Full text of DOI:10.1016/S0380-1330(02)70564-9

J. Great Lakes Res. 28(1):77–90
Internat. Assoc. Great Lakes Res., 2002
Transport of Sewage-Contaminated Sediment in
Northeastern Hamilton Harbour
John P. Coakley1,*, Michael G. Skafel1, Christopher H. Marvin1, and Tonny Bachtiar2
1National Water Research Institute
867 Lakeshore Road
Burlington, Ontario L7R 4A6
2Diponogoro University
Semarang, Indonesia
ABSTRACT. Dispersal patterns associated with sewage discharges from an STP outfall in northeast-
ern Hamilton Harbour were obtained in 1992 and 1996 using coprostanol as a tracer. Both surveys
revealed a distinctive shore-parallel pattern, characterized by a south-eastward trend and a second
transport direction toward the northwest and southwest. As physical support for the tracer patterns, verti-
cal profiles of currents near the outfall were monitored during four time periods from 1991 to 2000. Cur-
rent data indicated the flow at the outfall was complicated; there was flow in all directions at all depths.
However, the cumulative surface flow followed the prevailing wind (toward the east and northeast), while
the bottom flow was approximately in opposition to the surface flow (to the west-northwest). Combining
this circulation structure with the seasonal variation in buoyancy of the effluent discharge, the bi-direc-
tional transport pattern interpreted from the tracer distribution could be explained using a model in
which the STP effluent plume was advected by local current at a position in the water column where the
plume was stable. During January through August when the effluent was denser than ambient harbour
water, it tended to be advected by flows lower in the water column moving the plume toward the north to
west and into the central basin. During the rest of the year, it was lighter and more influenced by surface
currents, which advected the plume onshore (toward the east and northeast) where it was deflected by
shoreline geometry in a shore-parallel direction toward the southeast.
INDEX WORDS: Sediment, tracer, coprostanol, Hamilton Harbour, sewage, contaminants.
INTRODUCTION
main sources of nutrients such as ammonia (> 80%)
and phosphorus (~ 55%), STP outfalls are major
sources of persistent organic pollutants (POPs) and
heavy metal contamination (RAP 1992). Further-
more, these effluents are now considered to be the
major vectors for the entry of estrogen-mimicking
chemicals into the aquatic ecosystem (Purdom et al.
1994). When hazardous components of an outfall
are combined with proximity to nearby clean-water
amenities such as major bird rookeries (cormorants,
gulls, terns), municipal water intakes, and recre-
ational areas, their potential impacts assume impor-
tance for safe water management.
An understanding of the potential influences of
STP outfalls on aquatic systems requires the devel-
opment of methodologies for the study of their dis-
persal dynamics and mixing characteristics. These
characteristics can be studied using both physical
and chemical profiling techniques. Chemical tracers
Hamilton Harbour (Fig. 1) has been identified by
the International Joint Commission (IJC 1978) as an
Area of Concern (AOC), i.e., one of the 42 Great
Lake sites whose aquatic environment has become
so degraded that remedial action must be under-
taken. The harbor was also one of 41 sites where
contaminated sediments were cited as critical in the
assessment and remediation process. The harbor re-
ceives a large variety of contaminants from sur-
rounding industrial and municipal sources. Among
the municipal sources, four sewage treatment plants
(STPs) in the watershed discharged more than
400,000 m3/day of treated sewage effluent directly
or indirectly to Hamilton Harbour (Remedial Ac-
tion Plan (RAP) 1992). In addition to being the
*Corresponding author. E-mail: John.Coakley@cciw.ca
77

78
Coakley et al.
FIG. 1. Map of Hamilton Harbour showing the location of the Burlington Sky-
way STP outfall, outfalls of other STPs, and combined sewer overflows (CSO).
of pollution must be primarily sediment-bound and
stable such that gradients in concentration are due
to dilution and not to decay, consumption, or chem-
ical reaction. These compounds must also be related
to a unique source, in this case, treated sewage dis-
charged through the outfall. Coprostanol (5β-
cholestan-3β-ol) was selected as a sediment tracer
for this study. It is produced by gastric breakdown
of cholesterol, and is the major fecal sterol of hu-
mans, comprising 40 to 60% of the total neutral
sterols excreted. Virtually the only source of this
chemical is feces of humans and other mammals
(Venkatesan and Santiago 1989, Leeming et al.
