Coastal Flooding Risk Assessment for Prince William County

Flood Hazards analysis conducted by:

George Mason University, Department of Civil, Environmental and Infrastructure Engineering (CEIE) 445/645 Flood Hazards Engineering students

for

Prince William County, Department of Emergency Services

The Prince William County (PWC) Office of Emergency Management is seeking storm surge scenario data to validate and support evacuation planning assumptions for the coastal area. Specific areas to be studied are the Town of Occoquan, Veterans Park, and coastal region that sits between these two locations. Their project request asks that riverine, coastal, and sea level rise flooding hazards be incorporated to create a compound flood model. This project is intended to advise evacuation plan development, our task is to provide models that demonstrate which populated areas are at risk of flooding. The PWC Office of Emergency Management can assess the risk to critical building and transportation infrastructure in the region using the provided models and GIS Files. 

PWC Emergency Management Coordinator Brian Meisner and Deputy Emergency Management Coordinator Katie Kitzmiller provided feedback on some specific expectations for this model- the county is interested in models based on storms with moderate return periods and longer lasting storm durations (as opposed to flashy storm durations). 

The primary objective of this flood hazard project is to generate scenario data that will validate and support evacuation planning efforts for Prince William County, a region increasingly vulnerable to coastal flooding. The project aims to provide comprehensive models using HEC-RAS and ArcGIS software, addressing the specific areas of concern identified by the client. In addition to the client’s request for models focused on individual flooding hazards, our team proposes to develop compound flooding models that integrate both coastal and pluvial hazards for a more robust analysis. Different scenarios with varying return periods and intensities will offer the client a wider perspective of the risks that the region faces. This data will be invaluable for assessing the need for flood evacuation routes, thus enhancing overall preparedness in the region for inevitable flooding events. 

The Town of Occoquan currently faces challenges with coastal flooding, and the region anticipates increased future flooding due to rising sea levels and intensified storms.  Historically, the Occoquan and the Woodbridge areas have experienced inundation during hurricanes, particularly from coastal flooding:  

Aside from these extreme weather events, smaller storms frequently result in riverine flooding along the banks of the Occoquan River. It is anticipated that sea level rise will exacerbate flooding in the town, making rain events more likely to flood surrounding areas. There is a concern that the Town of Occoquan, Veteran’s Park, and Bay St. are at risk to coastal storm surge and sea level rise. Our client, the Prince William County Office of Emergency Management, has also observed recent flooding occurring after rainfall events. 

These DCR models served as the basis for this project’s models.  The flood modeling was performed with HEC-RAS 6.5, a flood modeling application produced by the US Army Corps of Engineers. Two separate DCR sub-watersheds captured the vast majority of the area specified by the client, leaving only a portion of the Occoquan Bay National Wildlife Refuge unanalyzed.

The meat of the modeling is the boundary conditions; this is what permits the input of riverine, coastal, and pluvial data into the model to be applied to the five aforementioned model components (terrain, geometry, etc.).   

• Riverine: Due to a lack of access to Occoquan River data, it was decided to represent the riverine boundary condition with a baseflow of 0ft because water is released from the Occoquan Dam only on occasions.

• Coastal: For 24-hour storms, the coastal elevation over time used a sine function such that the maximum was the MHW (0.87ft), the minimum was the MLW (-0.72ft), and two full tide cycles were achieved in the 24-hour period.  An elevation was calculated for and applied to the models every five minutes of the 24 hour period using this sine formula.  To model higher elevation patterns, a constant value was added to each 5-minute elevation.  The MHW and MLW are from the NOAA Tides and Currents Dahlgren station, taken October 2024. For 6-hour storms, a constant whole number coastal elevation was applied.

• Pluvial: The precipitation vs time data was taken directly from the DCR hyetographs 

To decide which storm scenario to provide to the client, a scenario table (shown below) was established based off the client’s request for storm scenarios with moderate rainfall intensity, a 6-hr or 24-hr storm duration, and a high coastal elevation. The group agreed that due to being restricted to uploading a single sub-watershed per map, more than two models would be too much to succinctly present to the client in this report. 

To understand the models, it is important to understand each part of the scenario table...

