BULLETIN OF MARINE SCIENCE, 73(2): 291–305, 2003

THE IMPACT OF COASTLINE CHANGE AND URBAN

DEVELOPMENT ON THE FLUSHING TIME OF A COASTAL
EMBAYMENT, KINGSTON HARBOUR, JAMAICA

Grant R. Bigg and Dale F. Webber

ABSTRACT
Kingston Harbour, Jamaica, has a history of eutrophication due to nutrient loadings
emanating from the city of Kingston. In this study a two dimensional, numerical, hydro-
dynamic model of the harbor is used to assess the impact of coastal development at the
western end of the harbor and on the evolution of this eutrophication. The model shows
that the development did not materially affect the flushing of the harbor; therefore, the
rapid increase in eutrophication since 1970 must be linked to concurrent population pres-
sure. However, the coastal modification has altered the circulation pattern within Hunts
Bay and has reduced the exchange between Hunts Bay and the remainder of Kingston
Harbour. Increased siltation and further eutrophication within Hunts Bay are predicted as
the impact of the modification. The model was also used to investigate the feasibility of
aiding the flushing of the harbor by adding a second entrance. The model showed that
such an opening is likely to have negligible impact on the present flushing time.

Kingston Harbour is the major port for the Caribbean island of Jamaica with the capital
of Kingston and its suburbs (with a population of 1.1 million people), spreading along the
northern and western side of the harbor (Fig. 1). Substantial industrial activity also takes
place along these shores. The harbor offers good protection for shipping, as the only
entrance to the sea is a channel in its southwestern corner, 3.5 km wide and 10–20 m
deep. The inner harbor is 12–18 m deep, sheltering behind the Palisadoes, a mangrove
and sand spit to the south, and is connected to the outer harbor via a shallow sill extending
in a southwesterly direction across the western end of the harbor.

Figure 1. Map of Kingston Harbour, showing principal regions, depth contours, 14 major river
inflows and location of the three current meters deployed.

291
292 BULLETIN OF MARINE SCIENCE, VOL. 73, NO. 2, 2003

The protected but deep water character of the inner harbor has, however, had a delete-
rious effect on its water quality. Wade et al. (1972) showed that Kingston Harbour had
begun to display eutrophic tendencies in the early 1970s because of large nutrient inputs
from the domestic waste of Kingston. During the course of their multi-year survey of
benthic species (conducted between 1968 and 1974) they recorded a rapid increase in the
area of abiotic seafloor, spreading from the eastern end of the harbor, and a similar de-
cline in species diversity on the harbor floor as a whole. More recent surveys of the
Hellshire coast south of Kingston Harbour (Morrison and Greenaway, 1989; Lindo, 1991;
Webber et al., 1992; Webber and Roff, 1996; Webber et al., 1996) have shown that the
impact of high nutrients flowing out of the harbor normally extends several kilometers
southwards, although little associated trace metal pollution has been found in this region
(Greenaway and Rankine-Jones, 1992).
Much of this export of nutrients is due to the riverine flushing of the western harbor
through Hunts Bay, into which the harbor’s major fresh water sources empty. During
flood conditions a plume extending from Hunts Bay along the Hellshire coast was ob-
served (Webber et al., 1992). Its width is defined by the width of the harbor entrance, and
even during flood conditions its impact is restricted to the upper 7 m of the water column.
During normal flow conditions a near-surface tongue of low salinity (< 33, compared to
36 off the Palisadoes) shows that a net southwards transport is a permanent feature of the
local circulation (Shurland et al., 1989; Webber and Roff, 1996).
The larger, eastern part of the harbor has almost no riverine flushing and a mean tidal
range of only 15 cm (maximum of 45 cm). The Blue Mountains to the north effectively
shield Kingston from the prevailing northeasterly winds, so that the local winds are strongly
controlled by the diurnal heating cycle, leading to a daily oscillation between southeast-
erly and northerly winds – a classic sea breeze system (Fig. 2). No mechanism for the

