Monica J. Holboke and Daniel R. Lynch
September 1996
Abstract
The Maine Coastal Current(MCC) is initiated off the eastern coast of Maine and circuits the Gulf of Maine in a counterclockwise direction transporting various nutrients, biological species, and pollutants. Its role in the ecosystem of the Gulf of Maine is of significance importance not only as a transporter, but also as an interface between remote Gulf waters and local coastal waters. The Gulf waters are mainly composed of Scotian Shelf water, that is relatively cold and fresh, and Slope water, that is saline and warm. The coastal waters are dominated by local river runoff. In an attempt to understand how the March-April MCC structure affects the ecosystem, the mean transport pathways, local plume circulation, and wind induced variability are investigated. The investigations are conducted with a three-dimensional nonlinear time-stepping finite-element model of the Gulf of Maine and specifically the MCC. March-April is chosen for its large river runoff, observed intensification of the MCC, and for dynamic atmospheric forcing. Combination of the numerical dynamic results with passive Lagrangian particle tracks makes it possible to infer pathways via the MCC. The results also demonstrate the importance of the Gulf-scale circulation, the impact of fresh water steering, and the intensification or dissolution of the MCC under severe wind stress.
Objective:
The cyclonic circulation of the Gulf of Maine is a widely accepted circulation feature and in particular over Jordan Basin(Brooks 1985; Brown and Irish 1992,1993; Bisagni 1996). This is consistent with throughflow from the Scotian Shelf which enters near Cape Sable and follows the bathymetry around the Gulf, a seasonal pattern of dense slope water intrusion into and through the deep basins via the Northeast Channel, and the attendant baroclinic circulation. The MCC is dominated by these physical influences seaward of roughly the 100m isobath. Inshore of this point, the dynamics shift to local influences, including locally-variable winds, the along coast frontal structure, and buoyancy inputs from freshwater runoff at the coast(Lynch et al. 1996b).
The transport pathways of the MCC are of intrinsic interest; toxic phytoplankton blooms, nutrients, and pollutants are all transported via the coastal current(Franks and Anderson 1992a,b, Brooks and Townsend 1989).
A conceptual model : the MCC is centrally located over the 100m isobath with reaches inshore to the 50m isobath and offshore to the 180m isobath. The region inshore of the 100m isobath are determined by local influences and the region offshore by remote influences. It is has been proposed in Lynch et al. 1996b to be composed of seven legs and 3 branch points as shown in the figure below.
There are three objectives:
Figure 1: This schematic represents the MCC in terms of legs and branch points.
E- eastern leg, J - Jordan leg, W - western leg, M - Massachusetts leg,
S - Stellwagen leg, N - Nantucket leg, and G - Georges leg. 1 - is the
Penobscot branch point, 2 - is the Cape Ann branch point, and 3 - is the
Great South Channel branch point.
Figure 2: AF9 mesh - it is has 6758 nodes and 11763 elements. The circles
represent the nodal locations of where the river sources were introduced.
M - Merrimack river, S - Saco river, KA - Kennebec/Androscoggin river,
P - Penobscot river, SC - St. Croix river, and SJ - St. John river.
3-D Circulation Model:
Governing Equations: Using a time-stepping finite element method algorithm, the Dartmouth Circulation Model solves the nonlinear, three-dimensional, shallow water equations with the conventional Boussinesq and hydrostatic assumptions [see Lynch et al. 1996a,b]. Forcing is comprised of surface pressure gradients, surface fluxes, 3-D sources, and the baroclinic pressure gradient computed from a prognostically evolving density field. Vertical mixing is represented by a level 2.5 closure scheme[Mellor and Yamada, 1982; Galperin et al., 1988; Blumberg et al., 1992]. Horizontal mixing is represented by a mesh- and shear-dependent eddy viscosity similar to Smagorinksy(1963).
Continuity Equation:
Horizontal equations of motion:
Conservation Equation for Heat:
Conservation Equation for Salt:
A constitutive relationship relates the density to the temperature and salinity:
Procedure: The Gulf-scale simulation for March-April was computed on the G2S.5B mesh, see Naimie, 1996 or Lynch et al., 1996b. The Lagrangian trajectories were obtained by Fourier decomposing the 3-D velocity field and storing the residual and tidal constituents at all nodes. Fourth-order Runge-Kutta integration of individual trajectories was then carried out(Blanton 1992).
