula (Hazen, 1892; 1911) as shown in Figure 6.

Figure 6: The permeability of soil in relation to its particle

66 u Journal of South Asia Disaster Studies

A CONCEPT OF A LOW-COST CONSTRUCTION METHOD FOR DEEP UNDERGROUND DAMS


CH: Hazen empirical coefficient
D10: particle size for which 10 per centof the soil is finer (cm)
Fine clay, which is applicable to the dam wall of low permeability, of the diameter
below 1 micrometer can be supplied with separation, with using the technologies for
making porcelain clay, of particles by size of waste soil discharged during the third process in the working tunnel. The remaining coarse sand and gravel after the separation are
applicable to the material for the buffer layers or for drain having high permeability as
shown in Figure 4.
(2) Carbonation Cementation
The strength of dam wall made of fine clay can be improved with compacting the fine
clay with the suitable rate of water, lime and sometimes oil, with the carbonation cementation among soil particles such as the seawall made with Choushichi-tataki technology
(Tanaka, 2004), artificial rock, in Japan in around 1900 and earthen walls, the predecessors of the Great Wall, made with the Ban-zhu technology, rammed earth, in China in
2000 BC (Onitsuka, 2002). The soil grouting technology, which have been recently available, with mixing portland cement is also applicable to the TEIM.
(3) Qanat
Many qanats are drying up in the world due to decreasing of groundwater level. The
additional construction of an underground dam with TEIM at the upstream side of the
mother well can renovate the drying up qanat as shown in Figure 7. The TEIM can construct the underground dam from the bottom to the top with the similar technology for
qanat with low cost and without using heavy machines.
The existing qanat is utilised for drain of ground water from the working tunnel during the construction period for the underground dam. Fine clay for the underground
dam wall material can be supplied by the excavation at the impermeable layer at the
upstream side of the underground dam wall. The water flow rate of renovated qanat can
be controlled by adjusting the permeability through the orifice packed with coarse sand
or pebble in the underground dam wall.
The additional underground dam may provide farm land at upstream side area of
qanat outlet by supplying the water from the shallow wells at upstream side of the underground dam.

Vol 3 No 1 June 2010 u

67

YOSHITO KOBAYASHI AND VIJAY KUMAR JATIYA

Figure 7: The combination of an underground dam with the TEIM and a qanat

(4) Low-height Check Dams Series


The combination of low-height check dams series, of which method was introduced by
S. Higashi (Takahashi and Higashi, 1987) aiming toward the compatibility with decrease
of flood damage and preservation of fish habitat at Hokkaido in Japan, and underground
dams series with the TEIM can brings the effects to stabilise the quantity of the reserving

68 u Journal of South Asia Disaster Studies

A CONCEPT OF A LOW-COST CONSTRUCTION METHOD FOR DEEP UNDERGROUND DAMS

Figure 8: The combination of low-height check dams series underground dams


series with the TEIM

water in the underground dam and to decrease the saline in the river water, because that
low height check dams series change the river bed configuration from gully, which erode
saline earth with high speed water flow, to flat river bed, where river water flows slowly
without erosion of saline river bed as shown in Figure 8. Low-height check dams having
0.5 to 2 m height can be constructed with piled gabions without applying high technologies with low cost and with the procurable materials in the vicinity.
The multiplicity of low-height check dams, which enable to raise the water level in
the river bed to widen the irrigated farmland with the combination with reservoir ponds,
terraced fields, etc., is effective to lower the total cost in comparison with a single big
dam and to drop the flow speed of river water flow to reduce the damage in flood and
the erosion of soil. The deposition of sediment at the upstream side of the low-height
check dams tends to raise and to flatten the river bed and can increase the quantity of
storage water in the underground dam with adding the water cut off wall at the top of
the underground dam wall with applying the traditional open cut excavation method as
shown in Figure 8(c).
(5) Vertical Desalinisation
In the late rainy season, when the saline of river water decreases and when the land at

Vol 3 No 1 June 2010 u

69

YOSHITO KOBAYASHI AND VIJAY KUMAR JATIYA


the upstream region of the underground dam wall is submerged, the surface region of the
land can be desalinised vertically with drawing groundwater with pumps from shallow
wells or through subsurface drainages.
(6) Directional Drilling
The directional drilling technology (Kourmaev, 2007) is applicable to excavate the working tunnel in the first process of the TEIM, the underground conduit connecting the
farmland at the downstream and the underground dam as shown in Figure 7 and the
subsurface drainage for the vertical desalinisation.

Advantages of the TEIM for the Construction of Underground Dams


The advantages of TEIM for the construction of underground dam wall unlike the conventional open cut excavation methods nor the soil cement underground continuous
wall method using heavy machines are:
(1) Utilising the Traditional Technologies and Experienced Knowledge on
Qanats and the Great Wall
The TEIM uses the working tunnel of unchanged small cross section. Adjusting the size
and shape of the cross-section of the working tunnel in the TEIM similar to those of qanat
enables to utilise the knowledge gained by experience in the construction of over 22,000
qanats having the length up to 70 km (Zarei, 2007) including the water bearing section of
6 km long (Beaumont, 1971), with installing reinforcing rings only in loose soil as shown
in Figure 9 (Wulff, 1968), the depth up to 400 m by hand labor of skilled workers, called
muqanni (Wulff, 1968) in the various geological environments. The TEIM may be able to
construct the underground dams of various configurations in the various geological environments with utilising the knowledge and hand labor technologies in the construction
of qanats, which originated in Iran in 800 BC (Yazdani et al 2006). The qanat technology
spreads in the world including in Iraq and in Afghanistan, as called Karez, in Oman and
UAE, as called Falaj, in India, as called Suranga (Padre, 2006), in Indonesia, in China, in
North Africa, as called Foggara, in Armenia, as called Kahrez, in Italy, in Spain, in Peru, in
Chile, etc. The utilisation of the knowledge on qanat construction shall minimise the work
for ground support in the working tunnel even though careful measure against the cave-in
at the roof of the working tunnel is necessary in using the TEIM because the excavation in
the upward direction in the TEIM remains fresh and loose surface at the roof in the working
tunnel differing from the excavation in the horizontal direction in the qanat.

70 u Journal of South Asia Disaster Studies

A CONCEPT OF A LOW-COST CONSTRUCTION METHOD FOR DEEP UNDERGROUND DAMS

Figure 9: Manual work in the qanat construction

For making the low permeable layer for water cut off wall of the underground dam,
the TEIM uses the traditional earth ramming technology with mixing lime for carbonation cementation, Ban-zhu technology, which originated in China in 2000 BC (Onitsuka,
2002). The technology has proven the long time durability as shown in the Great Wall.
(2) Minimum Ground Support
From the first process to the finish process of the construction work of the TEIM, the
shape and size of the cross-section of the working tunnel approximately keep its original
small ones. The construction work in the tunnel of the smaller cross section allows using
the lighter-duty ground support system because of the shorter span of its ground support
frame structure. The TEIM uses the light-weight strut supporting structure, which is removable and reusable, for ground support in the working tunnel. The cross section of the
working tunnel having the width of 0.8 m and the height of around 1.2 to 1.5 m, which
are similar to that of qanats, are chosen for manual work in the TEIM to minimise the
needs for the ground support and to utilise the abundant knowledge of qanats.
(3) Deep Wall Dam
In the first process as mentioned above, the working tunnel having a small cross-section

Vol 3 No 1 June 2010 u

71

YOSHITO KOBAYASHI AND VIJAY KUMAR JATIYA


similar to that of qanat may be applicable to excavate at deep underground up to 400 m
deep with ancient technologies for qanat and for mining without heavy machines.
(4) Low Energy
Soil tends to fall down with gravity during excavating at the roof of the working tunnel
in the second and third processes in the TEIM. Fine clay separated with sieving from the
excavated soil can be utilized as the materials for low permeable underground dam wall.
Residual sand, pebble, etc. derived from excavated soil can be utilised for buffer layer
without carrying out from the working tunnel. The utilisation of gravity in the roof excavation and the reduction of the amount of soil to be carried out from the working tunnel
reduce the total energy for the construction of the underground dam.
(5) Low Technology
The construction work, which comprises excavating with miners pick, sieving with sieve,
ramming with hand rammer and carrying in and out the materials in the TEIM, is simple
in the way and need not any high technologies neither special tools as shown in Figure
10.
(6) Simultaneous Execution of the Construction Works at Divided Working Faces
Dividing the construction work site into plural short sections, where the construction
works in the second and third processes can be executed simultaneously,

