Sunday, August 20, 2023

NMFS FINDS ERRORS IN RECREATIONAL FISHERIES DATA

 

It’s impossible to count every fish caught by recreational fishermen, yet accurate recreational catch and landings data are an essential part of the fisheries management process. Thus, saltwater fisheries managers are constantly seeking better estimates of anglers catch, landings, and effort. Currently, the National Marine Fisheries Service (NMFS) employs the Marine Recreational Information Program (MRIP) to provide the needed information.

 

MRIP estimates are derived from data developed by four different surveys, which may be supplemented by state recreational data programs. The two most important surveys are the Access Point Angler Intercept Survey, which interviews anglers returning from fishing trips, and counts and measures their catch, and the Fishing Effort Survey (FES), which mails questionnaires to both known anglers and to coastal addresses, asking how often household members have gone fishing in recent months.

 

Compiling the data generated by the surveys, and calculating the estimates of catch and landings, involves an intricate statistical process. To grossly oversimply how that process works, the average number of fish, broken down by species, caught by the interviewed anglers is multiplied by the estimated number of trips taken by every angler on the coast. The result is a relatively accurate estimate of recreational catch and landings.

At least, it was believed to be accurate until a pilot study sponsored by NMFS discovered an apparent flaw in the FES, which seems to be inflating estimates of shore and private boat anglers’ effort by as much as 30 or 40 percent. NMFS announced its discovery of the problem on August 7, 2023, during a call with members of the recreational fishing community.

Ironically, the problem seems to have arisen out of anglers’ desire to help NMFS’ surveyors.

The FES asks anglers a series of questions, the most important of which is how many times they went saltwater fishing during the most recent, two-month “wave.” NMFS also asks how many times the angler fished in the previous year. When researchers, conducting a pilot study on NMFS’ behalf, looked at anglers’ answers to those two questions, they found that about sixteen percent of responding anglers reported that they went on more fishing trips in the preceding two months than they did in the preceding year, an obviously impossible result.

 

Researchers found that some anglers are reluctant to provide a negative response, so such anglers report trips that they didn’t actually take. Such reports don’t, for the most part, represent intentional deception, but rather what is known as “telescoping error,” which, according to the pilot study, happens when “a respondent misplaces an event in time, usually placing the event more recently in time than it actually occurred.”

For reasons that remain unclear, shore fishermen seem more prone to telescoping error than private boat fishermen. Asking anglers how many shore-based trips they took, before inquiring about trips made aboard private vessels, also appears to inflate the reported effort. Surprisingly, errors due to “recall delay,” which occurs when respondents forget to report a trio, are relatively infrequent.

The pilot study suggested that much of the error can be eliminated by first asking anglers how many trips they took during the prior 12-month period, and afterward asking about the number of trips taken during the previous two-month wave. Apparently, the answer given to the former question creates a “boundary” that limits the number of trips reported for the shorter period.

Asking how many private boat trips an angler took before inquiring into the number of shore-based outings will probably further reduce the error. It’s impossible to know whether error can be totally eliminated but, as noted in a review of the pilot study, “Since the ‘truth’ is never completely known in these situations, the author [of the pilot study] makes plausible arguments to show the designs that are likely to lead to reduce management errors and estimates that closer [sic] to the unknown truth.”

In other words, while the possibility of error can never be completely eliminated, it can be substantially reduced. NMFS plans to undertake a larger scale study, expected to last throughout 2024, to determine whether reordering the FES questions will substantially eliminate the telescoping error issue.

In the meantime, NMFS needs to decide how to address the error in the FES.

Recreational catch and landings estimates are tightly interlaced with other aspects of a fishery management process in which such estimates are used in calculations of spawning stock biomass, the estimates of biomass are used to determine commercial and recreational catch limits, and such catch limits are used to set the necessary management measures. Then, at the close of each fishing year, the recreational estimates are used once again to determine whether the management measures successfully prevented overharvest.

Without accurate estimates of recreational catch and landings, it will be very difficult to successfully manage saltwater recreational fisheries, as managers could find themselves either allowing harvest to rise to unsustainable levels, putting the future of the fishery in peril, or making unnecessarily severe reduction in landings, to account for the current management uncertainty.

Despite the need for caution in the face of uncertainty, some members of the recreational fishing community will undoubtedly use the FES overestimates as grounds to call for less restrictive management measures. That was evident on the August 7th call, when one mid-Atlantic magazine editor complained that the Atlantic States Marine Fisheries Commission’s Atlantic Striped Bass Management Board was not told of the FES issue before it voted to extend emergency regulations intended to reduce recreational striped bass landings.


