Advection of Red King Crab Larvae in the Southeast Bering Sea: Interactions Between Changes in Spati

1
Advection of Red King Crab Larvae in the
Southeast Bering Sea Interactions Between
Changes in Spatial Broodstock Population
Structure and Physical Forcing Mechanisms. Timoth
y Loher David A. Armstrong University of
Washington, School of Fisheries and Aquatic
Sciences, Seattle, Washington Presently at the
International Pacific Halibut Commission,
Seattle, WA email tim_at_iphc.washington.edu web
http//www.iphc.washington.edu/staff/tim/tim.htm
Introduction The collapse of the Bristol Bay
red king crab fishery in the early 1980s has
been of concern to biologists, managers and
fishers alike. This population once supported a
lucrative fishery, but catches over the last 17
years have been modest in relation to the late
1970s. In our attempts to affect stock
recovery, we must understand the processes that
generate short-term and long-term fluctuations in
abundance. Models used to predict the rate and
likelihood of stock recovery tend to address the
problem numerically, based on stock recruitment
functions that treat the population as a single
unit. These models do not incorporate spatial
considerations, such as relationships between
components of the spawning stock and
oceanographic features or habitat. Yet, because
early juveniles have specific habitat
requirements, recruitment of early age-classes is
likely to be spatially structured, and the
contribution of spawners dependent upon their
physical location. Our understanding would be
enhanced by a greater knowledge of the
relationships between recruitment trends and
spatial stock structure, which could be
incorporated into spatially-explicit models of
stock function. Our research was aimed at
describing changes in adult and juvenile spatial
stock structure, and relating this structure to
changes in overall recruitment and potential
forcing mechanisms.
Larval drift predicted by the advection model
conformed closely to the underlying long-term
averaged flow patterns. Wind-driven advection
had little influence on larval displacement due
to the high degree of variability observed in the
wind data (i.e., summer winds rarely blew from a
consistent direction for extended periods), and
the tendency for mean wind direction to change
throughout the summer resulting in a drift loop
(lower right panel). Larvae hatched in spring
and subjected only to wind-drift were predicted
to travel relatively short distances and return
almost to their hatch location by the end of the
larval development period.
The Advection Model
A computer model was developed to simulate
advection of red king crab larvae in the
southeast Bering Sea, from hatch through
development to glaucothoe. Flow was driven by
observed long-term averaged currents, and by
wind-driven circulation. Larval development
proceeded according to a degree-day calculation
based on empirical observations of surface and
subpycnocline temperatures. All physical
parameters were modeled to fluctuate seasonally.
User-input allowed us to test the impact of
intermittent storms, variable thermal conditions,
and larval vertical migratory behavior on
predicted larval delivery patterns. The
flow-chart below outlines the steps involved in
calculating larval trajectories, given a
specified release location and hatch-date.
Early benthic phase (EBP lt age-3) red king crab
show a strong preference for complex substrates
in both controlled experiments (Stevens and
Kittaka, 1998), and field studies (Loher and
Armstrong, 2000). Highly fractal habitats thus
represent critical nursery habitat, and likely
provide refuge from predation as well as support
required prey populations. The availability and
quality of nursery habitat may be a critical
factor determining local recruitment strength.
As a result, the predicted regions in which
larvae must hatch in order to populate important
coastal nurseries (i.e., each coastal nurserys
larval envelope) occurred nearshore, generally
to the southwest of southern nurseries and to the
east of northern nurseries. There was no
evidence that larvae could be transported
substantially southward from hatch locations.
Furthermore, larvae hatched in the center of the
Bay were not expected to reach coastal nurseries,
even if larval envelopes were expanded by 50km in
all directions to account for likely dispersal
due to tidal energy.
However, complex benthic habitat is apparently
rare in the Bristol Bay region. Areas containing
substantial amounts of gravel are found in only a
few coastal areas (Armstrong et al., 1986), most
notably near Port Moller, Port Heiden and Kvichak
and Togiak Bays. Because red king crab larvae
are weak swimmers, they likely rely on currents
to ensure their delivery from hatch locations to
settlement sites. Long-term average currents in
Bristol Bay are characterized (Kinder and
Schumacher, 1981) by slow counterclockwise flow
inshore of 50m depth (Coastal Domain), with weak
and variable currents in the center of the Bay
(Middle Domain). These flow regimes are
separated by an oceanographic front. We
hypothesize that larvae must hatch at appropriate
upstream locations in order to reach suitable
nursery habitat, and that recruitment patterns
are sensitive to changes in hatch location and
oceanographic conditions.
