Thursday, July 3, 2014
AS LIKELY TO FAIL AS SUCCEED
When a federal appellate court decided Natural Resources Defense Council v. Daley fourteen years ago, the opinion sent a tremor through the fisheries management world.
For the first time, federal fishery management plans had to have at least a 50% chance of achieving their goals. That was pretty earth-shaking back then, a time when the Mid-Atlantic Fishery Management Council saw nothing wrong with a summer flounder management plan that had only a 17% chance of success—and so gave birth to the lawsuit—and the New England Fishery Management Council’s Northeast Multispecies (Groundfish) plan had less than a 10% chance of recovering some of its component species.
Today, the 50% standard is well-established.
At least, it is if you’re not Washington Congressman Doc Hastings, whose “Emptly Oceans” bill—H.R. 4742—would again subject some components of multi-species plans to overfishing in order to maximize the economic exploitation of somewhat healthier stocks.
It was pretty much a foregone conclusion that the House Natural Resources Committee, which in recent years has been generally hostile to natural resources of any kind—whether they walked, flew or swam, or even just rooted themselves into the ground and just stood there—was going to vote out bad Magnuson reauthorization bill, so the final form of H.R. 4742 didn’t come as a surprise.
However, what did come as a surprise, yet was not given much publicity, was an effort by Representative Niki Tsongas of Massachusetts, who tried to insert language into the bill that would have required every fishery management plan to have at least a 75% chance of succeeding.
Representative Tsongas had a very good idea, even if it was doomed to fail given the current makeup of the committee.
Because a 50% chance of success also means a 50% chance of failure.
In reality, the chances of failure are usually higher. That’s because there are a lot of unknowns in fisheries science.
Those unknowns fall into two broad general categories, “scientific uncertainty” and “management uncertainty”.
“Scientific uncertainty” is a very broad category that can include a multitude of unknowns relating to the abundance, recruitment and life history of a species. It can include environmental impacts, predation and the presence—or absence of a species own preferred prey.
“Management uncertainty” addresses the factors that people can theoretically control—fishing mortality and the impact of new regulations—but can be very difficult to assess. Recreational landings can be very hard to pin down, and for many species, commercial discards can be a statistical black hole. Illegal harvest, whether recreational or commercial, usually cannot be determined at all.
Such uncertainty leads biologists to hedge their estimates of stock size, recruitment, fishing mortality and other critical values with “confidence intervals” that acknowledge the inevitable errors that creep into such calculations. Thus, when a regional fishery management council is told that “there is a 50% probability that overfishing will not occur if the annual catch limit is 26 million pounds,” they are only hearing part of the story.
That 26 million pounds merely represents a “point estimate” that lies near the midpoint of a range of values. Any value within the range—whether higher or lower than the point estimate—is as likely to be the “right” harvest limit as any other.
Depending on the quality of the available data, the confidence intervals surrounding the point estimate can be fairly narrow, or distressingly large.
And so the dangers of using the 50% standard established by NRDC v. Daley begin to manifest themselves.
Under that standard, fisheries managers start out with a 50% chance of failure; the fate of fish stock may be determined by a statistical coin toss.
When you add an annual catch limit based on a point estimate, you create another 50-50 chance that the limit will be too high and the management effort will fail.
Next, compound those two opportunities for failure with a point estimate of recreational harvest, another of commercial discards, a third of recruitment and a fourth of population size…
It’s pretty clear that with all those coins flipped up into the air, there are going to be lots of times when too many come up “tails” and overfishing ensues.
I saw that happen with summer flounder back in 2003, when I sat on the Mid-Atlantic Fishery Management Council. The stock had been rebuilding nicely for a while, but then seemed to stall.
Biologists warned that the there might be problems with the model underestimating harvest, but folks still called for the Council to adopt a 50% chance of succeeding, as that let them to kill more fish than any other option.
During the discussion, one person even lamented the passing of the good old days when a 75% chance of failure was just fine…
The conversation, taken directly from the transcript of the Council meeting, went like this.
PAT KURKUL: … I’ve heard considerable concern over the last few years since the judge’s decision on summer flounder that indicated that we needed to achieve at least a 50 percent probability of attaining the target TACs, that the decision took away the managers’ discretion and prerogative, and I think you’ve heard me argue in the past that in fact it didn’t do that at all. What it did was set a sort of lower bound for the amount of risk that’s appropriate, yet there’s still a considerable range—there are still several different alternatives within a range available to managers where the decisions that we make have to do with the amount of risk we’re willing to assume.
I’m concerned that the tendency has become that we’re going to automatically just adopt the riskiest alternative without consideration of those other alternatives and whether those might be appropriate in different situations. And this is one of them. If you look at the special comments…in the Advisory Report it says given the persistent retroactive underestimation of fishing mortality, managers should consider adopting a lower TAL than that implied by the current overfishing threshold…
What might be a more reasonable way to deal with these fisheries is to look at a controlled expansion…If you look at the 75 percent probability, for example, the TAL would be…26.2 million pounds. So even at a 75 percent probability level, you’re still talking about a significant increase in the quota, and close to the highest TAL since this fishery management plan was implemented.
BOARD CHAIR DAVID BORDEN: Tom Fote.
TOM FOTE: I have very great difficulty with that, the fact that we were taking away from our job and our experience as managers and required to go to 50 percent, we’ll be looking at a fishery that we know there’s a lot more fish out there than we’re basically estimating. Most of us feel that way. Most of the data we see, most of the hook and release mortality shows the same thing. We are forced because of a lawsuit to go to this—well, the Atlantic States Marine Fisheries Commission wasn’t, but NMFS and the Council was, to go to 50 percent.
That 50 percent has provided a buffer. At 25 percent we made estimates years ago, we rebuilt the stock, it was going along fine. It was rebuilding, if you look at the graph and the chart as years progressed…
Just a year later, the Council recognized the error of their ways, and adopted an annual catch limit that had a 75% chance of success.
The resultant spectacular recovery of the summer flounder stock is now history.
Unfortunately, not everyone has gotten the message.
Down at the Atlantic States Marine Fisheries Commission, a new stock assessment determined that striped bass had been subject to overfishing for five of the last ten years, and warned that the stock was about to become overfished.
In response, ASMFC’s Striped Bass Technical Committee came up with a plan that had just a 50% chance of ending overfishing by the end of 2016.
As far as anyone can determine, the proposed plan includes no buffers at all for scientific or management uncertainty.
No plans that proposed a 60%, 70% or even 80% chance of success were considered; all of the chips are being bet on a maximum-risk strategy that is as likely to fail as succeed. (Although, speaking technically, it’s not correct to call the Technical Committee’s proposal to end overfishing a “maximum risk” plan; NRDC v. Daley is not binding on ASMFC, so it could adopt a plan with a 2% chance of success—or one with no chance at all—and still remain within the law.)
Just this week, I learned that the Technical Committee has reexamined commercial striped bass discards—which are nearly impossible to quantify—and decided that they may be smaller than previously thought. Thus, when the Striped Bass Management Board meets again in August, it is likely that the technical committee will recommend a smaller harvest reduction than they had recommended last May.
Such recommendation will ignore the uncertainty inherent in the commercial discard estimate, making it more likely that the effort to end overfishing will fail.
So we have to cheer on folks like Representative Tsongas, who realizes that fishery management plans that are as likely to fail as succeed are only marginally better than no management plans at all.
And until the political climate changes, and again makes it possible for folks such as Rep. Tsongas to improve and strengthen federal fisheries law, we have to try—at every opportunity—to convince managers that fishery management plans with a 50% chance of failure just aren’t good enough.