Figure 2. (A) Wind speed, in ms-1and (B) direction at Norman Manley Airport during the observation
period.
 BIGG AND WEBBER: TWO DIMENTIONAL NUMERICAL, HYDRODYNAMIC MODEL OF KINGSTON HARBOUR 293

Figure 3. Map of Kingston Harbour before construction of the Causeway enclosing Hunts Bay.

rapid exchange of water in the inner harbor with the open sea is clearly identifiable,
suggesting long residence times, perhaps of weeks. This conclusion was reached by Young
et al. (1993) using idealized numerical and analytical models of Kingston Harbour. Ex-
tended flushing times for the inner harbor would be consistent with the limited biological
evidence discussed above.
However, a significant change to the coastal morphology of the original western harbor
and Hunts Bay area was made as part of the development plans for the Portmore area in
the mid-1960s (see Fig. 3), which shows the harbor prior to this development. These
engineering works were carried out only a few years prior to the survey of Wade et al.
(1972), and involved substantial narrowing and focusing of the junction of Hunts Bay
and the outer harbor, as well as a construction of a 12 m deep, protected body of water
called Port Bustamante (see Fig. 1). Access to Hunts Bay from a smaller subsidiary inlet,
Dawkins Pond (see Figs. 1,3), was closed with a very narrow entrance channel dug at this
pond’s southern end, near the mouth of the harbor.
The objective of this study was to use a two-dimensional finite-element model to ex-
amine the circulation of, and particle trajectories within the modern harbor during both
normal and flood conditions, and to contrast this with the flow field prior to the 1960s
development. The consequences for the flushing time of the modern Kingston Harbour of
opening an additional entrance at its eastern end were also investigated.

MODEL DESCRIPTION

Stratification within Kingston Harbour is generally weak and transient, as a result of

mixing due to strong daily southeast Trade Winds (Fig. 2). There is rarely more than a
0.5º C variation vertically, and the typical vertical salinity gradients are less than 1.5,
except within Hunts Bay. Here, a shallow lens (< 1 m deep) of fresh water slides above
salty water (Ranston and Webber, 2003). With weak inflow from the Rio Cobre and the
Duhaney River, this lens is almost absent, while with normal inflow the vertical salinity
294 BULLETIN OF MARINE SCIENCE, VOL. 73, NO. 2, 2003

Figure 4. Distribution of elements used in the finite element model. See Fig. 7 for the Harbor
outline on its grid.

gradient can exceed 10. During flood conditions the major river inflow can increase by a
factor of 20, and subsidiary run-offs by a factor of 100 (Webber and Wilson-Kelly, 2003).
This would lead to significant freshening of Hunt’s Bay. Its effect on outflows from
Kingston Harbour proper is, however, restricted to a depth of 7 m (Webber et al., 1992).
The generally weak stratification within the harbor means that the flow equations can
be simplified to two dimensions with little loss of accuracy. Only within Hunts Bay, when
the river flow is significant, and in the western harbor during flood conditions, will this
assumption be violated. In addition, it is recognized that variations in the vertical struc-
ture of currents within the water body will not be represented in a two dimensional model.
The equations of motion in two dimensions are well known and can be found in Ninomiya
and Onishi (1991) or Young et al. (1993).
These equations were solved numerically by digitizing the complex geometry of the
harbor into 4514 finite elements (Fig. 4) and using the Galerkin finite element method to
construct a matrix equation linking the velocity and height fields of each of the 2415
nodes. This system is then stepped forward in time. For details of the solution technique
see Ninomiya and Onishi (1991) or Walters (1992). The depths used (Fig. 1) were mainly
taken by interpolating from Admiralty chart 454, with additional spot soundings taken in
Hunts Bay. This chart was also used to provide the geometry of the harbor shore, al-
though local investigation was required to correct some errors near Hunts Bay and Dawkin’s
Pond. The model used representative values of Chezy coefficient, C = 86.4 m1/2 s-1, and
eddy viscosity, me = 20 m2 s-1, with the wind stress drag being a function of wind speed as
in Gill (1982).
There are three types of boundary conditions for these equations. Along the seashore
the normal component of velocity is zero. At river discharge nodes, the normal velocity is
a prescribed flux. Out at sea, the astronomical, or observed tidal height, with the normal
velocity given by the local normal height gradient, is prescribed at the open boundary.
The open sea boundary condition assumes that any large-scale mean flow is negligible,
 BIGG AND WEBBER: TWO DIMENTIONAL NUMERICAL, HYDRODYNAMIC MODEL OF KINGSTON HARBOUR 295