1997). Coprostanol has been widely used as a
marker of sewage contamination (Eganhouse 1997,
Kirchmer 1971, Hatcher et al. 1977, Brown and
Wade 1984, LeBlanc et al. 1992, Grimalt et al.
1990, Holm and Windsor 1990, Coakley and Poul-
ton 1991, Coakley et al. 1992). A preliminary study
of sediment transport in Hamilton Harbour using
coprostanol as a tracer was carried out by Bachtiar
et al. (1996); results of this study are included in
the present report.
mixing characteristics of water masses. We have
previously used several current meter types for mon-
itoring the water flow in the vicinity of the STP out-
fall. Our earlier fieldwork made use of single point
meters. However, in 1999 and 2000, acoustic
Doppler current profilers (ADCP) were used in ad-
dition to a single point meter. The ADCP is a power-
ful tool in assessing the currents. These instruments
emit an acoustic signal that is reflected off small
particles moving with the water. The ADCP moni-
tors the reflected signal, which has a shift in fre-
quency depending on the direction of travel (the
Doppler shift). By timing or gating the reflected sig-
nal and measuring the Doppler shift, it is possible to
determine the average current velocities at specific
ranges from the instrument. Thus a velocity profile
can be acquired from a single instrument.
A combined physical and chemical profiling
methodology was applied to an STP discharging di-
rectly into northeastern Hamilton Harbour, the
Burlington Skyway STP (Fig. 1). This STP was
constructed in 1964 with an annual discharge vol-
ume of over 10,000 m3/day, compared to 1996 dis-
charges of over 90,000 m3/day of treated sewage
effluent (B. Seminario, Plant Manager, Burlington
Skyway STP, pers. comm.). The effluent discharges
Chemical tracer studies can be enhanced by phys-
ical measurements including current velocity pro-
files that cast light onto the dispersal dynamics and

Sewage Dispersal in Hamilton Harbour
79
through a diffuser pipe located in approximately 7
m of water and 1.5 m above the bottom. This study
demonstrates the applicability of combined mea-
surements of sediment levels of coprostanol and
measurements of current velocities using an ADCP
for defining net dispersal patterns and contaminant
transport pathways from a known STP point source.
These techniques serve as a complementary tool to
hydrodynamic modeling. The study also has water
management implications by providing a measure
of the capacity of northeastern Hamilton Harbour to
dilute sewage effluent volumes to water quality lev-
els acceptable for the clean-water uses. Moreover,
despite its restricted geographic focus, the work has
potential application in other restricted water bodies
and AOCs in the Great Lakes where long-term pat-
terns of contaminant dispersal from point sources
are required for remediation planning.
were collected on a radial grid (Fig. 2) centered on
the outfall of the Burlington Skyway STP in north-
eastern Hamilton Harbour. One set of samples was
collected in 1992 using a Shipek grab sampler,
while the other set was collected in 1996 using a
Mini Box Corer. As much as possible, samples
were collected at the same site, but in 1996 some
original sites were occupied by artificial islands
constructed the previous year (Fig. 2). A suite of six
samples was collected in the vicinity of the islands
in September 2000 to verify high values detected in
1996. Sediments making up the top 1 cm layer were
carefully collected and refrigerated aboard the sur-
vey vessel until storage in a cold room. Care was
taken to avoid loss of the soft surface layers. All
samples were freeze-dried, homogenized, and
passed through a 62 µm sieve prior to analysis.
Sterol Analyses
The analytical method for the 1992 samples was
based on gas chromatography with electron capture
detection as described in Bachtiar et al. (1996).
This method was modified to include quantitation
by gas chromatography-mass spectrometry for the
METHODS
Sample Collection
Two sets of 30 sediment samples each were ana-
lyzed for coprostanol and other sterols. Samples
FIG. 2. Location of sediment samples (see Table 2) and current meter moorings (see Table 1). Note pres-
ence of artificial islands to the east of outfall constructed shortly before the 1996 survey.

80
Coakley et al.
