Automated Urban Drainage Modeling with LiDAR DEM and GIS Algorithms

Introduction

Digital Elevation Models (DEMs) have long supported hydrologic analysis by enabling automated identification of watershed and drainage boundaries. Traditionally, these methods have been applied to river systems using relatively coarse elevation datasets, often at resolutions of 30 meters or greater. The San Francisco Department of Public Works (SF DPW) has advanced this approach by implementing high-resolution modeling techniques that operate at the scale of individual storm drain inlets. The resulting data now supports detailed hydraulic modeling for the city’s sewer infrastructure.

High-Resolution LiDAR as the Foundation

A one-meter horizontal resolution LiDAR survey conducted in 2007 provided the foundation for the modeling effort. Originally commissioned for security planning associated with a major national event, the dataset was later made available to the City. Because San Francisco’s topography is steep and complex, the fine spatial resolution offered by LiDAR proved especially valuable for accurately representing surface flow patterns. Using this dataset, SF DPW delineated drainage areas across the entire city, first at the level of individual inlets and then aggregated into larger basins for system-wide analysis.

Moving Beyond Manual Delineation

Historically, drainage boundaries were often defined manually by engineers interpreting contour maps, sewer infrastructure layouts, and property boundaries. Although effective, this process required extensive time and introduced variability when multiple analysts worked on different areas. Automated delineation using DEM-based methods provides a consistent, repeatable workflow that can be rerun whenever new data becomes available, reducing subjectivity while improving efficiency.

The “Urban Drainage” GIS Model

To automate the process, SF DPW developed a custom script within ArcGIS ModelBuilder known as the “Urban Drainage” model. The tool integrates multiple datasets from the Sewer Information System, including drain inlets, pipes, and manholes, and works alongside internally developed utilities such as Oracle Spatial network-tracing tools. Together, these systems enable the rapid creation of complete hydrologic and hydraulic models directly within the GIS environment.

The delineation workflow relies on a “steepest-path” algorithm that determines surface flow direction from the gridded elevation model. Using this approach, drainage areas for approximately 30,000 acres across San Francisco were automatically generated, with each drain inlet assigned its own subcatchment. Because the process is scripted, updates to elevation data or infrastructure layers can be incorporated quickly, allowing catchments to be recalculated whenever conditions change.

Hydro-Enforcement and Surface Flow Refinement

To improve model realism, hydro-enforcement techniques were applied to ensure that simulated flow paths reflect known physical conditions. For example, elevation adjustments are used to guide flow along established channels or ridgelines. Property boundaries are treated as elevated barriers to direct runoff toward adjacent streets, while street rights-of-way are represented as lower surfaces to prevent modeled runoff from reentering private parcels. Although a few localized exceptions may occur—such as areas where street slopes differ from sewer gradients—the approach is sufficiently accurate for basin-scale modeling and planning.

Drain inlets also play a role during depression-filling operations. Because certain depressions in elevation data represent real drainage points rather than survey artifacts, these locations are preserved by inserting null cells at inlet positions, preventing automated filling from eliminating valid flow sinks. This step significantly improves the accuracy of the resulting catchment boundaries and ensures that inlets function as proper loading points within the downstream network model.

Prior to delineation, smoothing is applied to the elevation model using focal averaging techniques to reduce small surface irregularities and produce more interpretable drainage boundaries. Additional datasets, such as multispectral imagery, are incorporated to estimate impervious surface coverage, while slope information derived from LiDAR data helps define hydrologic parameters for each subcatchment.

Integrating Catchments with Sewer Network Modeling

The city’s comprehensive Sewer GIS database, which includes the full inventory of pipes and manholes, allows the automatically generated subcatchments to be integrated directly into hydraulic simulation workflows. Engineers can identify all upstream catchments and network components associated with a specific location and export them into modeling software for detailed performance analysis. This capability significantly accelerates project preparation and ensures consistency across studies.

Benefits of GIS-Driven Automation

The Urban Drainage model demonstrates how modern GIS automation can replace labor-intensive engineering processes while improving data consistency and spatial resolution. Automated delineation not only shortens analysis timelines but also enables rapid recalculation when infrastructure or terrain datasets are updated. SF DPW continues to refine the model, working toward even higher-resolution simulations that separately represent household discharge and surface runoff flows. Aggregated catchment datasets are also used in broader citywide models that support long-term infrastructure planning and major capital improvement projects.

Through the integration of high-resolution LiDAR data, custom GIS algorithms, and sewer network databases, San Francisco’s approach illustrates how advanced spatial technologies can significantly enhance urban drainage analysis and hydrologic system management.