Many offshore anglers contend there are more red snapper now than in years past. A new, independent study assesses critically federal regulators’ red snapper historical catch numbers in the SEDAR 15 Red Snapper report.

August 2009 WebXtra

Many anglers call the Red Snapper SEDAR 15 assessment "bad science" or "junk science." Now a veteran marine biologist, Dr. Frank Hester, says federal authorities should reject the assessment entirely. In his thorough analysis below, Hester highlights perceived fatal errors that greatly overestimated the number of red snapper in previous decades. Now those guesstimated numbers are being used as proof positive that today's red snapper stocks are in severe decline.

Hester concludes that SEDAR 15 used insufficient and even erroneous catch data to produce an unreliable model and conclusions about the red snapper stocks, including conclusions about the numbers of spawning fish and the numbers of older fish in the stock’s population. Hester lodged serious reservations not only about the catch data, but about the indices of abundance, and the biological inputs about mortality, behavior, seasonal distribution, fecundity and substock identity location.

Most damning, Dr. Hester’s observations concluded that the SEDAR 15 statistical model failed to run any outcomes using a "domed selectivity component," meaning that as snapper age, they become less available to catch. The SEDAR model assumed that red snapper were consistently available throughout their lifetime, which might explain why the model indicates that older red snapper are lacking in the general population.

On July 6, the proposed Interim Rule to close red snapper fishing in Atlantic waters was published in the Federal Register (floridasportsman.com/casts/090706b/index.html). A complete year closure on East Coast red snapper is a very real possibility, starting in October.

Below is Hester’s report in full:

Independent Report on Red Snapper in SEDAR 15 Prepared for Southeastern Fisheries Association, Inc. (S.F.A.) East Coast Fisheries Section

May 8, 2009

By Frank J. Hester, PhD

The Southeast Data, Assessment, and Review (SEDAR) 15, Stock Assessment Report 1 (SAR 1) South Atlantic Red Snapper dated February 2008 revised March 2009 advises the Council, “The assessment indicates that the stock has been overfished since 1960 and overfishing is currently occurring”. The most remarkable feature of this determination is that the historical (“virgin” in 1945) stock size was nearly 30,000 MT and current stock size about 600 MT! How can this be? How was this determination made?

The usual SEDAR process was followed involving three workshops: Data Workshop, Assessment Workshop, and an “Independent” Review.

Data Workshop (DW) The workshop had among its tasks the determination of life history parameters for age, growth, fecundity, and natural mortality, historical catches, and abundance index trends. The workshop revealed the life history information and fishery data base for red snapper to be limited and/or of poor quality. Of particular concern are the facts that the DW was unable to:

• Provide observation based estimates of fecundity or natural mortality.

• Provide any fishery dependent or independent measure of recruitment.

• Provide early recreational catches prior to 1981 (headboat prior to 1972, with 1975 being the accepted year to start the series). However, three catch estimates for earlier recreational catches (US Fish and Wildlife Service (FWS) Saltwater Angling Survey 1960, 1965 and 1970) were discussed, but the DW recommended they not be used owing to the lack of reliability in the sampling methods.

The DW considered nine time trend series as possible indexes of abundance, three of which were recommended for use in the assessment. Two of these three have unacceptable levels of precision.

P30, Section III, SAR 1 (rev)

Assessment Workshop (AW) The AW employed a Catch-at-Age (SCA) model that permits benchmark estimation using very limited data and a production type model (ASPIC). The SCA model works the reverse of the more familiar virtual population type (VPA, SPA) models in that the SCA calculates forward in time from (in this case 1945) using life history values and catch data (and abundance indices when they enter the time stream) to estimate recruitment and biomass. The VPA type starts at the most recent year and back calculates. Both types of model suffer from problems when the data available are poor or missing. The VPA has the desirable feature in that errors tend to be self diminishing as it back calculates whereas the forward progression models may become unstable.

The preliminary SCA model runs caused the AW Report to offer this interesting statement on 14 of Section III SAR 1 (rev):

“The RWG [Recreational Working Group] applied a linear interpolation from 0 in 1946 to estimated values in 1981 for MRFSS and 1972 for headboat. During the Assessment Workshop, preliminary model runs suggested significantly higher landings in the early period (1946-1980) than reflected in the landings. Although the RWG dismissed estimates from the Salt-Water Angling reports (Clark 1962, Deuel and Clark 1968, Deuel 1973), the Assessment Panel agreed that these estimates were at least as reasonable as the linear interpolation to zero in 1946 used by the RWG. Therefore, recreational landings were interpolated between zero in 1946 to 1981 with intermediate landings estimates used for 1960 (Clark 1962), 1965 (Deuel and Clark 1968), and 1970 (Deuel 1973). In general, these values were assumed to include headboat landings for these years. Thus, when interpolating between 1970 and 1981, the headboat landings were subtracted for 1972-1980 (and listed separately for headboat). Headboat landings prior to 1972 were assumed zero (i.e., included in the MRFSS landings). Recreational landings (MRFSS and headboat) as estimated by the two approaches are compared in Figure 2.1”

“Several changes were made to landings to better conform to assumptions made by the primary age-structured assessment model. The beginning of the time series was set to 1946, with extra recreational landings included from 1946 to 1980 to reflect the lack of data on these substantial fisheries (§III(2)). In particular, saltwater angling surveys indicated that recreational landings amounted to 4.1 million lb in 1960 (Clark 1962), 5.7 million lb in 1965 (Deuel and Clark 1968), and 1.9 million lb in 1970 (Deuel 1973).”

Note (FJH): on page 1 in SAR 1, Section V this statement appears:

1 Revision and Corrections

1.1 Correction to recreational landings data

This section documents a correction to recreational landings data used in the stock assessment of South Atlantic red snapper.

As described in section 2.2 of the Assessment Workshop report, the assessment included observed recreational landings from Salt-Water Angling reports. These landings were reported to the level of species for red snapper in the years 1965 and 1970, and as unclassified snappers in 1960. Thus, the value in 1960 was estimated as the unweighted average ratios of red snapper to all snapper from 1965 and 1970. Linear interpolation was used to estimate the recreational landings stream in years surrounding the 1960, 1965, and 1970 point estimates.

After completion of the assessment, it was discovered that the recreational landings in 1965 and 1970 had been transposed when developing the recreational landings stream. Correction of these values affected not only the point estimates in 1965 and 1970, but also estimates in surrounding years that depended on the linear interpolations (Figure 1.1). Using the corrected recreational landings stream, the base assessment model was re-run, as described below.