1996 samples. Standard sterol compounds were ob-
tained from Sigma (Mississauge, ON): epico-
prostanol (95%); cholesterol (99+%); stigmasterol
(95%); dihydrocholesterol (95%); stigmastanol
(97.2%); coprostanol (98%). The derivatizing
agent N-methyl-N-trimethylsilyl-trifluoroacetimide
(MSTFA) was obtained from Pierce (Rockford, IL).
All solvents were of analytical grade. Water was
prepared using a Milli-Q purification system.
TABLE 1. Current meter deployments in north-
eastern Hamilton Harbour. Locations are shown
in Figure 2.
Water
Depth,
m
Station/Year/
Julian Days
Meter
Type
Measurement
Depth, m
A: 1991/184-295
B: 1993/162-314
C1: 1999/181-258 Hydra
5
7.2
8.0
8
8.2
8.6
Freeze-dried sediment samples were extracted in
dichloromethane using a Dionex ASE Model 200
accelerated solvent extractor. Samples (0.5 g–1 g)
were mixed with Ottawa sand to fill the entire vol-
ume of the extraction cell. Solvent extracts were
then subjected to an open-column cleanup proce-
dure employing 1.0 g of neutral alumina (Brockman
activity 1, 80–200 mesh); sterols were eluted with
12 mL of dichloromethane. Extracts were gently
evaporated to dryness under nitrogen and deriva-
tized with 100 µl MSTFA for 20 minutes at 130°C.
After cooling, extracts were reconstituted in 1 mL
of iso-octane for GC-MS analysis.
Analyses were performed using a Hewlett-
Packard 5890 gas chromatograph with a 30 m 0.25
mm i.d. 0.25 µm stationary phase DB-5 column and
a Model 5971A mass selective detector operated in
selected ion monitoring (SIM) mode. The injection
volume was 1 µL on-column with a carrier gas ve-
locity of 40 cm/sec. The transfer line temperature
was maintained at 300°C. The following ions were
monitored: m/z 388, 355, 231, 386, 316, 215, 370,
233, and 264. The following temperature program
was used: initial temperature 100°C; 100°C to
170°C at 20°C/min; 170°C to 300°C at 3°C.
Surrogate standards for the sterol analytes were
not available; sediment samples were spiked with
a laboratory standard containing four polycyclic
aromatic hydrocarbons (PAH) including naphtha-
lene-d8, fluorene-d10, pyrene-d10, and benzo[a]-
pyrene-d12. Spike recoveries were typically greater
than 75%. Method blanks were routinely carried
through the sample preparation and analysis proce-
dures. Between-run reproducibilities were typically
within 10%. Method detection limits were esti-
mated at approximately 60 ng/g (3:1 S/N) for indi-
vidual sterols in a 1 g dry weight sample in full
scan mode and 6 ng/g in SIM mode.
C2: 1999/181-258 ADCP 1.25 ≤ Depth ≤ 6.25 8.0
(1.0 m bins)
D: 2000/150-276 ADCP 1.25≤ Depth ≤ 6.25
8.0
(1.0 m bins)
where the vertical structure of the flow was mea-
sured in addition to the horizontal distribution. The
current data were examined in an effort to under-
stand the long-term circulation patterns in this por-
tion of the harbor. In Table 1, Stations A and B
were part of earlier studies of the overall harbor cir-
culation. Stations C1 and C2 represent one deploy-
ment in 1999; a Sontek Hydra meter and a RD
Instruments Acoustic Doppler Current Profiler
(ADCP) were placed out as a pair. Station D was
deployed in 2000. The Hydra meter is an acoustic
Doppler current meter that measures velocity in a
very small volume at a fixed location about 18 cm
distant from the meter head. It was deployed on the
bottom such that the measuring volume was 0.6 m
above the bottom, at a depth of 8.0 m. The ADCP
broadcasts a more powerful beam into the water (it
was also bottom mounted and “upward looking”).
The ADCP was set up so that the average horizon-
tal current was determined in one-meter sections, or
bins, of the water column. Because of the character-
istics of the ADCP and its deployment configura-
tion, the lowest bin was centered at a depth of 6.25
m, and the highest bin at a depth of 1.25 m. The
data from the Hydra were considered to represent
the bottom current, and the data from the top bin of
the ADCP were considered to represent the surface
current.
