Bangladesh Flood Forecasting and Warning Center adopts satellite-based flood forecasting system developed by SERVIR

Published: Jun 25 2014

Bangladesh, barely larger than the state of New York, is home to almost 160 million people. This densely populated country sits in the "sink" of the second largest river basin in the world, the Ganges-Brahmaputra-Meghna (GBM) river basin. And like a plugged sink drain, this country collects much of the water flowing in the three large rivers meandering toward the Bay of Bengal. Usually, by the end of the monsoon season (June through September), about one third of Bangladesh is under water.

Bangladeshi farmers rely on these seasonal floodwaters to prime their fields for rice crops by depositing rich soil washed down from the Himalayas. But the effects of climate change, including altered rainfall patterns, increased glacial melt, and other factors, are disrupting formerly predictable flood cycles. Floods destroy homes, decimate crops, and drown livestock. Thousands of human lives are lost.

The Flood Forecast and Warning Center (FFWC) in Bangladesh has adopted a new flood forecasting system for the 2014 monsoon season. The system was developed by SERVIR, a joint development initiative of NASA and USAID that helps developing countries use information provided by Earth observing satellites to address environmental issues, including natural disasters.

flooded house  

Flooded house south of Dhaka, Bangladesh (23°21’ N, 90°31’ E).  © Yann Arthus-Bertrand

"FFWC is like the River Forecasting Center of the National Weather Service in the U.S.," explains SERVIR Science Lead Ashutosh Limaye. "They provide data that is turned into flood warnings, evacuation notices, and the like. The FFWC-generated data is sent by disaster management agencies via sms text, radio, TV, daily bulletins."

However, current flood warnings in Bangladesh are issued only 3 days in advance.

"That's not enough time for families and farmers to prepare," says SERVIR Applied Sciences Team Member Faisal Hossain.

Longer lead times are needed for more accurate and timely warnings about rivers swelling ominously upstream during monsoon season. Unfortunately, although the rivers that feed the flooding in Bangladesh originate far beyond the country's borders, upstream countries do not share their data.

Hossain and his team of remote sensing experts, with active support from the SERVIR project and the International Centre for Integrated Mountain Development (ICIMOD), developed an innovative, satellite-based solution. In a nutshell, here's how the sytem works:

"A radar altimeter on Jason-2 measures the precise distance between the satellite and the river surface at points where the satellite crosses overhead," explains Hossain. "This data, available almost immediately, reveals the river's height at the point of crossing, so flood risks downstream can be assessed realistically for large rivers."

Jason-2, with its eye on headwaters, tracks levels of the Ganges and Brahmaputra Rivers more than 600 miles upstream of Bangladesh and produces daily 8-day flood forecasts of water levels for several water stations. (For the FFWC website, click here. For the prototype, click here.)

"Estimation of upstream transboundary flow river levels beyond the country's borders is a major issue in Bangladesh," says Arifuzzaman Bhuiyan of the FFWC/Bangladesh Water Development Board. "Until now we haven’t had a reliable boundary flow condition forecasting tool …. I feel this technique may have strong potential in improving the capacity of our existing flood forecasting system, especially to forecast transboundary flow.”  

Screenshot of river level data available at the FFWC website

Based on the JASON-2 data, FFWC can recognize impending floods and alert the public 8 days sooner than they could previously. Earlier warnings allow more time for the public to prepare for rising water or evacuate an endangered area. Such notice also helps farmers make a proactive decision on early harvest or delayed sowing.
"The JASON-2 system, which is simple and efficient, affords the longest lead time that FFWC has ever been able to produce," says Hossain. "The technique requires just 30 minutes of processing time -- mostly data downloading and height extraction -- and reading forecasts off a chart."

Thanks to SERVIR-provided training, FFWC mastered the use of this system and became completely independent in using the satellite technology and the processing tools.

"I think this sends a strong and encouraging message out to similar agencies with similar flood forecasting needs such as in Mekong, Indus, Niger, Zambesi and other large river basins," says Hossain. "We want such satellite systems to be sustainably adopted by as many stakeholder agencies as possible, and we will do everything we can to make sure it happens by transferring the ownership."

For the 2014 monsoon season, which begins mid-June, the JASON-2 system in Bangladesh will be called ‘experimental’.

"FFWC is generating warnings at several locations inside Bangladesh," explains Hossain. "They are all being archived for a post-flood season assessment."

After the system performance evaluation is completed, FFWC hopes to launch a full scale forecasting system for the public for 2015.
"It is critical that the currently prototyped system goes through a post-flood season performance evaluation so that various logistic issues that compromise forecast skill are addressed," explains Hossain. "Any needed corrections or adjustments will be identified and made so that the system can be comprehensively and effectively launched in 2015, providing the public information of the coming river-level scenarios 4, 5, 6, 7 and 8 days beyond the traditional 3 day forecasts."

Hossain believes that while this is an exciting first step towards ‘empowerment’ of an agency in a developing country, a lot more work remains to be done to make satellite missions deliver benefits on a more global basis. By 2020, it is expected that at least five satellite altimeters will be flying as a constellation in various orbits, providing an unprecedented global view of Earth’s surface water to enhance the capacity of conventional forecasting systems. These missions are AltiKa (launched in 2013), JASON-3 (launch date 2015), Sentinel 3A and 3B (launch date 2015), and ICESat-2 (launch date 2017). Hossain and his team, with support from NASA Applied Sciences Program, are already working on adding AltiKa to the JASON-2 forecasting system for FFWC. The long-term plan is to globalize the access to this satellite application concept for other agencies worldwide for their specific forecasting needs.


Dr. Hossain's team members: Dr. C.K. Shum, Ohio State; Dr. Francis Joseph Turk, JPL ; Dr. Hyongki Lee, University of Houston;  and Dr. Sylvain Biancamaria, Laboratoire d'Etudes en Géophysique et Océanographie Spatiales (LEGOS), France

The International Centre for Integrated Mountain Development (ICIMOD), SERVIR's partner organization in the Hindu Kush-Himalaya region, supported the development of the JASON-2 flood forecasting system.

The SERVIR Coordination Office played a highly supportive role in the adoption of the Jason-2 data for routine flood forecasting.
The US Department of State provided support to Hossain during the inception stages in 2013 with a Fullbright Faculty grant.

Related Articles