Study of early flood warning dissemination system in Bangladesh

Abstract

A study has been conducted with the objective to assess the existing early flood warning dissemination system (EFWDS) in Bangladesh and to suggest the suitable improvements in the same system based on review of literature, interaction with stakeholders of various organisations involved in flood forecasting and dissemination, and analysis of feedback from the flood-affected people of Dhobaura and Shibalaya subdistricts in Bangladesh. The existing set-up has been studied to assess the present activities and future expectations. The recommendation for active participation by all related organisations has been made in this study. Two studies have been conducted by surveying the opinion of flood vulnerable communities so that all elements of the EFWDS would provide useful flood warnings to all potential users.

Full text

Study of early flood warning dissemination system in Bangladesh
M.M. Rahman1, N.K. Goel2and D.S. Arya2
1 BWDB, Ministry of Water Resources, Dhaka, Bangladesh
2 Department of Hydrology, Indian Institute of Technology, Roorkee, India
Correspondence
Md. Mizanur Rahman, BWDB, Ministry of
Water Resources, Flat 102, Eastern Lotus,
House 7/A, Road 1, Dhanmondi R/A,
Dhaka 1205, Bangladesh
Email: mizanurbd2004@yahoo.com
DOI: 10.1111/jfr3.12012
Key words
Communities; early flood warning; flood
prone; opinion; survey; warning
dissemination.
Abstract
A study has been conducted with the objective to assess the existing early flood
warning dissemination system (EFWDS) in Bangladesh and to suggest the
suitable improvements in the same system based on review of literature, inter-
action with stakeholders of various organisations involved in flood forecasting
and dissemination, and analysis of feedback from the flood-affected people of
Dhobaura and Shibalaya subdistricts in Bangladesh. The existing set-up has
been studied to assess the present activities and future expectations. The rec-
ommendation for active participation by all related organisations has been
made in this study. Two studies have been conducted by surveying the opinion
of flood vulnerable communities so that all elements of the EFWDS would
provide useful flood warnings to all potential users.
Introduction
Dissemination of forecasts to the affected end-users is an
important link in any flood forecasting set-up. The flood
forecasts should be disseminated in a timely manner and in
understandable form. Moreover, the affected people, who
are mostly uneducated, poor, and simple should believe the
forecast and should act appropriately. Behavioural aspects
of the affected people also play an important role in the
success of any early flood warning dissemination system
(EFWDS). BMRB International (2001) carried out numbers
of post-event and annual public response surveys in
England and Wales, for effectiveness of early flood warn-
ings. Drabek (1999), Parker et al. (2009), Thrush et al.
(2005), and Fielding et al. (2007) have shown that an appro-
priate response to warnings may depend on the informa-
tion in the warning, the way it is disseminated, whether
recipients trust the source of the warning and perceive the
threat, environmental clues (e.g. observing heavy rain and
floods), personal attitudes (including previous experience),
and social networks. Drobot and Parker (2007), DEFRA/EA
(2008), and Twigger-Ross et al. (2008) engaged in the social
sciences for improving flood warnings and flood risk man-
agement in general. Their study was on for a greater under-
standing of the public’s behavioural response to flood
warning information and for effective and reliable ways of
enhancing this response.
Despite several advances in the flood forecasting system
in Bangladesh, the existing system often underperforms
because the warning, dissemination, and response of the
end-users are unsatisfactory (Chowdhury, 2005).
The present study has been taken up with the objective to
assess the existing EFWDS of Bangladesh and suggest suit-
able improvements in this system based on review of litera-
ture, interaction with officials of various organisations
involved in flood forecasting and dissemination, and inter-
action with the flood-affected people of Dhobaura and
Shibalaya subdistricts in Bangladesh.
Concept of early flood warning systems
Flood warning is the in-advanced information of floods,
which may occur in the near future. Emergency activities are
intended by warnings about potential dangers of onslaught
floods (Hossain, 2004).A warning of a flood event implies an
existing threat of danger to life and property, and it invites
responsive action to reduce threat. ‘EFWS’means immediate
alert to the vulnerable people in advance about the danger
about to happen after detection. Warning of forecasts must
reach the users without any delay and with sufficient lead
time to permit response actions to take place (ISDR, 2005).
The progressive development of forecasting system cannot
be a success without success of the dissemination phase of
the warning process. In spite of that, due to different reasons,
early flood warning dissemination receives less attention
than desirable (Parker, 1987). The EFWS has four interre-
lated stages (Smith, 1996). Rahman et al. (2008) framed an
early flood warning system by describing those four parts i.e.
Detection of onset flood; Warning, which is the method that
turns the forecast into a message; Dissemination, which is the
bs_bs_banner
J Flood Risk Management •• (2012) ••–•• © 2012 Blackwell Publishing Ltd and The Chartered Institution of Water and Environmental Management (CIWEM)