RESULTS AND DISCUSSION
Validation of Coprostanol as a
Sewage Tracer in Hamilton Harbour
Water Circulation and Nearshore Currents
The chemical stability of coprostanol under oxi-
dizing conditions in the water column or on the sed-
iment surface in the Great Lakes was questioned by
Dutka et al. (1974). However, studies by Hatcher
Records of local currents collected near the out-
fall (Fig. 2) were assembled for four periods from
1991 to 2000 (Table 1). Included were two periods

Sewage Dispersal in Hamilton Harbour
and McGillivary (1979) in the New York Bight
81
demonstrated that coprostanol concentrations in
cores were unchanged over 25 years. Elevated co-
prostanol concentrations were also found in sedi-
ment cores dated at more than 40 years taken near
the Burlington Skyway STP outfall (Coakley et al.
1996). Nishimura and Koyama (1977) observed no
decomposition of cholesterol and cholestanol under
anaerobic conditions during 450-day lake sediment
incubation experiments, while only minor decom-
position of cholesterol occurred under aerobic con-
ditions. Based on these studies, coprostanol appears
to be sufficiently stable to be adequate for tracer
applications in surficial sediments in Hamilton
Harbour.
Despite its wide use, some researchers urge cau-
tion in the use of coprostanol as a tracer, especially
in certain tropical areas where concentrations in
sediments are very low, for example, below 5 ng/g
(Sherblom et al. 1997), because of the possibility of
contamination by feces of marine mammals and by
the breakdown of algal sterols (Grimalt et al. 1990).
Studies by Venkatesan and Santiago (1989) in ma-
rine waters offshore of the southern California
coast, and by Leeming et al. (1997) in both fresh
waters and marine areas of Australia indicate that
confirmation of a human origin is obtained by the
ratio of some of the sterol epimers (e.g., co-
prostanol to epicoprostanol) in the sediments. Gri-
malt et al. (1990) used the ratio of coprostanol to
cholestanol to distinguish coprostanol derived from
diagenesis of algal sterols from that of human ori-
gin. In order to assess the possibility of coprostanol
being added by the bird rookeries in the study area,
samples of bird guano from the island northeast of
the STP outfall were analyzed for sterols. In addi-
tion to very low values of coprostanol (27 ng/g), the
results (Coakley, J.P., unpublished data) show
epimer ratios far outside the range for human ori-
gin. Therefore, the coprostanol measured within the
study area sediments is thought to have a primary
source (the STP), and is not significantly influenced
by proximity to local bird colonies.
FIG. 3. Annual trends in outfall effluent density
(dashed line) relative to ambient Hamilton Har-
bour water density (solid line).
(Fig. 3). From September to December the effluent
was significantly less dense than the receiving wa-
ters, while during the rest of the year it was more
dense (due to the elevated conductivity). These data
indicated that during September to December the
effluent would have had a tendency to rise to the
surface, while at other times it would have sunk to
the bottom. This process will be discussed later in
concert with the vertical current profile data in con-
structing a transport scenario compatible with the
tracer patterns.
Tracer Results and Transport Interpretation
The results of the analyses of sediment co-
prostanol concentration for the two sampling sur-
veys are presented in Table 2. Values ranged from
highs of 930 µg/g (1992) and 1,600 µg/g (1996), to
ambient background levels of less than 3 µg/g. The
maximum values corresponded to sites within
100 m of the outfall and were consistently lower
than that of raw human feces (5,600 µg/g, Leeming
et al. 1997). Coprostanol levels for both sample sets
were found to be generally similar. Sample loca-
tions exhibiting high coprostanol levels were the
same in both the 1992 and 1996 surveys. The aver-
age value of all samples in 1992 was 43 µg/g, com-
pared to 102 µg/g for 1996 that was primarily the
result of a value of 1,600 µg/g at one station. This
site had a value of 37 µg/g in 1992, before the sur-
rounding area was modified in 1995 and 1996 by
construction of artificial islands to the east and
Relative Density of STP Effluent
Effluent density (based on temperature and con-
ductivity) as a function of month of the year was
available from a study for a proposed new outfall
placement of the Burlington STP. Using the water
temperature and conductivity data for the upper 5 m
of Hamilton Harbour (Charlton et al. 1998), the
density of the receiving water was also estimated

82
Coakley et al.