Figure 5. Comparison of model (solid) and observed (dotted) current data. (A) KH-1, (B) KH-2,
(C) KH-3. See Figure 1 for location.

which, given the prevailing alongshore westward flow along the south coast of Jamaica,
is realistic. All experiments were initialized with zero velocity and height fields through-
out the model interior.
Five experiments are described below. Two of these take the existing geometry of
Kingston Harbour and examine the circulation patterns found at low (I) and high (II)
river inputs. The next experiments also consider low (III) and high (IV) river discharge
conditions, but with the geometry of Kingston Harbour as it was prior to 1960. The final
experiment (V) considers the impact on the circulation pattern of adding a 5 m deep
channel, 600 m wide, in the eastern end of the Palisadoes.
All experiments have the same tidal and wind forcing (Fig. 2) from a 7-d period. The
tidal record (Figs. 5,6) were from a tide gauge deployed offshore from Port Royal, in the
harbor mouth (Webber et al., 2003). During idealized experiments it was found that there
is a 1.3 hr delay, with little change in amplitude, between a tidal signal imposed at the
open ocean boundary of the model and the location of the tide gauge. Thus, the Port
Royal record was used to force the open boundary, with a time lag of 1.3 hrs. The hourly
wind record was provided by the Jamaica Meteorological Office at Normal Manley Air-
port (Fig. 1). Field experience suggests that the wind at this location is generally repre-
296 BULLETIN OF MARINE SCIENCE, VOL. 73, NO. 2, 2003

Figure 6. Comparison of model (solid) and observed (dotted) current data. (A) Meridional speed at
KH-1 (meter at 6 m), (B) Current direction at KH-2 (meter at 2.5 m), (C) Current direction at KH-
3 (meter at 2.5 m).

sentative of conditions over the entire harbor, so the same wind field was imposed, for a
given time step, over the whole model domain.
The river inputs were taken from Webber and Wilson-Kelly (2003). Twenty separate
input sites were located, although only 14 of these had a significant flow even at high
flow conditions. These are numbered on Figure 1. The differences in flow rate were be-
tween one and two orders of magnitude. At low flow only five of these sites were in-
cluded in the model (see Table 1).

MODEL VALIDATION

Experiment I was used to validate the model during low flow conditions. Instruments
deployed included three tide gauges and current meters at depths of 2.5 and 6 m near Port
Royal, which was used to force the numerical experiments. The three sites are labeled on
Figure 1 as KH-1, KH-2 and KH-3.
Comparison of the numerically generated and observed tidal records and current speed
and direction at the three sites is given in Figures 5 and 6. The observed currents are point
 BIGG AND WEBBER: TWO DIMENTIONAL NUMERICAL, HYDRODYNAMIC MODEL OF KINGSTON HARBOUR 297

Table 1. Depth-averaged velocities used as river inputs. Low flow sites with zero input are not
included in low flow experiments.

Site Location (in km grid) Low flow in cm s-1 High flow in cm s-1
1 (2.00, 8.33) 1.0 20.0
2 (2.00, 8.50) 1.0 20.0
3 (7.00, 7.50) 0.6 0.6
4 (4.58, 9.75) 0.0 1.0
5 (6.67, 7.50) 0.0 1.0
6 (6.83, 7.50) 0.0 1.0
7 (3.42, 10.00) 1.5 3.0
8 (3.33, 10.00) 1.5 3.0
9 (3.83, 10.25) 0.0 2.0
10 (3.92, 10.25) 0.0 2.0
11 (7.50, 7.00) 0.0 0.4
12 (7.67, 6.75) 0.0 0.4
13 (7.83, 6.50) 0.0 1.0
14 (11.83, 6.83) 0.0 1.0

