There’s nothing that says summer like a raw sewage leak into one of New York City’s major rivers.

There was another sewage spill in the Hudson, this time due to a broken pipe and resulting diversion into the Hudson at Tarrytown, New York.  A smaller part of the flow is also being routed into the river from Yonkers.  The volume is about 2.7 million gallons per day, about 30 times less than the North River Wastewater Treatment Plant spill in July 2011. The “product” is being chlorinated, which fortunately greatly lowers the dangerous pathogens put into the river.

One of the big questions is whether this will cause a cancellation of the swim leg of the national championship Ironman triathalon that takes place near George Washington Bridge (GWB) early Saturday morning (August 11th).

So, I dusted off our model’s pathogen tracer module, and here is a pathogen forecast through Saturday, covering the area’s waterways.  The “model” is the Stevens Institute of Technology Estuarine and Coastal Ocean Model (sECOM), as applied to and validated for the New York Harbor Observing and Prediction System (NYHOPS).

Maximum sewage pathogen concentrations in surface waters from Wednesday through Saturday, relative to a value of 1.0 in the region offshore of Tarrytown, NY. A full movie of the 4-day simulation is linked from the next figure, below.  The maximum is taken over time, separately at each location.

The simulation shows how dilution and mortality reduce the concentration of pathogens from an unknown concentration (assumed 1.0) in the Hudson near Tarrytown and (assumed 0.1) near Yonkers.  The sources are linearly ramped up from zero on Wednesday, to these constant values from the evening all the way through Saturday.

(UPDATE 8/10 1:20PM: To also take a more optimistic viewpoint, I’m have also now done a model run now where the sewage gets shut off at mid-day Friday — this helps reduce the concentrations roughly 5x lower for Saturday morning).

Here is the animation of the entire four-day period…

Click on the figure to go to Vimeo for a full animation of the 4-day period.

The main conclusions I draw from the simulation are:

• The transport is strongly influenced by the tides, which reverse every ~6.2 hours
• However, mixing spreads the plume out broadly after a few days, and the tidal effect no longer dominates – saying “the tide direction is all that matters”, which I’ve read in some news reports, is a bad idea
• The transport is also strongly influenced by the net river outflow, so the center of the plume on average moves more southward than northward
• The cross-channel transport from the east to the west (and the Ironman course) is intermittent
• Discrete water pathogen sampling may “miss” the highest concentrations and mislead organizers on the safety of the water, especially if one must wait 24 hours for incubation of the samples

At this point, this is a purely volunteered effort.  As we have no means for estimating the source concentrations, we cannot put absolute units or judge these results relative to safety guidelines.  The various counties and cities Departments of Health are in charge of making decisions on public safety, and they are taking observations that are the final word for determining where beach closures and swimming bans are necessary.  So one should consult the Westchester County DOH website, Rockland County DOH website, NJ DEP water sampling website, and NYC Dept of Health website for questions of water safety.

However, the ability of these organizations to inform the public of events as they happen, or as forecasts, is somewhat limited because observations of sewage indicators require 24-hour incubation periods and the water samples are collected at arbitrary times and locations that may not capture the actual exposure of swimmers on Saturday.  The most useful next step is to begin incorporating actual sewage outfall data into our NYHOPS model, enabling us to provide water quality nowcasts and forecasts for the region’s waterways and swimming areas.

Our poster from the “Clean Water Act at 40 Conference” at Vassar College last spring summarizes the sewage pathogen forecast concept and modeling of the North River treatment plant fire spill:  Orton_etal_CWA40_V4.