Resilience Scanner

UAV LiDAR and Photogrammetry for Flood Modeling Digital Terrain Models

Solution Overview

A World Bank-funded comparative study evaluated UAV photogrammetry and UAV LiDAR technologies for generating Digital Terrain Models to support flood modeling in the Msimbazi River basin.[1][2][3] This rigorous comparison aimed to provide guidance on when to apply each acquisition method based on project requirements, area characteristics, and operational constraints.[3][4] The initiative served dual purposes: producing reliable bathymetric and topographic data for riverine flooding studies while systematically comparing drone-based surveying technologies in a Sub-Saharan African context where such meticulous analysis had been lacking.[3][4]

Technical Components

The UAV LiDAR system consisted of a Velodyne HDL 32e laser scanner weighing 3.5 kg mounted on a DJI M600pro multi-rotor drone, with total take-off weight just under 15 kg.[2] The multi-rotor configuration enabled flights of 13-15 minutes duration before requiring battery changes and controlled landings, contrasting with photogrammetry fixed-wing platforms capable of hours-long flights covering hundreds of kilometers.[2] The survey campaign executed multiple technical tasks including setting out Ground Control Points and Benchmarks, RTK-GNSS River bathymetry profile measurements, UAV LiDAR Survey, UAV Photogrammetry Survey, and water level and salinity measurements.[4] DTMs were generated at multiple grid resolutions including 0.25 x 0.25 m, 0.50 x 0.50 m, 1.0 x 1.0 m, and 2.0 x 2.0 m, with no interpolation performed between points and grid cells.[2] The comparison methodology employed advanced GIS spatial analysis, with vegetation types mapped and metrics assessed for respective vegetation classes.[4] Performance evaluation covered fourteen different types of vegetated areas and three types of urbanized areas within the basin.[2]

Implementation Details

CDR International in association with Shore Monitoring and Research defined and executed the integral survey campaign for the lower basin of the Msimbazi River from February 14th until February 20th, 2019, with only water level logger retrieval occurring after this period.[2][3] The project targeted the Lower Basin of the Msimbazi River, a highly vulnerable area to flooding requiring accurate hydraulic flood modelling for identification and design of appropriate mitigation measures.[2] The Tanzania Urban Resilience Program deployed this innovative combination of LiDAR UAV surveying and bathymetric surveying to support flood modeling within the flood-prone basin.[5] More than 300 students and locals were trained to use drones, GPS, and mobile apps to collect data, transforming everyday citizens into urban planners.[7] UAV LiDAR survey costs were approximately 2 to 3 times more than UAV photogrammetry survey at the time of the report.[2] Preliminary insights were shared at a master class at the Land and Poverty Conference 2019 in Washington, DC (March 25-29, 2019) and at a workshop on DTM drone applications for Dar es Salaam stakeholders at the World Bank Office on April 11th, 2019.[3]

Benefits and Impacts

The survey of river cross-sections and hydraulic structures along the Msimbazi River supported creation of a comprehensive hydraulic model that led to development of a more realistic flood hazard map of the most vulnerable areas.[6] Flood modeling comparison using a 10-year return period rainfall event demonstrated that using the LiDAR DTM versus the 2016 DTM resulted in design level differences of 30-40 cm in the Jangwani area, which would have caused protective measures to be designed too low—a large impact.[2] In dense tree areas, the LiDAR ground point cloud contained approximately 10,000 points with uniform spatial distribution, while the photogrammetry ground point cloud had approximately 50,000 points but with poor spatial distribution concentrated in patches.[2] For mangrove and tree areas covering approximately 33% of the vegetated basin area, photogrammetry showed mean elevation errors of approximately 3 m, with maximum errors in DTMs between 5 to 10 m.[2] For tall reed and grass areas covering more than 30% of the vegetated basin area, photogrammetry showed elevation errors in the order of 1 m compared to LiDAR.[2] The analysis concluded that although LiDAR UAV survey was more expensive, it was considered a more suitable method for DTM generation for flood modeling purposes in Msimbazi-type basins with relatively large coverage of vegetation with variation in type, density, and heights.[2]

Climate Adaptation Relevance

The Msimbazi Basin, home to an estimated 27 percent of Dar es Salaam's population, has experienced increasingly severe flooding over the past decade, with major flood events in seven out of the last 10 years.[6][9] Each year, Msimbazi flooding leads to fatalities and destroys critical infrastructure.[6] Outputs from this surveying initiative will be used to guide river upgrades that will improve the city's resilience to flooding.[5]

Business Analysis

UAV-mounted LiDAR scanners potentially provide significant cost savings compared to both traditional plane-based LiDAR and field surveying methods as equipment becomes smaller and lighter while UAVs can fly longer and carry more weight.[4] The analysis recommended that removing uncertainties at a higher initial cost with LiDAR will easily get paid off in subsequent study and design stages by avoided additional study time and delays, and limitation of need to over-dimension designs of measures, increasing cost-efficiency.[2] The study provided recommendations on when to use which surveying methods with respect to project area size, land-use/vegetation cover, and challenges like import of equipment.[3][4] The project demonstrated deployment of real-time kinematic (RTK) mode using new, low-cost hardware capable of providing centimeter-precise surveys, with TU Delft student Kirsten van Dongen co-developing a method that delivered similar accuracy and precision as professional-grade survey equipment.[6] The DEM from 2000 may not accurately represent current topography of the basin, especially in areas with significant development, and the Pugu Hills are not well represented, highlighting the value proposition of updated high-resolution surveying.[8]

Sources