process that spreads onslaught flood to the potential end-
users; and Response to take action to mitigate the impact of
floods.
The EFWS should include public awareness and under-
standing of the forecast and warning information system
and its economic values (UN/ISDR, 2004). Under this
system, selection of concerned organisations is important
for effective dissemination. People’s participation at the
community level is essential in the response of early flood
warning. Community-based EFWS is an important method
and is emphasised worldwide.
Community-based early warning systems are ‘people-
centred’ system and empower individuals and communities
threatened by hazards to act in sufficient time and in an
appropriate manner so as to reduce the possibility of per-
sonal injury, loss of life, damage to property and environ-
ment, and loss of livelihood (Perez et al., 2007). Parker et al.
(2009) described that in the official EFWS, in most cases, for
various reasons, warning does not reach everyone prior to
flooding. Therefore, involvement of communities might be
an important tool to establish a successful EFWS.
The need for early warning system is recognised by the
Govt. of Bangladesh and the same has also been included in
the National Water Policy of Bangladesh which states,
‘......throughitsresponsible agencies, the Government will
develop early warning and flood proofing systems to manage
natural disasters like flood and drought’ (GoB, 1998).
Flood problems in Bangladesh
Bangladesh is a flood-prone country and extreme floods
inundate more than half of the country’s land mass almost
every year. The country is in the lowest ridge of Hindu Kush
Himalayan region, which makes the country hydrologically
very diverse, complex, and unique (Chowdhury and Ward,
2004; Sikka and Ringler, 2009). Economy, environment,
ecology, livelihood, and development are affected by devas-
tating floods every year.The country contains the confluence
of the Ganges, Brahmaputra, and Meghna rivers and their
criss-crossed tributaries as well distributaries, which dis-
charge into the Bay of Bengal. Heavy rainfall over the catch-
ment of these rivers can produce an average run-off of about
1 009 000 m3, and if the whole water were stored, the
country would have been flooded to a depth of 8 to 10 m
(Alam, 2000). Land is mainly flat, with 40% of its land mass
up to 10 m above mean sea level (Rahman et al., 2010).
Floodwater inundates around 20.5% of the country (3.03
million hectares) every year (Chowdhury, 2000; Mirza et al.,
2001). In extreme cases, it may cover 70% of the country
(Mirza, 2002), with varied effects. The trend of more fre-
quent large floods causing ever-increasing damages is likely
to continue due to factors such as global warming, sea-level
rise, changing land use, and increased urbanisation.
The climate of Bangladesh is tropical, influenced by the
Himalayan Mountains in the north and northeast, and by
the Bay of Bengal in the south. High monsoon rains associ-
ated with Bangladesh’s unique geographical location in the
delta of the world’s largest river basin make it extremely
vulnerable to recurring floods. Agriculture is the dominant
land use in the country covering about 59% of the land,
whereas rivers and other water bodies constitute about 9%
(BBS, 2002). Country has experienced seven highly damag-
ing floods in the 20th century. Since independence in 1971,
Bangladesh has experienced floods of a vast magnitude in
1974, 1987, 1988, 1998, and 2004 (FFWC, 2005a). Impacts in
terms of affected area, estimated economic loss (only direct
losses from damage of crop, infrastructures like roads, rail-
ways, embankments, bridges etc.), and deaths of human
beings during these severe floods are presented in Table 1.
The area affected and crop damage during major flood
events of the country since 1950 are presented in Figure 1
(Hossain, 2004).The largest recorded flood (both in terms of
depth and duration of flooding) in Bangladesh occurred in
1998 when about 68% of the country was under water for
several months.
Various structural measures such as flood embankments,
channel improvements, river training works, and coastal
embankments, and nonstructural measures such as flood-
plain zoning, flood forecasting, and early warning systems
have been adopted in the country to minimise the flood
losses.
Existing EFWDS in Bangladesh
Flood Forecasting and Warning Service (FFWS) of Bangla-
desh was established in 1972 as a permanent entity under
Bangladesh Water Development Board (BWDB).
Facing the devastating floods of 1987 and 1988, it was
conceptualised that the integrated hydro-informatics- and
geo-informatics-based forecasts and warning dissemination
system should be introduced. Amid the devastation of the
1998 flood, BWDB realised that the further improvement of
the FFWS was needed, thence a new project entitled ‘Con-
solidation and Strengthening of Flood Forecasting and
Warning Services (CSFFWS)’ (BWDB and DHI, 2006) was
implemented in cooperation with Danish International
Development Agency, during 2000–2004. This project
enhanced the capability of existing flood forecasting system
of Flood Forecasting and Warning Centre (FFWC), but it
was assessed that the dissemination of flood information to
the end-users could be improved.
A study was made by the FFWC under CSFFWS to assess
the response to its information by the people living in the
flood-prone areas. The flood-affected people have indicated
that they are receiving little information about flood onset
through the existing warning dissemination media like TV,
2Rahman et al.
© 2012 Blackwell Publishing Ltd and The Chartered Institution of Water and Environmental Management (CIWEM) J Flood Risk Management •• (2012) ••–••

