Operational Flood Forecasting for Bangladesh:
Tom Hopson, RAL-NCAR
Peter Webster, Georgia Tech
A. R. Subbiah and R. Selvaraju, Asian Disaster
Preparedness Centre
Climate Forecast Applications for Bangladesh (CFAB):
USAID/CARE/ECMWF/NASA/NOAA
Bangladesh Stakeholders: Bangladesh Meteorological Department, Flood Forecasting and Warning Center,
Bangladesh Water Development Board, Department of Agriculture Extension, Disaster Management Bureau,
Institute of Water Modeling, Center for Environmental and Geographic Information Services, CAREBangladesh
Climate Forecasting Applications for
Bangladesh (CFAB)
The CFAB project goals:
•
Develop resilient forecast schemes that capitalize on skillful
modeling techniques and advanced data sources at time-scales:
1-6 months, 20-25 days, 1-10 days (2000)
•
Develop an infrastructure within Bangladesh to:
a) make use of the forecasts -- establish pilot projects at selected
sites, showing measurable improvements (2006)
b) eventually own the prediction schemes -- facilitate a
technological transfer (2008)
NASA Aqua/Modis images
2004 dry season river flows …
… and during the July flooding event
Bangladesh background
About 1/3 of land area floods the monsoon rainy season
Size: sightly smaller than Iowa (144,000 sq km)
Border countries: Burma (193 km), India (4,053 km)
Population: 140 million
36% of population below poverty line
Within the top 5 of: poorest and most densely populated in the
world
Natural disasters:
Nov 1970 Bhola cyclone -- at least 300,000 died in 20 min
(12m)
April 1991 Bangladesh cyclone -- 138,000 died (6m)
River Flooding
Damaging Floods:
large peak or extended duration
Affect agriculture: early floods in May, late floods in September
Recent severe flooding: 1974, 1987, 1988, 1997, 1998, 2000, 2004, and 2007
1998: 60% of country inundated for 3 months, 1000 killed, 40 million homeless, 1020% total food production
2004: Brahmaputra floods killed 500 people, displaced 30 million, 40% of capitol city
Dhaka under water
2007: Brahmaputra floods displaced over 20 million
(World Food Program)
Overview:
Bangladesh flood forecasting
I.
CFAB History -- sea-level backwater effects
II. 1-10 day Discharge Forecasting
1. precipitation forecast bias removal
2. multi-model river forecasting
3. accounting for all error: weather and hydrologic errors
III. 2007 Floods and Warning System Pilot Areas
How important are (“forecastable”) interannual sea level variations on
country-wide extreme flooding events?
Answer: effects impact river heights roughly 200km upstream …
Upper-Catchment Flooding
disasterous flood year
“normal” flood season
Severe flood years
affect whole country,
with water depth
variations of O(1m)
=> Look at
precipitation-driven
effects on flooding
September 1998
August 2002 floods
CFAB Project: Improve flood warning lead time
Problems:
1. Limited warning of upstream
river discharges
2. Precipitation forecasting in
tropics difficult
Assets:
1. good data inputs: weather forecasts, satellite rainfall
2. Large catchments => weather forecasting skill “integrates” over large spatial
and temporal scales
3. Partnership with Bangladesh’s Flood Forecasting Warning Centre (FFWC)
=> daily border river readings
Merged FFWC-CFAB Hydraulic Model Schematic
Primary forecast boundary
conditions shown in gold:
Ganges at Hardinge Bridge
Brahmaputra at Bahadurabad
Transforming (Ensemble) Rainfall into
(Probabilistic) River Flow Forecasts
Rainfall Probability
Discharge Probability
0.2
0.35
0.18
0.16
0.3
0.14
0.12
0.25
0.2
0.1
0.15
0.08
0.1
0.06
0.05
0.04
0.02
0
0
1
2
3
4
Rainfall [mm]
5
6
0
10,000
30,000
50,000
70,000
90,000
Discharge [m3/s]
Above danger level probability 36%
Greater than climatological seasonal risk?
Daily Automated Operational Flood Forecasting Sequence
ECMWF 51-member Ensemble Precipitation Forecasts
5 Day Lead-time Forecasts
2004 Brahmaputra
Catchment-averaged Forecasts
Bias Adjustment
“Observed” Climatology
Model Climatology
Precipitation
Pmax
Pmax
Pfcst
Padj
25th
50th
75th 100th
Quantile
25th
50th
75th 100th
Quantile
Bias-corrected Precipitation Forecasts
Original Forecast
Corrected Forecast
Brahmaputra Corrected Forecasts
Daily Automated Operational Flood Forecasting Sequence
2003 Model Comparisons for the Ganges (4-day lead-time)
hydrologic lumped model
hydrologic distributed model
Resultant Hydrologic multi-model
Multi-Model Forecast
Regression Coefficients
- Lumped model (red)
- Distributed model (blue)
Significant catchment
variation
Coefficients vary with the
forecast lead-time
Representative of the each
basin’s hydrology
-- Ganges slower time-scale
response
-- Brahmaputra “flashier”
Daily Automated Operational Flood Forecasting Sequence
Significance of Weather Forecast Uncertainty
Discharge Forecasts
2004 Brahmaputra Discharge
Forecast Ensembles
7 day
9 day
8 day
10 day
Corrected Forecast Ensembles
7 day
3 day
5 9day
day
8 day
4 day
10 day
Step 1: generate discharge
ensembles from precipitation
forecast ensembles (Qp):
Probability
Producing a Reliable Probabilistic Discharge Forecast
1
PDF
1/51
Qp [m3/s]
Step 2: a) generate multi-model hindcast error time-series using precip estimates;
b) conditionally sample and weight to produce empirical forecasted error PDF:
forecast
a) 1000
b)
1
Residuals
PDF
horizon
[m3/s]
time
=>
-1000
Residual [m3/s] 1000
-1000
Probability
Step 3: combine both uncertainty PDF’s
to generate a “new-and-improved” more
complete PDF for forecasting (Qf):
1
Qf [m3/s]
2004 Brahmaputra Forecast Results
Confidence Intervals
2 day
Critical Q black dash
50%
7 day
3 day
59day
day
95%
8 day
4 day
Above-Critical-Level
Cumulative Probability
7 day
9 day
10 day
8 day
10 day
Overview:
Bangladesh flood forecasting
III. 2007 Floods and Warning System Pilot Areas
Five Pilot Sites chosen in 2006
consultation workshops based on
biophysical, social criteria:
Rajpur Union
-- 16 sq km
-- 16,000 pop.
