SMAST Dean Emeritus questions objectivity of "Terms of Reference" under which NEFMC Science and Statistical Committee will assess Massachusetts' request to Commerce Secretary

NEW BEDFORD, Mass. – March 30, 2011 - Today, the Science and Statistical Committee (SSC) of the New England Fisheries Management Council will review a report that Massachusetts Governor Deval Patrick submitted to Commerce Secretary Gary Locke last year after the Secretary visited Massachusetts in a high-profile pre-election visit and said he was prepared to use emergency powers to increase fishing allocations and provide financial relief.

The plea was rejected, after 63 days, which many in the industry, government, and academia considered an undue delay and an unwarranted slap in the face to New England officials and citizens. The conditions that spurred the report still exist.

More than half of New England’s groundfish fleet – 253 boats -- remains idle and, with between three and five jobs per vessel, a lot of working families have been hurt.

According to the Governor’s report is no scientific justification for this pain to be inflicted on working families.

The report, prepared by the Division of Marine Fisheries and the University of Massachusetts Dartmouth School for Marine Science and Technology, uses federal government data and applies scientific statistical analysis to reveal that an additional 14,500 metric tons of ground fish could safely be brought to market without hurting fishing stocks.

The Commerce Secretary and NOAA Fisheries Administrator Eric Schwaab rejected the Governor’s request saying they could not make changes to allowable catch levels without new scientific data. However, a 2009 request by Governor Patrick for funding to obtain new, independent data was ignored and eventually rejected, leaving the Commonwealth in a catch-22 with no data to present but the Fed’s own. Still, their analysis, using that data, showed that more fish could be safely caught.

Massachusetts officials hope that the SSC will agree with the findings in “A Report on Scientific and Economic Information that Supports Increases in Multispecies Groundfish Annual Catch Limits.” Some fishery scientists, including authors of the report, however, are concerned by the “Terms of Reference” (major discussion topics for the SSC meeting) that have been scheduled.

“Appropriate review by the SSC and subsequent action by the Council would be a step in the right direction to move New England fishery management closer to the intent of Congress,” said Brian Rothschild, Professor Emeritus at the University of Massachusetts Dartmouth School for Marine Science and Technology in an e-mail to the New England Fisheries Management Council. “This can only happen with this particular issue if the TORs appropriately span the problem. The TORs should enable and encourage innovation rather than being restrictive and retrogressive.”

Among the specifics, Rothschild and colleagues have the following concerns about the TORs:

· They do not address the socio-economic issues cited in the report.

· They set “best available science” as a standard, but do not require an explanation of what is meant by “best available science.”

· They miss the opportunity to inform the Council on serious fundamental difficulties assessing uncertainty and risk.

But in a letter to Rothschild, Executive Director of the New England Fisheries Management Council Paul Howard was positive. “Our Executive Committee approved the ToRs and felt that in doing so they had adequately addressed the Council motion on the issue,” he writes. “They also hoped the ToRs could lead to a wider-ranging discussion by the SSC if committee members expressed that sentiment.”

The text of Dr. Rothschild's letter to the executive director of the New England Fisheries Management Council follows:

 

From: Brian J Rothschild [mailto: This e-mail address is being protected from spam bots, you need JavaScript enabled to view it ]
Sent: Wednesday, March 23, 2011 10:34 AM
To: Paul Howard

Dear Paul,
 
I am looking at your March 14 memorandum outlining terms of reference for the March 30 SSC review of the report “A Report on Scientific and Economic Information that Supports Increases in Multispecies Groundfish Annual Catch Limits.”
 
The TOR should be modified.  Even if the TORs are fully addressed, the Council will not be informed on the facts associated with the report, “A Report on Scientific and Economic Information that Supports Increases in Multispecies Groundfish Annual Catch Limits.”
 
Appropriate review by the SSC and subsequent action by the Council would be a step in the right direction to move New England fishery management closer to the intent of Congress.  This can only happen with this particular issue if the TORs appropriately span the problem.  The TORs should enable and encourage innovation rather than being restrictive and retrogressive.  The SSC is a talented group and they have much to offer in reforming the scientific aspects of our fishery management system.
 
Regarding specifics, the TORs 1) do not address the socio-economic issues cited in the report, 2) set “best available science” as a standard, but do not require an explanation of what is meant by “best available science,” 3) seem to misstate the question concerning the choice of the Fx% criterion and the choice of x=40, and 4) miss the opportunity to point out to the Council very serious fundamental difficulties in assessing uncertainty and risk.
 
Regarding the socio-economic issues raised in the report, what is the Council’s opinion, or the opinion of any of its bodies, on the economic issues?  A striking feature of any economic analysis of the management system to date is the need to incorporate data on revenues and jobs.  Why doesn’t the Council facilitate a sample survey to obtain these data?  I have heard it argued that the SSC does not have eligible economists to serve on the SSC, so the SSC cannot consider economics.  The SSC has members that are not expert on stock assessment, but these non-experts contribute to stock-assessment decisions anyhow.  Why not co-opt someone like Lee Anderson to get involved?
 