newspaper, and radio as shown in Figure 2. This condition
persists despite the fact that the technology of flood forecasts
in Bangladesh, as in many other countries, has reached a
highly sophisticated level. This aspect of flood information
dissemination needs serious attention (Bhuiyan, 2006).
The early flood warning dissemination to various govern-
ment and non-government organisations (NGOs), media
groups, and other concerned parties is being done through
the FFWC. The FFWC of Bangladesh has the following limi-
tations in the area of EFWDS:
•The lead time of 48 h is not sufficient for disseminating
the information to effect timely response of flood-prone
communities because of the limited coordination
between associated organisations (Chowdhury, 2000).
•Early flood warning messages are prepared in technical
terms of ‘Water Level with respect to Danger Level’
(Paudyal, 2002) which is disseminated to the national
level and not for community level.
•Response of the flood warning messages is not well coor-
dinated by different levels of different organisations. The
community of flood-prone areas is not aware of the flood
forecasting and warning systems.
The present flood warning system can be improved sig-
nificantly by people’s participation and feedback from the
involved communities.
Two pilot projects for assessing the viability of further
improvement of EFWDS are described in subsequent
section.
Table 1 Some notable flood events after independence of Bangladesh and their impact
Event
Impact
Source
Inundated area (in terms of %
of the country’s geographical
area of 144 000 km2) No. of human deaths Estimated damages
1974 flood 36 Over 2000 deaths, followed
by famine with over 30 000
deaths
Not recorded FFWC (2005b)
1987 flood 40 Over 2055 deaths Approximately US$ 1.0 billion The World Bank (2002)
1988 flood 61 Persons affected – 45 million,
2300 deaths
Approximately US$ 1.2 billion The World Bank (2002)
1998 flood 68 Persons affected – 31 million,
1100 deaths
Damaged 500 000 homes,
23 500 km roads and 4500 km
embankment, crop damage
500 000 ha, damage worth
about US$ 2.8 billion
The World Bank (2002)
2004 flood 38 Persons affected – 36 million,
750 deaths,
Damaged 58 000 km roads
and 3 100 km embankment,
crop damage 1.3 million ha,
damage worth about US$ 2.2
billion
ADB-World Bank (2004)
Figure 1 Area affected and crop damage during major floods events since 1950.
Study of early flood warning dissemination system in Bangladesh 3
J Flood Risk Management •• (2012) ••–•• © 2012 Blackwell Publishing Ltd and The Chartered Institution of Water and Environmental Management (CIWEM)

Environmental monitoring and information
network (EMIN)
EMIN, a pilot project (2000 to 2007), is the information
network for coordination between key decision-makers by
monitoring the flood-related disasters in Brahmaputra-
Jamuna river basin. It functions in the national and local level
by analysing flood maps in the national extents and river
flood situation maps in the local extent. EMIN carried out
extensive information assessment (ICTDG and CEGIS, 2004)
and then developed a system to bring information of onset
floods to communities living in the floodplain. EMIN recom-
mended (1) use of high-resolution digital elevation model,
(2) development of flash flood warning system, (3) augmen-
tation of lead time of warning messages, (4) strengthening of
institutional linkage and information sharing among all con-
cerned organisations, (5) development of mode of dissemi-
nation to reach women, and (6) raising awareness of local
level institutions on how to use flood forecasts and warning
effectively for flood preparedness. This pilot project was
linked with FFWC during its tenure, but later, it was not
extended to be incorporated over all Bangladesh.
Community flood information systems (CFIS)
CFIS, also a pilot project (2001 to 2006), developed an
information software named WATSURF (Centre for Envi-
ronmental and Geographical Information System, Dhaka,
Bangladesh) (A software for water surface forecasting)
(CEGIS, 2005) to disseminate information on flood extent,
duration, and depth of water to the community before the
flood occurs. CFIS project disseminated flood warning from
WATSURF to three micro pilot areas (known as Mouzas). The
CFIS project raised awareness among local people about the
flood forecasts and warnings. Flood management committee
were formed in the community level involving local aged
elites, local government-elected representatives and officials,
and NGOs (Martin et al., 2006). Some individuals were
selected in the community to serve as the operators to receive
a daily text message with flood warnings and operate the flag
system and bulletin board to inform the community of the
flood warning. The message and symbols were designed with
active participation of the local people (BDPC, 2003). The
CFIS dissemination model varied slightly from the dissemi-
nation model used by EMIN. For example, CFIS relied
entirely on local volunteers to disseminate flood information.
CFIS showed that local organisations are capable of support-
ing the system and dissemination of flood warnings to all
socioeconomic groups. This is also a pilot project that was
linked with FFWC during its tenure, but later it was not
extended to be incorporated over all Bangladesh.
Institutions related with EFWDS
in Bangladesh
BWDB, under Ministry of Water Resources, maintains
FFWC which is the key entity for handling flood forecasts
and warning system. Hydrology Directorate of BWDB and
the Bangladesh Meteorological Department under Ministry
of Defence are prime organisations which furnish hydrologi-
cal and meteorological data, respectively for the forecasting
models. The Disaster Management Bureau (DMB) is an
important organisation for response and preparedness
against the onset of floods under Ministry of Food and
Disaster Management. These are some research and devel-
opment organisations working with floods and flood fore-
casting. The organisations were visited by the author to
review their activities related to flood forecasting and
obtained their views on possible improvements in existing
flood forecasting and warning system.
FFWC disseminates the flood warning messages to the
national level as well as local level through electronic media,
newspapers, and their own websites.FFWC sends the warning
message as bulletins to the highest authority of the govern-
ment, that is, Office of President of the Republic, concerned
ministries, DMB, some NGOs, concerned research institutes,
and district administrations. Local government and local
administration are not considered by FFWC.
The assessment of roles of different organisations indi-
cated the following:
Institutional problems:
Figure 2 Feedback from affected people about media of warning
dissemination conducted by an non-government organisation
during 2000–2002 under Consolidation and Strengthening of
Flood Forecasting and Warning Services.
4Rahman et al.
© 2012 Blackwell Publishing Ltd and The Chartered Institution of Water and Environmental Management (CIWEM) J Flood Risk Management •• (2012) ••–••

