Bangladesh has come up a new flood forecasting and warning system with several amazingly useful features of forecasting floods available on their website from June 2013. Flood forecasting is a vital non-structural measure to mitigate flood losses which can be very useful for a deltaic nation like Bangladesh which face brunt of floods annually. The Flood Forecasting and Warning Centre (FFWC) is under the aegis Bangladesh Water Development Board and is supported by UNDP through the Comprehensive Disaster Management Programme (Phase II) and Ministry of Disaster Management and Relief. The website can be accessed at http://www.ffwc.gov.bd/index.php.
While the features have become available on website since June 2013, many of them have been otherwise available since 2011. In fact, Interactive Voice Response [Calling from mobile (phone no- 10941), flood and weather messages can be heard in Bangla (charge applicable)] facility is available since 2011.
The Deltaic Bangladesh Geographically, Bangladesh is country located in one of the biggest active deltas in the world with an area of about 1,47,570 sq km. The climate of the country is subtropical-monsoon climate where annual average precipitation is 2300 mm which varies from1200 mm in the north-west to over 5000 mm in the north-east. Bangladesh, as stated in its Annual Flood Report of 2012, has a total of 230 rivers out of which 57 are Transboundary Rivers. 54 of these transboundary river flows from India to Bangladesh which includes the Ganges and the Brahmaputra. Bangladesh consists of flood plains of the Ganges, the Brahmaputra and the Meghna rivers and their numerous tributaries and distributaries. The FFWC carries out monitoring of 86 representative water level stations and 56 rainfall stations across Bangladesh.
The FFWC Website The homepage of the FFWC website presents a map of the whole country and its rivers, marked with flood forecasting sites in Bangladesh which is really useful for a number of reasons:
(i) For a first time visitor it gives a good idea of the rivers in Bangladesh and their flood forecasting sites.
(ii) When the cursor stops at a particular forecasting site, a pop up site appears with location details, water level, highest water level and danger level at that particular site.
(iii) with locations of forecasting sites on the same river clearly marked, one can get an overview of the situation across the river basin.
(iv) colour code of the symbol at each FF location gives an idea if the water level is above/ below warning or danger level or High Flood level.
The homepage of FFWC offers an option to view the whole website in Bangla. This is very welcome and useful since this important information can be accessed and used by people in local language.
The home page also provides information on the sites where water level is currently on the rise or above the danger mark. Below the line of tabs of the website, name of the sites where water level is on the rise keeps on scrolling. Clicking in any of the sites in the scroll, leads to a pop-up where the rise in water level is presented in graph for the previous one week. The rise in water level for each day is also showed according to time which appears when the cursor is kept on any of the dots of the graph. This pop up can also be saved either as a photo, as a pdf or as a SVG vector image.
The FFWC website provides rainfall data and water level data of every designated site in the four river basins of Bangladesh under the ‘data’ tab: Bramaputra (13 sites), Ganga (17 sites), Meghna (15 sites) and South East Hills (10 sites). The rainfall data table gives rainfall of the day, along with previous two days rainfall, normal monthly rainfall and cumulative rainfall for the month. But here FFWC can add the total rainfall data for the whole season which will make it more useful.
The water level data table provides water level of the day, previous day water level and projection of water level for the next three days along with name of the river and location name for 38 sites in Brhamaputra basin, 27 sites in Ganga basin, 26 sites in Meghna basin and 7 sites in South East Hill Basin.
The FFWC website under the ‘Forecasting and Warning’ tab provides very substantial information regarding floods under following heads: ‘Flood Summary’, ‘Flood Bulletin’, ‘3-day Deterministic Flood Forecast’, ‘5-day Deterministic (experimental) Flood Forecast’, ‘Medium Range (1-10 Days)Forecast’ ‘Structure Based Forecast’ (available for a few structures like embankments or bridges on experimental basis) and ‘Special Outlook’ (only in Bangla language).
Under the ‘Map’ tab, the website has ‘Rainfall Distribution Map’ and ‘Inundation Map’. The rainfall distribution map[i] provides at a glace picture of rainfall over the last 24 hours. The latter provide the options to view inundation maps[ii] for the given day and forecasts for next two days, both for Dhaka and for Bangladesh as a whole.
