Sources, Transport, and Nutrient Environment of Toxic "Red Tide" Populations in the Western Gulf of Maine

D.M. Anderson, B. Butman, D.A. Fong, B.A. Keafer, P.J.S. Franks, W.R. Geyer, T.C. Loder III, and R.P. Signell

November 1996 Progress Report to the RMRP

During the last six months, we continued to work on the synthesis of data collected during the 1993 and 1994 field seasons. Although some raw data is still being generated (e.g., cell counts of Alexandrium ), more effort has gone into merging datasets and working on interpretation of the results. For example, the nutrient fields were integrated with the cell counts and hydrographic data using Matlab software. This required that some project staff work closely with other team members who were not as familiar with the complexities of Matlab.

Results of the project continue to be presented at conferences and workshops. B. Keafer and C. Martorano attended the biannual Maine Biotoxin workshop in October at the Maine Dept. of Natural Resources in Boothbay Harbor, ME where shellfish managers and other regulatory officials were given updates on the project during open discussions. The animations of the hindcast simulations for both 1993 and 1994 (available on the Web) were presented on videotape for attendees. Other presentations on the project were by DM Anderson at a NOAA workshop on Management and Mitigation of Harmful Algal Blooms, held in Seattle in September, and at a NATO Advanced Study Institute on the Physiological Ecology of Harmful Algal Blooms, held in Bermuda in May.

Progress within each of the project elements is given below:

Nutrients and cell counts

Significant progress has been made in merging the Alexandrium cell count data with the nutrient data, enabling a comparative analysis. Nutrient samples (approx. 2000) from the 1993 and 1994 field seasons have been analyzed and finalized for nitrate, nitrite, phosphate, ammonium, and silicate (1993) and nitrate+nitrite, phosphate, ammonium, and silicate (1994). Contour plots of surface (3 m) nutrient concentrations and DIN have been created to match with the accompanying surface contours of Alexandrium . These plots and analyses of the nutrient data relative to the occurrence of Alexandrium blooms are the subject of C. Martorano's Master's thesis at the University of New Hampshire which will be completed early in 1997.

At this time, data analysis suggests that the major sources of nutrients to the surface waters of the southwestern coastal Gulf of Maine are: the Kennebec/Androscoggin freshwater plume, the Eastern Maine Coastal Current, wind-induced upwelling; tidally-driven wave-sill interaction on Stellwagen Bank, local river inputs (such as the Merrimack), and sewage effluent from Boston Harbor.

Initial analysis also indicates that the occurrence and persistence of red tide blooms in the southwestern Gulf of Maine is not a result of inorganic nutrient enrichment. The toxic Alexandrium blooms typically occurred during the late spring after nutrients in the coastal current were well depleted by early spring phytoplankton blooms mainly consisting of diatoms. In particular, the highest Alexandrium cell concentrations observed during the 1993 and 1994 sampling seasons (907 and 908 cells/liter) were measured in waters which contained DIN levels of 0.30 and 0.73 , respectively. These low nitrogen levels indicate that the nitrogen was already ``stripped'' from the surface waters by the earlier diatom bloom. It is unlikely that Alexandrium itself, as only a minor component of the phytoplankton community, could have a major impact on the nutrient concentrations. We hypothesize that Alexandrium survives due to its ability to grow at very low levels (0-1 ) of DIN concentrations, and/or by its ability to obtain nitrogen from the pycnocline region of the water column through vertical migration. We suspect that once established at low levels, Alexandrium could grow rapidly if it were transported into an area rich in nutrients caused by upwelling, vertical mixing, or anthropogenic inputs. However, upwelling is not conducive to the nearshore accumulation of Alexandrium (see below), so the latter two mechanisms may be more important to bloom development and maintenance. This dinoflagellate seems to have excellent capability for adaptation, and does not appear to be constrained by its chemical environment. Thus, the physical environment may have been the more dominant factor regulating the distribution of Alexandrium in the southwestern Maine coastal current during 1993 and 1994.

Hydrography and modeling

The hydrography and moored current meter data have been analyzed to better understand the alongshore transport of the freshwater and Alexandrium . (Geyer et al., 1997) The structure of the plume is strongly influenced at short time scales by fluctuations in wind stress; however the wind stress is not the dominant contributor to the net transport of the plume. A persistent, along-coast, barotropic current is the biggest contributor to the plume transport, exceeding the contribution of the baroclinic flow due to the density anomaly of the plume itself.

Another important implication of the physical observations is that the timing of forcing conditions is critical in determining the transport and fate of the plume. Meteorological records show variability in both wind speed and direction over time scales of a few days. Winds fluctuate between upwelling, downwelling, and neutral conditions; the longer time-scale (monthly) averaged wind stress is nearly zero. Analysis of the observations from the spring freshets of 1993 and 1994 (Fong et al. 1996; Geyer et al., 1997) emphasize that the local winds are responsible for plume variability over short time scales unresolved by the hydrographic surveys.

The effects of transient wind events on mixing the freshwater plume are being investigated using a three-dimensional numerical model. This is a simplified version of the Blumberg-Mellor model used by Signell et al. (1994) in studying flows in Massachusetts Bay. The model is forced by a point source of freshwater and different wind forcing conditions. These idealized modeling studies aim at learning about the plume's behavior in the presence of fluctuations in wind stress.

