Occasionally in the rush to get out the news, journalists miss the story. This was the case with our article of two weeks ago concerning a now widely-viewed private letter sent by Tim Ragen, director of the Marine Mammal Commission, to Corey Goodman, scientist and critic of the ongoing effort to paint Drakes Bay Oyster Company as an environmental menace.
Our story described what appeared to be the letter’s message—that Dr. Goodman had erred in his application of a statistical model and was biased toward the oyster farm’s position on disputed seal disturbances.
Regrettably we failed to dig deep enough, and as a result we missed puzzling statements in the letter that, among other things, reverse Dr. Ragen’s official position on the impacts of mariculture on harbor seals in Drakes Estero. In other words, we missed the story.
For even as Dr. Ragen slams Dr. Goodman he demonstrates that the latter has discovered the best model to explain seal activity in Drakes Estero. It is his first admission that Dr. Goodman’s model—developed alongside David Lewis, director of the University of California Cooperative Extension—is better than the best model offered by Point Reyes National Seashore scientists in their peer-reviewed study known as the Becker Report.
And, according to Dr. Goodman and Mr. Lewis, that model is not their best. They wrote six stronger models, the strongest of which explains 82 percent of what the Becker Report seeks to explain—in contrast to that report’s 50 percent. These models were submitted to Dr. Ragen last year.
Dr. Ragen also concedes in the letter that there is no actual evidence supporting the Becker Report. But then he waffles, stating that because such evidence could potentially exist, the study should not be dismissed.
Say what?
The Becker Report, published by Ben Becker, Sarah Allen and David Press under the misleading title “Evidence for long-term spatial displacement of breeding and pupping harbour seals by shellfish aquaculture over three decades,” attempts to prove a negative correlation between oyster production levels and seal counts in the estero. It does not attempt to prove causation, nor does it present more than 15 years of data from the three decades cited in its title.
It does use a problematic proxy for activity on the estero. As farm operator Kevin Lunny has long contended, annual production levels—the measure used in the report—don’t reflect the number of boat trips that could disrupt seals, as the farm disproportionately uses the cultivation sites located near seal haul-outs.
Those sites are the only ones the California Department of Health deems safe for harvest during the rainy season, so whether it’s a low or high production year, access to them remains constant.
Despite its flaws, the seashore report survived the final Marine Mammal Commission report, released last November.
In it Dr. Ragen concluded: “The Marine Mammal Commission believes that the data…are scant and have been stretched to their limit. Nevertheless, the analyses in Becker et al (2011) provide some support for the conclusion that harbor seal habitat-use patterns and mariculture activities in Drakes Estero are at least correlated.”
It is this correlation that Dr. Goodman and Mr. Lewis disputed in several reports submitted to Dr. Ragen last August, October and November. The two applied both the same statistical model used in the Becker Report, a generalized linear model, and another, a multiple regression model. Because the data set is so small, these models produced identical results.
Nevertheless, after criticism for using the second model, Dr. Goodman and Mr. Lewis ran the data using the first method in their October and November analyses. Had Dr. Ragen read those reports carefully he would have seen the data ran both ways, and the identical results.
Likewise he would have known that the linking of dependent and independent variables made no difference. Yet he wrote in his letter that, in doing so, Dr. Goodman rendered his conclusions “unreliable.”
Like the last, this criticism appears to have originated with Inverness resident Dominique Richard. As an engineer, perhaps Mr. Richard does not understand what Dr. Goodman and Mr. Lewis, biological problem-solvers, do: that in organic systems all variables are interrelated. It is the challenge of researchers to understand these relationships, not to eliminate them.
With its scant data, the Becker Report only achieves a correlation—and a statistically frail one at that—thanks to 2003.
That year, oyster production levels were low (the report lumps oyster production numbers as either high or low, represented by a one or zero). It was also the year an aggressive elephant seal killed 40 harbor seals at Double Point, likely pushing hundreds of others into the estero in the ensuing two years. During the same time, the regional seal population was peaking.
Dr. Goodman and Mr. Lewis believe that those two factors, the marauding elephant seal and the regional rise, largely account for the increased numbers in Drakes Estero. By 2005, when numbers decline again, the regional population had also shrunk and the effect of the elephant seal had fizzled out.
The Becker Report, however, correlates the waning Drakes Estero counts to an increase in oyster production beginning in 2005. Had the seashore scientists asked Mr. Lunny, they would have learned there was no increase in oyster production until 2007.
An informed reader cannot help but ask whether Dr. Ragen’s letter was written to bamboozle, or whether he himself is bamboozled.
Surely he considered the possibility that factors other than the oyster farm, particularly fluctuations in regional numbers, might better explain shifting numbers in Drakes Estero. Francis O’Beirn of Trinity College, a member of the 2009 National Academy of Sciences panel on Drakes Estero, suggested as much, as did Brian Kingzett in his 2011 report for the Marine Mammal Commission, which Dr. Ragen tucked into an appendix.
Dr. Goodman and Mr. Lewis’s top model combines the regional seal population, the effects of the aggressive elephant seal, and changes in oyster farm protocol. In 1992 a sensitive area of the estero was closed to boat traffic during the pupping season, and in 1996 that area was closed to kayaks during the same season.
Statistically this model is 10 times stronger than the Becker Report’s best.
For Mr. Lewis, a scientist trained in adaptive management, the report’s line of questioning suggests a bias: “Rather than asking, ‘How can we manage seal populations with mariculture and recreation in Drakes Estero?’ They are asking, ‘What is the impact of oyster activity on seals?’ And if they’re vested in finding an impact, are they equally committed to finding a solution? And what solution are they vested in finding?”
For his part, Dr. Ragen seems to have softened to the notion of an adaptive management solution for the estero—but how could he not, when he has come to the conclusion, albeit one he buried in a private letter?
He writes on page two of his letter to Dr. Goodman: “Given the uncertainty associated with the analyses, the results are not proof of a correlation, but they also do not provide a basis for dismissing such a relationship.” But even as he makes this remarkable about-face, he ends his letter stating that his “view of this case has not changed…”
Let’s review why Dr. Ragen’s change of mind—if that’s what it is—matters.
This November the 40-year reservation of use for the on-shore mariculture facilities at Drakes Estero expires. It is up to Interior Secretary Ken Salazar to terminate or renew a special use permit that would allow the farm to operate another 10 years. Mr. Salazar has stated that his decision will rely largely on a decision by the National Park Service, which has conducted an environmental review of the oyster farm and its supposed impacts. Under the draft review’s environmentally preferred alternative, the farm would be shuttered.
Released last September, the draft review was so heavily criticized that the National Academy of Sciences was called in to examine its scientific underpinnings. A panel selected by the academy commenced its investigation several weeks ago, and wilderness advocate Neal Desai of the National Parks Conservation Association submitted Dr. Ragen’s letter for their consideration. (The second part of Dr. Ragen’s letter addresses the remaining disputed seal disturbances, arguing that all four should remain on the record as potentially caused by the oyster farm.)
Dr. Ragen told various individuals that his communiqué was private, but he nevertheless passed it along to Point Reyes National Seashore Superintendent Cicely Muldoon the day after he sent it to Dr. Goodman. Mr. Desai quickly obtained it through a Freedom of Information Act request.
Dr. Goodman and Mr. Lewis asked the National Academy of Sciences panel if they could respond to the letter but were told the comment period had closed.
We are tempted to question whether Dr. Ragen’s letter was really intended to be private, or whether it was written with a broader purpose in mind. And, given its consequential but buried position, what exactly that purpose is.