The seasonal results from G2S.5B were then interpolated for use on the AF9
mesh as initial and boundary conditions. The AF9 mesh is used to focus on the
coastal region with increased dynamical and bathymetric detail. The increased
dynamic detail inputs a decadal average of river source strength and
climatological heat flux(Lynch et al. 1996b). The trajectories for
the AF9 mesh were computed similarly to the G2S.5B mesh. The climatological
wind stress for March-April is 
towards 
The variable wind simulation was run on the CSTB.1A mesh with initial conditions and boundary conditions from the Gulf-scale seasonal simulation(Holboke et al. 1995).
Gulf-scale:The Gulf-scale residual vertically averaged streamfunction with the residual vertically averaged currents are shown. The cyclonic circulation of the Gulf of Maine is well represented with a double gyre over Georges and Jordan Basin. Along the 0.1Sv streamline there is a meander at the Penobscot branch point. Although the MCC is shown in the currents it is not well resolved.
Figure 3a: Gulf-scale streamfunction in Sverdrups
and
Figure 3b:Residual vertically
averaged velocity for March-April.
Trajectories:
The drogues were placed along the 50m, 100m, and 180m isobaths in both domains. The drogue placements were fixed at depths of 10m, 30m, and 60m. The purpose of these placements is to test the conceptual model and suggest possible transport pathways. The drogues are allowed to advect for 60days. In the figures below drogues are displayed with:
Figure 4: Drogues at 10m depth.
At 10m depth the light color drogues stay near shore with some being pulled into local features not well represented on this mesh. The dark red meander along with the MCC until they move inshore in the Western Gulf or exit out of the domain around Nantucket Shoals. The black meander across Wilkinson Basin with only one entering Mass Bay and the rest exiting the domain with the dark red.
Figure 5: Drogues at 30m depth.
At 30m depth the light color drogues are entrained in the MCC in the Western Gulf and transported out of the domain, while in the east they are pulled by local features inshore. The dark red follow the MCC staying on the inshore branch around Penobscot and do not enter Mass Bay however are split at the third branch between exiting the domain and turning towards Georges Bank. The black have significantly more structure with most following the inshore branch at Penobscot, and those that get entrained into the MCC around Stellwagen going to Georges Bank.
Figure 6: Drogues at 60m depth.
At 60m depth the MCC is represented in the eastern branch, while in the west it is not well represented. Near Stellwagen the drogues in the MCC end up on Georges Bank.
The conclusion of the Gulf-scale results yields a coastal current in the east that extends from the 50m to the 180m isobath from a respective depth of 10m to at least 60m. In the western Gulf, the MCC extends from the 50m to the 180m with a depth of at least 30m, but no more than 60m. In the southwestern Gulf the coastal current extends from the 50m to 180m isobaths from a respective depth of 30m to at least 60m.
Coastal Region:The residual vertically averaged streamfunction with an overlay of residual vertically averaged currents are shown. In general the structure is well matched to the Gulf-scale results with small differences. This demonstrates that the modeling approach is well-posed. It also details a well resolved coastal current.
Figure 7a: Coastal streamfunction in Sverdrups and
Figure 7b:Residual vertically
averaged velocity for March-April.
Figure 8: Drogues at 10m depth.
At 10m depth the light color drogues are mostly entrained in the MCC, with a few being swept into the coast from plume water or wind stress at the surface forcing return flow at depth. The dark red stay within the coastal current being swept to the western Gulf with some entering Mass Bay and others bypassing. The black look very similar to the Gulf-scale results. The difference here is around the Penobscot outflow, the rivers are on and this increases the cross-shore pressure gradient increasing along-shore flow and also spreading the MCC closer inshore.
Figure 9: Drogues at 30m depth.
At 30m depth the light color drogues are involved in complicated flow patterns associated with upwelling and river flow, in Mass Bay however they circuit the Bay and return to the MCC exiting to the south. The dark red flow with the MCC in the east, with some turning offshore at the Penobscot branch point. In the west they are involved in local features except near Mass Bay where they transit both the inshore and offshore legs here. The black again look similar to the Gulf-scale results.
Figure 10: Drogues at 60m depth.
At 60m depth the results become similar to the Gulf-scale with a definite eastern leg, no structured western leg, but again a clear Stellwagen leg.
The conclusions of the coastal region results demonstrate that the Gulf-scale response controls the 180m extent of the MCC, further verifying the modeling approach. The general structure is not changed, although there are differences near the river plumes, and in the western Gulf the MCC is shallower.