Figure 10: Manual work in the working tunnel with the TEIM for
construction of underground dams

72 u Journal of South Asia Disaster Studies

A CONCEPT OF A LOW-COST CONSTRUCTION METHOD FOR DEEP UNDERGROUND DAMS


it enables to shorten the total construction term. Full length of the working tunnel is
partitioned rationally into short working faces between adjacent vertical shafts.
(7) No Requiring of Heavy Machines
The construction work of underground dams with the TEIM can be carried out with
manual work, which comprise the excavating at the roof, sieving the excavated earth
to separate fine clay for a low-permeable underground dam wall material and ramming
earth of fine clay to compact the filler in the working tunnel as shown in Fig. 10. The
workers are protected from cave-ins by removable shield panel system. Fine clay for low
permeable layer material can be prepared with using the water-sieving technology for
potters clay in advance in the rainy season. The application of the light-weightstopper
drill for mining and the hand held vibration ramming machine and the mini size shovel
loader for road work are effective to shorten the construction term.
(8) Low Cost
Above mentioned features of the TEIM including minimising the ground support work,
low energy, low technology, no requiring of heavy machines as well as the utilisation of
excavated soil for the dam wall material induce to decrease the construction cost.
(9) Multi-layer Wall
The infilling work in the TEIM is basically manual work on site unlike the construction
methods with using heavy machines such as chainsaw, bucket, auger, etc., so the layer
structure of underground dam wall and buffer layer can be composed with freedom. Buffer layers as shown in Fig. 4 equalise the ground water level, which correlates the ground
pressure on the underground dam wall, because of its high permeability. Buffer layers of
pebble and boulder can flatten the concentration of displacement of dam wall causing
by the localisedmovement, such as the action of geological fault, of the adjacent ground
during operational period of the underground dam, because of the stress dispersion effect
like that of the ballast under rail. Thus, buffer layer of high permeability and high stress
dispersion brings the effect to improve the durability of dam wall. The application of
impermeable sheet, such as plastic sheet, reduces the needs for soil sieving work and for
soil carrying out work.
(10) Applicable in Remote Region
The TEIM can use affordable materials such as fine clay in the impermeable layer in the

Vol 3 No 1 June 2010 u

73

YOSHITO KOBAYASHI AND VIJAY KUMAR JATIYA


vicinity of the construction site. So the TEIM is applicable for the construction of underground dams at remote region where any approach road is not provided for carrying
heavy machines, dam wall materials, fuel, etc. into the construction site.
(11) Creation of Employment in Barren Area
The way of the manual work for the TEIM is simple, so the construction of underground
dams with the TEIM can contribute to the creation of employment as the construction
workers for the inhabitants in the vicinity of the construction site, where have been barren and destitute of water, without long-time job training.
(12) Applicable to Various Geologies
The utilisation of the abundant knowledge on qanats and hand mining at underground
may assist to apply the TEIM to various geologies.

Safety Measure for Minimizing the Risk of the Tunnel Caving in


To assure the safety of the workers in the construction site of the underground dam,
several safety measures should be provided in the TEIM. Figure 11, which is modified by
present authors from the original figure (Houghton, 1980), shows the stability of the unsupported tunnel in the relationship between the width of unsupported tunnel and the
Q value, which is a classification index on the stability of the ground. To keep the width
of working tunnel smaller is important to avoid caving in of the working tunnel. Drain
system is necessary in the wet ground, because that Q value decreases with the increasing
of the water content of the ground.
Figure 12, which is modified by present authors with using the correlation between
RMR and Q value (Palmstrom, 2008) from the original figure (Bieniawski, 1989),shows
the time dependency of the cavability of the unsupported tunnel in the relationship between the width of unsupported tunnel and the Q value.
To keep the raising speed of the working tunnel level in the second and third processes
in the TEIM higher is important to avoid caving of the working tunnel. To increase the
raising speed of the working tunnel level, it is necessary to increase both of the excavating
speed and the infilling speed. The dividing the whole construction site to many working
sites having small construction zone can effect to shorten the whole term of construction.
At the preliminary process, the digging the vertical shafts along the planned underground dam wall site gives the useful information on the geology and hydro-geology in the
vicinity of the planned underground dam wall site and so the construction manager can

74 u Journal of South Asia Disaster Studies

A CONCEPT OF A LOW-COST CONSTRUCTION METHOD FOR DEEP UNDERGROUND DAMS


draw up the plan of the construction process and provide the necessary ways and means
before starting the excavation of the working tunnel in the first process of the TEIM.
Both ends of the short construction sites in the working tunnel have the connecting
ducts, which secure the breathing of the worker and secure the escape and rescue routes
in case of accident, to the ground.

Figure 11: The stability of unsupported tunnel in relation to Q value

Figure 12: The time dependency of the capability of the unsupported tunnel in relation to Q Value

Vol 3 No 1 June 2010 u

75

YOSHITO KOBAYASHI AND VIJAY KUMAR JATIYA


The application of the pole supported stopper drill of long rod secures the safety of
workers because of the increase in the excavation speed in the second and third processes
of the TEIM and because that the worker can operate the drill apart from underneath the
excavating face, where is most dangerous for caving.
The application of the hand held vibration tamper secures the safety of workers because of the increase in the infilling speed in the second and third processes of the TEIM.
To shorten the time for filtering work to get fine clay for the low permeable layer material, fine clay can be prepared in advance with applying the water sieving technology for
making the porcelain clay from the mud in the bottom of the pond with using the excess
water in the rainy season.
The excavation at the working tunnel face in the TEIM produces a loosen earth zone,
of which area is depend on the property of the adjacent ground, in the ground above the
working tunnel in the same situation at the usual tunnel excavation as reviewed by A.
Sahebi et al (Sahebi, 2010).
To keep the roof of the working tunnel lancet shape reduces the danger of caving in
the loosen earth zone as shown in Figure 4 (f).
Estimation of the Quantity of Storage Water and Construction Cost of the
Underground Dam with Applying the TEIM

Figure 13: A typical underground dam

76 u Journal of South Asia Disaster Studies

A CONCEPT OF A LOW-COST CONSTRUCTION METHOD FOR DEEP UNDERGROUND DAMS


Table 1: An estimation of the quantity of storage water and construction cost of
underground dam

Figure 13 shows the definition of the factors on the estimation. Table 1 shows the
factors and results on the quantity of storage water and the construction cost on the assumption that the construction cost of the underground dam wall per unit area is equal
to that of a shallow underground dam of hand made with applying the traditional open
cut excavation method as shown in the case 1 and 2. The construction cost of the underground dam wall per unit area with applying the TEIM, which excavates upward at the
roof of the working tunnel where soil tends to fall down with the action of the force of
the gravity in the loosen soil zone mentioned above, may be lower than that with applying the traditional open cut excavation method, which excavates downward at the bottom of the trench against the force of the gravity in the tight soil zone. The case 1 may
cause the problems of salinisationor over wetting the land above the area of the storing
water because the shallow water level induced by its shallow depth of the top of the dam
wall. The case 3 and 4 show the cheaper construction cost per unit volume of the storage
water for the deeper underground dam with applying the TEIM. The investigation on the
geology, hydro-geology and meteorology is necessary to make the plan of underground
dams, such as the selection of the location and the design of underground dams.
Construction, Operation and Maintenance of Deep Underground Dams by
Community Participation
Stakeholders or community participation is an important area that needs special attention for the successful implementation of any development programme. People who live

Vol 3 No 1 June 2010 u

77

YOSHITO KOBAYASHI AND VIJAY KUMAR JATIYA


together in a settlement share common interest, common resources and feel that they
belong together and form a community. Construction of deep underground dams by
involving the community participation will help not only in creating job opportunity
but also in educating the community towards the degrading ecology. Maintenance of
deep underground dams is to be done by involving stakeholders or local communities by
educating them the guidelines of operation andmaintenance of TEIM.

