I met Mohammad Jufran first time in 2014 while conducting a baseline survey at Purbogram under Nikli Upazila in Kishoreganj. The small farmer was full of energy and enthusiasm at his 28. He owned two hectares of land that he received from his ancestors. He borrowed money and took lease two hectares more for cultivating Boro. His wife also didn’t sit idle. She reared duck, a suitable income generating activity in haor. The couple with two children was dreaming of a better future for them ahead. After the severe flash flooding in April this year I found a different Jufran with no lights in his eyes. It damaged his standing crops. It also killed 100 ducks his wife was rearing. He is now facing troubles with his cattle. He can’t feed them. Tearful Jufran informed this was not the first time he lost everything. He was also devastated by flash floods in 2010. He had to struggle hard to overcome that loss. He will have to go through the same situation again!
The recent flashfloods inundated vast tracts of land in the Northeastern wetlands in Bangladesh and jeopardises lives of thousands of farmers like Jufran. The Department of Agriculture Extension (DAE) estimated that it destroyed more than 200,000 hectares of mature Boro paddy (a rice variety cultivated from December to May). Another significance of this event is that it came much earlier than usual. Does it indicate a sign of climate change?
Flooding that begins within six hours, and often within three hours, of a heavy rainfall (or other cause) is considered flashflood which is one of the main natural disasters of the Northeast region of Bangladesh. Every two or three years’ interval, flash flooding devastates the region. It affects the primary production sector of agriculture and thus threatens the lives and livelihoods of the inhabitants.
The ‘haor’, a very low lying river basin area, is covered by water almost a half of the year starting from the monsoon season. Most of the lands here remain inundated six to seven months from June to November. People here are used to living in this condition although it causes numerous sufferings to their lives. It almost confines them to their small homes, limits their mobility, and practically eliminates the possibility of finding jobs, if not in the fields in other regions of the country
 Geographically, most of the haors are situated in seven districts, Sunamganj, Kishoreganj, Netrokona, Sylhet, Habiganj, Maulvibazar and Brahmanbaria of the North-East Bangladesh. These haors are an important source of agricultural production for a large number of vulnerable people. During winter, cultivated land of the area produces paddy with minimum efforts while during monsoon the same is turned into breeding place for open water fishery, hosting a wide range of water biodiversity. Haors are important areas for Boro cultivation which is a large mono-cropping agriculture system here. Almost 80% of these haor areas are covered by Boro rice while 10% area is covered by T-Aman production including other hybrid rice varieties. . Therefore, a dedicated early flashflood warning system is not only urgent for the survival of the people living here but also necessary for saving the economy of the country.
Flash floods generally occur in the north-eastern haor and southeast (Chittagong) regions of Bangladesh. However, the northeast region of Bangladesh is highly vulnerable to flash floods due to surrounding hilly areas and presence of numerous large, deeply flooded depressions. When heavy rainfall occurs in the Assam and Meghalaya hilly region of India, flood water quickly moves towards the haor area of Bangladesh through the trans-boundary rivers and eventually enters into the haor. In most of the cases, farmers don’t get enough time to harvest their standing Boro crop. Thus flash floods before harvesting of Boro crops not only create negative impact on the food security of the northeast region but also damage the national economy to some extent.
Flood Forecasting and Warning Centre (FFWC) under Bangladesh Water Development Board (BWDB) issues flood warnings by simulating a numerical model of the Bangladeshi river network named as “Super Model”. But the fact is this model is not capable to issue a strong and dedicated early warning system of flash flood with a considerable leading time for haor region. Furthermore, it resumes its operation during monsoon in June. Therefore, farmers of haor don’t get any warning in April-May when they need most to protect their harvest.
Currently ‘Climate Adaptation and Livelihood Protection (CALIP)’, a supplementary project supported by International Fund for Agriculture –IFAD has been working for developing a community based early flash flood warning system (FFEWS) in the Sylhet Basin. Under this project FFWC, Institute of Water and Flood Management (IWFM) at Bangladesh University of Engineering and Technology (BUET), Institute of Water Modeling (IWM) and Bangladesh Meteorology Department (BMD) are working together. But still it is in experimental stage.
There are a number of challenges for developing an early flash flood warning system in Bangladesh according to experts. Professor AKM Saiful Islam, who is also the Principal Investigator of this project at IWFM, identified the first challenge is inadequate data. The existing system relies heavily on the information of flood water coming from the upstream countries. About 93 per cent of the catchment areas of the major rivers – the Ganges, the Brahmaputra and the Meghna lies outside the country. Unfortunately, these river data are not available. Existing model uses data from Global Forecast System (GFS). Inside the country BWDB has 79 stations across the country including 18 stations in haor region. But only eight stations in the area are currently providing monsoon flood forecasting. However, BWDB will increase its number of forecasting stations from 8 to 25 in Sylhet basin under CALIP.
Regarding weather data Bangladesh Meteorology Department (MBD) is also not capable to generate pre-monsoon RADAR data which is very important to catch the flash flood event. One of the reasons behind this is Moulvibazar Doppler RADAR station works few hours of the day. BMD has shortage of manpower to keep RADAR stations active round the clock. Uninterrupted electricity supply is another issue according to the BMD.
Forecasting floods using numerical models such as hydrological and hydrodynamic models also require several geospatial and hydro-meteorological data. At present, the only source of these vital information is collected by the satellite based remotely sensing which suffers from inaccuracy and inconsistency.
Setting up danger level is another challenge for flash flood forecasting. Haor region is different from other parts of the country. There are a number of submergible dams here. Danger level of the rivers here should be set differently according to a recent study conducted by IWFM. It recommended additional danger level for the pre-monsoon flash flood forecasting for haors.
During the last one year IWFM tested several models for identifying potentials and limitations of existing models. Incorporating all the findings it has already developed a model which is at experimental stage. It can provide warning with 90 percent accuracy two days before it occurs, with 91 percent accuracy one day before and with 93 percent accuracy 12 hours earlier. Generally, a standard flash flood forecasting system is capable of forecasting three to five days earlier the flood occurs. Earlier than five days increases error according to the experts.
IWM also improved another model and started experimental forecasting in its website. After comparison of two systems FFWC would adopt the most appropriate one which would warn the farmers at least 10 days earlier (probabilistic) and confirm three days earlier (deterministic) about the flash flood. It will cover five haor districts of Netrakona, Habiganj, Brahmanbaria, Kishorganj and Sunamganj. FFWC expects to launch the EFFWS by 2019. If it succeeds no more Jufran will lose hope.

The writer is Communication & Knowledge Management Specialist, International Fund for Agricultural Development