There
is natural variation in most, if not all, aspects of red
snapper life history, even among the things fishery scientists
and managers know well, that can generate statistical uncertainty
in assessment outcomes. For example, the assessment models
are very sensitive to age at harvest. Even though individual
red snapper can be aged with great accuracy, it is impossible
to age every fish that is brought to the dock. However,
a great many harvested red snapper can be measured for length
in a relatively short period of time. Fish biologists have
developed age at length relationships (also called age/length
“keys”) based on thousands of otolith-aged red
snapper that, based on probability, assign ages to harvested
red snapper from their lengths. For example, let’s
say that based on the age at length relationship, a 24 inch
long red snapper has a 10% probability of being age 3 years,
a 20% probability of being age 4, a 40% probability of being
age 5, a 20% probability of being age 6, and a 10% probability
of being age 7. Among 1000 red snappers for which we only
have a 24 inch length, 100 would be assigned an age of 3
years, 200 would be 4 years, 400 would be 5 years, 200 would
be 6 years, and 100 would be 7 years. This is, in all likelihood,
is not the true distribution of ages within our sample of
1000 red snappers of 24 inches length, but it is likely
very close.
Because
of the wide variation in red snapper length at age, this
assigning of ages from age/length keys generates statistical
uncertainty within the computer model. To overcome this
problem, the computer models allow the life history variables
of interest, such as age at harvest, to fluctuate over ranges
normally observed in nature. In addition, the models are
run over and over again, usually as many as 500-1000 times
(called bootstrapping), until a set of models emerge that
provide the best fit to the life history data, the historical
landings, and the FI and FD indices of abundance. Most of
the time spent on assessments is devoted to this first step,
since every subsequent calculation hinges on these calculations.
This is the base model and it is used to predict allowable
biological catches (ABCs) over the next couple of years,
which is all that stock assessment models are intended to
do.
Predictions
of long-term management objectives (stock recovery) use
results from the assessment above and project them forward
in time based on options for management alternatives (season
length, minimum sizes, total allowable catch, etc.) that
the Gulf Council wishes to consider. For example, a run
of the model might be made with a 40 percent reduction in
bycatch, two fish bag limit for recreational fishery, 15
inch size minimum for the recreational fishery, 13 inch
minimum for the commercial fishery, seasonal closures, and
each fishery with its specific regulatory discard rate and
discard mortality. The next run could be a 50 percent reduction
in bycatch, three fish bag limit and 20 inch minimum size
for the recreational fishery, 14 inch minimum for commercial
sector, slight changes in discards, shorter fishing season,
etc. These are projected forward in time, perhaps 1000 times
under different assessment uncertainties, both to estimate
a harvest rate and to determine if the management goal can
be reached by the year at which recovery is to occur, i.e.
2032 in the case of red snapper. A realistic estimate of
the probability that management objectives can be achieved
under a certain set of management criteria is generated
from the uncertainties both in the assessment estimates
and in future levels of recruitment. Thus, one set of management
criteria may produce only a 50% statistical probability
of successfully achieving the management objectives while
another set may have a 90% probability of success.
Back to Top |
The
Gulf Council must choose a total allowable catch (TAC) from
the range of ABCs for red snapper for the next fishing year
(and sometimes for several years) as the final step of the
stock assessment process. The chosen TAC should have a high
probability (see above) of ensuring that the SFA mandated
stock rebuilding schedule will be achieved. By law, the
Gulf Council must pick an annual TAC from among those sets
of management alternatives that have a 50% or higher probability
of success in achieving stock recovery. For the last 15
years, the Gulf Council has picked TAC from the low probability
end (closer to a 50% chance than to a 100% chance of rebuilding
the stock by 2032) of the allowable biological catch (ABC)
range. These alternatives all were based on an anticipated,
but never achieved, significant reduction in shrimp bycatch
mortality. Each year that shrimp bycatch remained high,
the less likely it became that a 9.12 million pound combined
harvest levels could be sustained. Because it was possible,
albeit improbable, that recovery could occur with the lenient
regulations, the hard decisions were postponed until the
shrimp bycatch numbers had been fully resolved. When it
became overwhelmingly obvious that a technological solution
to decreasing shrimp bycatch of red snapper was not possible,
no options other than major catch restrictions were feasible.
Back to Top |
