Modeling the Gulf of Maine

Daniel R. Lynch, Dartmouth College

Over the past several years, I and several colleagues (see the bibliography) have been busy creating computer models of the Gulf of Maine. The funding and purpose has been multi-faceted, including individual projects with NSF, USGlobec, RMRP, and SeaGrant. The basic theme, however, has been constant: to provide a comprehensive 3-D model of the Gulf circulation, and to use it in pursuit of specific, focused scientific investigations with colleagues; and to transfer mature products to operational agnencies. Here I will try to provide an overview and progress report, emphasizing first the hydrodynamic circulation, and then the biological application of the computed circulation fields in specific studies.

Circulation Models

We use the finite element method. This involves dividing the ocean into triangular, rather than square, computational elements. Its principal advantage is that it allows variable local resolution, i.e. small triangles where there a lot of detail is needed, and big ones where less is needed. In this way we are able to achieve the wide geographic coverage indicated in Figure 1, including of the entire Gulf, beyond Georges Bank and Browns Banks, and the Scotian Shelf, while maintaining high local resolution (e.g. 1 km on Georges Bank).

There are two families of circulation models: diagnostic and prognostic. The diagnostic family is simpler; it depends on observations for the temperature and salinity, and computes the circulation implied by the observations. Because it has less work to do, it runs fast and we have used it extensively in early and exploratory studies. The prognostic model is more complex. It accounts for tidal-time transport of heat and salt; incorporates advanced turbulence closure, and generally represents the state-of- the-art fusion of finite element technology with the last two decades of research in coastal ocean turbulence closure. Running the prognostic model is our chief occupation today, and the results herein all come from it.

Circulation Model Results

There are three illustrative features of our solutions: the Gulf-wide or "Bigelow" circulation; the Maine coastal current; and Georges Bank.

The Bigelow Circulation: Figure 2 depicts the computed Gulf-wide tidally-averaged circulation, representative of March-April conditions. A dominant feature is the cyclonic gyre (recirculation approx. 0.3 Sv) over Jordan and Georges Basins. This feature has been identified in essentially every circulation study of the Gulf since Bigelow's seminal 1927 work. Its dynamical origin lies in the dense bottom water present in the deep basins. At depth there are separate gyres over the two basins, which merge near the surface. Figure 2 shows a cyclonic gyre in the "Scopex" region between Cape Cod and Georges Bank, also of baroclinic origin. Unrelated to these is a separate gyre over Grand Manan Basin, which originates in barotropic tidal rectification.

Along the shoreward boundary is a well- developed coastal current. In the eastern Gulf this current is the northern extent of the Jordan Basin gyre described above. It departs from the coast south of Penobscot Bay with only a portion returning to the coast. Further west, additional branch points are found at Cape Ann, where a portion of the coastal current enters and circuits Massachusetts and Cape Cod Bays; and east of Cape Cod, where a portion of the flow (0.1-0.2 Sv) is diverted east to Georges Bank, with the balance exiting the Gulf through Great South Channel. (For a more complete description, see [8].)

The Maine Coastal Current: Several of the above coastal current features have been studied at greater resolution, using a refined mesh of the northern and western Gulf coast. Representative results appear in Figure 3, wherein the branch point at Cape Ann is illustrated under March-April conditions. We are studying the details of this and the other coastal branch points with an emphasis on the transport and fate of contaminants and planktonic species to/from the various estuaries. In Figures 4 and 5 we show representative "numerical drifter" tracks, illustrating the diverse pathways in the Gulf. (See [12] for details.)

Georges Bank: To the south of the central Gulf, the familiar partly-closed gyre surrounding Georges Bank dominates the circulation. Such a gyre figures prominently in all studies of the Bank (e.g. [3], [7]). The central Gulf gyre merges with this feature along the northern flank and enhances the along-bank transport there. As part of the USGLOBEC program, we have constructed detailed circulation patterns on and around Georges Bank. In Figure 6 is a representative cross-frontal circulation pattern on a transect across the steep northern flank of the bank. The transport of nutrirents, zooplankton and fish larvae in the bank's frontal system is a primary concern of this aspect of our modeling (see [11], 14]).

Overall, these features are qualitatively very realistic by comparison with consensus opinion and with contemporary moored and drifting measurements of the circulation. Quantitative comparisons are ongoing. In addition, the model domain has been extended across the Northwest Atlantic shelf, to include the Scotian Shelf in greater detail, plus Cabot Strait and the Newfoundland Shelf ([4], [10], [16]).

Coupled Physical/Biological Studies

These circulation fields have been used in a series of studies of cod, haddock, and sea scallops during early-life stages [13]. A principal focus has been on identifying the constraints placed upon these populations by the combination of the physical environment, life cycle, and individual behaviour. There are two types of models: Individual-Based and Concentration-Based.

Individual-Based Models: These models follow the trajectories of individual organisms, as they move through the variable environment. The base state of the IBM is the passive particle, which involves only Lagrangian particle-tracking in the computational flow field. To this is added behavioural simulation, including buoyancy and swimming; growth and feeding relationships; and hydrodynamic dispersion via random walk. All of these features are in general sensitive to the current state of the individual (e.g. age, weight); to the ambient fluid state (temperature, salinity, stratification, turbulence); and to the biological environment (e.g. prey concentrations, predation). IBM's have been used on Georges Bank to study the interactions among 3 commercial aggregations of sea scallops [2]; advective influences on cod and haddock during the egg and larval stages [1, 6]; and the interaction of advection and trophodynamic influences on cod and haddock [5, 9]. A new study is just beginning which will focus on IBM simulation of lobster larvae in the Gulf.

