FUMBLING THE NUMBERS ON SUMMER FLOUNDER

The way things are supposed to go, fisheries managers compile their data.  Then they wade through the numbers, and with the information provide, adopt a policy for managing the relevant fishery.

I’ve long said that if you clear your mind, and let the numbers do the talking, they’ll tell you just what to do.

However, there are folks out there who want to turn things upside down.  Instead of creating a policy based on the data, they try to create data that supports their favorite policy, whether or not it makes sense to anyone else.

We saw that recently with summer flounder.

Just last week, an article appeared in The Fisherman titled “A Case for Change in Fluke Management.”  It was written by someone named Tom Smith who, according to information accompanying the article, is a corporate numbers-cruncher, but seems to lack any training or experience in the science of fisheries management.  

It’s telling that the piece was subtitled

“New and exciting data could indicate that NOAA Fisheries efforts to save the summer flounder fishery may actually be contributing to the recent decline, an argument that could bear fruit in [Save the Summer Flounder Fishery Fund’s] ongoing efforts,”

since Mr. Smith is apparently a devotee of one of the newer fads in summer flounder management, the notion, promoted by the aforementioned Save the Summer Flounder Fishery Fund, that a decline in the summer flounder biomass is directly attributable to higher size limits that focus recreational harvest on female fish, and thus cause the female spawning stock to shrink below sustainable levels.

Mr. Smith had already advanced that position in a comment letter that he sent to the Atlantic States Marine Fisheries Commission earlier this year, when he wrote, with regard to proposed 2017 regulations

“…So if we were to maintain a 5 fish limit similar to ’16 but change the mix, establish a slot limit of maybe three smaller fish combined with 2 fish at the existing 18” limit.  Give breeders a few more years to help the overall biomass…”

In such comments, he didn’t mention how reducing the minimum size for three of the fish in his proposed 5-fish bag limit--and thus make it easier for anglers to find legal-sized fish--would help reduce 2017 landings by 41% compared to landings in 2016, as ASMFC was then attempting to do, without radically shortening the summer flounder season.

Such omission suggests that Mr. Smith either overlooked that critical issue, which is somewhat troubling, or didn’t fully understand the implications of the proposal that he was submitting, which is even more troubling, given that he is writing magazine articles on fisheries management.

However, his comments do suggest that in writing the recent article for The Fisherman (which had previously been published in the newsletter of the Recreational Fishing Alliance, another advocate of such reduced size limits), he was attempting to justify a position that he already held, rather than trying to analyze available data in order to discover the proper management policy.

Such supposition is supported by the fact that the premise of the article in The Fisherman is highly dependent on a variable that Mr. Smith refers to either as “SSB recruitment strength” or “reproductive strength”. 

Such variable, which Mr. Smith seems to have created solely to support his argument, is apparently derived by dividing the annual recruitment of summer flounder, expressed in numbers of Year-1 fish, by the size of the spawning stock biomass, expressed in metric tons.  Mr. Smith notes that

“The [resulting] graph shows that the reproductive strength of the summer flounder SSB has been decimated over the last 20 years or so.  The ratio peaked in 1993 [when the SSB was well below the “overfished” threshold] at ~3,243 [Year 1] fish per metric ton of SSB and reached its low in 2015 of ~644 [Year 1] fish per metric ton.  That constitutes an approximately 80% decrease in recruitment strength, the result of which has been a 13-year decline in overall SSB because as recruitment strength declines the number of fish maturing to create a sustainable SSB declines as well.”

While that conclusion may seem superficially attractive, in the end it merely serves as an illustration of why non-scientists who play with the data should always be aware of the statistician’s warning that

“Correlation does not imply causation.”

Before making any attempt to draw a conclusion from seemingly related numbers, it’s necessary to first make sure that there are no so-called “lurking variables” that might give rise to both sets of figures, and so create a false correlation.  

For example, there may well be a correlation between people who go to sleep with their shoes on and those who awake with a headache.  However the shoes have far less to do with their morning malaise than the fact that they went to bed stone drunk the night before, which is why they left their shoes on in the first place…

Mr. Smith fell victim to such false correlation.  Simply because he has inadequate knowledge of fisheries science, he assumed that there was a straight-line relationship between the size of the spawning stock biomass and recruitment.  In fact, no such simple relationship exists, and any stock/recruitment relationship that might occur is far more subtle.

To quote from Fisheries Ecology and Management, a standard fisheries biology text written by Carl J. Walters and Steven J.D. Martell,

“Beverton and Holt [who developed the basic population model currently used to assess summer flounder, and many other species] examined early data on the relationship between parental spawning stock and subsequent recruitment, and data on the early mortality rates of juvenile fish.  They noticed that the spawning stock is generally a very poor predictor of recruitment (recruitment being independent of parental abundance) except at relatively low parental stock sizes…  [emphasis added]”

Which suggests that the “SSB recruitment strength” or “reproductive strength” variable created by Mr. Smith is a meaningless figure.

That general comment by Walters and Martell is reinforced by research specifically addressing the relationship between summer flounder spawning stock biomass and recruitment.  

In a paper entitled “Larval abundance of summer flounder (Paralichthys dentatus) as a measure of recruitment and stock status,” a team of biologists from Rhode Island, New Jersey and North Carolina stated that

“For summer flounder there appears to be no direct relationship between larval supply [which does seem to be related to the size of the spawning stock] and recruitment at Beaufort Inlet or Little Egg Inlet [the two study sites].  This finding implies that recruitment strength may be determined by factors later in the life cycle, likely during the estuarine juvenile stage.  [emphasis added]”

Once again, we find that “recruitment strength” is not determined by the size of the spawning stock, but “by factors later in the life cycle, likely during the estuarine juvenile stage.”  So clearly, Mr. Smith’s calculation of the so-called “reproductive strength” variable an exercise in futility.

More to the point, the entire argument that high size limits hinder recruitment by removing too many large females from the population is shown to be absolute bunk.

But we already knew that.

A while ago, I wrote about the concept of “steepness,” which is the right way to calculate the impact a reduction in the spawning stock has on recruitment.  Steepness is calculated by dividing the recruitment when the stock is reduced to just 20% of its spawning potential, and comparing that to the recruitment that would be expected from an unfished stock.

The less difference there is between the two values, the higher the steepness.  High steepness indicates a low correlation between recruitment and the size of the stock.

The steepness calculation for summer flounder was discussed in the last benchmark stock assessment.  Although biologists disagree on just what the precise value is, all agree that such value is high, meaning that it doesn’t take very many females, relatively speaking, to produce an average year class.

Which again means that the argument presented in Mr. Smith’s article just doesn’t fly.

That’s hardly surprising, given that he got the whole process backwards.

To make sense of the numbers, you first have to learn a little about the science, and how the biology of a particular species actually works.  Then you look at the data, and relate it to the biology of the relevant species.  Only after you can do that are you ready to try to make sense of the numbers, and use them to formulate a policy that can be used to sensibly manage the stock.

Do it the other way around, setting the policy first and then looking for data to support your conclusions, and you can come up with some strange ideas, the kind of mistaken notions that can give fisheries managers unneeded headaches, even if they do take off their shoes before going to sleep for the night.