•Shortage of funds and lack of leadership for disseminat-
ing meaningful flood warnings to communities and
infrastructure managers;
•Inadequate coordination and feedback; and lack of lead-
ership of the overall EFWS.
Technical Problems:
•Inadequate maintenance and upgrading of existing flood
forecasting and warning system,
•Limited coverage of the dissemination of warning
messages;
•Flood warnings not understandable by many potential
end-users.
Case studies
Two case studies have been conducted for two different types
of flood-prone areas in Bangladesh. The studies were under-
taken to determine whether or not the communities were
aware of flood and receive early warning of floods. The
objective was also to prepare local level authorities to dis-
seminate the warning to the communities as well as to
prepare the people for flood-fighting response. The survey
was conducted during the flood season of 2009. Two flood
vulnerable areas namely Dhobaura, a flash flood-prone area
in Mymensingh district, and Shibalaya, a river flooding area
in Manikganj district, were selected. The location of these
two areas is shown in Figures 3 and 4.
Area used in case studies of EFWDS
Two case studies for Dhobaura and Shibalaya areas have
been conducted for incorporating the feedback of local com-
munities in the EFWDS.
Dhobaura area
Dhobaura subdistrict is located at latitude: 25°5′N and lon-
gitude: 90°31′E (Figure 3) under Mymensingh district, has a
population of 3 915 068 (according to census in 1991) in the
northern part of Bangladesh.
The total area of Dhobaura Upazila is 251.05 km2.
Dhobaura has a population of 157 027 (according to census
in 1991). The Nitai and Kongsho rivers control the hydrol-
ogy of Dhobaura. These rivers originate from Meghalaya hill
tracts and cause frequent and high-magnitude flash floods in
the flat terrain in Dhobaura immediately after steep hill
slopes. The area is characterised as heavy rainfall area. The
area is in the vicinity of the Cherapunji (latitude: 25°18′N,
longitude: 91°42′E), Assam, India.
Dhobaura is vulnerable to flash floods caused by intensive
rainfall in upstream hilly catchments. During flash floods,
water level rises quickly and the areas are flooded rapidly.
Flash floods cause extensive damage to agriculture, infra-
structure, and culture fisheries. Homesteads close to the
main river channel are also subject to erosion during flash
floods. At present, there is no early flood warning system in
Dhobaura and adjacent areas.
Shibalaya area
Shibalaya subdistrict is located at latitude: 23°50′N and
longitude: 89°47′E (Figure 4) under Manikganj district, has
a population is 1 274 829 (according to census in 1991), in
the central part of Bangladesh. Shibalaya has an area of
199 km2and a population of 143 842 (according to census in
1991). The area is vulnerable to flooding as most of the land
is low lying.
The central part of Bangladesh is highly vulnerable to
floods caused by over bank flow from the Brahmaputra-
Jamuna river. Shibalaya is one of the most vulnerable areas,
and floods result in significant damages to crops and liveli-
hoods, as experienced in 1998 and 2004.
Floods and erosion cause damage to crops, infrastructure,
culture fisheries, and homesteads. Floods in Shibalaya are
influenced by flows in three rivers: the Ganges, Jamuna, and
Padma. At present, Shibalaya does not have any EFWS. The
onset of flood is severe when the Ganges and Brahmaputra
face high peak; flood recession is slow because of backwater
pressure by the river Padma due to spring tide or depression
in the Bay of Bengal.
Methodology for the case studies
The methodology of this study is based on the survey of the
opinion of the communities. For this purpose, an easily
understandable questionnaire was prepared (Appendix A).
The questionnaire was first pretested in the same area to
understand whether people can respond easily with interest
or not and whether there is any shortfall in the question-
naire. A total of 12 questions were placed before the vulner-
able communities to get the information regarding the early
flood information system. The questionnaire was prepared
in Bengali language. However, the English version of the
questionnaire is given herein. The survey was conducted
during August and September 2009.A total of 217 question-
naires were collected. These questionnaires were analysed
using SPSS (IBM Corporation, Armonk, NY, USA).
Analysis of survey results
In Dhobaura and Shibalya subdistricts, the number of
respondents were 117 and 100 respectively. The age of
respondents varied from 17 to 90 years in Dhobaura and
from 17 to 78 years in Shibalaya. The mean and standard
deviation of the age of the respondents are 50.65 and
16.91 in Dhobaura and 46.13 and 13.72 in Shibalaya. The
Study of early flood warning dissemination system in Bangladesh 5
J Flood Risk Management •• (2012) ••–•• © 2012 Blackwell Publishing Ltd and The Chartered Institution of Water and Environmental Management (CIWEM)