Second, the TORs dwell on the requirement to use the best available science.  This is the right thing to do, but in a matter as serious as this, all of us have to understand what we mean by “best available science.”  I think the guidelines on this subject are not sufficiently specific to guide the Council.  Is it as the courts determine, anything the Agency produces?  Is it “best available” because it is subject to peer review?  Both of these standards are inadequate because they are what logicians call “appeals to authority.”  In other words, an assertion is either “true” or “false.” Because the assertion is proclaimed by a federal agency or an expert (i.e., a peer) does not in itself make the assertion true or false.  Look at the peer reviewed stock assessment process in its entirety, or the peer reviewed pollock assessment, or the peer reviewed claim that all stocks would become extinct by 2050, just to cite a few examples.
 
There is a big difference between “available science,” “best science,” and “best available science.”  “Available science” could be anything.  “Best science” means the “best” of “anything that is available.”  So how do we know if we are looking at “best available science?”  We know that science is “best available” if we examine everything that is available and then we apply criteria to everything that was available to determine what is “best.”
 
What are some of the criteria that can be used to select the “best science” among the alternatives?  A simple list would include simplicity and parsimony, explanatory power, adherence to axioms, implementing unwarranted assertions.  It goes without saying that there must be a clear and lucid description of the available science, the criteria, and methodology for selecting the best science.
 
Let us look at the stock assessment process and see whether it meets the test of “best available science”:
 
Simplicity and Parsimony—It is relatively simple to develop few-parameter estimates of Bmsy and Fmsy in an appropriate non-equilibrium setting.   Instead of demonstrably going through this step, as good scientific methodology would require, much more complex models are used involving many parameters.  How are the complex approaches justified as being better than the simpler approaches?
 
Explanatory Power—The data need to fit the model.  It is pretty clear that the data and models are wildly divergent (e.g., retrospective patterns).  The explanatory power is further diminished by a considerable degree by the fact that the sampling theoretic basis of the data that enters the model is not established.  Stock assessment analyses are not accompanied by a clear explanation of the sampling properties of shore side sampling, nor of the construction of age length keys.  A particular fracturing of explanatory power arises in the adoption of the least known component of fishery theory—stock and recruitment—as the foundation or basis of stock status determinations.
 
Implementing Unwarranted Assertions—A great example of an unwarranted assertion is the assertion that Fx% (with x set equal to 40) is somehow good, and that it is required by the fact that the stock and recruitment curve is not clearly defined.  The use of the F40% criterion means that we assert that F40% is somehow better than Fmsy.  Speaking of “science,” the criterion is arbitrary.  Look at the large range in F generated by the arbitrary choice of a value of M and an arbitrary choice of x.  F40% is contrived to be the lowest possible level of F, rather than the more rational average lowest value of F (determined by averaging across “curvatures” of the stock and recruitment relationship).  In addition to this, a choice of F40% will result in much bigger rebuilding goals than x=20, 30%, rebuilding goals that may be impossible to meet.
 
Adherence to Axioms—Beverton and Holt demonstrated a long time ago that disconnecting stock and recruitment in a population would lead to instability.  Since populations, the way we understand them, are stable, we wouldn’t expect that simulations intended to produce the variance in population abundance at some time in the future would be realistic.  Yet this is what is done when projecting populations into the future to predict uncertainty.
 
Regarding TOR 2, I really do not understand the logic behind this TOR.  First, I don’t think direct estimates of Fmsy and Bmsy were in fact chosen!  Isn’t it a fact that F40% was the chosen criterion?  So if Fmsy and Bmsy were the best available science, then why were they not used?  Why would one not want to declare specific values for Fmsy and Bmsy and then make a reduction for uncertainty?  Why would one want to compute an entirely different criterion (F40%), claim that it is somehow better than the Fmsy, and then take additional discounts from F40% to account for uncertainty?  This TOR makes it sound like the SSC computed Fmsy and Bmsy but didn’t like the results, so they contrived another procedure where the results were more consonant with their liking.  This may not be what was intended in TOR 2.
 
Regarding TOR 4, the SSC should be put in a position to level with everyone and let them know that we really do not know how to calculate uncertainty and risk, particularly with regard to understanding the bivariate distribution of the chance variables F and the biological reference point.  This should draw Council members attention to the great difficulties in assessing uncertainty and risk.
 
All of us are looking forward to learning of the deliberations of the SSC.  We all recognize these are hard problems.  Our purpose for forwarding this critique is to inform Council members that 1) these are difficult problems, and 2) ensuring “best available science” requires, at the minimum, a lucid explanation of the steps that were taken to sample the data and to decide on an analysis and to justify the use of more complex techniques over simpler techniques.  I do not believe that Council members have clear and lucid explanations of the scientific base for the decisions they have to make.  I believe that the Council would be better served if the terms of reference were expanded so that the SSC could also consider 1) the socio-economic issues, 2) review the standards for best available science and how these specifically apply to reviewing the report, and 3) leveling with the Council to tell them explicitly what can and cannot be certified with regard to assessing uncertainty and risk.  I also think the Council would be better served if TOR 2 was modified to request the thoughts of the SSC directly on F40% versus Fmsy rather than ask them to speculate on who thought what when?
 
It seems as well that the SSC has not allocated sufficient time to address these very important issues.
 
I understand that this is a lengthy and somewhat technical communication.  This is warranted by the need to record viewpoints on the subject.
 
I hope you can articulate this communication to the Council.
 
Sincerely,
Brian


-- Brian J. Rothschild Montgomery Charter Professor of Marine Science and Technology School for Marine Science and Technology University of Massachusetts Dartmouth 706 South Rodney French Boulevard New Bedford, MA 02744-1221 USA