Contribution of The Teim


The applicability of the TEIM at deep aquifer, at remote region and at various geologies
provides a great possibility for constructing low-cost underground dams at semi-arid regions in the world. The global problems that can be resolved by TEIM are:
(1) Water Shortage
(2) Food Shortage
(3) Global Warming
(4) Rural Poverty

Further work for the Improvement in the Teimtechnology


For the advancement and spread of the TEIM technology as a low cost and widely applicable construction method for deep underground dams, further works are necessary:
(1) Accumulation of experiences of the construction of underground dams with the TEIM
in the various cases in cooperation with high technologies to select the suitable site of
construction with using satellite information and geological surveying techniques
(2) Development in the special devices for the reduction of cost and term of construction,
for safety of workers such as ground support frame, for separation of excavated soil, for
drainage for seeping groundwater in the working tunnel, etc.
(3) Collaboration with currently running programmes for water resource, irrigation, ant
desertification, afforestation, the reclamation of farmland, etc.

References
ASAL Consultant Ltd., Water Supply by Rural Builders, Water for Arid Land.http://www.waterforaridland.
com/publications/Book8/book8chapter6.asp.
Beaumont P., Qanat Systems in Iran, Bulletin of the International Association of Scientific Hydrology,
XVI, 1. 3, 1971, pp. 39-50.
Carman P.C., Flow of gases through porous media, (London: Butterworths ScientificPublications,
London, 1956).

78 u Journal of South Asia Disaster Studies

A CONCEPT OF A LOW-COST CONSTRUCTION METHOD FOR DEEP UNDERGROUND DAMS


Carman P.C., The determination of the specific surface of powders, J. Soc.Chem. Ind. Trans., vol. 57,
1938, pp. 25.
Carrier W.D., Goodbye, Hazen; Hello, Kozeny-Carman, J. Geotech.&Geoenvir.Engrg.,vol. 129, no. 11,
2003, pp. 1054-1056.
Foster S., and Tuinhof A., Brazil, Kenya: Subsurface Dams to Augment Groundwater Storage in Basement
Terrain for Human Subsistence, The World Bank, Sustainable Groundwater Management, Lessons
from Practice, Case ProfileCollection No 5, 2004,pp. 1-7.http://siteresources.worldbank.org/
INTWRD/Resources/GWMATE_English_CP_05.pdf
Hazen A., Discussion ofdams on sand foundation by A. C. Koenig, Trans.Am. Soc. Civil. Eng..,vol. 73,
1911, pp. 199-203.
Hazen A., Some physical properties of sands and gravels, with special reference
Houghton D.A. and Stacey T.R., Application of Probability Techniques to Underground Situations,
Proc. 7th Regional Conf. for Africa on Soil Mech. andFoundation Engineering, Accra, vol. 2,
Balkema, 1980, pp. 879-883.
Jatiyaand V.K., Hosokawa T., Role of Rainwater in Aglo-Economic Growth of India, Journal of Rainwater
Catchment Systems, vol. 8, no. 1, pp. 13-19, July 2002,ISSN 1343-3646.
Kobayashi Y., Japan Patent no. 4150063 (in Japanese), 2008.
Kobayashi Y., WO 2009/011254, a patent application published by the World Intellectual Property Organization
(WIPO), 2009.
Kourmaev R., Makarievand M., Teplitskiy A., Using directional drilling in underground construction,
TRIZ J., pp. 1-6, 2007,http://www.triz-journal.com/archives/2007/09/03/
Kozeny J., Ueberkapillareleitung des wassersimboden, Wien, Akad.Wiss.,no.136, 1927, pp. 271-306.
Ministry of the Environment, Japan and Overseas Environmental Cooperation Center,Japan (2004),
Model Project to Combat Desertification in Nare Village , BurkinaFaso, Technical report of the
Subsurface Dam, 1-77.http://www.env.go.jp/earth/report/h16-08/eng/PDF/full_e.pdf
Nilsson A., Groundwater Dams for Small-scale Water Supply, (London: IntermediateTechnology
Publications Ltd., 1988). .
Nishigaki M., Kankam-Yeboah K., and Komatsu M., Underground dam technology in some parts of
the world, J. Groundwater Hydrology, 2004, 24(2), 113-130.
Okinawa General Bureau, Governmental Project for Irrigation and Drainage, Ie District,Japan(in
Japanese)http://maff.go.jp/nouson/hyouka/h15/zizen/kokuei_h150829/tiku/1-09.pdf
Onitsuka K., Lu J., Tang W.W. and Kai D., A Study on the Ancient Han-Chiku Technique in China, Soils
& Foundations, 2002, 50(5), 26-28, (in Japanese).
Padre S., Surangas, Manmade Caves to Tap Underground Water, India Together,(March 02, 2006),http://
www.indiatogether.org/2006/mar/env-ssuranga.htm

Vol 3 No 1 June 2010 u

79

YOSHITO KOBAYASHI AND VIJAY KUMAR JATIYA


Palmstrom A., Comparing the RMR, Q, and RMi Classification Systems, 2008.http://www.rockmass.
net/class/Correlation%20Q-RMR-RMi_T&T-part%202.pdf
Sahebi A., Jalalifar H., Ebrahimi M. and Abdolrezaee A., Stability Analysis of Tabas Coal Mine Roadway
using Empirical and Numerical Methods 2010 Underground Coal Operators Conference, (February
11-12, 2010).http://ro.uow.edu.au/cgi/viewcontent.cgi?article=1990&context=coal
Salih A., Qanats a Unique Groundwater Management Tool in Arid Regions:The Case of Bam Region
in Iran, International Symposium on GroundwaterSustainability (ISGWAS), 79-87, January 2006,
Alicante, Spain.
Shyh Ting Chen, 4-2-2 Pingtung, (in Chinese), 2009.http://163.26.52.242/~river/modules/tad_book3/
page.php?tbdsn=17.
Takahashiand G., Higashi S., Conservation offish habitat in streams by themethod of low dams series,
Erosion and Sedimentation in the Pacific Rim(Proceedings of the Corvallis Symposium, August,
1987, IAHS Publ. No. 165,pp.409-410, 1987, http://www.cig.ensmp.fr/~iahs/redbooks/a165/
iahs_165_0409.pdf
Tanaka S., ChoushichiHattori and Artificial Rock, Japan Lime Association, pp.1-13,2004.http: //www.
jplime.com/bunkaisan/005/index.html, (in Japanese).
to their use in filtration, 24th Annual Rep., Massachusetts State Board of Health, Pub.Doc. no. 34,
1892, pp. 539-556.
Wulff H.E., The Qanats of Iran, Scientific American, 1968, 218(4), 94-105.
Yazdani M.R., Shirani K., Rahnama F., Khodagholiand M., Azad M., Application of Underground Dams
for Groundwater Aquifer Recharge and Control ofQanats Water, November 20-23, 2006, Istanbul,
Turkey.
Zarei G. and Nasseri A., Historical, Structural and Environmental Features ofthe Qanat in Iran,
International History Seminar on Irrigation and Drainage, 451-456,May 2-5, 2007, Tehran, Iran.

80 u Journal of South Asia Disaster Studies

Assessment of Drought
Vulnerability in North Western
Region of Bangladesh
Shams Al-Amin, Umme Salma Rima, and Md. Monsurul Huda

Abstract
North-western region of Bangladesh, commonly known as the Barind Tract, is the largest Pleistocene physiographic unit of the Bengal Basin, which is comparatively at a higher elevation than
the adjoining floodplains. During monsoon, the Barind Tract remains free from flooding; it is
drained by a few small streams. Landscape modification has minimised groundwater recharge
through rainwater, and shifting of the Tista and the Atrai and their distributaries has greatly
influenced the climatic conditions of the area too, turning it into a hot region. This paper presents Aridity Index map of North-western region of Bangladesh for their vulnerability of agricultural drought following UNESCOs (1979) proposed rainfall-evapotranspiration ratio approach.
Through Historical analysis of precipitation and evaporation data over 36 years in the area,
annual rainfall and PET has been calculated. Aridity Index (AI) has been used as a numerical
indicator of the degree of dryness of the climate in the area. Maps have been generated by GIS for
analysing trend of change in precipitation.

