© Fabeha Monir pour Action contre la Faim
Headline
SURF-IT: Introducing an AI Solution to Enhance Disaster Preparedness in Bangladesh
Bangladesh has long been battling the negative effects of climate change. While eastern areas suffer from massive floods, especially during the rainy season, the coastal areas along the Bay of Bengal experience devastating cyclones and tidal surges, which put communities at high risk.
While the country has made significant progress in saving lives through early action programs, such as disseminating information, safeguarding material assets and minimizing losses, it remains a work in progress. Current cyclone forecasting models only cover cyclones and storm surges — they don’t predict tidal surges during non-cyclonic periods. To address this gap, Action Against Hunger launched a SURF-IT pilot project in collaboration with Uttaran, its strategic partner and project lead, and North South University in Dhaka*. The goal is to develop a surge forecasting model utilizing artificial intelligence (AI) and machine learning that can provide early warnings and prompt timely actions.
Rethinking Forecasting in Disaster-Prone Areas
Satkhira sits thousands of kilometres from Dhaka and is one of the most climate-vulnerable areas in Bangladesh. Located near the Bay of Bengal, its coastal areas are frequently hit by devastating tropical cyclones and storms. The tropical cyclone warning system alerts communities through mobile networks, the internet, and radio broadcasts. However, the early actions warning system is still developing, and people continue to face displacement, material losses and multiple invisible impacts on their health and well-being.
Coastal Satkhira is protected by a network of embankments of various sizes and lengths. These structures connect villages and roads, while helping to keep floodwaters out. “In Bangladesh, most of the embankments are made of earth”, explains Nibraz Bahar, research assistant, Uttaran. “These embankments were built between the 1960s and 1980s, and every 3 to 4 years, we see them deteriorate, with erosion and occasional breaches”. The embankments were originally designed to protect tidal floodplains, enabling agriculture and households to be established safely within them. However, it didn’t take into account tidal or storm surges, which sometimes reach up to 8 meters high in this area, or breaches in the embankments.
Sumon Homaun Kabir, Program Manager at SURF-IT, recalls the early discussions that led to the project resulting from these observations and findings: “Unlike other initiatives that rely on existing cyclone forecast models for information sharing and early action, SURF-IT stands out. It aims to develop a surge forecasting model powered by artificial intelligence — a first-of-its-kind approach in Bangladesh”.
© Fabeha Monir pour Action contre la Faim
© Fabeha Monir pour Action contre la Faim
© Fabeha Monir pour Action contre la Faim
Identifying Weak Spots Through Drone Mapping
One of the key innovations of the project is the development of a 3D map of the embankments in the study area that will enhance the accuracy of identifying weak embankments. “The embankment isn’t the same throughout — its shape and height vary in different places”, explains Nibraz Bahar, in charge of the drone footage for the SURF-IT project. “We’re working on creating a 3D map of the embankment using a LiDAR drone. The drone captures a digital elevation model, which we feed into the artificial intelligence system.”
The drone enables the creation of a detailed 3D map of the entire embankment area, capturing variations in height and structure. This high-resolution mapping helps the team better understand the condition of the embankments and identify potential weak points. “In addition to the mapping, we conduct soil testing on the embankments to assess their strength. We also use water sensors to predict the height of incoming surges.”, shows Nibraz. The team also collects water velocity data from international dashboards, helping to complete the analysis.
Each drone flight captures approximately 200 images, generating around 5 gigabytes of data. All this information is then transferred to the research cell at North South University in Dhaka, where it is processed and analyzed using artificial intelligence to support the development of the surge forecasting model.
When Local Knowledge Meets Science
From the very beginning of the project, it became clear that developing an effective forecasting model would not be possible without the valuable support of local communities who are directly affected by these disasters. For Prof. Jakariya, who works at the Department of Environmental Science and Management at North South University, this became an important pillar of the data collection. “People have been facing climate disasters for generations, and over time, they’ve developed remarkable local knowledge”, says Prof. Jakariya, who brought together researchers from various disciplines to work on this project. “We spoke with the community to identify weak points in the embankment and understand why they’re fragile — whether it’s water pressure or other factors. That insight comes directly from them”, he explains. That’s why scientists work closely with humanitarian and local organizations to gather all relevant data and ensure no critical factor is overlooked in the research.
Prof. Jakariya is confident in the project’s importance and potential, which is set to move into the design phase early next year. “At the start, I suggested incorporating machine learning to make the concept even more powerful, as it offers great potential. However, I would describe this project more as community science-based, which will ultimately help us develop a highly accurate warning system for local communities”.
Artificial intelligence is a key component of the project, used to develop a machine learning algorithm that analyzes large volumes of data from sensors, drones, soil tests, and river mapping. Following a series of tests and simulations, this pilot project will be integrated into the national forecasting system. From there, it will be deployed to other regions of the country, helping to enhance disaster preparedness on a larger scale.
*SURF-IT: Spatial Surge Forecasting Using Artificial Intelligence and Community Knowledge for Inclusive and Transformative Early Actions (financed by FCDO through International Development Research Centre – IDRC).
Bangladesh
Related news
All the news of our Action: articles, events, testimonials, press releases…
