DUMPING ON THE DATA

I’ve just finished reading the summary of the recent joint meeting of the Mid-Atlantic Fisheries Management Council’s and Atlantic States Marine Fisheries Commission’s Summer Flounder, Scup and Black Sea Bass Advisory Panels.

The panels are dominated by members of the commercial and for-hire fishing industries, and when recreational measures are being discussed, as was the case at the recent meeting, the for-hire industry dominates the discussion.

That being the case, the summary can almost be written before the meeting takes place.  

Advisors will find that the bag limit needs to be higher, the size limit needs to be lower and the season needs to be longer.  All those changes are justified because the scientists are undercounting the fish in the water and that National Marine Fisheries Service is overcounting the number of fish that are caught.

At least, that’s what the advisors usually say…

The December 2014 Advisory Panel meeting pretty well stuck to the script.  NMFS’ recreational data-gathering program got dumped on again. 

While there’s certainly room for improvement, some of the comments made were so out of line that it seemed like a good time to take a look at the issue.

Complaints fell into two broad categories.  There were those who criticized the Marine Recreational Information Program (“MRIP”) itself, and those who had an issue with the estimates that it produced.

It’s pretty clear that the latter group of critics either attacks any results that they find inconvenient, or just doesn’t understand how the survey works.

Or both.

This comes out in the meeting summary, which notes that

“Advisors agreed that MRIP catch estimates for black sea bass are very problematic, with some comments noting that the numbers are very far from reality and should not be taken seriously or used for management. Specific examples were cited, including MRIP estimates that describe New York and Massachusetts private boats as harvesting 1,250,000 lb of black sea bass through wave 4 in 2014, which is more than all party/charter and commercial landings combined for black sea bass in the entire U.S. fore that same time period. Several advisors agreed that this was highly improbable or impossible…”

If you actually take the time to put the data referred to in context, it doesn’t seem out of the ordinary.  Nor, standing alone, is it particularly relevant.  It must be viewed as part of a whole.

Querying the NMFS data on recreational landings  quickly reveals that there is nothing anomalous about the private boat numbers for either state. 

Preliminary 2014 estimates show that New York’s private boats landed about 455,000 pounds of black sea bass through the end of August 2014.  That’s actually less than the 511,000 pounds that they landed during the same period in 2013; 2012 landings for the same period were a little smaller, about 293,000 pounds. 

However, it should be noted that the “Percent Standard Error” used to measure the precision of the estimate was about 61% in 2012, about twice as high as the 34.7% in 2013 and 30.4% in 2014, meaning that when the error inherent in the calculation is taken into account, the New York private boat landings in 2012, 2013 and 2014 were not very different in  size.

We see about the same thing in Massachusetts, with the 635,000 pounds of black sea bass landed through August 2012 not too much below the 795,000 pounds landed through August 2014.  The 292,000 pounds landed through August 2013 was significantly less.  However, it should also be noted that anglers took significantly fewer trips targeting black sea bass in 2013, which could explain a lot of the difference.

But there is another, bigger reason that some advisors’ efforts to play “Gotcha!” by disputing isolated bits of data should be ignored.  MRIP just isn’t intended to work with such small parts of the whole.  It is designed to provide estimates at a relatively high level, and the more detailed the analysis, the less precise those estimates become. 

Thus, if we examine black sea bass landings in the entire Northeast (Maine through Connecticut) for the years 2012-2014, we find a Percent Standard Error that ranges from 13% to 22%.  A similar examination for the Mid-Atlantic (New York through North Carolina) would return PSEs of 14.7% to 18% (PSEs for the entire Atlantic coast are even lower, between 10% and 12%, but can’t be used in this example, as black sea bass north of Cape Hatteras are managed separately from those farther south). 

Such precision is more than adequate for managing harvest.

But when we try to manage at a finer scale, precision deteriorates pretty quickly. 

Break harvest estimates down from the regional to the state level, and PSEs for the same three years can be as low as 17.1% (Rhode Island, 2013) or as high as 86.4% (New Hampshire, 2012).  PSEs over 50% are generally deemed to be worthless.

Break state landings down into two-month “waves” and the PSEs spread out, from 12.6% (Delaware, November/December, 2013) to 100.2% (Virginia, May/June, 2013).

Try to break that down even farther, into sectors of the angling community, and the precision gets even worse, with PSEs ranging from 11.9% (Rhode Island, September/October, 2013, Party Boats) to 120.9% (Rhode Island, July/August, 2014, Shore Fishermen).

Thus, it is clear that those who try to use very small pieces of data, broken down into state, sector and two-month wave, to impeach a coastwide landing estimate don’t have much of an argument.

Instead of demonstrating that the estimate in question is faulty, all they are proving is that they understand neither how MRIP works nor the statistical framework that MRIP is built on.

Folks will say almost anything to kill a few more fish. 

An advisor who operates a boat out of Maryland is attacking MRIP data from here in New York, even though he doesn’t fish here and doesn’t have a clue about New York’s private boat black sea bass fishery (although I will grant him that the disparity between private boat and for-hire catch might not be as great as MRIP depicts; the for-hire landings could well be undercounted). 

Of course, then he goes on to rhapsodize about times gone by when there was only

“a [very small] size limit that forced sea bass to start spawning young and a recreational release ratio that ran about 30%”

which pretty much tells you where the guy is coming from. 

On the other hand, the folks who criticize the recreational data-gathering program do have some very valid points.

NMFS recognizes that, and is working very hard to make sure that the MRIP program is a significant improvement over what went before. 

Anglers who really want to understand how MRIP works, and avoid the kind of foolish mistakes that appear in the Advisory Panel summary, are well advised to visit the NMFS webpage at http://www.st.nmfs.noaa.gov/recreational-fisheries/in-depth/outreach-resources/index, which links to enough information on MRIP, how it works and work still to be done, to keep anglers occupied on many cold winter nights.

Anyone who don’t want to understand, and just wants to make noise and kill fish, need to do nothing at all.

In fact, the less that sort does, the better…