occupation distributions of the respondents of both areas
are given in the Figure 5.
The survey has illustrated that the people are aware of the
floods in their areas. Of the respondents, 93.2% in Dhobaura
and 98% in Shibalya have given their opinion that flood is a
regular phenomenon in their area. They are also very much
aware of the causes of floods. About 71% of persons from
Dhobaura have opined that flood in their area is flash flood
and 90% of Shibalaya respondents knew that it is river flood
in their area (Table 2). More than 93% and 99% respondents
in Dhobaura and Shibalaya respectively have assessed that
their area is vulnerable to enormous flood damages. Figure 6
shows vulnerability statistics of life and properties. Of the
respondents, 16% and 7% feel danger for their life in
Dhobaura and Shibalaya, respectively. More than 50% of the
respondents of both areas (Dhobaura 52% and Shibalaya
58%) are worried for the loss of their agriculture due to
flood every year. Response against no damage was absent in
the collected data.
•Under this study, respondents have been asked whether or
not they are receiving any warning message from any
corner. Response against this question is not satisfactory,
whereas FFWC has been running since 1972.Among total
respondents, 93% in Dhobaura and 79.4% in Shibalaya
do not know about the warning messages from FFWC.
This may be because the FFWC warning messages are
being disseminated to the national level with less impor-
tance to facilitate warning messages for flood mitigation
Figure 3 Dhobaura area.
6Rahman et al.
© 2012 Blackwell Publishing Ltd and The Chartered Institution of Water and Environmental Management (CIWEM) J Flood Risk Management •• (2012) ••–••

preparedness in Dhobaura and Shibalaya. Also, warning
messages that are prepared in technical terms of ‘Water
Level with respect to Danger Level’ may not be under-
standable to the local community level or local people of
these areas, as they are not aware about the flood forecast-
ing and warning system. Table 3 shows the statistics of
collected data about the availability of warning messages.
People of the areas have been found to be very keen to
receive the warning messages to save their properties and
take safety measures against floods. Figure 7 shows the
response of respondents in favour of warning message
dissemination, 96% and 98% of them in Dhobaura and
Shibalaya respectively think that warning messages would
be beneficial for them to save the damages.
Figure 4 Shibalaya area.
Others, 3%
Student, 6%
Service, 14%
Busine ss, 10% Agricu lture,
68%
Others 10% Agriculture 35%
Busin ess 20%
Servic e 28%
Student 7%
Dhobaura Shibalaya
Figure 5 Occupation of respondents.
Study of early flood warning dissemination system in Bangladesh 7
J Flood Risk Management •• (2012) ••–•• © 2012 Blackwell Publishing Ltd and The Chartered Institution of Water and Environmental Management (CIWEM)

Dhobaura and Shibalaya are rural areas and Internet is not
easily accessible. Moreover, people know only local Bengali
language. Most of the respondents are related to agriculture.
So, the format of flood warning messages should be designed
in Bengali which is understandable by the local people. The
results of survey about magnitudes and extent of flooding
are presented in Tables 4 and 5.
The mode of dissemination of warning messages is an
important component of any early flood warning system. It
depends on the end-users, i.e. whether the end-user is a
Table 2 Response of local people about the causes of flood in their area
Causes of flood
Data quality
Selection of event causing
flood
Frequency of response Per cent of response
Dhobaura Shibalya Dhobaura Shibalya
Valid Rain water 6 6 5.1 6
Flash flood 83 4 70.9 4
River flood 9 90 7.7 90
Total respondents 117 100 100 100
House and Roads
32%
Agr i c u l tu r e
52%
Human
life 25%
Poultry and
Cattle
75%
Life
16%
House and
Roads
35%
Agr i c u l tu r e
58%
Human
life 14%
Poultry and
Cattle
86%
Life
7%
Dhobaura Shibalaya
Figure 6 Vulnerability statistics of life and properties.
Table 3 Response statistics of dissemination of present warning messages by Flood Forecasting and Warning Centre
Present situation of warning message dissemination
Data quality
Availability of warning
messages
Frequency of response Per cent of response
Dhobaura Shibalaya Dhobaura Shibalaya
Valid Available 8 20 6.9 20.6
Not available 108 77 93.1 79.4
Total 116 97 100 100
Missing 1 3 – –
Total respondents 117 100 – –
Nee d , 96. 58
No Ne ed , 3. 42
Nee d, 98
No Ne ed, 2
Dhobaura Shibala
y
a
Figure 7 Response (in %) of respondents in favour of warning message dissemination.
8Rahman et al.
© 2012 Blackwell Publishing Ltd and The Chartered Institution of Water and Environmental Management (CIWEM) J Flood Risk Management •• (2012) ••–••