Concentration-Based Models: CBM's are the more conventional description of marine ecosystem dynamics. Concentrations of species are the basic variables. Unlike the IBM, the identity of individuals is lost, so the direct link to biological behaviour is not possible. Instead, the coupled interaction among aggregates of several species (e.g. nutrients, phytoplankton, zooplankton) is represented. Generally speaking, we utilize the IBM's for simulating the target species as above; and the CBM's for the lower trophic levels.

At the simplest level, the transport of clouds of passive IBM's is reproduced by a CBM for a conservative substance (Figure 7). Beyond this simple level, the conceptual differences in the two approaches come into play and allow different types of investigations. Creative, hypothesis-focused use of both approaches is necessary, as a generally-accepted canonical description of the biological aspects of marine systems is likely to elude us for some time.

Communications and Collaboration

Collaboration on these studies beyond Dartmouth has been essential. In particular, colleagues at the Bedford Institute of Oceanography; the Univ. of New Hampshire; the NMFS Northeast Fisheries Science Center; St. Andrews Biological Station; the Pacific Biological Station in Nanaimo; and the Univ. of North Carolina. At each of these sites we have installed some modeling capability and are receiving both feedback on model performance and additional software. This community of model users is essential in broadening the participation of the scientific community in the modeling enterprise, and in providing proper scritiny of our model solutions. Ultimately, we hope to transfer all operational models to operating agencies on both sides of the border, where we hope their evolution can continue.

The internet has been an essential medium of information exchange in this project. All finalized software and circulation results are kept in a public archive which is accessible as a World Wide Web directory (http://www-nml.dartmouth.edu/gom/gom.html). We are busy now in developing an interactive graphical interface to this data via the combination of web browsers (e.g. Mosaic, Netscape) and graphical tools (e.g. AVS, MatLab). These services will be incorporated into the Gulf of Maine information management system under development within the RMRP. Finally, summaries of recent and ongoing studies using these models are maintained in World Wide Web servers at DFO [16], UNC [17] and Dartmouth [15].

For More Info:

A selected project bibliography follows. A complete bibliography and reprints are available upon request. See also the WWW document [15] for a more complete graphical overview.

Selected Bibliography

1) Influences of mean advection and simple behavior on the distribution of cod and haddock early life stages on Georges Bank. F.E. Werner, F.H. Page, D.R. Lynch, J.W. Loder, R.G. Lough, R.I. Perry, D.A. Greenberg, and M.M. Sinclair . Fish. Oceanogr. 2:2, 43-64 (1993)

2) Drift of sea scallop larvae Placopecten magellanicus on Georges Bank: a model study of the roles of mean advection, larval behavior and larval origin. J.M. Tremblay, J.W. Loder, F.E. Werner, C.E. Naimie, F.H. Page, M.M. Sinclair. Topical Studies in Oceanography (special issue on GLOBEC physical/biological modeling). Deep Sea Resch II, 41, 7-49 (1994).

3) Seasonal variation of the 3-D residual circulation on Georges Bank. C.E. Naimie, J.W. Loder, D.R. Lynch. J. Geophys. Resch. 99, C8, pp 15,967-989 (1994).

4) Seasonal variation of the baroclinic circulation in the Scotia-Maine region, C.G. Hannah and J.W. Loder. Proc. 7th Int. Biennial Conf. of the Physics of Estuaries and Coastal Seas, Woods Hole, MA, 28-30 November 1994.

5) A Coupled Individual-Based Trophodynamics and Circulation Model for Studies of Larval Cod and Haddock on Georges Bank, F.E. Werner, R.I. Perry, R.G. Lough, D.R. Lynch. U.S. GLOBEC News , Sept. 1994. http://www.opnml.unc.edu/manusc ripts/mans.html)

10) Hydrography and Baroclinic Circulation in the Scotian Shelf Region: Winter vs Summer. J.W. Loder, G. Han, C.G. Hannah, D.A. Greenberg and P.C. Smith. Canadian Journal of Fisheries and Aquatic Sciences Supplement. Submitted, 1995.

11) Georges Bank Residual Circulation during weak and strong stratification periods - Prognostic numerical model results, C.E. Naimie. J. Geophys. Resch. (in press 1995).

12) Dynamical Influences on the Maine Coastal Current. D.R. Lynch, M.J. Holboke, C.E. Naimie. Continental Shelf Research (submitted 1995b). (An early version is at http://www-nml.dartmouth.edu/nml/mcc_html/mc c2.html)

13) Cod Fishery Collapses and North Atlantic GLOBEC, M. Sinclair and F. Page, U.S. GLOBEC News , March 1995. (http://www.usglobec.berkeley.edu/usglobec /news/news8/news8.sinclair.html)

14) USGLOBEC Progress Report: Importance of physical and biological processes to population Regulation of Cod and Haddock on Georges Bank: a Model- Based Study. D.R. Lynch et al., http://www-nml.dartmouth.edu/gom/progress95/r eport.html (1995).

15) Numerical Methods Laboratory -- Environmental Program. http://www-nml.dartmouth.edu/nml/environmenta l.html (1995).

16) Northwest Atlantic Shelf Hydrodynamics, C.G. Hannah et al. http://biome.bio.dfo.ca/~channah/nas1.html (1995).

17) Ocean Processes Numerical Modelling Laboratory, F.E. Werner et al. http://www.opnml.unc.edu (1995).