measurements, from a depth of 2.5 m at KH-2 and KH-3 and 6 m at KH-1, while the
model currents are depth-averaged. The comparison of tidal heights at each site is good
(Fig. 5). However, because the current comparisons are between single depth observa-
tions and depth-averaged computations, the velocity comparisons show only moderate
correlation of peaks and dominant tidal directions. This moderate correlation is seen at
site KH-1 in the dominant meridional velocity (Fig. 6A) where the zonal velocity is much
smaller in both the model and observations. The site KH-2 was close to a secondary
shipping channel in the harbor and was subjected to extensive wash and other physical
disturbance (the record was short because the gauge was ‘rescued’ by a local fisherman).
Therefore, current directions could only be compared at KH-2 (Fig. 6B). KH-3 was in the
mouth of the Hunts Bay, where the flow is dominantly two layer and the depth-averaged
model is less applicable; sometimes the current meter was in the surface layer and some-
times in the subsurface layer. Thus, only the observed current direction is shown for KH-
3 as well (Fig. 6C).
As a consequence of the above issues, the trajectories reported here must be considered
as being representative of the bulk flow within the harbor, rather than that of the surface,
or any particular depth. Observations suggest that water in the upper few meters in deeper
reaches of the harbor may be flowing much faster than the depth-averaged velocity in
extreme circumstances. This is partly due to surface wind forcing and partly due to a
pronounced tendency for the harbor circulation to have a reversal of flow direction be-
tween the surface and deep waters (Webber et al., 2003). Thus, in Figure 6A the negative
of the observed 6 m KH-1 velocity is compared to the model depth-averaged velocity
becasue observations suggested that this current meter was in the shallow upper layer
(Webber et al., 2003). Discharges into the surface layer in such deep areas, particularly if
they are relatively buoyant (that is, warm and/or fresh), will follow similar paths (if with
a different tidal phase) to those discussed below, but much faster. This is also likely to be
true of water in the top meter of Hunts Bay, where for high flow conditions the stratified
nature of the Bay weakens the validity of the model.
298 BULLETIN OF MARINE SCIENCE, VOL. 73, NO. 2, 2003

Figure 7. Experiment I (low river flow): typical ebb tide (hr 60).

PARTICLE TRAJECTORIES IN THE PRESENT DAY KINGSTON HARBOUR

The tidal flow within Kingston Harbour is principally aligned along the length of the
harbor, except in the vicinity of the bays on the western side. The typical circulation patterns
for the peak flow during ebb and flood tides are shown in Figures 7 and 8, and for low river
flow conditions (experiment 1), corresponding to hrs 60 and 54, respectively, in Figures 5
and 6. The fastest depth-averaged tidal flows, approaching 5 cm s-1 occur over the complex,
but mostly shallow bathymetry across the entrance to the inner harbor. The relatively large
volume of water tidally forced into the harbor is restricted to flow through the sill of the
narrow shipping channel, and consequently this region experiences larger velocities. Note,
however, that the depth-averaged tidal flow within the majority of the inner harbor and the
docks along the north side of the harbor is very weak, less than 1 cm s-1.

Figure 8. Experiment I (low river flow): typical flood tide (hr 54).
 BIGG AND WEBBER: TWO DIMENTIONAL NUMERICAL, HYDRODYNAMIC MODEL OF KINGSTON HARBOUR 299

Figure 9. Trajectories for a typical tidal period in experiment I (low river flow). The water’s point
of origin is indicated by large dot; final point is indicated by small dot.