decision-maker or local level flood victim. However, the
survey has been conducted through the local level flood
victims, and their opinion is that the radio and TV would be
the suitable mass media to receive the information in their
area. Some people have selected SMS through mobile tel-
ephones rather than voicemail. Actually, they have selected
the said media according to their knowledge and practices.
Voicemail might be easier for the villager, but because of the
lack of knowledge about the voicemail, they have not
selected it. Their response is shown in pie chart in Figure 8.
The respondents are familiar with the local institutions/
organisations in their area. They have easy access to
institutions/organisations like ‘Union Parishad’, NGOs,
Upazila Parishad, religious key installation, member of par-
liament etc. In this respect, they are interested to strengthen
the local institutes according to their own perception and
interest. Probably, the role of these institutes in their society
has also been influenced by their decision, and this is
reflected in Tables 6 and 7. Table 6 illustrates the mandate to
the institutions/organisations for dissemination activities
and Table 7 illustrates the mandate to the institutions/
organisations for response and disaster management like
flood fighting, relief, rescue etc., according to the warning
messages. Regarding dissemination activities,majority of the
respondents opted for local government (Union Parishad)
and NGOs.
The Dhobaura and Shibalaya areas are vulnerable to
flooding which causes enormous damages. No localised
Table 4 Response in % in favour of different proposed formats for flood depth warning message dissemination
Opinion for format of warning messages for flood depth
Data quality
Format of message for
flood depth
Frequency of response Per cent of response
Dhobura Shibalaya Dhobura Shibalaya
Valid Flood depth in feet 37 82 32.5 24.5
Flood depth in hands or
fingers
32 10 28 10.3
Flood depth by marking
on vertical post or pillars
45 5 39.5 65.2
Total 114 97 –
Missing 3 3 – –
Total respondent 117 100 – –
Table 5 Response in % in favour of different proposed formats for flood extent warning message dissemination
Opinion for format of warning messages for flood extent map
Data quality
Format of message for
flood map
Frequency of response Per cent of response
Dhobaura Shibalaya Dhobaura Shibalaya
Valid Thana map 39 9 34 9.1
Union map 74 80 64.3 80.8
Mousa map 2 10 1.7 10.1
Total 115 99 100 100
Missing 2 1 – –
Total respondent 117 100 – –
Radio & TV
57%
Newspaper
3%
SMS
39%
Voicemail
1%
Radio & TV
82%
Newspaper
3%
SMS
10%
Voicemail
5%
Dhobaura Shibalaya
Figure 8 Response about suitable mass media for dissemination of warning messages.
Study of early flood warning dissemination system in Bangladesh 9
J Flood Risk Management •• (2012) ••–•• © 2012 Blackwell Publishing Ltd and The Chartered Institution of Water and Environmental Management (CIWEM)

EFWDS exist at present in these areas. FFWC disseminates
the warning messages to the national level bodies but lacks in
disseminating warning message for flood preparedness to
the people in Dhobaura and Shibalaya areas. The local
people feel that it would be beneficial if an EFWDS could be
developed for their areas. They have suggested the format
and language of the warning messages. The flood depth
marking on a post near the rivers would be understood by
them and flood extent maps shown on Union Parishad map
might make it easy to understand the probable damages of a
flood. The participating people of the two areas feel that the
Union Parishad, a micro level local government body, might
be good local platform to disseminate the warning messages
to the communities of the end-users. At present, in FFWC
policy, Union Parishad is not included in the list where flood
bulletins are sent. Warning messages can be disseminated
through radio, TV, and SMS.
Conclusions and recommendations
The present study has been taken up with the objective to
assess the existing EFWDS of Bangladesh and suggest suit-
able improvements in this system based on review of litera-
ture, interaction with personnel of various organisations
associated with flood forecasting and dissemination, and
interaction with the flood-affected people of Dhobaura and
Shibalaya subdistricts in Bangladesh. Two survey studies
have been conducted to collect the opinion of flood vulner-
able communities to make the EFWDS more effective. Based
on the review of functions of the organisations associated
with EFWDS, the following recommendations are made for
better functioning of EFWDS:
•Organisations and institutions are required to interlink
for well coordination under a common platform.
•Upgradation of knowledge and skill of the staff, and
maintenance of logistic technologies is very important.
•The leadership and responsibilities should be mandated
by the government to a key institute properly for dissemi-
nating meaningful flood warning messages.
•The feedback from the community level about the EFWS
can improve the system to be an economic service. This
community involvement should be routine work to make
warning messages understandable and have responses
from the end-users.
The survey conducted among the flood-affected people of
Dhobaura and Shibalaya indicated that the Dhobaura and
Shibalaya areas are vulnerable to flooding and no localised
EFWDS exists at present in these areas. FFWC disseminates
Table 6 Response in favour of the institutions or organisations to disseminate the warning messages to the end-users
Response to institution/organisation for dissemination of warning messages
Data quality Institutions/organisations
Frequency of response Per cent of response
Dhobaura Shibalaya Dhobaura Shibalaya
Valid Member of parliament 19 3 16.4 3.0
District administration 3 2 2.6 2.0
Upazila Parishad 7 6 6 6.0
Union Parishad 28 70 24.1 70.0
NGOs 43 12 37.1 12.0
Religious key stations 16 7 13.8 7.0
Total 116 100 100 100
Missing 1 – – –
Total respondents 117 – – –
NGOs, non-government organisations.
Table 7 Response in favour of the institutions or organisations to respond for disaster management according to warning messages to
the end-users
Response to effective authority too flood response
Data quality Institutions/organisations
Frequency of response Per cent of response
Dhobaura Shibalaya Dhobaura Shibalaya
Valid Member of parliament 22 12 18.8 12
District administration 2 2 1.7 2
Upalla Parishad 5 2 4.3 2
Union Parishad 29 62 24.8 62
NGOs 59 22 50.4 22
Total 117 100 100 100
NGOs, non-government organisations.
10 Rahman et al.
© 2012 Blackwell Publishing Ltd and The Chartered Institution of Water and Environmental Management (CIWEM) J Flood Risk Management •• (2012) ••–••

