\
Not to be cited without prior reference to the author
lCES
C.M~
1993/0:10
P
Statistics committee/
Ref. Theme Session P
Sess~
,
,
Advice under uncertainty to managers for two fisheries
of the northeastern Pacific
William H. Lenarz
National Marine Fisheries ~e~vice
3150 Paradise Drive
Tiburon, california, USA 94920
lntroduction
, Flsher!es managers must base ttieir decisions on information
that hascorisiderable uncertainty. lmportarit populationdyriamic
processes such .. as . the spawn/recruit relationship are highly
stochastic •. The accuracy.of quantitative descriptions of these'
processes is often low. Confidence limits about.parameter
estimates derived fram fisheries dependent and fisheries
independent data are often wide. While usually overlooked in the
technical advice given managers, there also is considerable
uncertainty in the political outcome cif management decisions~
For example, .environmental.grciups may object to adecision and . ,
through the judicial or political system cause changes that could
be costly tci governmerit agericies and fishing iridustry. Fisheries
managers are very aware that they work iri an.üncertairi.field and
desire, but often do not receive techriical advice on the risks of
alternative managemerit options.
.
,In this paper I describe results of two studies of Northe'ast
Pacific fisheries, that convey risk information to.fishery
managers. The first study was conductedon Pacific sardine,
(Sardinops sagax) fishery in 1970~ Thesecorid study is on the
widow rockfish (Sebastes entomelas) fishery and still is in
progress.
Pacific Sardine Study
. I n 1970t:he. nort~erri .stoc~ ofthe Pacific sardi~e was badiy
The.b~omass was est~matedto have decreased from .
almost 4,000,000 tons in 1932 tö between 1000 and5000 tons in
1970~
Landings were restricted to low levels; Reported 1970
landings were 221 tons~ However, sardines had become very
valuable for use as bait by recreatiorial fishermen and there was
anecdotal evidence cif significarit non-reported landirigs.
depleted~
My colleague Paul smith . (Nationai Marine Flsheries service,
La Jolla ca.) and I became .concerned that poteritial,for stock
recovery could be reduced by the fishery and decided to examine
II
'2
!
I
the possible effects of the fishery ,with a simulation model~ I
based the model on the results of Murphy's (1966) study of
Pacific sardine, population dynamics~IThe model used standard
catch and growth equations andthese were, assumed to be
deterministic. Since there was considerable uricertainty irithe
spawn/recruit relationship (SR) and the SR is extremely important
in projections of future populations, II used a stochastic SR. I
employed an autoregressiveequation with a lognormal error term
to model deviations from expected values of a Ricker SR.
I
We presented results of the itiodeling exercise tC,managers in
the form of probability ofrecovery as a function ofbiomass in
1970 and average future, catches, Figura 1~" TWo,recovery levels
were presentedi 20,000 ,tons in ten years arid 950,000 tons in 25
years. Simulatiorislndicated that it I would be safe to have, ci low
interisity bait fishery,if the population recovered to 20,000 tons
and that maximum sustainable productien occurs at about 950;000
tons. 'The results, indicated that if the biomass was orlly 1000
'to~s that a ,reduction ,in;c~tch~s frcim!ev~n the,reported laridirigs
would,improve th~ probab1l1ty of recovery~ The results also
showed that if catches werekept at lew levels there was a good
chance for recovery within
reasonable time frame.
,
"
a
,
,'.'"
I
,.
,. ",
"
We, presented the" results to various levels of state
government (the regulatory authörity) laridincluded the
recommendation that a regulation be enacted to prevent use cf
Pacific sardines for bait; The admiriistrators" sent "the paper out
for review; Some scientific assumptions of the,sttidywere .
questioned and Paul smith did an excellent,job of preseriting data
that convinced t~e re,:iewers that the Ias.~uitipti.o~s ~e~e ' "
'
reasonable., ,Cal1forn1a enacted.~h~ r~comm~ndat10ns 1n,~974~
Landings averaged less than 10 tons for the next 5 years. We
will, riever ,knoweit1'.?-er the,exactlevels of,19?0 c~tcl:l arid bi,omass
nor the exact nature of the SR, but the population is recovering.
By 1986 . the population had recovered sufficiE~ntly to allow,a
small directed fishery. The 1992 quota was set at 25,000 tons.
I
widow Rockfish study
I
.'"