As a generalization of the wind conditions present in the western Gulf of Maine, initial experiments were conducted by forcing the model with sinusoidally varying (in time) wind stresses of different amplitudes. These experiments demonstrated that even with a zero temporal mean wind stress, wind fluctuations result in significant mixing of the plume. As illustrated by Figure 1 , moderate winds of amplitude 0.5 are capable of causing significant ``dispersal'' of the freshwater in comparison to the zero wind forcing case. Strong wind fluctuations of amplitude 1.0 dramatically mix the freshwater with the ambient coastal water; in particular, strong sustained upwelling fluctuations appear capable of strongly mixing the plume such that the freshwater signal of the plume is largely removed.

More detailed and quantitative analyses are being conducted to investigate the role of individual upwelling and downwelling favorable wind events. Strong upwelling and downwelling winds are both capable of initiating significant mixing between the plume and ambient coastal waters. Downwelling winds compress the plume against the coast which results in a reduction in overall stratification; the reduced stratification, coupled with enhanced shears in the alongshore velocity field promotes enhanced mixing through shear instability processes. Upwelling winds, in contrast, spread and advect the plume offshore consistent with Ekman drift theory (see Fong et al., 1996); the associated shears with the cross-shore Ekman transport work against the stratification and enhance mixing.

Detailed analyses of the cumulative vertical buoyancy flux and salinity distributions show that upwelling winds are more effective at mixing the plume than comparable magnitude downwelling favorable winds. (Figure 2). The differences in mixing rates between the two wind conditions is attributable to differences in wind energy available for mixing the plume. Upwelling winds broaden the plume in cross-shore extent thereby providing a large surface area over which the wind's energy can be input into the mixing the plume. In contrast, the downwelling winds which push the plume tightly against the coast, decrease the surface area over which the wind can act to mix the plume and ambient coastal waters. The large differences in plume surface area for the different wind conditions result in at least twice, often three times, greater mixing rates for upwelling favorable wind conditions.

While these modeling exercises are simplified idealizations of the real western Gulf of Maine plume system, many important physical processes have been identified. The model runs clearly demonstrate that the temporal variability of winds is responsible for significant changes in plume structure and transport over time scales as short as a few days. Strong wind events, particularly upwelling favorable, appear capable of enhancing mixing to levels such that the alongshore transport of plume waters is strongly retarded, and the freshwater and Alexandrium are significantly dispersed in both the cross-shore and vertical directions. Given the combination of both elevated mixing rates and temporarily arrested of the downcoast transport, upwelling winds seem most capable of dispersing concentrations of Alexandrium cells. It seems likely that for sufficiently strong upwelling wind events, the downstream transport of freshwater and toxic cells may be greatly reduced; mixing will dramatically reduce the baroclinically-driven flow and make the ambient current the only major agent capable of net plume transport. The idealized simulations also emphasize that the timing of cell release with respect to wind events is critical in predicting algal bloom dynamics.

Studies are continuing with the idealized model to identify the locations of strongest mixing during large wind events. Preliminary results suggest intensified mixing occurs at the offshore edge of the plume. This behavior would suggest that red tide cell growth may be largest in the frontal regions of the plume given the elevated fluxes of nutrients associated with the amplified mixing. However, any increases in the population density due to growth at the offshore mixing zone may be outweighed by subsequent dispersion of the population during strong upwelling conditions. This also emphasizes the need to understand the timing and source conditions associated with release of toxic cells. The combined timing of buoyancy and wind forcing and cell release will determine the ultimate toxicity and location of high concentrations.

In addition to the idealized modeling, we have also numerically hindcasted the 1993 and 1994 field seasons using realistic bathymetry, measured wind, river and heat flux data, and boundary conditions derived from in situ observations. In these simulations we seek to determine the degree to which three-dimensional models can ``recreate'' the observed field data and, based on this determination, use the models to address issues that are unresolved by the field data.

One of the main hypothesis we tested with the hindcast simulations was that it was the differences in local wind forcing and river discharge between 1993 and 1994 were sufficient to explain the large differences in Massachusetts toxicity that was observed between these two years. In both years, however, the hindcast simulations show freshwater lenses containing Alexandrium cells working their way into Massachusetts Bay where the cells experience in situ growth due to warm water temperatures and relatively long residence times. This suggests that toxicity in Massachusetts cannot be attributed to physical variability of the wind and river discharge alone, but is dependent on biological variability as well.

One of the certain causes of biological variability is in the actual process by which the cells move from the cysts in the sediment into the fresh plume waters. In the hindcast simulations, however, cells were introduced in the river water with the same fixed concentration for both years. Had the fresh water lens that moved into Massachusetts Bay during 1994 contained no cells, the model would not have predicted toxicity. The simulations show that our present formulation of the source condition for the cells is much too simplified. A more accurate source condition, based on a better understanding of the process by which cells germinate and are transported to the surface waters, would allow a time variable source condition for cells that could explain the toxicity differences between 1993 and 1994. Without better understanding of the source conditions, however, it is not clear that further progress can be made in the hindcasting or prediction of red-tides in this region. We expect to pursue this level of detail in a forthcoming proposal to study the initiation zone in another large-scale Alexandrium study in the Gulf of Maine.


Fong, DA, Geyer, W.R. and RP Signell, 1996. The wind-forced response on a buoyant coastal current: Observations of the western Gulf of Maine plume. Journal of Marine Systems, in press.

Geyer, W.R., Signell, R.P, Fong, DA, Wang, J., Anderson DM, and BA Keafer, 1997. The Kennebec River plume and the coastal current of the western Gulf of Maine, in prep.

Signell, RP, Jenter, H.L., and A.F. Blumberg, 1994. Modeling the Seasonal Circulation in Massachusetts Bay, Proceedings of the 3rd International Conference on Estuarine and Coastal Modeling III, Publ by ASCE, NewYork, NY, USA., p 578-590.