The ‘Reports’ tab holds annual reports on floods in Bangladesh for last five years. In the ‘Hydrograph’ tab, the website first provides of forecast of water level for each site through a graph. Under ‘Forecast’ the website provides observed water level at each selected site for the day, previous six days (with all the observations available on the graph) and forecast for the next three days. This is provided for 89 sites which include 8 sites from South East Hills basin, 26 sites in Meghna basin, 21 sites in Ganga basin and 34 sites in Brahmaputra basin. Division wise break up is: Borisal-1; Chittagong-11; Dhaka-30; Khulna-6; Rajshahi-14; Rangpur-12; Sylhet-15. It also provides monsoon hydrograph for 104 sites, where the water level throughout the monsoon is given, some of the graphs seems to show flat levels, though, raising questions if these are providing correct information.
Here for some sites several selected years’ monsoon season data is also represented through graph for some of the sites but it becomes bit confusing since no rational for choosing a particular year is provided. Providing the basis for choosing a particular year monsoon will clarify the picture. In both ‘forecast’ and ‘monsoon’ options the sites are presented not only river basin wise but also on the basis of regions making it more easy to use. Hydrograph also includes ‘Real Time Data’ of three sites. The real time data of these three sites presents the water level from one day to last 60 days through graph.
GoI-CWC’s Flood Forecasting website If we compare this with the flood forecasting website of Government of India’s Central Water Commission (http://www.india-water.com/ffs/index.htm), the CWC site home page has two options ‘list based’ and ‘map based’ but the latter surprisingly never worked for this whole monsoons season, it has never worked, it seems. Clicking on the map based option leads to page which asks for plug in download. There is also an option for map browser installation but that never worked.
The ‘list based option’ has the flood forecasting sites under three options – state, basin and region-wise. But there is no page where all the flood forecasting sites can be found at one place. CWC flood forecasting site provides forecasts available at a given point of time, only for one day and which is removed as soon as the time of forecasting gets over. The forecasting in table format is made only for those sites where the water level has crossed the warning level. The CWC flood forecasting website does not provide three day, or five day or ten day forecasts. This is in contrast with the flood forecasting done by Bangladesh because the website provided water level data for each site in the country in one place and it is fast, more responsive and user friendly. Besides, on CWC site there is no option of keeping a record of previous forecasts. All the forecasting data available in the CWC website is available in English and not even in Hindi.
In case of CWC, there were also instances when water level even after crossing the danger level did not appear in the flood forecasting table. This has been observed for example in case of the Sahibganj site on the GangaRiver. Sahibganj did not appear in the flood forecasting table even for once even though the site was witnessing low to moderate floods as shown in the list base option of the same website.
There are also several other issues with the CWC flood forecasting website e.g. wrong flood forecasting, no forecasting of the floods during the Uttarakhand disaster and inadequate flood forecasting sites which had been pointed out to CWC by SANDRP[iii].
CWC need to do serious homework to improve its performance in flood forecasting for and in doing so it can take some good lessons from Bangladesh.
Extending the forecasts to GBM basin There is a huge scope for Transboundary cooperation in extending the flood forecasting across the countries sharing Ganga-Brahmaputra-Meghna basins, including Bangladesh, India, Bhutan, Nepal and China. In fact Md Amirul Hossain, Executive Engineer at FFWC wrote to SANDRP in response to a question, “The FFWC of Bangladesh has been receiving little (rather very little) upstream data/ information on water level or flood situation. This little data helped us very much for generating everyday flood forecasting. If we would be in a position to share hydro-met data of the upstream, most of within India, few in Nepal & Bhutan, then the Flood Forecasting and Warning Lead time could further be increased/ extended. FFWC-Bngladesh likes to share the information, skill and experience and would be happy to participate in an effort to develop “BASIN FLOOD FORECASTING AND WARNING” for Brahmaputra-Ganges and Meghna basin (GBM-Basin).” It is heartening to see this offer from FFWC, Bangladesh. We hope steps will be taken to realize this potential and more and transparent sharing of flood forecasting information across the countries will be a reality soon.
Himanshu Thakkar and Parag Jyoti Saikia
5 thoughts on “Bangladesh’s Amazing and New Flood Forecasting: A Tip for India’s CWC to Improve its Flood Forecasting Performance”
This is excellent stuff. CWC must learn from it.
Cooperation on water can enter into meaningful stage with a basin level flood forecasting model.