'
,
•
•
• I,
• ,
,
,',"
Since 1970, fishery data have improved somewhat and our
analytical techniques everi more so. While the level of
uricertainty in management advica is still high, the improvements
in data andtechniques allow,us to describe better,the
uncertairities associated with stock assessments and management
recoitimendations. Another change thatlhas occurred is ,that groups
with streng conservation ethics have become wellorganized and
aggressive in, attempts to meat their goals;" In the past
fisheries decisions ,were sometimes challenged in the jUdicial or
political arena by user groups willingto take long term risks to
obtain short term gains; Now fishery;managers also encounter
efforts by groups that are primarilyinterested iri,avoiding,long
term risks to natural populations~ Management decisions made by
I
,i
I
•
I
,
•
.•..
e
.~,
!'
~.
. • . ." .' ; ,;.
.,: .. ( ....
"~
.• 1
3
•
I,
I
I
'. I
4
.
.,
" '.
".
.
recruitment is independent of spawn~ng, stock. It .was also,
.'
~s~umed.that fishing mortalit~ and population fecundity estimates
followed a log normal distribution~ Each 200 year simulation
began with an unfished population.at equilibrium~ 2000
.
replicates were run for each combination of management policy and
2 SR'j.
I
.
-'
:
I will first present results as i might tO,a technical.
Both the management.policYland SR bad considerable
influence on yield and populationfecundity~ For example, .under
policy P fishing mortality would be 0~21 and relative yield would
be aboutO;7 when recruitment"is stock independent, and 0~6,under
the Beverton-Holt SR; Figure 3. Relative population fecundity at
the end of the 200 year pericid would be about 0.4 when
.'
recruitment is stock independent and 0.3.under the Bevertön~H6lt
SR. There are everi greater differences in the results when
management policy A is followed (F=0.50). Relativeyield would
be,about l~Owhen recruitment,is density independent and 0.2 ,
under tbe Beverton~Holt SR. Relativelpopulation fecuridity at tbe
end cif the simulation would be about 0~15 when recruitment is
stock independent and less than o;OOilwhen .under the Beverton~
Holt SR~ .When management policy C islused, I assume that, F=:=0.5
when the fishery is open. Relative yield would beabout 0~8when
recruitment is density independent and 0.6 under.the BevertöriHolt, SR; Under both SR assumptions relati~e yield would_decreas~
by about 0.05 when. adjusted for cost of years closed to . fishing
and the fishery was closed about 50% of tbe time. Fishery
closures ranged from 1 tri more than 10 years with typical . '
closures being ~bout 4 years. Under, .both SR assumptions relative
popUlation fecundity at the end of thc simulätion would be
between 0.30 and 0.35.
!,
,
audience~
.
,
I
!
!
I
'
.
The"results shown in Figures 3 and 4 indicate that if
mariagement .used policy A.h~cause it.assumedthat recruitment was
density, independent and the assumptiori was correct, yield wöuld
be relatively high, 1.0, but population fecuriditY.would,be lower
than 0.2. Ifmanagemerit had guessed wrongly,yield would bevery
low, 0;25, and the,population would have collapsed by.the end of
the simulation period. However if management policy P was üsed,
relative yieldwould only be 20% lessithan under management
policy,A when recruitment is density, independent and yield would
by 300% higher than under policy A when there isthe assumed
Beverton-Holt SR. PopUlation fecundity at the erid of the
_
simulation period would be considerablyhigher under policy P
t:ban underpolicy A regardless of SR. J Management policy C '.
produces results that are similar'to policy P except that the
fishery would be closed about 50% of the time.
. .
. . '. '. I
.
.
I think that for even a technical audience it would be .
easier to comprehEmd the important outcomes of the study, . if tl'iey
were pres~~ted as show~ in Figure 5. !It can easily be se~n ..that
the aggressive policy would result inllow population fecundity
!
II
I
••
;
-~
..,
.
~:
-,
'"
.
.,'
,
.'
;
,
.p -
"
",
~
:
,.
'.
.
~
.,
'::.
5
under either SR and would'also'result:in very poor yield under
the Beverton-Holt SR. It is also easy to see that management
policies P and C produce similar yields and population
fecundities, but that fishery would be closed about 50% of the
time under policy C.
Acknowledgements
•
I thank Paul smith for many interesting discussions and for
carrying more than his share of the burden of communicating the
results of our sardine study. I also thank Alec MacCall for
discussing the widow rockfish study with me and reviewing this
paper.
Literature cited
Clark, W. G. 1991. Groundfish exploitation rates based on life
history parameters. Can. J. Fish •. Aquat. Sei. 48(5):734750.
Hightower, J. E. and W. H. Lenarz. 1990. Status of widow
rockfish fishery in 1990. Appendix F In status of Pacific
coast groundfish fishery through 1990 and recommended
acceptable biological catches for 1991. Pacific Fishery
Management Council, Metro Center, Suite 420, 2000 S.W. First
Ave., Portland OR 97201-5344.