In order to couple coastal nurseries with larval
hatch occurring near the center of the Bay, it
was necessary to impose strong winds upon the
system that blew from a constant direction for
long periods of time. However, no such wind
conditions were actually observed at St. Paul
Island from 1984-1999.
During the late 1970s and early 1980s, changes
were observed in the spatial distribution of
early juvenile (left panel) and mature female
(broodstock right panel) red king crabs.
Settlement appears to have been common in the
southern Bay early in the time series,
particularly near Port Moller and the Black
Hills no settlement was observed in Togiak Bay
and to the north. This recruitment pattern
reversed in the 1980s. Concurrently, broodstock
distribution shifted northward. Once common
along the Alaska Peninsula from Unimak Pass
through Port Moller, females have been largely
absent from these regions since the early 1980s.
As a result, the annual location of egg-hatch was
displaced to the northeast. Was the spatial
shift in settlement the direct result of shifts
in broodstock distribution? In particular, were
southern nurseries decoupled from larval supply
as females moved north? Larval advection
modeling was conducted to investigate these
questions.
Conclusion Links between spatial stock dynamics
and long-term recruitment seem apparent. The
shift from southern to northern nursery
productivity appears consistent with coupling /
decoupling of nurseries from spawning stock via
larval advection. The shift may explain some
lost recruitment in the stock if southern
nurseries were relatively more productive. Also,
different broodstock components probably have
different value in generating recruitment,
determined by their spatial relationship with
nurseries and regional oceanography. Still, our
understanding of larval advection falls short of
fully explaining recruitment patterns. In
particular, there is little evidence that the
surveyed spawning stock can seed coastal
nurseries with larvae based on long-term averaged
currents and wind-driven advection alone. Yet,
there is evidence that these regions support
populations of EBP crabs. This suggests that
either important segments of the broodstock
population lie nearshore, outside of the annual
survey, or that mechanisms such as
tidally-mediated transport are important in
larval delivery. A greater understanding of
larval distribution and transport, and of larval
behavior, are required in order to address these
issues. A better understanding of the processes
governing crab distribution will improve our
ability to predict long-term recruitment trends.
To some extent, these processes may be
environmentally determined and reversible, but
changes in spatial stock dynamics are not likely
to be governed by only one or two forcing
functions. Overall population structure is
likely the result of complex interactions between
physical, biological, and anthropogenic factors.
Acknowledgements We thank P.J. Stabeno, J.D.
Schumacher and A.J. Hermann for discussions
critical to the development of the model. AVHRR
temperature data were obtained from NOAA-NESDIS
(Suitland, MD), near-bottom temperature data from
G. Walters (NMFS, Seattle, WA) and S.R. Hare
(IPHC, Seattle), and wind data from the National
Climate Data Center (Asheville, NC). NMFS trawl
survey data were obtained with help from B.G.
Stevens (NMFS, Kodiak) and C.B. Dew (NMFS,
Seattle). The research was funded in part by the
University of Washington, and by NMFS
Saltonstall-Kennedy Grant NA76FD0036. References
Armstrong, D.A., Incze, L.S., Wencker, D.L.
and Armstrong, J.L. 1986. Distribution and
abundance of decapod crustacean larvae in the
southeast Bering Sea with emphasis on commercial
species. US Department of Commerce, NOAA, NOS,
Anchorage, Alaska. Kinder, T.H., and Schumacher,
J.D. 1981. Circulation over the continental shelf
of the eastern Bering Sea. In Hood, D.W. and
Calder, J.A. (eds.), The Eastern Bering Sea
Shelf Oceanography and Resources. University of
Washington Press, Seattle. Pp. 53-75. Loher, T.
and Armstrong, D.A. 2000. Effects of habitat
complexity and relative larval supply on the
establishment of early benthic phase red king
crab (Paralithodes camtschaticus Tilesius, 1815)
populations in Auke Bay, Alaska. Journal of
Experimental Marine Biology and Ecology
24583-109. Stevens, B.G. and Kitakka, J. 1998
Postlarval settling behavior, substrate
preference, and time to metamorphosis for red
king crab, Paralithodes camtschaticus. Marine
Ecology Progress Series 67197-206.