Poster Presented at International GLOBEC Open Science Meeting in June 1998

Lewis Inzce and Christopher Naimie

INTRODUCTION

In the Gulf of Maine (southeastern Canada/northeastern U.S.), the American
lobster (Homarus americanus) recruits to the benthos as a small (ca. 10 mm
carapace length), cryptic  infaunal burrower during late summer-early fall.
 Sublittoral cobble habitat, which is generally coastal and shallow (<10 m
below mean low water) is a major recruitment substrate.  Reproductive
populations are broadly distributed around the Gulf and are abundant at
depths from 50-120 m.  Eggs are attached to the females for a 9-month
incubation period and hatch to produce planktonic larvae which swim in the
upper mixed layer.  Larvae develop through three stages before molting into
a postlarva, which is neustonic (mostly in the upper 0.5 m).  The duration
of each planktonic stage is strongly temperature-dependent and total
development may exceed 50 d.  

Circulation in the Gulf of Maine is energetic.  The combination of  slow
development and strong residual flows may lead to significant distances
between reproductive sources and recruitment sites.   Temperature and
currents vary seasonally and spatially, thereby making dominant recruitment
pathways difficult to predict without modeling.  Both the along-shore and
the across- shelf mechanisms of transport are of interest for a general
understanding of population dynamics and for management.

We are using a climatological-average 3-D circulation model of the Gulf of
Maine coupled with a simple individual-based model of larval and postlarval
depth regulation and temperature- dependent growth to examine the
contribution of the average circulation and temperature properties to
patterns of recruitment in the early life of the lobster.


METHODS

Model:  We use a 3-D finite element hydrographic model with advanced
turbulence closure (Lynch et al. 1995).   The model uses bimonthly
climatological averages of the temperature and salinity fields and wind
forcing.  We have interpolated between the bimonthly averages to yield
conditions at weekly intervals in order to smooth transitions between
previously archived runs.   The model is run in tidal time with electronic
drifters representing larval and postlarval stages.  Depth is fixed (2.5 m
for larvae, 0.5 m for postlarvae) and development is regulated by
temperature.  The model generates average transport values for each node in
the finite-element grid.  Because stream functions are smoothed (small- to
medium-scale eddies are not resolved), transport regimes can be depicted
with a relatively small number of drifters which we deploy from a series of
 transects.  

We use inverse solutions of the model to explore potential sources of
larvae which recruit to an area which we have studied for several years
(Incze et al. 1997, Wahle and Incze 1997), and forward solutions to
describe expected average trajectories of larvae from various source
regions around the Gulf.

A wind model of the diurnal sea breeze is used to examine its role in
onshore transport of the final, neustonic, postlarva stage along the
central coast of Maine.  Here we have strong evidence that postlarvae are
advected in from offshore (Incze et al. 1997), we have established that
wind transport is significant for coastal recruitment patterns (Wahle and
Incze 1997), and we have a reliable long-term record of wind-forcing from a
coastal buoy.  The model results we show here are from a modest sea breeze
which reflects local conditions: velocity = 5 m/s for 4 h with a one hour
period each for initial building and decline; offshore extent = 20 km).
Larvae, which are not neustonic, are moved by the general circulation
patterns in our calculations, but not by the wind.  We do this by moving
particles at 5 m, below the Ekman layer, but growing them at temperatures
of the upper layer (2.5 m).  We do not attempt wind calculations over the
entire Gulf because of a lack of spatially resolved data. The
climatological average bimonthly wind-forcing is unidirectional and
inappropriate for near- surface particles which are long-lived.  This is an
area in need of development.  In the meantime, we focus on along-shore
transport rates for larval stages and across-shelf transport for postlarvae. 

Results and conclusions are explained in the figure captions.

Fig. 1a. Top panel shows bathymetry (shaded, in meters) and transects (numbered) used for modeling drifter trajectories. Ten asterisks per transect mark drifter release points (forward runs) or end-points (inverse runs).

FIg. 1b. Middle panel shows the finite element grid used in the hydrodynamic model. Note increased resolution near shore, in shallow areas, and in regions of steep bathymetry.

Fig. 1c. Bottom panel shows the July-August climate-averaged residual circulation at 2.5 m depth. Note regions of enhanced flow in the northern part of the Gulf of Maine, and divergence and weakening of flow in the mid-coast region.

Fig. 2a. Bimonthly average temperature field ( C) at 2.5 m depth for May-June. Steep spatial gradients of temperature and seasonal warming along with the spatially-varying residual currents require a coupled biological-physical model to predict population development and transport of the pelagic stages.

Fig. 2b. Bimonthly average temperature field ( C) at 2.5 m depth for July-August. Steep spatial gradients of temperature and seasonal warming along with the spatially-varying residual currents require a coupled biological-physical model to predict population development and transport of the pelagic stages.

Fig. 3. Larval transport from eight transects (Fig. 1) using a release date of June 10 at all points. Lengths of the pathways are determined by transport rates and temperatures encountered along the way (because of their effect on development rates). At the end of each trajectory (open circles), we predict the end of the 3rd larval stage and beginning of the postlarva stage. Note that along-shore transport lengths differ markedly and that offshore transport, in the average, is predicted in several locations.

Fig. 4a. The average temperature encountered by larvae over the course of their development is quite uniform (transect numbers 1-8 are given along the x-axis. Transects are oriented as in Fig. 1 so that transect numbers are at the seaward end of each transect and each individual histogram is a release point). This result was not anticipated, and we conclude that larvae are moved rapidly from cold regions to warmer water because the coldest areas tend to have the highest transport rates (see residual velocities in Fig. 1; temperatures in Fig. 2; stage-specific durations in Fig. 5).

Fig. 4b. Differences in the net advective displacement of drifters show a strong east-west gradient in transport length-scales which was anticipated, but could not be predicted accurately without integration of current and temperature conditions.

Fig. 5. The duration of each stage and total larval development times from all release sites. Increased skewness in the distributions from Stage 1-3 appears to result from transport and seasonal warming. Spatial differences in the peak hatch (release) dates will be incorporated next.

Fig. 6. Inverse runs of the model show the anticipated sources (hatching) of larvae which arrive ready to molt to postlarva stage in the central coast of Maine, where we have studied postlarval abundance and recruitment for 7 years. Instead of using a single transect as the end-point, we have used the offshore perimeter of a polygon in order to take a more regional approach. The movement of larvae from the eastern to the central coast coincides with repeated observations of relatively low and high recruitment, respectively. Transport is lowest near shore. Note in Fig. 3 that the nearshore region of the eastern coast is not repopulated with larvae from transects 6 and 7 (southern Nova Scotia). Closer scrutiny of the oceanography of this region is warranted. Shading here and in Fig. 7 shows bathymetry, same as in Fig. 1.

Fig. 7a. Inverse runs of the model for postlarval stages arriving in the central coastal region (same region as in box described for Fig. 6). End-points are an along-shelf transect in 40 m of water and a single across-shelf transect in the center (transect #3 in Fig. 1). This figure shows trajectories from the climate-averaged calculations using transport values from 5 m depth (below the Ekman layer, same as larvae).

Fig. 7b. This figure shows that a modest diurnal sea breeze (see Methods) can capture a significant portion of the along-shelf moving postlarvae and transport them near shore. Thus, the sea breeze appears to have a local effect along shorelines as well as a population-level effect. Onshore flow during the postlarval season can vary by a factor of 2 or more between years and may vary along the length of coastline in the Gulf of Maine. Regional and interannual surface pressure differences (generated by land-sea temperature differences) and shoreline orientation are two factors that could account for this.