Climate Forecast Applications
in
Bangladesh (CFAB-I) 2000-2003
Overview
The Climate Forecast Applications project in
Bangladesh (CFAB-I) was managed by the Program
on Atmospheric and Oceanic Sciences (PAOS) at
the University of Colorado and the Earthwork and
Atmospheric Sciences of the Georgia Institute of
Technology (EAS/GATECH), USA, which engaged in
research aimed at increasing the lead-time of
flood forecasting in Bangladesh. As the primary
implementing partner, the Asian Disaster
Preparedness Center (ADPC) identified broader
forecast application opportunities and sought
for the institutionalization of the project in
Bangladesh. The US Agency for International
Development’s Office of Foreign Disaster
Assistance (USAID/OFDA) supported the CFAB-I
project
Develop forecasting schemes for floods in
Bangladesh that would add predictive skills
to current efforts within Bangladesh
Develop resilient schemes that will be able
to take advantage of improvements in data
availability, predictive modeling, data
assimilation, etc. as they occur in the
future
Promote international cooperative efforts
that will provide data and tools necessary
for the prediction schemes, noting that
flood forecasting techniques require
substantial technological investment and
infrastructure
Develop an infrastructure within Bangladesh
that will make use of the forecasts and will
eventually own the prediction schemes
Work actively with partners in Bangladesh to
facilitate a rapid technological transfer
Infrastructure development.
A strong infrastructure has been set up
within Bangladesh to facilitate the
development of flood forecasting schemes and
their application. A Steering Committee,
consisting of Disaster Management Bureau (DMB)/MFDM,
Department of Agriculture Extension (DAE),
Flood Forecasting and Warning Center (FFWC),
Bangladesh Meteorological Department (BMD),
Center for Environmental and Geographic
Information Services (CEGIS), Institute of
Water Modeling (IWM), ADPC, and
CARE-Bangladesh was formed. The Steering
Committee meets regularly to guide project
implementation. At the international level,
collaboration has been forged by the PAOS at
the University of Colorado and the Earth and
Atmospheric Sciences (EAS) at the Georgia
Institute of Technology and the European
Center for Medium-range Weather Forecasts (ECMWF)
for sharing of forecast products.
Testing of long-term (1-6 months) forecast
schemes. Historical data
were used to initialize the long-range flood
forecasting scheme. Results of the test
indicate that major flood years could be
predicted months before their occurrence.
For example, the 1998 floods would have been
predicted three months before their
occurrence and hence giving sufficient lead
time for anticipatory actions.
Development of new medium-term (20-25 days)
forecast scheme.
A new
statistical scheme for the prediction of
rainfall and river discharge into Bangladesh
has been developed, providing potential
application for disaster management,
particularly for the management of floods
and drought.
Establishment of skill of short-term (1-6
days) flood forecasts. The
project has showed that short-term forecast
can be extended from the current 2 days to
nearly 8 days. This increase in lead time
will allow emergency planning and selective
planting or harvesting to reduce potential
crop losses at the beginning or end of the
cropping cycle.
Forecast application.
A
method to bridge the gap between producers
and users of probabilistic forecasts has
been developed through the generation of a
User Metric – a tool that provides an
aggregated risk analysis to aid user
community in making absolute decisions, e.g.
whether to harvest early to obtain 80% of
the potential yield or wait for two weeks
for 100% yield or entirely lose the harvest.
Delivery of experimental forecast during the
summer monsoon 2003. The
project performed and validated experimental
real-time long, medium, and short-term
forecasts of river discharge into Bangladesh
for the summer of 2003