the warning messages to the national level bodies but do not
disseminate the same to the people in Dhobaura and Shiba-
laya areas. The EFWDS is recommended to be developed for
these areas with warning messages in local language through
radio, TV, and SMS. The flood depth marking on a post near
the rivers and flood extent maps shown on Union Parishad
map is recommended for better understanding and minimi-
sation of flood damages. The Union Parishad, a micro level
local government body, is perceived to be a good platform to
disseminate the warning messages to the end-users. At
present, in FFWC policy, Union Parishad is not included in
the list where flood bulletins are sent.
Acknowledgements
The authors would like to acknowledge the help extended by
the BWDB in providing literatures on early flood warning
system. Dr Shafia Shaheen, Institue of Public Health, Bang-
ladesh, Mr Dinesh Bala, Research Scholar, Department of
Civil Engineering, IIT, Roorkee, India, Mr Nasir Uddin
Khan, Chief, water management, BWDB, and Mr Md. Saiful
Hossain, Executive Engineer, BWDB, are also acknowledged
for their cooperation. The authors acknowledge the con-
structive suggestions received from anonymous reviewers.
References
ADB-World Bank. Emergency flood damage rehabilitation
project. Joint ADB-World Bank damage and need assess-
ment, Dhaka, November 2004.
Alam J.B. The concept of flood shelter and its planning to cope
with flood. In implementing Hazard-Resistant Housing pro-
ceedings of the Third Housing and Hazards International
Conference held in Dhaka, Bangladesh, 24–26 November
2000 and Exeter, UK, 4–5 December 2000, edited by
Hodgson, Seraj and Choudhury, 2000, 217–224.
Bangladesh Bureau of Statistics (BBS). Ministry of planning.
Dhaka: Government of Bangladesh, 2002.
BDPC. Community needs assessment: floods vulnerabilty and risk
reduction through community based flood information systems
(CFIS). Dhaka: Bangladesh Disaster Preparedness Centre,
2003.
Bhuiyan M.S. Flood forecasting, warning and response system.
Proceedings of National Workshop on Options for Flood
Risks and Damage Reduction in Bangladesh, 7–9 September
2004. 2006.
BMRB International. Flood warning dissemination post event
survey autumn 2000. Bristol: Environment Agency, 2001.
BWDB & DHI. Consolidation and strengthening of flood fore-
casting and warning services. Final report volume-V –
project extension report, In association with IWM, NCG,
Mott Macdonald, Aqua, BDPC, GKSS, Prip Trust, December
2006.
CEGIS. Draft technical document on WATSURF Version 2.2.
Bangladesh: Ministry of Water Resources, 2005.
Chowdhury M.R. An assessment of flood forecasting in Bangla-
desh: the experience of the 1998 flood. Nat Hazards 2000, 22,
(2), 139–163.
Chowdhury M.R. Consensus seasonal flood forecasts and
warning response system – an alternate for flood manage-
ment in Bangladesh. Environ Manage 2005, 35, (6), 716–725.
DOI: 10.1007/s00267-004-0010-4.
Chowdhury M.R. & Ward N. Hydro-meteorological variabilty
in the greater Ganges-Brahmaputra-Meghna basins. Int J
Climatol 2004, 24, 1495–1508.
DEFRA/EA. Supporting the development of a social science
strategy for FCERM R&D. FD2604 technical summary.
London: DEFRA/EA, 2008.
Drabek T.E. Understanding disaster warning responses. Soc Sci J
1999, 36, 515–523.
Drobot S. & Parker D.J. Advances and challenges in flash flood
warnings. Environ Hazards 2007, 7, (3), 173–178.
Fielding J., Burningham K., Thrush D. & Catt R. Public response
to flood warning. R&D technical report SC020116. Bristol:
Environment Agency, 2007.
Flood Forecasting and Warning Centre (FFWC). Annual flood
report. Dhaka: Bangladesh Water Development Board,
2005a.
Flood Forecasting and Warning Centre (FFWC). Consolidation
and strengthening of flood forecasting and warning services.
Final report, volume II – monitoring and evaluation, Dhaka:
BWDB, 2005b.
Govt. of Bangladesh (GoB). National water policy. Dhaka: Min-
istry of Water Resources, 1998. http://www.warpo.gov.bd/
(accessed 24 October 2008).
Hossain A.N.H.A. The impact of flood on Bangladesh and
options for mitigation: an overview. Keynote Paper, Proceed-
ings of National Workshop on Options for Flood Risks and
Damage Reduction in Bangladesh, 7–9 September 2004.
ICTDG and CEGIS. Flood and Erosion Information Needs
Assessment Report. EMIN Implementation Phase, Ministry
of Water Resources, Bangladesh, 2004.
International Strategy for Disaster Reduction (ISDR). Integrated
flood forecasting, warning and response system. Guidelines
for reducing flood losses. 2005. http://www.unisdr.org/eng/
library/isdr-publication/flood-guidelines/isdr-publication-
floods.htm (accessed 16 November 2008).
Martin T., Hassan A., Rahman S. & Wirthlin K. Vulnerbility and
risk reduction through a community-based system for flood
forecasting and monitoring. Paper presented to the 4th
Annual Mekong Flood Forum, Siem Reap. 18–19 May. 2006.
Mirza Q.M.M. Global warming and changes in the probability
of occurrence of floods in Bangladesh and implications. Glob
Environ Change 2002, 12, (2), 127–138.
Mirza Q.M.M., Warrick R.A., Ericksen N.J. & Kenny G.J. Are
floods getting worse in the Ganges, Brahmaputra and
Meghna Basins? Environ Hazards 2001, 3, (2), 37–48.
Study of early flood warning dissemination system in Bangladesh 11
J Flood Risk Management •• (2012) ••–•• © 2012 Blackwell Publishing Ltd and The Chartered Institution of Water and Environmental Management (CIWEM)