Model Properties

• Storm Duration: Simply, it is the duration of the storm in the model. 6-hr and 24-hr were chosen over other storm durations because these are the durations that the DRC precipitation datasets last for.
• Rainfall Intensity: This is the collective depth of rainfall during the entire storm durations. The rainfall intensity at any exact moment within the storm duration varies just as it does in a real storm.
• Coastal Elevation - NAVD88: All coastal water elevations in the model are with reference to the NAVD-88 datum. For 6-hr storms, we chose a static coastal boundary, meaning whatever elevation "y" is equal to in the table is what the coastal elevation was modeled as. For 24-hr storms, the aforementioned non-static, sinusoidal coastal boundary was imployed. Whatever "y" is equal to in the table is the maximum elevation of the coastal boundary throughout the storm duration. Because the maximum elevation is relative to MHW (0.87ft) and the lowest elevation is relative to MLW (-0.72ft), the lowest 24-hr coastal boundary gets is y-0.87ft-0.72ft. For example, a scenario with a coastal boundary of non-static y=3ft, the highest the coastal boundary gets is 3ft, and the lowest the coastal boundary gets is 1.41ft.

Design Scenario

Design storms are the "once every 100 years" storms. A design storm is defined by its rainfall intensity, storm duration, and frequency (25 years, 50 years, etc.). Durations and frequencies are standardized; it is the intensity for any specific duration and frequency that varies from region to region. An example from the scenario table: a 24-hr, 25-yr storm in the Town of Occoquan has an intensity of 6.04in. In our models, a 24-hr, 6in storm is 0.04 inches short from replicating a 25-yr storm. Therefore, the first highlighted scenario in the scenario table can be referenced as being approximately a 25-yr storm. Which models that were chosen for this presentation was also motivated by how closely they resembled specific design scenarios.

It is important to note that no considerations for existing stormwater systems in the Town of Occoquan were incorporated into our models. This likely results in an overestimate in the flooding that the Town of Occoquan region experiences. Ballywhack Branch Creek, a stream that approaches the town from the west, has a noticeable impact on the flooding that the town experiences. Beginning with the team's earliest model runs, it was noted that Ballywhack Creek was the Town of Occoquan's main source of excessive flooding. As depicted in the results, Union Street suffers the most from Ballywhack overflowing.

Flood Model Simulation Results: Scenario 1

This visual representation illustrates the impact of Scenario 1 on buildings and roadways. Scenario 1 simulates a storm with a 24-hour duration and a 25-year return period, combined with a 2-foot non-static coastal elevation. This scenario was selected to align with our client's expectations for mid-range climate events.

Key Findings:

• Affected Buildings: 6,833 buildings intersect with the flood extents across the entire model domain.
• AOI 1 has more buildings affected than AOI 2 and 3

An important note about our models:

Due to the nature of the DCR Terrain and Geometry files, some man-made infrastructure projects, like bridges and culverts, are not incorporated. The most noticeable example of this is the Ox Road Bridge spanning the Occoquan River. Although the models show this bridge to be flooded, this actually just represents water flowing under the bridge.

• Roadways: We have assumed that any flooding depth greater than 0.5 feet can cause roadblocks.
• AOI 1 (Historic Town of Occoquan): Parts of Union Street, Mill Street, Commerce Street, Polar Alley and Center Lane seems to be experiencing roadblocks at portions highlighted in red.

• AOI 2 (Veteran's Park): This area experiences fewer roadblocks on the park's frontage. One significant portion is on Veteran's drive south of the park. However more roadblocks are visible on the west side of the AOI.
• AOI 3 (Bay Street): No Significant Roadblock in this area

Please note, this simulation does not account for the existing storm drain network and represents an exaggerated version of the actual conditions.

We have also included a second Scenario; a 24-hour storm with a 50-year return period in conjunction with 3 ft non-static coastal elevation. We included this scenario because we wanted to see a more severe case, keeping in mind that the area has previously encountered such heavy storms.

The result from this scenario follows the same pattern as the previous one, with slightly bigger impact.

Historically, the Town of Occoquan, Veterans Park, and the coastal area in between have been subject to major flooding from coastal and pluvial hazards. With this known risk, the Prince William County Office of Emergency Management has requested flood models for moderate storms to better anticipate infrastructure damage and plan evacuation routes.

Our models use a combination of pluvial and coastal conditions, which model current 25-50 year storms with varying static coastal elevations, show flooding through the County’s identified Areas of Interest. The results obtained from these simulations can now be used to plan emergency evacuation maps as roadblocks and affected buildings have been identified for each storm.

According to the results, out of the three study areas, the Town of Occoquan experienced the most infrastructure damage due to flooding since it is the most densely populated/developed area. However, during a 50-year storm, the roads and developed areas surrounding Veterans Park were also heavily impacted.

In the future, the accuracy of our model can be further enhanced by including the existing storm drain infrastructure in our model. Other models may also consider riverine flooding downstream of the Occoquan Dam, or take into account a non-static coastal boundary for a more robust simulation.