Open the model
Worked example: Montour Run PFAS
Montour Run drains part of Moon Township, Pennsylvania, past Pittsburgh International Airport and on to the Ohio River - a drinking-water source for millions downstream. On the airport sits an aircraft fire-training area where firefighting foam (AFFF) was used for decades. That foam is loaded with PFAS, the “forever chemicals,” and when it rains the PFAS washes off the ground, into the storm drains, and out to the stream.
Where Warkworth asks is the model right, this example asks a question you cannot answer by looking at a pipe: how far above the safe limit is the water, and does treatment bring it back? The answer is a number with a lot of zeros in it.
Download the files
montour-run.inp · montour-run-gac.inp (with GAC treatment)
What is in the model
Section titled “What is in the model”| Network | 354 subcatchments, 356 nodes, 354 conduits, two outfalls (OF1, OF2) |
| Source | S_N63 - a 1.6-acre fire-training subcatchment on the OF1 branch, toward the airport |
| Pollutant | PFAS, in µg/L. 1 µg/L = 1,000 ng/L = 1,000 ppt - the easy factor-of-1000 to lose |
| Wash-off | Event-mean concentration (EMC), no build-up curve - the right choice for PFAS |
| Event | A roughly 1-inch design storm over about an hour, routed dynamically for 6 hours |
| Limit | The EPA drinking-water MCL for PFOA and for PFOS: 4 ppt each |
The goal
Section titled “The goal”Three questions, in order:
-
Did it conserve mass? Quality continuity first - a pollutant model that leaks mass is not worth reading. This one reports 0.26%.
-
How concentrated is the source, and how far does dilution take it? This is what the storm network computes: one contaminated subcatchment, then acre after acre of cleaner runoff joining the trunk.
-
Does treatment close the gap to the limit? And if not treatment, then what?
The dilution cascade
Section titled “The dilution cascade”The fire-training subcatchment’s wash-off EMC is set to 62.9 µg/L, which makes its own runoff leave the source at 62,900 ppt - the concentration documented at a real stormwater outfall here, and about 15,700 times the 4 ppt limit. Everything after that, SWMM computes: as the pipe runs downhill and clean urban runoff joins it, the pollutograph falls.
Follow the trunk from the source to the outfall (peak concentrations SWMM computes):
| Node | Peak PFAS | vs 4 ppt MCL |
|---|---|---|
N63 - the fire-training source | 62,900 ppt | 15,700× |
MH22_C51 - two blocks down | 26,100 ppt | 6,500× |
N21 - a large tributary joins | 5,740 ppt | 1,400× |
MH27_C52 - last junction before the outfall | 1,300 ppt | 320× |
OF1 - the outfall to Montour Run | 1,166 ppt | 290× |
OF2 - a branch with no source (urban background) | 15 ppt | ~4× |
Two things worth sitting with. First, the network dilutes the source 54-fold on its own - that is real hydraulics, not an assumption, and it is why an outfall sample tells you much less than a sample at the source. Second, look at the last row: a branch with no PFAS source at all still runs at 15 ppt, because ordinary urban runoff carries a trace of PFAS. Even the background is nearly four times the drinking-water limit.
Seeing it in the app
Section titled “Seeing it in the app”Loading the model, running it, and seeing the PFAS are all free. The Water Quality · SWPPP panel adds the analysis workflow - loads against a limit, first-flush, treatment and the compliance report - and that is the Pro layer.
-
Run the model. The green Simulate button; the run finishes in about a second.
-
Colour the map by PFAS. Right panel → Map tab → Color nodes by → PFAS. The source lights up against a low-concentration network; scrub the time slider to watch the plume move through the pipes as the storm peaks.
-
Graph the pollutograph. Graph ▸ Quick graph → click the source node, then
OF1, and pick PFAS from the variable list. The source spikes to 62,900 ppt; the outfall peaks near 1,166 - the same water two orders of magnitude apart. -
For the compliance case, open Water Quality · SWPPP (the toolbar droplet icon, a Pro feature). Its Setup tab already carries the
PFASpollutant and theFireTraining/Urbanland uses; Simulate WQ, then Results gives per-node loads against a limit, the mass balance with quality continuity, and the treatment/compliance workflow.
Mitigation: the treatment that actually works
Section titled “Mitigation: the treatment that actually works”Open the second file, montour-run-gac.inp. It adds one thing - granular activated carbon
(GAC) treatment at the outfall - written in [TREATMENT] as OF1 PFAS C = 0.01*PFAS, a 99%
removal. Run it and graph the OF1 pollutograph (or colour the map by PFAS): the outfall drops
from 1,166 ppt to 12 ppt.
Even 99% GAC leaves the outfall at 12 ppt - still three times the limit. The gap to 4 ppt is closed not by treatment alone but by treatment and the receiving water, and the real remedy is upstream of both: stop putting PFAS on the ground. Source control (phasing out AFFF, remediating the training-area soil) is the only thing that lowers the 62,900 ppt at the source.
The receiving water
Section titled “The receiving water”The documented data gives two numbers: 62,900 ppt at the stormwater outfall, and 430 ppt in the tributary a mile downstream. The model reproduces the first and computes the path to the outfall. The mile downstream is receiving water - a different body from the pipe - so the model reaches it by mass balance rather than by routing:
The two outfalls together discharge about 0.008 lb of PFAS in 3.2 million gallons over the
storm (from the Outfall Loading Summary), a flow-weighted mean
near 310 ppt. The measured 430 ppt a mile down sits between the peak at OF1 (1,166 ppt) and
that storm-averaged discharge - exactly the range receiving-water mixing produces. That is a
consistency check, not a routed result: the natural stream past the outfall is not in this model.
Why this site, and why the numbers matter
Section titled “Why this site, and why the numbers matter”The EPA set the first federal PFAS drinking-water limits in April 2024: 4 ppt for PFOA and 4 ppt for PFOS, with a compliance deadline of 2029. Pennsylvania set its own state standards in 2023. The airport reports outfall sampling to PA DEP and Three Rivers Waterkeeper monitors the stream independently, so the concentrations here are on the public record. And because Montour Run drains to the Ohio River, dilution is not an academic question - it is the difference between a local problem and a regional drinking-water one.
What this example does not show
Section titled “What this example does not show”Being clear about the edges is part of the example:
- It is representative, not as-built. The network comes from OpenStreetMap and terrain; the airport’s internal drainage is not public data. Do not read a specific pipe off this map.
- The source concentration is calibrated, not measured here. 62.9 µg/L was chosen to reproduce the documented 62,900 ppt outfall value. The model computes what happens downstream of it, not the source strength itself.
- The mile-downstream 430 ppt is a mass-balance check, not a routed result. The receiving stream beyond the outfall is not part of the storm network.
- The 15 ppt urban background is representative, not measured for this catchment - a plausible figure for ordinary urban runoff, there to make the point that background alone exceeds the limit.
- PFAS is a single lumped tracer here, not speciated into PFOA, PFOS and the rest. The 4 ppt
line is the per-compound MCL; the model’s
PFASstands in for the AFFF signal as a whole. - Only the stormwater pathway is modelled. At real PFAS sites the chemical also moves through groundwater; that pathway is out of scope for a surface storm model.
- PFAS water quality - why PFAS breaks the usual pollutant defaults, and how to set it up
- Regulatory compliance - turning concentrations into a compliance case
- Reading results - continuity, loads and the pollutograph
- Worked example: Warkworth - the sibling example, on calibrating against measured data