Parker D.J. Flood warning dissemination: the British experience.
Flood Hazard Management: British and International Per-
spectives. Geo Books, Norwich, 1987, 169–190, a fig, 33 ref,
append.
Parker D.J., Priest S.J. & Tapsell S. Understanding and enhanc-
ing the public’s behavioural response to flood warning infor-
mation. Meteorol Appl 2009, 16, 103–114.
Paudyal G. Forecasting and warning of water-related disasters in
a complex hydraulic setting – the case of Bangladesh. Hydrol
Sci J 2002, 47, (S), S5–S18. August 2002, Special Issue:
Towards Integrated Water Resources Management for Sustain-
able Development.
Perez R.T., Espinueva S.R. & Hernand H. Community-based
early warning systems. Briefing Paper: Workshop on The
science and practice of flood disaster management in urban-
izing Monsoon Asia/4–6 April 2007, Chiang Mai, Thailand.
Rahman M.M., Arya D.S. & Goel N.K. Approach of early flood
warning system for urban flood risk mitigation. International
Conference on Disaster and Development –Bridging the gap
between theory and practice, Kathmandu: IC0DAD, 2008.
Rahman M.M., Arya D.S. & Goel N.K. Limitation of application
of 90m SRTM DEM in hydrological application/analysis
using D8 method in the flat terrain of Bangladesh. 2010.
Applied Geometrics (Springer Link), DOI: 10.1007/s12518-
010-0020-2.
Sikka A.K. & Ringler C. Introduction to the special issue. Int J
River Basin Manag 2009, 7, (2), 107–110. © 2009 IAHR,
INBO & IAHS.
Smith A. Anthropological research on hazards and disasters.
Annu Rev Anthropol 1996, 25, (1996), 303–328.
Thrush D., Burningham K. & Fielding J. Flood warning for vul-
nerable groups: a qualitative study. Science report SC990007/
SR3. Bristol: Environment Agency, 2005.
Twigger-Ross C., Tapsell S., Davoudi S., Fielding J., Sheate B.,
Warburton D. & Fernandez-Bibao A. Demystifying social
science research in flood risk management. Paper presented
at the Annual Flood and Coastal Management Conference,
Manchester. 2008.
UN/ISDR. Living with risk: a global review of disaster reduction
initiatives. United Nations Inter-Agency Secretariat of the
International Strategy for Disaster Reduction. Early warning,
2004, p.358.
The World Bank. Bangladesh disaster & public finance. World
Bank Paper no-6. Dhaka, 2002.
Appendix
Questionnaire
Please put the right mark (✓) on your option(s) among/
between the answers of questionnaires for people’s partici-
pation to ‘Develop the Early Flood Warning Methodology’ as
given below:
1. Would you face any flood in your area?
Answer: Yes regular/Irregular/No
2. Please mention the type(s) of flood(s) in your area
Answer: Rain fed flood/Flash flood/River flood/Please
mention if there is/are any other reason(s) for flooding
inyourarea:....................................
3. Provide the quantitative magnitude of the damages in
your area.
Answer: Immense damage/Normal damage/Not at all
4. Which type(s) of property (ies) is (are) under threat by
flood (if occurred) in your area?
Answer: Agricultural products/Houses and
communications/Poultry and cattle/Human life
5. Is there any early flood warning in your area?
Answer: Yes/No
6. Would you need any early flood warning in your area?
Answer: Yes/No
7. Would you believe that early flood warning would be
benefited in your area?
Answer: Yes/No
8. Which type(s) of information will be easy understand-
ing to magnify the flood depth in your area to response
the floods?
Answer: Flood depth in feet above the river bank
level/flood depth in hands or fingers above the bank
level/marking in a reference vertical post or pillar/
Please mention if there is any other option by
you:..............................
9. Which type(s) of information will be easy understand-
ing to magnify the flood extent in your area to response
the floods?
Answer: Flood map for sub-district level/flood map
for union level/flood map for village (Mouza) level/
Please mention if there is any other option by
you:..............................
10. Early flood warning will be useful to you if it is dissemi-
nated through
Answer: Radio and TV/Newspaper/SMS by mobile
phone/Voice mail by mobile phone
11. Early flood warning will be disseminated properly if it is
sent to the authority of
Answer: Member of Parliament/District
Administration/Upazela Chairman Parishad/NGO/
Imam of Mosque/any other authority by your
opinion:.........................
12. Flood damage would be reduced if the early flood
warning messages could be sent to
Answer: Member of Parliament/District
Administration/Upazela Chairman Parishad/NGO/any
other authority by your opinion:..................
Your Name:..............................Age:
Your occupation:..................Your father’s
Name:.........YourAddress:..................
12 Rahman et al.
© 2012 Blackwell Publishing Ltd and The Chartered Institution of Water and Environmental Management (CIWEM) J Flood Risk Management •• (2012) ••–••