The impact of the two major rivers draining into Hunts Bay, even at low flow, is clearly
seen within that bay. Note that the circulation in the western part of Hunts Bay has anticy-
clonic curvature, re-circulating water from the Rio Cobre within this part of the Bay,
rather than draining it through the connection with the main harbor (Andrews et al., 1999).
The Duhaney River reinforces the main tidal circulation within the Bay, although during
the flooding tide the circulation is deflected into the northeast corner of Hunts Bay.
Weeklong trajectories for particles starting at a representative sample of points through-
out Kingston Harbour (experiment 1; Fig. 9) shows that the net depth-averaged move-
ments of pollutants at any point within the harbor is very small under low flow conditions
(i.e., initial and final particle locations cannot be distinguished even when tidal oscilla-
tions > 100 m). Nowhere does it exceed 50 m for the week simulated. Careful inspection
of the trajectories shows a net movement into the harbor from the open ocean, and an
even smaller movement out of Hunts Bay. Thus, for this fairly typical wind pattern and
riverine input there is a very small net accumulation of water within the harbor. There-
fore, it is likely that the flushing of the inner harbor for much of the year is driven by
diffusive forces, and the flushing time may be of the order of several months, if not
longer. Such stagnant flow would explain the biological desert found in the early 1970s
over much of the harbor benthos and the accompanying reduction in water quality (Wade
et al., 1972; Grahame, 1976; Webber and Wilson-Kelly, 2003).
High river flow conditions make only a minor alteration to these findings. Typical peak
ebb and flood tide circulation patterns for such conditions (experiment II) are shown in
Figures 10 and 11. The much stronger river flow, most marked in the Rio Cobre, necessi-
tates a new arrow scale. However, the impact of the increased river input is restricted to
Hunts Bay and the outer harbor. The weak circulation of the inner harbor is practically
unaltered from experiment I–II. This is reflected in the trajectories of seeded particles
(Fig. 12). Only in Hunts Bay, or its entrance, is there clear net movement of the particles
out from the Bay (recall that this is reflecting the depth-averaged circulation; the surface
300 BULLETIN OF MARINE SCIENCE, VOL. 73, NO. 2, 2003

flow will be significantly greater). However, within the Bay, the net flow tends to be
trapped in the southwest corner, suggesting that the Bay may not be completely flushed
even during high flow conditions. Note that because of this increased flow, there is a
weak net movement of water from the outer harbor towards the open sea. However, the
inner harbor remains a convergence zone.
These experiments show that the tidally induced, depth-averaged motion of much of
the harbor is very slight, for typical wind conditions, irrespective of the river flow state.
Only in Hunts Bay itself is there appreciable net depth-averaged outflow at high flow,

Figure 10. Experiment II (high river flow): typical ebb tide (hr 60).

Figure 11. Experiment II (high river flow): typical flood tide (hr 54).
 BIGG AND WEBBER: TWO DIMENTIONAL NUMERICAL, HYDRODYNAMIC MODEL OF KINGSTON HARBOUR 301

Figure 12. Trajectories for a typical tidal period in experiment II (high river flow). The water’s
point of origin is indicated by large dot; final point is indicated by small dot.

and even then it would take several weeks for the entire Bay to be flushed. However, it
must be remembered that the near surface water flows significantly faster than the depth-
averaged value because of the reversal of current direction with depth found in recent
observations (Webber et al., 2003). For example, the current meters deployed off Port
Royal frequently show a five-fold surface amplification over mid-depth flow speeds, as
the latter were near the level of no motion. Thus, during high flow states in Hunts Bay,
much of the surface-trapped, fresh river water will be swept out of the harbor within a
few days (Shurland et al., 1989). The inner harbor, encompassing the Dock area, Petro-
chemical plants and much of the harbor’s industrial activity will, nevertheless, experi-
ence very weak flushing even at the surface. The residence time for sub-surface water
within this region is likely to be significantly longer than the time scales considered in
this study, perhaps in units of years.

THE IMPACT OF DEVELOPMENT ON THE CIRCULATION OF KINGSTON HARBOUR

Prior to 1968, the shape of the western end of Kingston Harbour was very different
from today (see Fig. 3). More recent development has nearly closed off Hunts Bay; it
formerly had a mouth 2–3 km wide. Dawkin’s Pond has been sealed off from Hunts Bay
and given a new, very narrow exit (10 m wide) at its southern end. A new dock in the
northwest corner of the main harbor has been built, reducing flow in that area signifi-
cantly.
To examine the impact of these changes on the circulation and flushing time of the
harbor, the numerical model was adapted to the old geometry, with depths of former sea
areas interpolated from adjacent locations at present day. Two experiments were carried
out: a high river flow ebb tide (III) and high flow flood tide (IV) state and using the same
flow data, tidal and wind data as for experiment II.
302 BULLETIN OF MARINE SCIENCE, VOL. 73, NO. 2, 2003