TABLE 2. Sterol determinations in sediments from northeastern Hamilton Harbour.
UTM Coordinates
(NAD27)
1992
Coprostanol
µg/g
1996
1996
Sample Identification
Coprostanol Epicoprostanol C/E
1992
1996
North
East
µg/g
µg/g
ratio
HH 2-1
HH 2-2
HH 2-3
HH 4-1
HH 4-2
HH 6-1
HH 6-2
HH 6-3
HH 8-1
HH 8-1A
HH 8-2
HH 8-3
8-4
HH 10-1
HH 10-2
HH 10-3
HH 10-4
HH 10-5
HH 12-1
HH 12-2
HH 12-3
HH 12-4
HH 12-5
HH 13-1
HH 13-2
HH 13-3
HH 14-1
HH 14-2
HH 14-3
HH 14-4
HH STP
2-1
2-2
2-3
4-1
4795492
4795679
4795952
4795614
4795774
4795386
4795292
4795287
4795242
4794963
4794856
4794417
596640
596554
596433
596773
596813
596774
596936
597000
596782
596916
596972
597052
7.2
1.6
0.2
2.9
0.1
37
0.1
< 0.1
150
14
0.5
9.3
6.3
13
0.3
1.6
2.8
4.8
47
3.1
0.9
3.2
2.7
3.0
15
4.4
30
1.5
1.9
0.2
930
43
3.3
18
0.4
1.7
2.1
6.3
ND
0.7
1.6
2.8
2.4
6-1
6-2
1,600
1,600
0.4
1.0
6.5
2.6
8-1
8-1A
8-2
190
57
51
5.3
2.2
85
0.1
7.1
6.8
18
63
3.7
10
16.3
13
6.4
12
110
53
38
1.6
1.1
1.4
2.1
8-3
2.5
2.9
58
ND
2.0
2.7
4.0
72
1.0
3.0
4.6
2.5
3.1
5.9
0.7
15
10-1
10-2
10-3
10-4
10-5
12-1
12-2
12-3
12-4
12-5
13-1
13-2
13-3
14-1
14-2
14-3
14-4
STP
4795295
4795117
4794756
4794055
4793106
4795355
4795257
4795070
4794762
4794326
4795409
4795381
4795366
4795458
4795559
4795644
4795753
4795403
596670
596588
596422
596093
595699
596603
596432
596075
595326
594426
596365
595999
595491
596577
596397
596210
595943
596717
Avg. Copr.
1.5
3.5
2.5
4.5
0.9
3.7
3.4
3.5
5.3
2.1
2.1
5.0
2.7
6.8
1.7
3.5
41
2.7
8.8
ND
740
102
49 *
0.4
5.1
0
600
1.7
*(Avg. without outlier)
southeast of the outfall (Fig. 2). Analysis carried
out on additional samples from around this site in
September 2000 showed background levels. This
result suggested that the high 1996 value repre-
sented a transient or very localized hot-spot, whose
significance cannot be extrapolated to the surround-
ing areas. If this anomalous value is removed, the
1996 average is reduced to 49 µg/g, which is only
slightly higher than the 1992 average. There were
also significant increases in concentrations at sites
farthest from the outfall between the 2 study years.
This could have been a reflection of greater impact
of the effluent associated with the steady increase in
the STP discharges between 1992 (79,000 m3/day)
and 1996 (91,000 m3/day). In addition, ratios of co-
prostanol to epicoprostanol were, with the excep-
tion of one value of 0.9, always greater than 1. This
confirmed the earlier assessment that, despite the
large gull and cormorant populations on nearby is-
lands, the coprostanol was primarily of human ori-
gin and predominately associated with the effluent
of the Burlington Skyway STP.
The tracer concentrations were plotted onto the
radial grid and contoured by hand at intervals se-
lected to enhance the definition of the concentration
plumes. The singular hot-spot was not included in
the 1996 plot. The tracer distribution patterns for
both the 1992 and the 1996 sampling surveys are
presented in Figure 4. The contoured tracer concen-
tration pattern is characterized in both surveys by

Sewage Dispersal in Hamilton Harbour
83
FIG. 4. Spatial concentration patterns for the sediment tracer, coprostanol,
in the northeastern sector of Hamilton Harbour. Top panel shows concentra-
tion patterns for the survey done in 1992; bottom shows similar data for 1996.
