Bridging Satellite and UAV Technologies for High-Resolution Hydraulic Simulations: A Case Study in Iran’s Marun

Bridging Satellite and UAV Technologies for High-Resolution Hydraulic Simulations: A Case Study in Iran’s Marun Basin
Abstract
This study focuses on evaluating and comparing the effectiveness of high-resolution Digital Elevation Models (DEMs) derived from UAVs and satellite (ALOS) data for hydraulic simulations. Conducted in the Marun Basin in Iran, the research assesses the accuracy of these DEMs in modeling flood events using the HEC-RAS 2D simulation framework. By integrating rainfall data and streamflow measurements, the study underscores the potential of UAV-derived data for precision hydraulic modeling while exploring the utility of freely available satellite data for broader applications. This dual comparison offers valuable insights for flood management, especially in regions where precise data acquisition and timely response are critical.

Background and Objective
Floods are one of the most significant natural disasters globally, causing substantial economic and human losses. Climate change exacerbates these risks. Central to flood simulations are Digital Elevation Models (DEMs), which provide the foundational data on terrain and topography.
The study examines the capabilities of UAV-derived DEMs, known for their high spatial resolution, and ALOS satellite DEMs, which offer extensive coverage at a lower resolution. UAVs have revolutionized flood modeling by enabling precise data acquisition, especially in small, localized areas. In contrast, ALOS data is widely available, cost-effective, and better suited for large-scale applications. By employing both sources for 2D hydraulic modeling, the study provides a comprehensive evaluation of their strengths, limitations, and potential for integration.

Materials and Methods
The research was conducted in the Paskuhak region of Shiraz, Iran, encompassing a 4.3 km² section of the Marun watershed. Rainfall and streamflow data were collected using local gauges, while DEMs were derived from UAVs and ALOS satellite. The drone was used to capture high-resolution imagery. The data was processed to produce DEMs with a spatial resolution of 5 cm and a vertical accuracy of 2 cm. ALOS data, with a spatial resolution of 12.5 meters, was calibrated using UAV data to ensure comparability and reliability.
The HEC-RAS 2D software was employed for hydraulic simulations. Precipitation was used as the boundary condition, a novel approach compared to the traditional discharge-based boundary conditions. Calibration and validation of the model were performed using observed hydrographs, with Manning’s roughness coefficient optimized for accuracy.
Mesh sizes for the simulations were carefully selected to balance computational efficiency and result precision. A 2 m x 2 m mesh was used for the UAV DEM, while a 5 m x 5 m mesh was applied to the ALOS DEM.

Results
The UAV-derived DEMs outperformed ALOS DEMs in accurately representing terrain features. Their higher spatial resolution provided a more detailed and realistic depiction of channel meandering, slope variations, and floodplain characteristics. This precision translated into more accurate hydraulic simulations, particularly in predicting peak discharge and time-to-peak metrics. In terms of peak discharge, the UAV DEM estimated peak discharge within 0.85% of the data observed, while the ALOS DEM overestimated it by 5.2%. The UAV DEM's predictions were nearly identical to the observed data, whereas the ALOS DEM underestimated the time to peak by 8.6%. The UAV DEM consistently simulated lower maximum flood depths compared to the ALOS DEM, aligning more closely with real-world observations. For instance, the UAV DEM predicted depths 14.2% lower than the ALOS DEM on average. These differences highlight the superior ability of UAV data to capture fine-scale terrain details, which are essential for accurate flood depth estimation.
The inclusion of rainfall as a boundary condition enhanced the dynamism and accuracy of simulations. This method contrasts with traditional practices that rely on discharge time series and demonstrated the potential to eliminate the need for separate hydrological studies. The rainfall-driven simulations provided a more comprehensive understanding of watershed response, contributing to improved predictive capabilities. Both UAV and ALOS DEMs produced hydrographs that closely matched observed data, with notable differences in peak intensity and timing. The UAV model, with its higher temporal resolution (6-minute intervals), captured rapid flow changes more effectively than the hourly interval data from hydrometric stations. This capability is particularly valuable for real-time flood forecasting and emergency response. Error metrics validated the superior accuracy of UAV-derived data so that Root Mean Square Error (RMSE) resulted in UAV (0.022) vs. ALOS (0.024) and Relative Error (RE) in Peak Discharge depicted UAV (10.9%) vs. ALOS (14.6%). These findings reaffirm the potential of UAV technology for precision hydraulic modeling and emphasize the trade-offs between high-resolution data and computational requirements. The study highlights the complementary roles of UAV and satellite data. In brief, UAV Data is Ideal for localized studies requiring high precision. Limitations include operational constraints, higher costs, and limited coverage. However, Satellite Data is Suitable for large-scale applications, offering cost-effective and widely available solutions despite lower spatial resolution. These insights guide decision-making in selecting appropriate data sources for specific hydrological applications.

Conclusion
This study underscores the efficacy of UAV-derived DEMs in enhancing hydraulic simulation accuracy, particularly for flood management and risk assessment. While UAVs excel in precision, ALOS satellite data provides a cost-effective alternative for broader applications. Key findings include:
UAV-derived DEMs deliver superior performance in predicting hydraulic parameters, offering lower maximum depths and reduced error margins compared to ALOS data.
ALOS DEMs, despite lower resolution, are sufficiently accurate for peak discharge predictions, making them viable for cost-sensitive projects.
The implementation of rainfall as a boundary condition demonstrates the potential to simplify hydraulic modeling by eliminating the need for separate hydrological studies.
Higher temporal and spatial resolution in UAV simulations enables more accurate representation of flood dynamics, particularly at peak flows.
Integrating UAV and satellite data offers a balanced approach to achieving accuracy and scalability in hydraulic modeling.
The research paves the way for future advancements in hydraulic modeling, emphasizing the need for innovative data acquisition methods and enhanced computational techniques. Recommendations include deploying advanced UAV sensors, utilizing multiple UAVs for larger coverage, and leveraging machine learning algorithms to streamline data processing and improve predictive accuracy.
By addressing the limitations of both UAV and satellite data, the study provides a roadmap for optimizing hydraulic simulations, contributing to more effective flood risk management and decision-making.
Keywords: Flooding; Numerical Simulation; Precipitation; ALOS data; Drone