The principle difference between the experiments with present day and pre-1960 ge-
ometry is in the flow between Hunts Bay and the outer harbor (cf. Figs. 10, 11, 13, 14).
The tidal flow at the western end of the inner harbor is also increased by about 10%.
However, these changes are not sufficient for significant alteration to the depth-averaged
trajectories in either low or high river flow states; trajectories for experiments III and IV
are nearly identical to those in Figures 9 and 12, thus were not presented. The long flush-
ing times inferred in the last section for the inner harbor will have also been characteristic

Figure 13. Experiment IV (high river flow in pre-1968 harbor): typical ebb tide (hr 60).

Figure 14. Experiment IV (high river flow in pre-1968 harbor): typical flood tide (hr 54).
 BIGG AND WEBBER: TWO DIMENTIONAL NUMERICAL, HYDRODYNAMIC MODEL OF KINGSTON HARBOUR 303

Figure 15. Experiment V (low river flow with Palisadoes entrance): typical ebb tide (hr 60).

of the pre-1968 harbor; environmental degradation of the harbor has proceeded indepen-
dently of the coastal morphology.

BREAKING THE PALISADOES

One possible remedy to the eutrophication of the inner harbor is to add another connec-
tion to the open ocean through the Palisadoes itself. The resulting additional tidal forcing
might shorten the flushing time, either through the new exit, or through the existing har-
bor entrance. This possibility was tested numerically by adding such a channel, 500 m
wide and 5 m deep, at the eastern end of the Palisadoes. The same tidal and wind forcing
was applied to this experiment (V) as for the others, and the river input was taken as the
low flow state because of the lack of effect of river flows in the inner harbor in the
previous experiments (II and IV). No tidal phase difference was imposed between the
new entrance and the open ocean boundary; in reality the hypothesized entrance would
tidally lead the main harbor by about 10 min, but this difference was considered negli-
gible.
As can be seen from a map of the circulation during the ebb tide (Fig. 15; the flood tide
map shows essentially the reverse circulation), there is little impact on the circulation
more than 1 km from the new entrance. The one exception to this is west of the airport,
where the addition of another entrance produced a node in the flow field, locally restrict-
ing the circulation. Trajectories over the whole harbor, however, were essentially identi-
cal to those of experiment I (Fig. 7).
The low tide amplitude experienced in Kingston Harbour, coupled with the harbor’s
bathymetry, which isolates the inner harbor effectively from the open ocean, act to reduce
the flushing of the harbor. Neither the changes in the coastal morphology accompanying
the development of the 1960s, nor the addition of another connection to the open ocean
makes a significant difference to the depth-averaged flow in the inner harbor.
304 BULLETIN OF MARINE SCIENCE, VOL. 73, NO. 2, 2003

Therefore, the eutrophication of Kingston Harbour can only be reversed by control of

the domestic and industrial waste presently released into it. The slow flushing time of the
harbor suggests that even with such waste being diverted, the return of the harbor benthos
to a pre-1968 state would take years rather than months.
The numerical experiments have also shown that the impact of flood conditions on the
depth-averaged circulation for much of the harbor is small, although the surface signature
can be dramatic. The result is consistent with the findings of Webber et al. (1992). Note,
however, that Hunts Bay and its near vicinity form an exception. This region is also the
one location where the 1960s coastline change has lead to weaker exchange with the
open ocean.

ACKNOWLEDGEMENTS

We wish to thank the British Council for supporting the link between our respective universities,
which resulted in this paper. We also wish to thank the Center for Marine Sciences at the University
of the West Indies (CMS-UWI), the Center for Engineering and Environmental Management of
Bays and Coastal Areas (CIMAB) and the Natural Resources Conservation Authority (NRCA) for
providing the expertise and equipment. The 1993 survey was funded by a United Nations Environ-
mental Programme (UNEP) Technical Cooperation Plan.