Plume Circulation
A horizontal slice at 10m, 30m, and 60m are taken around the river plume areas of the Merrimack, the Penobscot and Kennebec/Androscoggin, and the St. John. The figures display salinity at the respective depths along with the horizontal velocity structure as overlays. The results are consistent with the differences noticed in the Gulf-scale and coastal region drogue figures. The results also show the impact of fresh water steering on the MCC, especially near Mass Bay
Merrimack Plume: At 10m depth the influence of the river source on salinity is clearly seen expanding both upstream and downstream. The river water hugs the coast and bulges around Cape Ann causing a high pressure at Cape Ann. This forces the MCC to follow the pressure lines around Cape Ann and enter Mass Bay.
Figure 11a: Salinity and
Figure 11b: Residual velocity
at 10m depth near Merrimack River Outflow.
At 30m depth the influence of the river plume is still noticeable with the lowest salinity found at the southern side of Cape Ann. The MCC is again pulled into Mass Bay by the river plume hugging to Cape Ann as it expanded south. Upstream from the source at this layer we see a flow reversal caused by the gravitational spreading of the river plume and wind stress.
Figure 12a: Salinity and
Figure 12a: Residual velocity
at 30m depth near Merrimack River Outflow.
At 60m depth the river plume is barely noticeable and the MCC can be seen offshore outside of Mass Bay.
Figure 13a: Salinity
and
Figure 13b:Residual velocity
at 60m depth near Merrimack River Outflow.
Penobscot Plume and Kennebec/Androscoggin Plume: At 10m depth the Penobscot plume moves eastward down the bay through the islands until it connects up to the MCC. The MCC can be seen to be flowing out at the offshore boundary coincident with the branch point. Further downstream there is additional flow coming in through the boundary representative of the inshore branch returning to the coastal region. The Kennebec/Androscoggin plume moves upstream toward the Penobscot plume and downstream towards Casco Bay. It creates a complicated flow structure which is related to the gravitational spreading of the plume and wind stress.
Figure 14a: Salinity
and
Figure 14b:Residual velocity at 10m depth near
Kennebec/Androscoggin and Penobscot Rivers Outflow.
At 30m depth the Penobscot plume influence can be seen in the deep channel on the westward side of the bay. Here the flow is into the bay caused by the flow out at the surface and around the eastward side of the bay. The MCC can be clearly seen further offshore than at the previous depth, but with significant more shoreward flow on the return leg of the Penobscot branch point. The Kennebec/Androscoggin plume influence can still be seen with return flow at depth to counter the outward flow at the surface.
Figure 15a: Salinity and
Figure 15b:Residual velocity at 30m depth near
Kennebec/Androscoggin and Penobscot Rivers Outflow.
At 60m depth there is a signature of freshwater directly offshore from the Penobscot bay. However, it is difficult to ascertain if this is related to the Penobscot plume or the branch point which transports freshwater offshore here. However, the MCC can still be seen with an outgoing and incoming leg. The Kennebec/Androscoggin plume has very little influence to be seen at this depth.
Figure 16a: Salinity and
Figure 16b:Residual velocity at 60m depth near
Kennebec/Androscoggin and Penobscot Rivers Outflow.
St. John plume: At 10m depth the plume spreads horizontally related to the strong tidal mixing here sloshing the fluid back and forth. The currents flow anticyclonically around the fresh water signature, describing a high pressure. The flow then continues down the coast toward Grand Manan circuiting the island in both directions.
Figure 17a: Salinity and
Figure 17b:Residual velocity at 10m depth near
St. John River Outflow.
At 30m depth the plume is more uniformly distributed and the currents display more structure with inflow at the boundary now affecting the flow down the coast around Grand Manan.
Figure 18a: Salinity and
Figure 18b:Residual velocity at 30m depth near
St. John River Outflow.
At 60m depth the plume signature is still present, however now the flow field is moving up Bay of Fundy. This is to balance the outflow in the upper layers. Inflow at the boundary is still visible at this depth and the cyclonic circulation near Grand Manan that joins the MCC.
Figure 19a: Salinity and
Figure 19b:Residual velocity at 60m depth near
St. John River Outflow.
Observed Wind:The wind fields used for this simulation were assembled by Hui Feng and Wendell Brown at UNH. They used the NOAA buoy stations marked below within the Gulf of Maine and objectively analyzed the wind speed and direction onto our domain mesh. Their wind fields were then used to force our numerical model. The model was initialized from climatology on the coastal region mesh with river sources. The results shown here are vertical transects taken near the Kennebec/Androscoggin plume and the Merrimack plume. The transects are displayed in this image as lines emanating from the coast.
Figure 20: Wind Stations(dots) and Transect locations(lines).