Introduction
Increased dryness has long been recognised as a major environmental problem affecting
the living conditions of the people in the affected regions in many countries of the world.
The studies indicated that over the preceding 20 years, the problem of land degradation,
due to increased aridity, had continued to worsen. In the past, dry lands recovered easily
following long droughts and dry periods. Under modern conditions, however, they tend
to lose their biological and economic productivity quickly unless they are managed in a

Vol 3 No 1 June 2010 u

81

SHAMS AL-AMIN, UMME SALMA RIMA AND MD. MONSURUL HUDA


sustainable manner (GOI, 2001). Once the dryness of an area increase beyond a certain
level, it becomes difficult to recover. Therefore, to study the extent of dryness of an area is
essential to combat land degradation and desertification. Measurement of dryness is also
necessary for evaluating drought vulnerability, measuring drought severity, monitoring
climatic change, assessing bio-environment, monitoring soil moisture, and planning agricultural. Therefore, it is very important for a country to have an updated dryness maps.

Methodogy
Study Area
The study area (Figure 1) represents the drought prone areas of
Bangladesh, which is the North-western and Central-western
zone of Bangladesh. Increased dryness has long been recognised as a major environmental problem, affecting the living
conditions of the people in these affected regions.

Annual Rainfall Estimation


There are about 84 rainfall stations in North-western region of
Bangladesh operated by BWDB. Among them, about 53 rainfall stations have been selected
for the study area due to data
availability, which is shown in
Figure 1: Study Area
Figure 2. The daily rainfall data
have been collected from BWDB and based on that, point
map of annual precipitation has been generated from 1972
to 2008.The missing data have been filled up by the data of
previous year, and by the data of the surrounding neighbouring stations.

Annual Evaporation Estimation

Figure 2: Rainfall stations


of North Western Zone
of Bangladesh

There are about 11 evaporation stations in North-western


region of Bangladesh operated by BWDB shown in the
Figure 3. The daily potential evaporation data have been
collected from BWDB and based on that, annual potential
evaporation of each of the stations has been determined

82 u Journal of South Asia Disaster Studies

ASSESSMENT OF DROUGHT VULNERABILITY IN NORTH WESTERN REGION OF BANGLADESH


from 1972 to 2008. The missing data have been filled up by the data of previous year,
and by the data of the surrounding neighbouring stations. Based on the data, thiessen
polygon map of evaporation has been generated for the whole region.
Rainfall-evapotranspiration ratio based method for the estimation of dryness. In this
paper, UNESCO AI is used for the aridity mapping of Bangladesh. For this, thiessen polygon extension has been used to describe the area of influence of each of the evaporation
stations (Figure 2).
The values of potential evaporation under these polygon areas have been considered
as constant. Annual precipitation and evaporation maps have been integrated using the
equation: P/PET. Finally, AI maps have been generated in GIS by generating surface area
on the basis of their classification.
The UNESCO (1979) proposed a method for aridity mapping from the ratio of precipitation (P) to potential evapotranspiration (PET), that is:
AI=P/PET (1)
where, PET is the Annual Potential Evapotranspiration and P is the Annual Precipitation based on the AI values, five climatic regions of aridity are proposed as given in Table
1.
Table 1. The classifications of the climate region based
on UNESCO Aridity Index
AI Value

Climate Class

<0.03

Hyper Arid

0.03-0.2

Arid

0.2-0.5

Semi-arid

0.5-0.65

Dry sub-humid

>0.65

Humid

Determination of Aridity Index


Precipitation-temperature based method proposed by De
Martonne (1926) is widely used for measuring aridity of
an area. Another widely used method is Thornthwaites
method (Thornthwaite, 1931) based on precipitation and
temperature. Recently UNESCO (1979) has proposed a

Figure 3: Evaporation stations


of North-western Zone
of Bangladesh

Vol 3 No 1 June 2010 u

83

SHAMS AL-AMIN, UMME SALMA RIMA AND MD. MONSURUL HUDA

Statistical Analysis using GIS


For the statistical analysis of AI values for the individual district, Arc View GIS 3.2 has
been used. The surface map of these values of each of the districts have been generated
individually by interpolating grid values using inverse distance weighting method and
then the minimum, maximum values are calculated to find out the range of AI values for
each of the districts.

Result and Discussion


From the annual precipitation values and annual evaporation values of the study area, it has
been found that the total annual evapotranspiration of this zone is lower than or equal to
annual rainfall. In fact, the amount of annual
rainfall varies generally from 1,500-1,600 mm,
which is very low compared to the annual rainfall of whole Bangladesh. It means that the climate of this region can be said to be very close
to dry climate. The rainfall changing patterns
of Rajshahi and Rangpur throughout the last 36
years (1972-2005) are shown in Figure 3.

Statistics of Aridity Index Values


For north-western and central-western zone of
Figure 3: Annual rainfall (1972-2005) in
Bangladesh, AI has been calculated from 1971
Rajshahi and Rangpur
to 2007. AI in this area, as the index shows, has
been increasing over years. However, it is still very low compared to the rest of the country. The improvement of drought condition can be attributed to the large-scale initiatives
taken by the government. A project named Barind Integrated Area Development Project
(BIADP) was initiated in mid-1980s to develop groundwater irrigation in the area. Apart
from providing irrigation, there have been other programmes, such as tree planting and
excavation of ponds and khals to arrest the degradation of the environment. At the beginning, the project generated a lot of concern among environmental scientists about
sustainability of the groundwater resources. However, 15 years of its operation has proved
it a success. Environmental degradation has also been checked, and positive results such
as higher rainfall and higher vegetation coverage are evident. The project has been re-

84 u Journal of South Asia Disaster Studies

ASSESSMENT OF DROUGHT VULNERABILITY IN NORTH WESTERN REGION OF BANGLADESH


named Barind Multipurpose Development Authority (BMDA) since the early 1990s, and
now covers a large part of the Barind Tract.
Table 2. Statistics of Aridity Index Values
Aridity Index Values
1994

1995

1996

1997

1998

1999

2000

Min

0.57

0.96

0.72

0.60

0.04

1.04

1.07

Max

3.82

3.27

2.80

2.49

3.40

4.07

3.33

Mean

1.40

1.92

1.38

1.48

1.94

2.05

1.79

Std Dev

0.51

0.56

0.42

0.46

0.57

0.62

0.53

Aridity Index Values


2001

2002

2003

2004

2005

2006

2007

Min

0.90

0.76

0.88

0.68

1.19

0.75

0.61

Max

3.11

3.36

4.07

4.84

4.07

3.91

3.16

Mean

1.62

1.69

1.76

2.01

2.04

1.74

1.70

Std Dev

0.46

0.56

0.65

0.80

0.66

0.54

0.45

Extreme drought years in Northwestern


Region
North-western and central-western zone of
Bangladesh suffered extreme agricultural
drought in 1972 and 1994. The drought
can be attributed to the extreme low rainfall (as shown in Figure 4) associated with
prolonged summer in those years. The average rainfall in 1972 and 1994 were 1,360
mm and 1,355 mm respectively. The average Rainfall in 1982 was also very low (1,522
mm). The north-western and central-western zone suffered heavy agricultural loss in
those years.

Figure 4: Extreme drought years in


Northwestern Region

Vol 3 No 1 June 2010 u

85

SHAMS AL-AMIN, UMME SALMA RIMA AND MD. MONSURUL HUDA

Aridity index map for north-western region


The AI map of different years including the drought years (1972 and 1994) and recent
year (2007) have been shown in Figure 5. From the map of 1972, it has been found that
all the parts were humid zone except Naogaon, Nawabganj, Rajshahi and some parts of
Thakurgaon, Dinajpur and Rangpur where the climates were close to dry sub-humid zone,
especially in some parts of Naogaon and Nawabganj where the conditions were even
worst. From the map of 1982,
Aridity Index Map of
Aridity Index
it has been found out that the
1972 (Drought Year)
Map of 1982
dry sub-humid zone extends towards Natore and Bogra too, but
Dinajpur and Rangpur show better condition. In case of 1994,
the dry sub-humid zone extends
to Joypurhat although the other
parts show some better condition than what was in 1982.
From the map of 2007, it is clearly visible that the dry condition
seems to reach better condition
in all these parts of the study
area. However, Lalmonirhat is
Aridity Index Map
Aridity Index Map of
of 2007
1994 (Drought Year)
the new addition, which is going
to become dry zone slowly.

