Background: Accurate elevation data of network junctions are fundamental parameters in hydraulic modeling of water distribution systems. Traditional surveying methods including optical leveling and ground-based GPS, despite their high accuracy, are prohibitively expensive and time-consuming for extensive urban networks. This study evaluates the capability of airborne LiDAR data for decimeter-accuracy elevation extraction of water distribution network junctions and assesses the implications for hydraulic modeling.
Methods: Airborne LiDAR data with a density of 4.2 points/m² were acquired from the Taman Mutiara Rini residential area, Johor Bahru, Malaysia (2.8 km²). A 1-meter resolution Digital Elevation Model (DEM) was generated using the natural neighbors interpolation algorithm in ArcGIS. Elevation values were extracted for 4,823 network junctions (valves, fittings, hydrants, and meter-boxes) using the Extract Values to Points tool. Validation was performed using 32 independent ground control points surveyed by GPS-RTK with sub-2 cm accuracy. Statistical indicators including Mean Error (ME), Root Mean Square Error (RMSE), and Maximum Absolute Error (MAE) were calculated. The implications for hydraulic modeling were assessed through comparative analysis with standard elevation data sources.
Findings: The generated DEM exhibited a mean error of +0.03 m, RMSE of 0.11 m, and maximum absolute error of 0.24 m. The minimum detectable elevation difference was 0.08 m, equivalent to a slope of 0.08% over 100 m, which is sufficient for determining natural pipe slopes in PVC networks with 150 mm diameter (recommended minimum slope: 0.1-0.5%). Spatial error analysis revealed that 81.2% of control points (26 points) had errors below 0.05 m, concentrated in open areas, while 15.6% (5 points) showed errors between 0.05-0.10 m, and only 3.1% (1 point) exhibited an error of 0.24 m, located in areas with dense vegetation cover. Comparative analysis with conventional elevation sources (1:50,000 topographic maps) indicated RMSE reduction from 2.5 m to 0.11 m, representing a 95% improvement in vertical accuracy.
Conclusion: Airborne LiDAR data with moderate point density (4 points/m²) provides sufficient vertical accuracy (RMSE = 0.11 m) for extracting elevation values of water distribution network junctions in flat urban areas. This accuracy meets the requirements for hydraulic modeling applications including pressure zone analysis, leakage detection, and EPANET model calibration. The methodology can reduce field surveying costs by 90-95% and enable rapid updating of GIS databases for water utilities. The findings have direct implications for improving the accuracy of hydraulic models and supporting evidence-based decision making in urban water infrastructure management.