Wind Event: The wind event occurred on March 6, 1986 and lasted for sixty hours. The wind vectors during this time period are plotted below for station 44005 and 44011. The wind at 44005 is representative of the wind at the other stations while on Georges Bank the wind is relatively calm although similar in shape. There are four periods shown at the two transects. The first period, 'A', is averaged over the spin-up from climatological wind, these results therefore are closely representative of the climatology. The second period, 'B', is averaged over the northwestward event. The third period, 'C', is averaged over the transition from northwestward to southeastward. The fourth period, 'D', is averaged over the southeastward event.
Figure 21: Wind Velocity(m/sec) starting on March 6,1986 and lasting for 60hours
at two buoy locations for Station 44005(top) and 44011(bottom).
Merrimack Salinity In period A the near shore field is representative of the plume and the coastal frontal structure. In B mixing from downwelling has made the isohalines more vertical. In period C upwelling has spread the isohalines far offshore, with a significant amount of fresher water evident. In D upwelling is still apparent with the isohalines increased horizontal extension.
Figure 22a: Isohalines along Merrimack Transect, Period A.
Figure 22b: Isohalines along Merrimack Transect, Period B.
Figure 22c: Isohalines along Merrimack Transect, Period C.
Figure 22d: Isohalines along Merrimack Transect, Period D.
Kennebec/Androscoggin Salinity In period A the near shore field details the existence of a river plume and along the coastal shelf a coastal frontal structure can be seen. Period B shows little difference with the isohalines becoming more vertical. Period C again shows the isohalines becoming more vertical probably due to the shearing of the fluid mixing the water vertically. In the final period, D, the water is spread offshore at the surface and upwelled in the bottom layers.
Figure 23a: Isohalines along Kennebec/Androscoggin Transect, Period A.
Figure 23b: Isohalines along Kennebec/Androscoggin Transect, Period B.
Figure 23c: Isohalines along Kennebec/Androscoggin Transect, Period C.
Figure 23d: Isohalines along Kennebec/Androscoggin Transect, Period D.
Kennebec/Androscoggin Normal Velocity Positive values indicate that the current is going into the page or up the coast of Maine, while negative values are directed out of the page or with the coast on its right. Period A shows the classic signature of the coastal current with most flow going out of the page and flow reversal at the bottom. In the next period, B, the current has increased in strength and spread deeper. Period C shows a reversal of flow, the surface is now flowing up the coast and the bottom layers are going down the coast. In the last period the coastal current is completely reversed with most of the flow going up the coast.
Figure 24a: Contours of Normal Velocity along Kennebec/Androscoggin Transect,
positive is into the page(coast on right) and negative is out of the page(coast
on left), Period A.
Figure 24b: Contours of Normal Velocity along Kennebec/Androscoggin Transect,
positive is into the page(coast on right) and negative is out of the page(coast
on left), Period B.
Figure 24c: Contours of Normal Velocity along Kennebec/Androscoggin Transect,
positive is into the page(coast on right) and negative is out of the page(coast
on left), Period C.
Figure 24d: Contours of Normal Velocity along Kennebec/Androscoggin Transect,
positive is into the page(coast on right) and negative is out of the page(coast
on left), Period D.
Conclusions: The first objective, to test the conceptual model, was achieved with it being upheld. The MCC is centrally located on the 100m isobath and it reaches into the 50m isobath and out to the 180m isobath. The branch points are evident most clearly in the trajectories, but can also be seen in the Penobscot and Merrimack plume figures.
The second objective, to determine transport pathways based on the conceptual model, was obtained. Near the surface particles are mainly affected by local features of river plumes and wind stress. Below the Ekman layer most particles are centrally located along the 100m isobath. At a greater depth Gulf-scale circulation determines the pathways.
The third objective, investigate wind-induced variability, was successful. The use of observational winds to force a numerical model achieved realistic results with interesting affects of intensifying the MCC during downwelling winds and dissolving it during upwelling events.
Acknowledgments: John Loder and Peter Smith provided the climatological wind and (T,S) fields; David Greenberg provided the original finite element mesh and boundary conditions; Francisco Werner and Brian Blanton provided the Lagrangian particle tracking software; and Wendell Brown and Hui Feng provided the observed wind fields and objectively analyzed them. Wendell Brown, Chris Naimie, and Wendy Gentleman have all provided additional insight relative to this work. Financial support from the Gulf of Maine Regional Marine Research Program and the New Hampshire Sea Grant College Program is gratefully acknowledged.