Comparison of no rainfall days:

Figure 5: Aridity Index Map of Various Years

86 u Journal of South Asia Disaster Studies

Number of no rainfall days, as


expected, is found to be higher
in all the stations in the four extreme drought years. Nawabganj
station is the highest in all four
cases with a total days in excess
of 300. Other than only a few
exceptions, most of the stations
have no rainfall days excess of
250 days.

ASSESSMENT OF DROUGHT VULNERABILITY IN NORTH WESTERN REGION OF BANGLADESH

A.

Figure 6: No rainfall days in 1982 and 2007 Number of


no rainfall days are found to be increasing at a steady
rate in stations like Dinajpur, Rajshahi and at an alarming rate in stations like Natore.

Figure 7: Trend of No rainfall days in


selected stations

Comparison of AI with the most drought prone year


Comparison of AI between the Most Drought prone Year and Recent Year has been shown
in Table 3. It can be found from the maps as well as the table that except Lalmonirhat,
the other districts show greater AI Values, which means that the amount of rainfall is increasing to an extent. Nevertheless, most of the rainfall occurs in monsoon season. If we
consider the dry season rainfall, it will be clearly visible that the rainfall is not sufficient
at all. This is the reason for agricultural drought due to shortage of rainfall in dry period.
Since the area is free from flood, rainwater is the main source of groundwater recharge.
A decrease in rainfall has caused the successive lowering of the groundwater table of
the region, which has eventually been greatly affecting the environmental parameters.
Besides lowering of the water table, there is a noticeable change in forest area. Almost 70
per cent forestland of the region had been changed into cultivable land during the past

Vol 3 No 1 June 2010 u

87

SHAMS AL-AMIN, UMME SALMA RIMA AND MD. MONSURUL HUDA


few decades (Ahmed, 2001). Due to the dry nature, relatively low rainfall and massive
deforestation, the vegetation cover in the area has decreased distinctively. If this trend
continued, the area might become almost an arid region.
Table 3: Comparison of AI between the Most Drought prone Year and Recent Year
Name of District
Thakurgaon
Dinajpur
Nilfamari
Naogaon
Nawabganj
Rangpur
Gaibandha
Joypurhat
Rajshahi
Natore
Lalmonirhat

AI

AI

in 1972

in 2007

0.69-1.50

1.37-2.02

Near Dry subhumid

Humid

0.69-1.53

1.42-3.16

Near Dry Subhumid

Humid

1.17-2.04

1.59-2.21

Humid

Humid

0.5-1.42

1.13-1.74

Dry Subhumid

Humid

0.55-0.98

1.15-1.62

Dry Subhumid

Humid

0.83-1.9

1.38-2.31

Near Dry Subhumid

Humid

1.09-1.45

1.24-2.22

Humid

Humid

0.97-1.31

1.56-1.91

Humid

Humid

0.66-1.09

1.15-1.93

Near Dry Subhumid

Humid

0.9-1.39

1.15-1.78

Near Dry Subhumid

Humid

1.72-1.92

0.61-2.17

Humid

Near Dry Subhumid

[2] 2. Shahid, Chen, Hazarika (2005). Assessing Aridity in Bangladesh using GIS,
Bangladesh, 2005
[3] UNESCO, 1979. Aridity definition (UN documents), UNESCO, New York

88 u Journal of South Asia Disaster Studies

ASSESSMENT OF DROUGHT VULNERABILITY IN NORTH WESTERN REGION OF BANGLADESH

Conclusion
An attempt has been made for the mapping of dryness of north-western region of Bangladesh using geographic information system for assessment of vulnerability of drought in
north-western region of Bangladesh. The most recently used model is used for this purpose. The result shows that there exists no climatically dry zone in Bangladesh although
some parts in the western side of the country have AI that is close to dry zone. Deforestation and other environmentally harmful activities in the region might cause a gradual
declination of rainfall, inclination of temperature, and consequent increase of aridity of
the region. If the aridity of the area tends to increase, it might lose its productivity and
become an arid region. Necessary steps should be taken to manage the situation in a sustainable manner

References
Ahmed K.M., Barind Tract, In Banglapedia, S. Islam edited, Asiatic Society of Bangladesh, Dhaka, 2003.
http://banglapedia.search.com.bd/HT/B_0309.htm.

Vol 3 No 1 June 2010 u

89

Quantifying Disaster Risk and


Vulnerability at State Level: A Case
of Gujarat, India
Shital Hardik Shukla

Introduction
Disaster management (DM) is not only relief and response activities but is also inclusive of
pro-active strategies like preparedness, mitigation, and development of various resources.
A strong need has emerged today for mainstreaming DM into the development goals.
Mainstreaming DM into developmental goals means systematic and more conscious ways
of integrating disaster risk reduction into the development process. In other words, one
needs to see that the process of development and the kind of development choices do not
create or increase the risk of disasters. This has reflected in the recognition that disaster
risk reduction strategies must be addressed as one of the developmental goals. Therefore,
it is of an immediate necessity to develop tools which help effectively in mainstreaming
disaster risk reduction strategies into developmental goals.
In this context, quantifying vulnerability due to the natural disasters is the first necessary step for any serious contemplation of disaster reduction strategies of development
policies. Quantifying vulnerability indicators is an emerging technique, which can systematically assess and analyse vulnerability and risk due to future possible climatic hazards. It can reveal many adverse effects, which are led by climate-induced disasters on
society. Therefore, it helps state as well as society for preparing and adopting adequate
and successful measures for reducing the risk factors. In other words, quantifying vulnerability can show direction for preventing losses. It leads towards preparedness to reduce
recovery time rather than dependence on relief and reconstruction. Therefore, measuring
the vulnerability can help decision-makers in prioritising their intervention for pro-active

Vol 3 No 1 June 2010 u

91

SHITAL HARDIK SHUKLA


disaster mitigation strategies and climate change policies. This paper makes an attempt
to identify the vulnerability indicators at state level. It further goes beyond identifying
the indicators of vulnerability and endeavours to quantify those indicators at district
level. It then explores the regional model for assessing the pattern of vulnerability within
the state and identifying the drivers of vulnerability in the state. Such an exercise is also
useful for identifying the most vulnerable district in the state for regional planning. Although indicators are not comprehensive like the concept, yet they tend to help policy
makers by deriving relative numbers and helping in prioritising the planning process in
a relative ranking.
The paper is divided into three partsthe first section outlines the global as well as
national level efforts to work on vulnerability measurement; the second section provides
the methodological approach to measure the risk and vulnerability indicators for Gujarat,
and the last section assesses the risk due to disaster and explores the regional model of
vulnerability for Gujarat.

Global and National Level Efforts on Vulnerability Measuring


Internationally, the Hyogo framework for action is a major landmark in analysing the
disaster risk and vulnerability pattern at various levels. The Hyogo framework of action
is a strategic outcome of the world conference on Disaster Risk Reduction which was
organised by the UN. The framework outlines major course of actions for the period of
2005-15. One of the major actions of the plan declares that the starting point for reducing
disaster risk and promoting a culture of disaster resilience lies in the knowledge of the hazards
and the physical, social, economic, and environmental vulnerabilities to disasters that most societies face. One of the key activities under this course of action directs towards developing
a system of indicators for disaster risk of vulnerability at national and sub national levels.
This can help decision-makers in their planning for risk reduction. In short, the conference of disaster risk reduction stresses not only on identifying indicators of vulnerability
but also asks for measuring them systematically.
Disaster risk and vulnerability analysis is relatively new area of research and has received attention for the last few years. The most prominent and landmark study in this
area is carried out by the Centre for Hazards and Risk Research at Columbia University
in collaboration with the World Bank. The study entitled Natural Disaster Hotspots: A
Global Risk Analysis is the first attempt to assess the risk potential at global level. This
study has assessed the potential risk in terms of two major outcome indicators of disasters
namely; mortality losses and economic losses. However, addressing the social vulnerabil-

