Prediction and the FONSI: Why Subsidize Hog Operations on Flood Plains?

Statistical inferences shouldn't be so hard to pin down in NEPA documents.

Our story on the loan guarantees to be made by two federal agencies to a hog operation on the banks of Big Creek—a major tributary to Arkansas’s Buffalo River—reported on a federal court order that those agencies actually study and consider the risk of environmental harm from such operations.

hog farm lagoons

Lagoons at a Hog Operation

Hog farms and concentrated feeding operations in the southeast have a terrible way of flooding. When that happens, even well-designed and carefully operated containment systems will fail.  So has been the issue in the NEPA litigation surrounding the Big Creek/Buffalo River operations.

After all, even if the designed waste treatment systems function well and never leak to ground water, the “lagoons” involved could be inundated in the event a 50- or 100-year rain event or flood occurred.  With that, an enormous quantity of hog waste could wash down Big Creek and from there into the Buffalo River.  The recently completed “finding of no significant impact” (FONSI) by USDA’s Farm Services Agency and the Small Business Administration for the loan guarantees minimizes those risks into oblivion, though.  Indeed, this is the sum total of their analysis:

It is highly unlikely there would be an overflow of waste from the holding ponds in a significant rainfall given the ponds over‐capacity. There is the potential for an accidental spill of waste. There are extensive safeguards and BMPs to minimize the risk of discharges off‐site and into waterways. During its operation, [this plant] has not had an accidental discharge. Given the volume, and depending on location and distance to surface water, accidental discharges would not result in long‐term (chronic) or significant impacts to surface water quality.

Final Environmental Assessment & FONSI at 3-24.

Now consider that the operation will be permitted to grow and have on site up to 2,503 sows and 4,000 piglets at any one time.  The lagoons will have capacity for some 2 million+ gallons of manure and wastewater storage.  For some perspective on what a spill that size might entail, the Gold King Mine spill into the Animas River last year was estimated to total about 3 million gallons.

How unlikely is it that a flood or rain event could inundate these treatment systems, cause them to overflow or otherwise fail?  How did the government derive the probability estimate(s) underlying its conclusion (“high unlikely”)?  For all it disclosed, the spatial and volumetric calculations were an ad hoc mixture of precision and vagueness.  For example,

[b]ecause the ponds are designed to divert stormwater from entering the ponds, during a rainfall event the amount of water entering the ponds would be limited to what is falling directly over the ponds. This further minimizes the potential for the ponds to overflow. The ponds have much greater capacity than a 25‐year, 24 hour flood event (6.96 inches). The amount of rainfall in a 100‐year, 24‐hour storm event would be approximately 8.48 inches (NOAA 2015). The 1 foot of freeboard above the 25‐year, 24‐hour storage level has a volume of greater than 425,000 gallons. Based on the total square footage at the top of the ponds, 1 inch of rainfall would equate to approximately 35,000 gallons. Therefore, including the freeboard, the ponds have sufficient storage to hold the volume generated by a 100‐year, 24‐hour storm event.

Id. at 3-22.

But is the 100-year, 24-hour storm event the right upper bound on the stresses this plant may face?  NOAA and the National Weather Service, the agencies that generate these averages, are the first to explain that the recurrence interval notion itself can be deceiving.  The expected recurrence interval is, firstly, derived from past experience (something that, in a climate-disrupted future, is hardly as reliable as it once was).

Secondly, it is a statistical expression regarding a single event’s probability—not an expression of overall risk to a location like this hog operation.  We might expect 8.5 inches of rain to fall in a single 24-hour period only once every century.  But what about three consecutive 4 inch rain falls in a 72-hour period?  What if the first of those saturates the surface soil and ground water in the region?  What has the calculated average to say about that threat?  (Exactly nothing.)

And then consider the vaguer bits of the agencies’ forecasts:

[a]ny discharge during a rainfall event would be restricted to an overflow minimizing the volume of discharge; the entire contents of the ponds would not be discharged. An overflow during a significant rainfall event could have short‐term impacts to surface water quality since nutrients and bacteria concentrations would dilute or be available for biological uptake during downstream transport through the system. However, stream volumes and velocities would be greater during a severe weather event and nutrients and sediment would dilute and disperse at rates much greater than base flow and would move rapidly through the system.

Id. at 3-23.

How do they conclude that the dispersion and dilution factors involved would adequately diminish the adverse effects from such an event?  How do they conclude that “[t]here is no potential risk from Big Creek flooding the ponds because of the distance to the creek and the ponds are elevated well above the creek”?  Id. at 3-23 (emphasis added).  How much is “well above”?  How much distance is enough to reduce the risk of Big Creek flooding the facility to zero?  None of this work is shown in the FONSI.  And it is supposed to be the document informing the decisionmaker(s).

These forecasts are precisely the problem with federal government subsidies of such operations in our great river valleys.  Why was it so imperative to subsidize this risk taking?  The agencies’ FONSI topped 80 pages and, on its face, looked earnest, informative, even complete.  But beneath that surface it lacked any specificity regarding the most crucial analytical inputs.  It papered over risk-taking with precise-sounding calculations and self-assured, dismissive conclusions.  The document admitted that FEMA floodplain maps are not available for this location.  And yet it still confidently predicted that a truly significant flood event is simply impossible.  And that should be unacceptable in NEPA compliance.

The local environmental group, the Buffalo River Watershed Alliance, recently announced it won’t pursue the NEPA suit further, choosing instead to concentrate its efforts on the underlying permits for the operation.  It remains to be seen whether its partner in that case, EarthJustice, will do so.

{Image: The Buffalo River, courtesy of Wikipedia}

I teach environmental, natural resources, and administrative law at Penn State Law. Before teaching I was an enforcement lawyer at U.S. EPA. Along the way I've done work for environmental nonprofits and written a fair bit about NEPA.
No Comment

Leave a Reply