Assessment of UAV-based topographic mapping for flood risk management in urban communities

Thesis Abstract

Urban communities are increasingly vulnerable to flood disasters, necessitating accurate and timely topographical data to enhance flood risk assessment and management strategies. Conventional surveying methods often fall short in providing high-resolution, cost-effective, and rapid topographic mapping, especially in densely built-up areas where real-time data acquisition is critical. This study aims to evaluate the effectiveness of unmanned aerial vehicle (UAV)-based photogrammetric techniques in generating precise topographic maps to inform flood risk management in urban settings. The specific objectives include (1) assessing the accuracy of UAV-derived topographic data in comparison with traditional ground survey benchmarks; (2) examining the utility of UAV-derived models in identifying flood-prone zones; and (3) providing a framework for integrating UAV data into existing flood management systems. The research adopts a mixed-methods approach, combining quantitative accuracy assessment with qualitative stakeholder interviews. The study setting is the urban district of Riverside City, with a population of approximately 250,000 residents. A total of 50 UAV flights were conducted over selected flood-prone zones, utilizing a quadcopter equipped with a high-resolution RGB camera and GPS module. The sample for validation comprised 100 ground control points (GCPs) measured through differential GPS (DGPS) surveys, serving as the standard for accuracy assessment. Data collection instruments included UAV photogrammetric software (Pix4D and DroneDeploy) for generating digital surface models (DSMs), targeted DGPS devices for GCP measurement, and semi-structured interview guides for stakeholder perceptions. The primary method of data analysis entailed descriptive statistics, root mean square error (RMSE) calculations, and spatial accuracy assessments to evaluate the UAV-derived models' precision. A comparative analysis using paired t-tests determined the significance of differences between UAV and ground survey data. Furthermore, flood zone identification was achieved through overlay analysis of UAV-derived DSMs with historical flood extent data in ArcGIS, supported by regression analysis to correlate elevation variations with flood occurrence frequency. The anticipated findings indicate that UAV-based topographic mapping achieves a horizontal RMSE of less than 15 centimeters and a vertical RMSE of less than 10 centimeters, demonstrating high positional accuracy suitable for urban flood modeling. The study is expected to reveal that UAV-derived models significantly enhance the resolution and timeliness of flood risk assessments compared to traditional methods, facilitating more precise identification of flood-prone zones. Stakeholder insights are expected to confirm the operational feasibility and cost-effectiveness of integrating UAV data into existing flood management frameworks. The study will also identify limitations associated with atmospheric conditions, flight regulations, and data processing challenges. This research makes a significant contribution to knowledge by providing empirical validation of UAV photogrammetry as a reliable tool for urban flood risk management, offering a scalable and adaptable framework for similar contexts globally. The findings will inform policymakers, urban planners, and disaster response agencies on the advantages of UAV technology over conventional survey methods, promoting its adoption in flood mitigation strategies. The main conclusion underscores that UAV-based topographic mapping enhances the precision, efficiency, and accessibility of geospatial data critical for flood risk mitigation in urban environments. The study recommends the integration of UAV-derived topographical data into existing hazard modeling and urban planning practices, alongside the development of standardized protocols for UAV operations in flood-prone areas. Future research is suggested to explore the application of multispectral sensors and machine learning algorithms for real-time flood prediction and response. This study thus advances the methodological paradigm in urban flood risk management and presents a comprehensive validation framework for UAV-based topographic mapping in complex environmental settings.