Applications of Medium Range Probabilistic Flood Forecast for Societal Benefi ts: Lessons Learnt from Bangladesh - Reducing Disaster: Early Warning Systems for Climate Change - page 171

Table 9.3 Sources of early
warning dissemination
system identifi ed

Source Where How

• Teacher • School/college • Miking

• Imam • Mosque • Poster

• Upzila member • Bazaar • Door to door

• Door to door • Flag

•

Signboard

Table 9.4 Ranking and time
of fl ood effect on each sector
in study area

Time of effect
During Sectors Ranking Post

Agriculture 1 H L

Fisheries 11 H L

Livestock 9 H L

Poultry 12 H L

Household 4 H M

Homestead 10 M No effect

Business 6 H M

Health and life 7 H H

Drinking water sources 5 H No effect

Sanitation 8 H H

Communication 3 H H

Other (handloom) 2 H H

H high, M medium, L low

different water levels. A level survey was conducted to defi ne their perception to
exact water level from the msl. Thus for the fl oods, the thresholds that have been
identifi ed are:

Low fl ood: 17-18.5 m
Medium fl ood: 18.6-19.5 m
High fl ood:

≥ 19.6 m

9.5 Testing the System for the 2007 Floods

Approximately 26,000 square kilometres of 46 districts within Bangladesh's terri-
tory were inundated during the fl ood of 2007, affecting almost 13.3 million people
(DER 2007 ). The 2007 fl oods came in two waves. The fi rst wave commenced
around 24 July 2007 and receded on 6 August which affected 39 districts. The sec-
ond wave commenced on 5 September and continued up until 15 September 2007.
In both the cases, the model was able to predict the fl oods well in advance.
The warning information was disseminated through the FFWC's traditional
dissemination process (e.g. fax, telephone, e-mail). In the fi ve pilot areas, the
overall process of fl ood forecast dissemination was divided into three parts. First,
the requirements of the flood forecast dissemination were analysed through
consultation between the FFWC and community. Second, the message production