92 u Journal of South Asia Disaster Studies

QUANTIFYING DISASTER RISK AND VULNERABILITY AT STATE LEVEL: A CASE OF GUJARAT, INDIA
ity remained a challenge for this study. The UNDPs effort dealt with this challenge and
derived relative index of vulnerability at country level in a study entitled Reducing Disaster Risk: A Challenge for Development. As UNDPs focus is mainly on human well being,
the vulnerability indicators related to socio-economic vulnerability included economic
growth, diversification of economic activities, demographic scenario, health andsanitation, early warning capability, education and knowledge, and human development
index. However, inadequacy in data at country level made the scope of the work carried
out by UNDP very limited in deriving a comprehensive index of vulnerability.
India is also one of the important signatories in Hyogo action plan. The finance commission in India devotes a whole chapter on disaster relief in India from the 6th finance
commission. The 13th finance commission very distinctly addressed this issue by introducing two types of fundsdisaster response fund and disaster mitigation fund (in place
of calamity relief fund). Further, the 11th five-year plan has very clearly called for a strong
action for integrating disaster risk reduction into development planning. The preparation of Vulnerability Atlas (VA) is one of the actions carried out to address the issues of
vulnerability and risk due to natural disasters. The VA considered the vulnerability as
major function of physical condition and structural aspects in a society. Although it has
considered some of the socio-economic variables, the fundamental question has to be
elucidated that whether the characteristic of the vulnerability remain analogous over
space. As the characteristic of vulnerability depends upon a variety of factors, it needs to
be specified on a basis of location. Therefore, it is appropriate to measure the vulnerability
at regional level. In this context, regional vulnerability is the vulnerability of a specified
region incorporating its people, society, economy, ecosystem and environment. Regional
model of vulnerability tries to address the issues related to the development of a region.
As the regional vulnerability looks into its human development, economic growth and
equity, sustainable livelihood, infrastructural development and environmental resources
within the region, it can be useful for administrative planning.
Being one of the forefront states of India, Gujarat has achieved significantly in terms
of economic growth. It recorded the third highest per capita income (Rs. 63,961 or US$
1351) in 2010. Favourable industrial strategies have facilitated the economic growth of
the state in recent decade. At the same time the state is also considered as one of the most
disaster prone states in India. It faces several natural disasters such as droughts, famines,
cyclones, earthquakes and floods. The pace of development put the state more exposed
towards the potential risk due to disaster. The statistics of the disaster scenario in Gujarat
indicates that the state faces a high risk due to various disasters. Therefore, the state has

Vol 3 No 1 June 2010 u

93

SHITAL HARDIK SHUKLA


to have an integrated regional planning to deal with the potential losses due to disaster
in future.
Further, the state face substantial financial burden of disaster losses. The fund collected
under calamity relief fund remains far less than what usually required for relief and reconstruction. The state has to rely on external funding. DM in Gujarat remained reactive
by responding disasters through provision of relief during and after calamity. In other
words, the state used to adopt a relief driven DM system. In recent times, there has been
a paradigm shift and the state has become or is becoming more proactive with emphasis
on disaster prevention, mitigation and preparedness. The Gujarat State DM Authority has
been set up as a nodal agency to plan for various strategies related to disasters. GSDMA has
not only worked extensively on relief and rehabilitation activities but has also focused on
risk reduction strategies such as preparation of pre DM plan at various levels. The state has
shown a sincere commitment towards DM activities. Therefore, it is very important that
the state looks into its regional planning for dealing with any future disaster.

Risk Assessment for Gujarat: An Approach


It is well established that the risk assessment reduces risk more effectively and helps
achieving development goals at regional level. Risk assessment requires a systematic
measurement of risk over particular geographical location. As the risk is a function of
the hazard and vulnerability, both the factors need to be measured separately, and then
interlink them for planning purpose. The study has assessed the risk due to disaster by
estimating a probability and intensity of disaster derived from the past trend. The study
has considered three major climatic variability induced disasters namely; flood, drought
and cyclone. The past trend of these hazards has been prepared by collecting information
from EM-DAT series from OFDA/CRED International Disaster Database. This has been
updated and improved by crosschecking the information collected form the relief department of government of Gujarat. The data on hazards have been obtained for the last 80
years. Such a large data set can provide very useful information on past trend of hazards
in a state. An annual probability has been calculated from this time series of data for three
selected disasters. The information collected on hazards incorporated the immediate impact in terms of death toll, people affected and economic damage.
It is assumed here that the risk due to disaster depends on the probability and intensity. The annual probability can be estimated by calculating the frequency of these three
major disasters. As far as intensity is concerned, it can be estimated by assessing the impact of disaster. The three different indicators of impact are derived for assessing intensity

94 u Journal of South Asia Disaster Studies

QUANTIFYING DISASTER RISK AND VULNERABILITY AT STATE LEVEL: A CASE OF GUJARAT, INDIA
of the disaster. These are coefficient of mortality, coefficient of social vulnerability and coefficient of economic vulnerability. The methodology for calculating these coefficients is
adopted from UNDPs approach. The coefficient of mortality is derived by calculating the
total death toll divided by the population of the state. The 2001 data has been selected as
the base year. The coefficient of social vulnerability has been derived by dividing the total
affected people with the total population of the base year. The coefficient of economic
vulnerability is derived by dividing the total damage with the gross domestic product of
the selected base year. Further, the risk also depends on the vulnerability of the region
which is exposed to the selected disasters. Therefore, the vulnerability index is carried out
separately. Vulnerability, by its simple definition, refers to the susceptibility of a person,
group, society, region, or system to any external risk. In this context, the external risk is
in terms of disasters. Although there are number of attempts made to define vulnerability
it lead to the simple meaning explained in above definition. In fact, it is not possible to
define vulnerability precisely. The concept of vulnerability provides flexibility to accommodate a number ofphenomena under it. Therefore the term vulnerability is relative
and explanation of the term depends upon various factors. United Nations International
Strategy for Disaster Reduction (UN/ISRD) defines vulnerability as the condition determined by physical, social, economic, and environmental factors, which increase the susceptibility
of a community to the impact of hazards(UN/ISRD, 2004).
Vulnerability indicators need to be robust, evident, self-expressive, replicable, comparable and easy to understand. The selection of the indicators is a technical challenge.
Based on UNDPs approach and other literature, the indicators are selected. For instance,
Human Development Index is an important indicator of vulnerability which explains
the status of human condition in a given region. This exercise is limited to the indicators
which could be derived based on availability of information at district level in the state
of Gujarat. Some of the indicators are inversely related to vulnerability. For instance,
the area under forest can protect the region and community from the intensive effect of
disasters.
In order to derive the vulnerability index in this paper, various demographic, physical, economic, social and environmental indicators have been identified and clubbed
together. The indicators selected under each category of factors are shown in Table 1.
About 22 indicators have been selected under four major factors, which are responsible
for vulnerability of Gujarat.

Vol 3 No 1 June 2010 u

95

SHITAL HARDIK SHUKLA


Table 1: List of indicators derived for quantifying vulnerability
No.

Indicators
Demographic Indicators

1.

Population Density

Population growth rate

Urban population growth rate

Dependency ratio

Pace of urban growth


Economic Indicators

Poverty rate

Unemployment rates

Percentage of labour (Agri labour + other labours)

Industrial Investment per 1000 people

10

Un-irrigated land

11

No. of banks per lakh population


Social Indicators

12

HDI

13

GDI

14

No. of hospitals for 1000 people

15

Availability of schools per 1000 people

16

% of drinking water coverage

17

Availability of transportation facility (index)

18

Availability of communication facility (index)

19

Pakka road per sqkm


Environmental Indicators

20

Degraded land (waste land)

21

Area under forest in %

22

Water resource development

As the number of indicators is large and overlapping, the principle component analysis was used to reduce the number of indicators and derive the major drivers of vulnerability. The SPSS (Statistical Programme for Social Science) was used for carrying out
principle component analysis. The major aim to use this technique was to reduce the
complexity of information and reduce the observed variables by clubbing them together

96 u Journal of South Asia Disaster Studies

QUANTIFYING DISASTER RISK AND VULNERABILITY AT STATE LEVEL: A CASE OF GUJARAT, INDIA
in a hidden factor which represents the number of variables. The variables were scaled
up in order to compare with each other. The formula used for scaling up remained X = X
-(Mean-2SD)/SD for each data set.
The vulnerability index has been also calculated by giving equal weightage to each
driver of vulnerability. The four major regional drivers of vulnerability derived from the
exercise are exposure to hazards, vulnerable people, vulnerable surrounding, and vulnerable resources. These are explained in detail in the following section.

