Strengthening the capacity of satellite-based flood forecasting using near real time Jason-2 satellite altimeter data under SERVIR-Himalaya Small Grants Programme, ICIMOD

TheInstitute of Water Modeling (IWM) in Bangladesh has developed a cost-effective information system that provides 8-day flood forecasting. The system provides near real-time information through web GIS with local-level flood forecast maps for flood events. The Jason-2 satellite altimeter data collection system enables the forecasting of trans-boundary flows in downstream countries sharing international river basins. The study was carried out in partnership with the Tennessee Technological University (TTU) and supported by ICIMOD under the SERVIR-Himalaya Initiative’s Small Grants Programme.

Bangladesh’s location makes it extremely vulnerable to natural disasters. Huge cross-border inflows, low-lying topography, and the deposition of sediment in rivers have made Bangladesh highly vulnerable to floods. Flood forecasting is an important tool for reducing damage to existing resources and livestock.

Flood forecasting in the downstream part of any hydrological basin is extremely difficult due to lack of basin-wide hydrological information, unavailability of near real-time data, and the absence of a data sharing treaty among countries sharing trans-boundary basins. Furthermore, the accuracy of forecasts emerging from hydrological models can be compromised if there is no knowledge of the day-to-day flow regulations at different locations upstream of Bangladesh. Only satellite altimeter monitoring allows the identification of actual river levels upstream that reflect the human intervention at that location. Such information is critical for accurate downstream forecasts.

The project used a number of virtual stations, a combination of satellite estimates of water level, and modelling, which provided a forecast of stream flows in Bangladesh. Jason‐2, a joint NASA‐French Satellite mission that measures water levels and provides data relatively promptly (within 24-hours), was used for this purpose. Seven virtual stations at suitable locations along the satellite passing path were selected on the desired river reach.

Forecasting rating curves (FRC) were developed in correlation to observed river-level data at Bahadurabad and Hardinge Bridge, with virtual station data collected over the period 2009–2013. The FRC were used to forecast water levels for 2014 at Bahadurabad and Hardinge Bridge. The 1D HEC-RAS hydraulic model forecasts water levels on all river cross-section locations on the major rivers in Bangladesh. ArcGIS extension HEC-GeoRAS is used to delineate the inundation map on fine resolution digital elevation model (DEM) for Kulkandi Union. The boundaries of homesteads and the alignment of local rivers in and around the project area are shown in these inundation maps.

The project also carried out a field study to identify infrastructure such as schools, roads, agricultural fields, rural businesses, and market places located in flood prone areas. The project identified five flood prone areas in Kulkandi, where the majority of people depend on agriculture, livestock, and fishing for their livelihood. A fisherman from Kulkandi, Nurul Islam, described how the flood forecasting system installed in his village could improve his livelihood: “If we receive an early warning about a flood, we can take timely action to save our paddy and livestock.” Farmers and residents of Islampur upazila, Jamalpur expressed similar views.

During the two stakeholder consultation meetings organized to discuss the project’s findings, a commitment was solicited from members of non-government organizations and government officials to further improve the forecasting system and disseminate information to members of communities living in flood hazard zones. These meetings were attended by representatives of the Bangladesh Water Development Board (BWDB), Flood Forecasting and Warning Centre (FFWC), Water Resources Planning Organization (WARPO), Bangladesh Agriculture Research Centre (BARC), and I/NGOs such as Oxfam, Islamic Relief, and Regional Integrated Multi-hazard Early Warning System (RIMES). During the meetings, the satellite-based forecast system with inundation maps and web-GIS portal prepared by the project was presented and explained to the participants. “It is important for communities to know how vulnerable they are; we will disseminate the findings among communities to help them take necessary precautionary actions”, said a representative from Islamic Relief.  “We can improve and replicate this method in other unions in Bangladesh”, said Md Amirul Islam, Executive Engineer of FFWC. Representatives from NGOs also agreed to cooperate in disseminating flood information to villagers.

In addition to developing the tool itself, the project was successful in educating communities about the satellite-based forecasting approach. Farmers like Tajul, Abdur Rob, and Razzak of Kulkandi were astonished to find their village identified on the inundation maps as vulnerable with the extent and depth of possible flooding shown. The people of Kulkandi are now better informed and willing to participate in the forecasting process. After introducing a flood forecast system and learning about inundation depth and extent in maps at an early stage, stakeholders now understand the importance of preparedness measures and how to safeguard their homesteads, including by moving household goods, harvesting crops early, and protecting fish by covering fish ponds with a net during floods.