Kaijuri Union
-- 45 sq km
-- 53,000 pop.
Average Damage (Tk.) per Household in Pilot Union
70,000
60,993
Average Damage (Tk) per
Household
Uria Union
-- 23 sq km
-- 14,000 pop.
64,000
60,000
50,000
40,000
28,745
30,000
20,000
10,000
7,255
4058
0
Uria
Gazirtek
Kaijuri
Rajpur
Union
Gazirtek Union
-- 32 sq km
-- 23,000 pop.
Bhekra Union
-- 11 sq km
-- 9,000 pop.
(annual income:
30,000 Tk; US$400)
Bekra
2007 Brahmaputra Ensemble Forecasts and
Danger Level Probabilities
7-10 day Ensemble Forecasts
7 day
9 day
7-10 day Danger Levels
8 day
7 day
8 day
9 day
10 day
10 day
Response of National Institutions for 2007 flood forecasts
•
Flood Forecasting and Warning Center (FFWC) incorporated the
CFAB forecasts to produce water level forecasts for many locations
along Brahmaputra and Ganges well in advance
•
National level Disaster Emergency Response Group prepared
emergency response plans, logistics for preparedness and relief in
advance
Selvaraju (ADPC)
Response of local institutions for 2007 flood forecasts
•
Local project partners used community vulnerability
maps to assess the risk of flooding
•
Local NGOs and CBOs mobilise boats to rescue people
and livestock from the “char” areas
Selvaraju (ADPC)
Community level decision
responses for 2007 flood
forecasts (Low lands)
•
Secured cattle, poultry birds, homestead vegetables,
protected fishery by putting nets in advance
•
Planed to evacuate and identified high grounds with
adequate communication and sanitation facilities
Community level decision responses
for 2007 flood forecasts (High lands)
•
Protected homestead vegetables by creating adequate
drainage facilities
•
Livestock was protected in high lands with additional
dry fodder (paddy straw)
•
Early harvesting of B.aman rice and jute anticipating
floods in Gaibandha and Sirajganj, respectively.
Selvaraju (ADPC)
2007 ADPC Warnings issued …
“We were able to inform the people in advance and on 25th July we started communicating the information to as many
people as possible about the certainty of exceeding danger levels along the Brahmaputra…The local partners, nongovernment organization (NGO) networks and DMC members were advised to inform the poorest of the poor, especially
those people living in river islands (“chars”)...”
“On the 28th and 29th, meetings were organized in villages near Rangpur (northern Bangladesh), where the Teesta River was
flowing just a few inches below the rim... However, they perceived that the river water level would fall, but our forecasts
showed a rising trend…We informed them the significant chance of overflow and breaches, as the embankments are weak in
certain places. We engaged the local partner NGOs to prepare an evacuation plan urgently…”
“We communicated the forecast to another pilot union DMC chairman (Uria Union in Gaibandha District) directly on July
26th so he could arrange to mobilize resources for evacuation through DMC members and volunteers. All the six villages in
the union were later flooded to a height of 4-6 feet on July 29th. We contacted him again on the 29th to know more and he
informed us that about 35% of the people in the union were evacuated in advance.”
“The communities in Rajpur Union of Lalmunirhat (relatively medium lands), were able to use the forecast for preparedness
activities like mobilizing food, safe drinking water for a week to 10 days, protecting their T. aman rice seedlings, fishing nets,
and raising and protecting their fish pods.”
“For the first time we have communicated 10-day in-advance official forecasts of significant chances of exceeding danger
level in all the gauge stations along the Brahmaputra River through the FFWC (and not just in the pilot regions) and directly
through all local NGOs, and the DMC members obviously had 8 days extra lead-time they otherwise would not have had.”
Conclusions
2003: CFAB forecast went operational
2004:
-- Forecasts fully-automated
-- CFAB became an entity of Bangladesh government
-- forecasted severe Brahmaputra flooding event
2006:
-- Forecasts incorporated into operational FFWC model
-- 5 pilot study dissemination areas trained
2007: 5 pilot areas activated during two severe flooding
events
Future Work
Dartmouth FloodWatch Program river discharge
estimates
Improved river routing
Fully-automated forecasting scheme applied to other
river basins in Africa (and elsewhere)
Climate change impacts …
Thank You!