Technical assistance combined open data, community mapping and participatory modelling to develop urban flood risk models for Dar es Salaam, Tanzania, strengthening local capacity and enabling more accurate, locally owned flood hazard information in a data scarce environment

Dar es Salaam, Tanzania’s largest city and economic centre, faces high and growing vulnerability to pluvial and fluvial flooding, driven by rapid, unplanned urbanisation and a chronic lack of formal data to schematise and calibrate hydrodynamic flood models. To address this gap, the “Participative Open Data and Model for Flood Risk Assessment” project was carried out under the World Bank’s Global Facility for Disaster Reduction and Recovery (GFDRR) Challenge Fund, building on the existing Ramani Huria (“Dar Open Map”) community mapping initiative and the OpenStreetMap (OSM) platform.

In its first phase, the project developed PARTERRA GEE, a first of its kind terrain synthesis tool that fuses OSM mapped features (roads, waterways, culverts, buildings) with freely available SRTM elevation data to produce high resolution (sub 1 metre) digital terrain models suitable for urban pluvial flood simulation, an approach that depends entirely on data and local knowledge contributed by citizen mappers. The tool was piloted in Dar es Salaam using the 3Di hydraulic model and InaSAFE impact simulations.

Building on this, the project developed and applied a novel participatory modelling and mapping framework, piloted in the flood prone Manzese ward, that integrates community-based data collection directly into the development of a coupled 1D 2D urban flood model (D Flow FM), rather than only involving stakeholders after a model is built. Twenty-five participants from 13 organisations, including community mappers, disaster managers, university researchers and government officials, were engaged through a circles of influence structure and a three-day stakeholder workshop covering data evaluation, model building exercises and field reconnaissance surveys. This iterative process of community mapping, data collection, quality control and model refinement improved the accuracy of the flood model, corrected mapping errors such as non-existent or missing drainage channels, and validated assumed river cross sections against ground conditions.

The project contributed to strengthening participatory, open data-based flood risk management in Dar es Salaam by:

• Developing PARTERRA GEE, an open source terrain synthesis tool fusing OSM and SRTM data for high resolution urban flood modelling, made available via GitHub.
• Designing and testing a new iterative framework for participatory urban flood modelling, integrating community mapping directly into model development and improvement, in the Manzese ward pilot.
• Building local technical capacity through stakeholder workshops, training mappers, modellers and community members in flood processes, OSM data collection and hydrodynamic model use.
• Establishing a 1D 2D flood model (D Flow FM) for Manzese ward, improved through field validated community data, demonstrating measurable gains in model accuracy.
• Assessing cascading impacts of critical infrastructure failure, such as loss of evacuation routes, water points and power, on communities, using the Deltares Circle tool with Red Cross volunteers.
• Preparing guidelines for further OSM based data collection to support future automated flood model schematisation.