Risk due to Disaster:


The risk of disaster in Gujarat has been
increasing over the last few decades.
The observation made by Centre for
Research on the Epidemiology of Disasters (CRED) under international disaster database very clearly indicates
that the disaster in Gujarat state has
increased more than three times during the last two decades. The following table shows the frequency of three
selected disasters (flood, drought and
cyclones) faced by Gujarat over a period of time.
It indicates that the frequency of disasters has increased sharply over a period of time.
This implies that there will be an increase in risk due to potential disasterin the state.
Further the probabilities of the various natural disasters differ in the state. The following table presents the annual probability of three major disasters faced bythe state.
Table 2: Annual Probability of three selected disasters in Gujarat
Type of Disasters

Frequency of three major disasters


during the last 80 years

Annual Probability

Drought

7.5

Flood

20

25

Cyclones

14

17.5

Source: EM-DAT: The OFDA/CRED International Disaster Database

Vol 3 No 1 June 2010 u

97

SHITAL HARDIK SHUKLA


It is interesting to note here that the annual probability of the flood is the highest
among all the three selected disasters. This is followed by cyclones and drought with 17.5
per cent and 7.5 per cent respectively. Drought is also very frequent phenomena in the state
with arid and semiarid climate. However, it is not considered as disasters when there is partial drought. In fact, one of the estimates in Gujarat shows that the state is facing drought
in every three years. Having a quite high annual probability, the flood causes high financial
burden on the state, which ultimately has severe implications on the state economy.

Impact of Disasters on the state


The immediate impact of the disasters is considered by the death toll it created from its
impact. Table 3 portrays the death toll due to various disasters over a period. It is clear
from the table that disasters have significant social adverse effects. The trend in causalities shows that the trend for flood up to1980s is increasing. The decade of 1990s shows
the decline in death toll. This is in spite of the fact that this decade has experienced a
number of floods. The preparedness and mitigation actions have played major role in
protecting life of local people from flood. The irrigation department of the state government devotes a substantial proportion of its budget for flood mitigation strategies. Such
strategies mainly include structural measures. Further, the death toll due to cyclone has
increased over a period of time. The sudden onset of this disaster does not give chance to
protect the lives of coastal community. Although there is early warning system working
very well, the local community people do not follow the warning mainly because it is the
question of survival of their livelihood. Death due to drought is uncommon. Unless and
until, there is severe drought, the death toll does not happen. The state has experienced
Table 3: Death toll due to various disasters in Gujarat
Decade

Drought

Flood

Cyclone

1921-30

56

1941-50

45

1951-60

17

1961-70

100

1971-80

1348

172

1981-90

2139

1171

1991-2000

1948

3149

2000-03

20

233

Total

20

5,830

4,553

Source: EM-DAT: The OFDA/CRED International Disaster Database

98 u Journal of South Asia Disaster Studies

QUANTIFYING DISASTER RISK AND VULNERABILITY AT STATE LEVEL: A CASE OF GUJARAT, INDIA
death due to drought mainly because of
heat wave along with the drought or suicide due to debt burden in the period of
drought.
Although there in an inconsistency in
data on people affected by various disasters, the overall pattern is very clear showing that the reported number of people
affected by various disasters has increased
sharply over a period of time. Furthermore, the intensity of effects is the highest observed
in the decade of ninety in which the state faced severe negative social implications of
disasters such as flood, drought, cyclone and epidemics.
As far as an adverse economic impact of disasters is concerned, the data
shows that the economic loss due to
various disasters has dramatically increased since 1970. The increase in
decadal loss from 1970s to 1990s is
very high. The economic loss during
1990s is 30 times higher than those
during 1970s. Further, the economic
loss during the years of 2001 and 2002
was around US$ 4.2 million, which is
almost 88 per centof economic loss
due to the damage during 1990s. This indicates the severity of damage is increasing over
a period of time in the state. This confirms the fact that the trend of human suffering and
economic loss is worsening in Gujarat.
Looking from the disaster point of view, Gujarat emerges at global level by an earthquake hit area where earthquake seems to be the most dangerous disaster. Or in other
words, Gujarat is always illustrated as earthquake hit state in the global picture. In fact, the
DM activities started only after the event of an earthquake. However, this does not mean
that other disasters are less dangerous for the state. Comparing the impact of various disasters, cyclones are the worst disasters which have made tremendous economic loss (US$
4.3 million) in the state. This is in spite of the fact that many minor disasters (flood and
cyclones) with less economic impact are not taken into estimation of economic loss.

Vol 3 No 1 June 2010 u

99

SHITAL HARDIK SHUKLA


Intensity Indicators: Three indicators of impact are derived in order to measure the
risk intensity of disasters.
Table 4: Indicators of risk intensity due to disasters
Risk Intensity Indicators

Flood

Cyclone

Drought/Heat Wave

Average Death toll (M)

13,463

5,069

685

Coefficient of mortality (CM)

26.6

10

1.4

Number of people affected (SV)

25,61,18,523

1,07,33,839

39,11,75,000

Coefficient of Social Vulnerability (CSV)

5.1

0.2

7.7

Average Economic loss Eh in US US$ (EV)

19,01,669

43,23,056

6,76,359

Coefficient of economic vulnerability (CEV)

12.2

27.7

4.3

It is assumed here that the higher impact of disaster is mainly due to greater intensity.
Table 4 denotes the coefficient of vulnerability. The coefficient of mortality is the highest
for flood, which is followed by cyclone and drought. Paradoxical to this the coefficient of
social vulnerability is the highest for drought and lowest for cyclone. This shows that the
social impact of drought is prevalent in the state. The majority of the rural population
is dependent on rain fed agriculture for their livelihood maintenance in the state. The
drought affects the livelihood of the people at a great extent. It is interesting to note here
that the coefficient of economic vulnerability is the highest for the cyclone. The coastal
area of the state is highly prone to cyclone. Further, the port led economy has developed
the coastal area with higher economic growth. Densely populated and economically well
developed coastal region put the area more exposed to cyclones, and therefore, the probability of economic damage due to cyclone is exponentially high in the state.

Vulnerability Index
As discussed in the earlier section, vulnerability indicators are identified and measured
through principle component analysis. Indicators are the operational representation of
attributes showing qualities or characteristics of vulnerability. Identification of indicators
is carried out by refereeing various studies on disaster indicators. The first step towards
analysis was to club the overlapping indicators and identify the drivers of vulnerability
pattern in Gujarat. The analysis resulted in four major drivers of vulnerability which persuaded the vulnerability pattern in the state of Gujarat. These four drivers are exposure,
vulnerable people, vulnerable surrounding and vulnerable resources. Figure1 represents

100 u Journal of South Asia Disaster Studies

QUANTIFYING DISASTER RISK AND VULNERABILITY AT STATE LEVEL: A CASE OF GUJARAT, INDIA
the conceptual model for all the drivers
of vulnerability. Each driver is explained
here in elaboration.
1. Vulnerable people: The first important driver of vulnerability is vulnerable people. The variable under these
drivers mainly represent the marginalised groups of the society. These
are dependents, percentage of SC/
ST, persons below poverty line, unemployed, female agricultural labour,
destitute,illiterate females. Higher the numbers in these variables make the districtmore
vulnerable. Further lower human development index, gender development index, gender empowerment measure and infant mortality rates are also the variables falling into
this category. These indicators are negative indicators. Lower the values of these indices make the district more vulnerable. The vulnerability explained by these variables
has major role indetermining vulnerability pattern of the state. These variables have
played asignificant role in making backward districts, such as Dahod, Panchamahal,and
drought prone districts, such as Kachchh and Surendranagar more vulnerable.
2. The second driver of vulnerability is named as an exposure which incorporates the
growth of population and economy. Densely populated areas with high industrial
growth and even higher economic growth are more vulnerable. Most of the districts
having higher exposure are located in the coastal areas. For instance, Jamnagar, Rajkot,
Porbanadar, and Amreli.
3. The third set of variables, which shapes the vulnerability, is called vulnerable surrounding. An easy access to basic services for human being provides better surrounding. Lack
of such services makes people of that region more vulnerable towards the selected disasters. The important basic services in this category involve access to drinking water
and sanitation, access to health and education and access to pakka road and electricity.
When disaster strikes, an inadequate access to such basic services drastically increases
the impact of disasters and reduces the chances of survival. These variables have played
major role in determining vulnerability in backward districts such as Dohad, Dangs,
Panchmahal and Surendranagar as well as coastal districts such as Kachchh Jamnagar
and Valsad.