Such complaints only demonstrate individuals’ ignorance of both the effect of the FES errors and the relationship between recreational catch and landings data and the estimates of spawning stock biomass.

As Dr. Evan Howell, Director of NMFS Office of Science and Technology, noted on the August 7th call, while the FES issue creates an error in the magnitude of recreational landings estimates, causing them to be too high, it does not impact trends in those estimates. Thus, managers can rely on the fact that 2022 recreational striped bass landings were nearly twice as high as those in 2021, even though the magnitude of both 2021 and 2022 landings were really less than originally believed.


There is a similar link between recreational landings estimates and estimates of spawning stock biomass, particularly in fisheries which are dominated by the recreational sector, as the population models used to assess the health of such stocks typically treat higher recreational landings as an indicator of greater abundance.

For example, in June 2018, MRIP estimates were calibrated to reflect adoption of the FES as a replacement for the less accurate Coastal Households Telephone Survey; the calibrated estimates indicated that striped bass landings were 160% higher than originally believed while striped bass catch, which includes both landings and releases, was 140% higher. Due in part those increases, the last benchmark striped bass stock assessment, released in 2019, set the target spawning stock biomass at 114,295 metric tons, roughly 160% of the 72,032 metric ton target set by the previous assessment, completed five years before.

 

If recreational striped bass catch-and-landings estimates are revised downward as a result of the newly discovered FES error, both the estimate of spawning stock biomass and the spawning stock biomass target will almost certainly be lowered as well. As a result, the FES error will probably have little effect on the regulations needed to manage the recreational striped bass fishery.

The same should be true for most recreationally important fish stocks.

What probably will be affected are the commercial quotas for federally-managed fished species, which could be impacted in two different ways.

First, because such quotas effectively permit commercial fishermen to catch a fixed share of a stock’s estimated biomass, if the FES error led managers to overestimate such biomass, the resulting commercial quotas will also be too high, and could lead to overfishing; reducing commercial quotas will be necessary to prevent that result. On the other hand, FES data has been, and is being, used to reallocate catch between the commercial and recreational sectors, based on the belief that recreational catch has, historically, been underestimated. Since it now appears that recreational catch has been overestimated, an argument can be made that such reallocations take too many fish away from the commercial sector, and need to be revisited.

Decisions on such issues probably won’t be made until NMFS has had an opportunity to fully evaluate the impacts of the FES error on fishery management efforts.

In the meantime, recreational industry advocates, who already have a long history of condemning MRIP while supporting state recreational data programs, will undoubtedly try to use the FES error to support their arguments. Such intent was telegraphed by stakeholder comments on the August 7th call. Yet FES-based attacks on MRIP ignore the fact that the recent discovery of the survey’s flaws is evidence of MRIP’s strength, not its weakness.

 

The FES error was only discovered because NMFS is constantly striving to improve the MRIP process, and conducts continuing studies on how that process can be improved. Even though the National Academy of Sciences gave MRIP a very positive review, NMFS didn’t become complacent, but instead continued to review the process and search for possible flaws. Now that one of those reviews has discovered the apparent FES error, NMFS can take the actions necessary to correct it and make MRIP a better tool for estimating recreational catch, effort, and data. Had NMFS never sponsored the study, the error would not have been found, and managers would have continued to rely on inaccurate information.

 

While state data programs are being actively promoted by some recreational fishing organizations, such organizations fail to provide assurances that such programs are subject to the same sort of rigorous study and review that characterizes MRIP. They fail to provide assurances that state programs are free of errors at least as serious as the FES error, which was only discovered because NMFS was looking for it.

But now that the FES error was discovered, the next question is, “What happens next?”

In the short term, NMFS will have to figure out how the error might impact stock assessments and the annual management measures that govern federal fisheries. It must decide what actions might be needed to prevent harm to fish stocks while the issue is being resolved. Such discussions have already begun.

Over the next two years or so, the agency will need to determine the true extent and magnitude of the error; the current belief that effort is overstated by 30 to 40 percent may or may not be right. Once the extent of the error is known, the agency will need to come up with a way to fix it.

In the long term, discovery of the FES error will only strengthen MRIP and make it a better tool. That is how science works. There is no dogma, no absolute truth. Instead, new information, supported by convincing evidence, continually reshapes perceptions of reality.

Now that the FES’ estimates have been proven wrong, NMFS can begin working on getting them right.

-----

This essay first appeared in “From the Waterfront,” the blog of the Marine Fish Conservation Network, which can be found at http://conservefish.org/blog/

 

No comments:

Post a Comment