•
Kimura, D. K. 1988. Stock-recruitment curves as used in the
stock-reduction analysis model. J. Cons. int. Explor. Mer
44:253-258.
Murphy, G. I. 1966. Population biology of the Pacific sardine
(Sardinops caerulea). Proc. Calif. Acad. sei. XXXIV(1):84p.
1.0
0.9
-e
ca
~
0
....
.6
CI)
0.8
0.7
0.8
r:::
0.5
g
0.4
.s
4000
•
Ö
N
A
Q.
0.3
2000
0.2
0.1
(a)
0.0
0
500
1000
1500
2000
Catch (tons)
1.0
0.9
-
l!!
ca
~
LO
N
.5
l!!
.s
§
g
0)
-
0.8
0.7
0.6
•
4000
0.5
0.4
2000
0.3
A
a.
0.2
(b)
0.1
0.0
0
500
1000
1500
2000
Catch (tons)
Figure 1. Probability of recovery of Pacific sardine population tO 20,000 tons
in 10 years (a) and 950,000 tons in 25 years (b). 1970 population size (tons)
isopleths are shown with annuaJ catch and probability of recovery.
25
-
r---------------------------,
20
( f)
c:
.2
.- 15
-'E
E
Q)
E
:t:
...o
:::]
10
Q)
CI:
5
OL...L.
o
---l-
0.2
~
0.4
.....l..._
...L.._
0.6
0.8
Relative Fecundity
•
Figure 2. Simulated recruitment of 5 year old widow rockfish as a function
of relative population fecundity.
...I_I
1
1.2
~1
"C
r----------------------------,
,"
c:
:::s
Yield (Density Independent RecruitmentJ-.....
\"'~"..
"'"
"'"',>,
~ 0.8
'.".
"'~"'"
'.'.' . ,......
'.
...o
U.
"C 0.6
ä)
..
>=
~ 0.4
~
ä)
a:
0.2
". :~.>.,<.. .
....
.
.............
••••••
··~
/Yield (BH R.ecruitment)
.__.:-._._._._._. .
........
,
.
/
Fecundity (Ind. Rec.) ••••••.
~~~~=:
Fecundity (BH Rec.)/
-.....
'-'-
~
o
L-"-
-"
o
-'--
0.1 .
0.2
-'--
- ..- ..- ..- ....:.__..- .._._.__
........
-"-"-.-.--'--..;;;.-=.-=.
0.3
0.4
-._..••••••
_-_..-'--'
•
0.5
Fishing Mortality
Figure 3. Relative yield and population fecundity of widow rockfish
as a function of fishing mortality.
1.2
~1
"C
r------------------------------,
"'~
'~.
c:
:::s
...o
:2
~~,
Yield (Density Independent Recruitment)"'-.
','.
~ 0.8
.
.~>,.
'<>,.,
.
Adjusted Yield (BH Rec.)
,.,.<~:> ._. ~~~~~~~~~ ..:-.:-.:-:~~~1~~~~~~.:·.:-.:·.:-.:-.:·::.:·.:·.:·.:·.:
0.6
Q)
..
-
:~ ~~<2_:
v......
. . . ....... -...-...- .._
Adjusted Yield (BH Rec.)
.
...
·_--...::--_..- ..- .._~ecundity Ond. Rec.)
· (BH R -).~._.-.::.::..-.._._..-._..- ..- ..- ..ec.
-'-'-'--'-'_'_. .-.._.._-_.._-_..Fecundity
. . .
. . . ..- .._.._.
. .
>=
~ 0.4
_- - _.- - ___ _-
~
Q)
a:
-.:.:.: .:
0.2
O~-----"-----..L------'-------L--
o
0.1
0.2
0.3
0.4
Fishing Mortality
Figure 4. Relative yield and population fecundity of widow rockfish
as a function of fishing mortality. Simulated fishery was c10sed if
population fecundity dropped below 20% of unfished level and reopened
when fecundity recovered to 35% of unfished level. Yield is relative to
maximum yield of unclosed fishery. Adjusted yield reflects cost of closure.
L-I
0.5
•
AI-Q
1.0
0.9
0.8
CI-{)
0 -PI
0.7
-c
Q)
0.6
PS-Q()
>=
-eS
~ 0.5
~
Q)
a:
0.4
0.3
0.2
o-AS
0.1
0.0
0.0
0.1
0.2
0.3
0.4
0.5
Relative Population Fecundity
Figure 5. Relative yield and terminal population fecundity of simulated
widow rockfish fishery under three types of management; agressive (A),
court-ordered (C), and present (P); and two spawn/recruitement
models; recruitment density independent (I) and 8everton-Holt (8).
Filled portions of circles are proportional to portion of time that fishing
is allowed. Yield under court-ordered management reflects additional
costs.