Vol 3 No 1 June 2010 u

101

SHITAL HARDIK SHUKLA


Table 5: Ranking of by major drives of vulnerability
District

Exposure

District

Vulnerable
people

District

Vulner- District
able surrounding

Vulnerable resources

Gandhinagar

15.2

Dohad

26.5

Dohad

16.4

Sabarkantha

15.6

Ahmedabad

14.3

Dangs

24.8

Dangs

15.3

Valsad

15.6

Jamnagar

14.1

Navsari

21.8

Navsari

13.9

Vadodara

14.3

Rajkot

13.2

Panchmahals

21.1

Surendranagar

12.4

Junagadh

14.3

Patan

12.9

Surendranagar

15.0

Panchmahals

11.9

Panchmahals 13.3

Porbandar

10.5

Narmada

13.4

Kachchh

10.0

Patan

12.9

Amreli

10.4

Kachchh

12.6

Jamnagar

9.6

Ahmedabad

12.9

Mehsana

10.2

Vadodara

10.6

Valsad

9.6

Bharuch

12.7

Bhavnagar

8.5

Mehsana

9.1

Vadodara

8.8

Kachchh

11.5

Junagadh

8.0

Banaskantha

8.8

Narmada

8.6

Surat

11.0

Bharuch

7.9

Sabarkantha

8.5

Surat

8.4

Amreli

9.9

Sabarkantha

7.7

Jamnagar

8.0

Bharuch

8.2

Anand

8.9

Banaskantha

7.7

Junagadh

8.0

Junagadh

8.2

ICheda

8.9

Anand

7.5

Valsad

7.4

Bhavnagar

8.1

Porbandar

8.9

Narmada

7.3

Amreli

7.1

Banaskantha

8.0

Narmada

7.9

Vadodara

7.2

Surat

S.I

Mehsana

8.0

Gandhinagar

7.7

Navsari

7.0

Gandhinagar

5.8

Porbandar

7.7

Rajkot

6.9

Kheda

S.8

Bhavnagar

5.2

Anand

6.9

Dangs

6.4

Kachchh

S.8

Bharuch

4.1

Amreli

6.7

Navsari

6.4

Panchmahals

S.7

Kheda

1.8

Sabarkantha

6.5

Mehsana

6.0

Dangs

S.5

Patan

1.4

Kheda

6.3

Surendranagar

6.0

Valsad

S.3

Anand

0.8

Rajkot

5.5

Banaskantha

5.1

Surendranagar

5.5

Porbandar

0.7

Gandhinagar

4.8

Dohad

3.0

Surat

5.1

Ahmedabad

0.2

Patan

4.2

Bhavnagar

2.3

Dohad

3.1

Rajkot

0.1

Ahmedabad

2.3

Jamnagar

1.8

102 u Journal of South Asia Disaster Studies

QUANTIFYING DISASTER RISK AND VULNERABILITY AT STATE LEVEL: A CASE OF GUJARAT, INDIA
4. The fourth category derived from the analysis was named as vulnerablere sources of
the region, which make the region more vulnerable towards the selected disasters. The
contributory indicators in this category are degraded forest land, waste land, over used
ground water and deteriorated quality of drinking water sources. The degraded resources
do not support the community livelihood at the time of disaster such as drought, flood
and cyclone and therefore make the people of that region more vulnerable towards disasters. In Gujarat, the poor base of natural resources in Sabarkantha, Valsad, Vadodara,
Junagadhad; Panchmahal put this districts in vulnerable region.
Table 6: Relative vulnerability Index at district level in Gujarat
District

Vulnerability Index

District

Vulnerability Index

Panchmahal s

13.3

Mehsana

8.3

Dangs

13.2

Bharuch

8.2

Dohad

12.3

Patan

7.9

Navsari

12.3

Surat

7.7

Vadodara

10.2

Banaskantha

7.4

Kachchh

10.2

Porbandar

7.0

Surendranagar

9.7

Jamnagar

6.7

Valsad

9.7

Anand

6.1

Junagadh

9.6

Kheda

6.0

Sabarkantha

9.6

Ahmedabad

6.0

Narmada

9.3

Rajkot

4.8

Amreli

8.5

Bhavnagar

4.3

Gandhinagar

8.4

The above analysis clearly indicates that the pattern of vulnerability can be explained
by four major drivers of vulnerability. Therefore, in order to reduce the vulnerability of
the region, it is important to address the drivers of vulnerability. Addressing these drivers can essentially reduce disaster risks by reducing vulnerability. The vulnerability index
derived from these drivers represents the relative position of districts within the state.
Vulnerability index, in this context, helps in prioritising the district for development
planning. The following table presents the vulnerability index.
As far as vulnerability pattern in Gujarat is concerned, the backward districtsPanchmahal, Dangs and Dohadare at higher risk due to disaster. Given a similar hazard scenario within the state, these districts have higher chances of potential losses at the time

Vol 3 No 1 June 2010 u

103

SHITAL HARDIK SHUKLA


of disaster. The district level planning can further look into the block level planning for
further prioritisation for development activities. Such an exercise is quite useful for state
level as well as district level planning. The development activities at district level have
to incorporate the disaster risk reduction strategies. It is very important to see that the
development activities do not increase the vulnerability towards disasters.

References
Adger W.N., Social Vulnerability to Climate Change and Extremes in Coastal Vietnam, 1999. World
Development, 27: 249-69.
Birkmann J., et al., Measuring Vulnerability To Natural Hazards, TERI Press ISBN- 8179931226, New
Delhi, 2006.
Blaikieet P., al., At Risk: Natural Hazards, Peoples Vulnerability and Disasters, Routledge, 1996.
Bohle H.G., Downing T. E., and Watts M. J., Climate Change and Social Vulnerability: The Sociology
and Geography of Food Insecurity, 1994, Global Environmental Change 4:37-48.
Cannon T., Twigg J., et al., Social Vulnerability, Sustainable Livelihoods and Disasters, Report to DFID
Conflict and Humanitarian Assistance Department (CHAD) and Sustainable Livelihoods Support
Office, London, 2005,DFID: 63
Dilley et, al.,Natural Disaster Hotspot; A Global Risk Analysis, the World Bank and Columbia University,
New York, 2005.
GOI, District Census Handbook, Census of India, New Delhi, 2001).
GOI, The Gujarat VA, prepared by Building Materials & Technology Promotion Council, GOI, New Delhi, 2005.
HirwayInidira and Mahadevia Darshini, (2004). Gujarat Human Development Report, Mahatma
Gandhi Labour Institute, Ahmadabad.
Kapur Anu, Vulnerable India: A Geographical Study of Disasters, (New Delhi: SAGE Publication, 2010).
Oliver-Smith A., Theorizing Vulnerability in a Globalized World: A Political Ecological Perspective,
Mapping Vulnerability: Disasters, Development and People, Bankoff, (London: Frerk and Hilhorst,
2003), Earthscan: 10-24.
UNDP, Reducing Disaster Risk: A Challenge for Development, Bureau for Crisis Prevention and Recovery,
New York, USA, 2004.
United Nations, Hyogo Framework for Action 2005-2015: Building the Resilience of Nations and
Communities to Disasters, UNISRD, 2005, A/CONF.206/6
Watts M. and Bohle H. G., The Space of Vulnerability: the Causal Structure of Hunger and Famine.
Progress in Human Geography 17(1), 1993.
Wisner B., Blaikie P., Cannon T. and Davis I., At Risk - Natural Hazards, Peoples Vulnerability and
Disasters, (Wiltshire: Routledge, 2004), ISBN 0-415-25216-4.

104 u Journal of South Asia Disaster Studies

You might also like