Last Thursday the Prince William Conservation Alliance held a community presentation on the newly completed 2040 Occoquan Land Use Report, titled “Bubbling up: What Did the Occoquan Model Find Beneath the Surface?” The presentation had an introduction by Normand Goulet, the Director for the Northern Virginia Regional Commission (NVRC) Division of Environmental and Resiliency Planning, a short presentation by Dr. Stanley Grant, Director of the Occoquan Reservoir Monitoring Laboratory and featured Dr. Siddharth Saksena, one of the report’s primary authors. The below discussion is taken directly from their talks and the report itself available for review on NVRC website.
Mr. Goulet gave the background and history of why this
report exists. Dr. Grant covered what are the current water quality challenges
faced by the Occoquan Reservoir as a water supply for 1.2 million people. Finally, Dr. Saksena covered the limitations
of the model and the findings.
When the Digital Gateway rezoning was being considered, Fairfax
Water took the unusual step to ask that Prince William County convene the
Occoquan Basin Policy Board and oversee a Comprehensive Study of the proposed
PW Digital Gateway Comprehensive Plan Amendment (CPA) and the 2040 Comprehensive Plan Update to
evaluate their impact on water quality in the Occoquan Reservoir before any
action is taken. Instead, the PW Digital Gateway CPA and 2040 Comprehensive
Plan Update that eliminated the Rural Crescent were approved in 2022 and the
request was sent to the NVRC, though the funding to undertake the assessment
was not approved until February 2024. It took until late last year to complete
the work.
The Occoquan Watershed Model was developed to evaluate the
impact of land use decisions and compare potential land use scenarios and their
impact on the Occoquan Reservoir water quality. It turns out that the model is
old, does not have a groundwater component,
and the framework used is no longer being supported and it has significant
quality challenges. The Occoquan Model is the Frankenstein result of attaching a
Hydrological Simulation Program–FORTRAN model (that is rather ancient in
computer modeling terms -note the FORTRAN programing language) to a CE-QUAL-W2
model that is not fully calibrated.
A CE-QUAL-W2 model is a two dimensional water quality model that
is extensively used by government agencies like the USGS and EPA for water management tasks.
Its applications are centered on simulating how waterbodies respond to
operational and environmental changes. Researchers apply the model to predict
how rising air temperatures or changes in watershed land use will trigger algal
blooms or increase sediment phosphorus release. It is centered on runoff
and temperature.
Dr. Saksena reported that the team calibrated the model using data from the years 2013 to 2017, with land use conditions benchmarked to 2016. The future simulations were conducted for the period 2039–2041 and compared to the baseline period of 2013–2015. The team consisting of the Department of Civil and Environmental Engineering and the Occoquan Watershed Monitoring Lab at Virginia Tech that conducted the study found that the location accuracy was around 70% and the ability to predict whether streamflow would increase or decrease was also around 70%. Other model quality parameters were lower.
Though the entire goal was to understand how future urban
growth and future climate might affect water quantity and quality (for selected
nutrients) in the Occoquan Watershed, they used a regression model to predict
the key input- rate of development in the basin. They assumed that the rate of
building in the past would continue into the future. The location of
construction in Prince William County was informed by the Small Area Plans and
Activity Center areas where potential denser redevelopment would occur
including Yorkshire, Digital Gateway, Old ARC land data centers, and other
areas.
Because this key variable- rate of development is assumed to
stay the same as the baseline period the usefulness of this work to evaluate
potential risks the Occoquan Reservoir, a key drinking water source, is limited.
Within the above discussed limitations the study simulated
how projected developments (such as roads and buildings) and changes in
rainfall patterns could influence streamflow and water pollution levels. It is
to be noted that the original development of the FORTRAN portion of the model
predates knowledge of the connection of streamflow to groundwater. So the
impacts to groundwater and the groundwater streamflow interface is not
addressed beyond noting that more impervious surfaces result in higher peak
flow (flooding).
Multiple scenarios were explored, including a future with
just urban growth, and combinations of urban growth, climate influence, and
changes in average treated (reclaimed) water discharged from the Upper Occoquan
Service Authority (UOSA). These simulations covered the five sub-watersheds and
the reservoir area. Key findings show that by 2040, urban land is expected to
grow from about 29% to 39% of the watershed area, increasing the amount of
impervious surface and reducing natural areas. This is likely to lead to more
runoff and higher flows in local streams, especially under more extreme
climate, where flow increases of up to 70% were simulated in some areas.
In order for the model to calculate impervious area, the zoning from the comprehensive plan was used for underdeveloped and undeveloped lands. This resulted in the following values that were used to calculate impervious area acreage increases for each transect category in the comprehensive plan; T-0/A-1 = 3.5%, T-1/SR-1 = 11%, T-2/R-4 = 29% , T-3/R-6 = 38%, T-4/R-16 = 43.5%, T-5 was calculated at 55%, and T-6 was calculated at 65%. The problem with this is that it is not reflective of what is happening in Prince William County now. The rezoning of Longleaf at Kettle Run, Alderwood at Kettle Run and Hawthorn at Kettle Run was over 1,200 acres alone. Four more rezoning are in progress in just the middle Broad Run area alone. The other areas are experiencing both the changes made to the Comprehensive Plan and continual up-zoning requests.
![]() |
| These are the development inputs assumed in the report |
The higher streamflow the model predicted will carry more pollutants into the reservoir. The study found that in extreme scenarios, nutrient pollutants like phosphorus and ammonia may increase by more than 50% and 60%, respectively. While some variability was seen, the amount of pollution entering the water system is expected to rise because of stronger stormflows. This makes clear that the volume of water moving through the system, not just pollutant concentrations, is a key driver of future water quality conditions. If you don’t have enough forested lands and you increase rainfall (on average to 120% of historic average rainfall) you get flooding and more nutrient pollution into the Occoquan Reservoir. The model is incapable of predicting any other type of pollution.
The study concludes that both urban growth and climate
variability could significantly impact the watershed and its water quality by
2040. Even if some scenarios don’t show large changes in pollution levels on
average, seasonal events or heavy storms could cause meaningful impacts on
reservoir health.
To protect the Occoquan Reservoir we need better and more
detailed modeling, the ability to test changes in outcome if building rates
change or zoning changes are requested along with continued monitoring. There
are also important limitations to consider. There is tremendous uncertainty in
the exact locations where new development will occur and the actual rate that
it will occur at. The difference in impervious surface. Additionally, only a
handful of pollutants (such as nutrients and sediment) were modeled;
contaminants such as PFAS and salts, which are increasingly of concern, were not
addressed by this model and the assessment.
To improve future assessments, a next-generation modeling
framework is being developed by Virginia Tech, called the Occoquan Watershed
Modeling Framework (OWMF).