LITERATURE CITED

Andrews, J. E., A. M. Greenaway, G. R. Bigg, D. F. Webber, P. F. Dennis and G. A. Guthrie. 1999.

Pollution history of a tropical estuary revealed by combined hydrodynamic modeling and sedi-
ment geochemistry. J. Mar. Syst. 18: 333–343.
Gill, A. E. 1982. Atmosphere-ocean dynamics. Academic Press, New York. 662 p.
Grahame, S. 1976. The phytoplankton of Kingston Harbour, Jamaica. Scientific Report of the UWI
ODA Kingston Harbour Project 1972-1975. Zoology Department Research No. 4, University
of the West Indies, Kingston, Jamaica. 101 p.
Greenaway, A. M. and A. I. Rankine-Jones. 1992. Elemental concentrations in coastal sediments
from Hellshire, Jamaica. Mar. Poll. Bull. 24: 390–397.
Lindo, M. K. 1991. The effect of Kingston Harbour outflow on the zooplankton populations of
Hellshire, southeast coast, Jamaica. Estuar. Coast. Shelf Sci. 32: 597–608.
Morrision, B. and A. M. Greenaway. 1989. Nutrient dynamics. Pages 19–31 in I. Goodbody, ed.
Caribbean coastal management study: the Hellshire Coast, St. Catherine, Jamaica. University
of the West Indies, Kingston.
Ninomiya, H. and K. Onishi. 1991. Flow analysis using a PC. CRC Publishers, Southampton.
190 p.
Ranston, E. R. and D. F. Webber. 2003. Phytoplankton distribution in a highly eutrophic, estuarine
bay, Hunts Bay, Kingston Harbour, Jamaica. Bull. Mar. Sci. 73: 307–324.
Shurland, D., A. J. Bowen and M. Hendry. 1989. Hydrodynamics. Pages 19–31 in I. Goodbody, ed.
Caribbean coastal management study: the Hellshire Coast, St. Catherine, Jamaica. University
of the West Indies, Kingston.
Wade, B. A., L. Antonio and R. Mahon. 1972. Increasing organic pollution in Kingston Harbour.
Mar. Poll. Bull. 3: 106–110.
Walters, R. A. 1992. A three-dimensional finite element model for coastal and estuarine circulation.
Circ. Shelf Res. 12: 83–102.
Webber, D. F., M. K. Webber and J. C. Roff. 1992. Effects of flood waters on the planktonic com-
munity of the Hellshire coast, southeast Jamaica. Biotropica 24: 362–374.
 BIGG AND WEBBER: TWO DIMENTIONAL NUMERICAL, HYDRODYNAMIC MODEL OF KINGSTON HARBOUR 305

___________, ___________, and D. D. Williams. 2003. The relative importance of meteorological

events, tidal activity and bathymetry to circulation and mixing in Kingston Harbour. Bull. Mar.
Sci. 73: 273–289.
___________ and P. Wilson-Kelly. 2003. Characterization of the sources of organic pollution to
Kingston Harbour. Bull. Mar. Sci. 73: 257–271.
_______ and J. C. Roff. 1996. The influence of Kingston Harbour on the phytoplankton commu-
nity of the nearshore waters of the Hellshire coast, south-east, Jamaica. Bull. Mar. Sci. 59: 245–
258.
Webber, M. K., J. C. Roff, L. A. Chisholm and C. Clarke. 1996. Zooplankton distribution and
community structure in an area of the south coast shelf of Jamaica. Bull. Mar. Sci. 59: 259–
270.
Young, E. F., G. R. Bigg and J. S. A. Green. 1993. Prediction of pollutant transport in Kingston
Harbour using analytical and numerical models. Jamaican J. Sci. Technol. 4: 16–26.

ADDRESSES: (G.R.B.) School of Environmental Sciences, University of East Anglia, Norwich, NR4 7TJ
United Kingdom. PRESENT ADDRESS: Department of Geography, University of Sheffield, Winter Street,
Sheffield S10 2TN, United Kingdom. (D.F.W.) Department of Life Sciences, University of the West
Indies, Mona, Kingston 7, Jamaica. E-mail: <[email protected]>.