Heavy solid lines represent the interpreted southeast dispersal trend; the
dashed lines represent the southwest trend.
two main lobes, a southward trend (containing the
higher concentration values) and a second trend
curving from north to west to southwest. Although
the radial grid samples were deficient in covering
large areas to the west and southwest, a region of
very low values immediately to the west of the out-
fall was confirmed in both surveys. The persistence
of this low-concentration region relatively close to
the outfall is surprising, and is an indication of min-
imal transport (and deposition) in that direction.
The possibility of scour by waves was discounted
because the bottom sediment was similar to that at

84
Coakley et al.
other sites (mud) and the depths were as deep or
deeper. The area was probably one consistently out-
side the transport plume for STP effluent. The su-
perimposed arrows included in Figure 4 show the
principal sediment transport directions interpreted
from the tracer results.
Current Regime
Wu et al. 1996 documented that several driving
processes could be identified from current data but
that wind was dominant. They found that drogues at
2 m followed the wind, and at 5 m, some drogues
followed the wind and some moved against the
wind. The use of the ADCP in the present study al-
lowed a detailed examination of the vertical struc-
ture in the vicinity of the STP outfall.
The characteristics of the flow were examined
using progressive vector diagrams which give
“pseudo” displacements (Emery and Thomson
2001) and current roses calculated from the velocity
time series. The currents in 1999 and 2000 showed
considerable vertical structure, particularly in terms
of direction. In Figure 5, progressive vector dia-
grams are shown for depths of 1.25, 3.25, 6.25, and
8.0 m for the July to September 1999 deployment.
The net surface displacement, as indicated by the
trajectory at 1.25 m depth, was to the east-north-
east. At increasing depth the net displacement di-
rections were rotated counterclockwise down to the
bottom, where the direction was to the west-north-
west. All the trajectories were very irregular, indi-
cating that the flow was constantly changing
direction. There was displacement in all directions,
and hence dispersion in all directions, but greatest
in the dominant directions mentioned. These pro-
gressive vectors did not provide as useful an esti-
mate of water particle movement as Lagrangian
tracers. Nevertheless, when used in conjunction
with a map of the local shoreline, they proved to be
useful indicators. The easterly surface displace-
ments were intercepted by the shoreline, likely de-
flecting the flow parallel to the shore to the
southeast. Similarly the north and northwesterly
bottom displacements will eventually by deflected
westerly and southwesterly along the north shore-
line. The displacement at 3.25 m depth may well be
bifurcated, some flowing southwest along the north
shore and some southerly along the east shore. The
displacement at 6.25 m depth would eventually be
forced to the southwest.
FIG. 5. Progressive vector diagrams calculated
from the velocity time series at four depths for the
July to September 1999 deployment.
The Hydra meter was not deployed in 2000, so there
were no data collected at 8.0 m. These plots confirm
that the current at the site was highly dispersed at all
depths. The flows near the surface (1.25 m depth)
were very similar in direction, distribution, and
magnitude for both years. The largest number of oc-
currences was toward the east and significant flow
occurred in all other directions with the least toward
the south. The flows near the middle of the water
column (3.25 m) were also similar in the 2 years,
with the largest number of occurrences to the north-
west and the least to the south. The flows near the
bottom (6.25 m depth) were rather different in the 2
years, being generally higher in magnitude in 2000
compared to 1999, and more directed toward the
west, in contrast to the northerly flows in 1999.
Closest to the bottom (8.0 m depth) in 1999, most of
the occurrences were to the west and southwest, and
the least toward the southeast.
The direction patterns of the roses for the 2 years
suggested that the flow structures were similar. The
most important direction at the surface was to the
east, and at the bottom to the west, and northerly at
intermediate depths. The shapes of the current roses
indicated a current regime similar to that described
based on the progressive vector diagram. However,
the roses did not illustrate the persistence in direc-
The current roses for depths of 1.25, 3.25, 6.25,
and 8.0 m for 1999 and 2000 are shown in Figure 6.

Sewage Dispersal in Hamilton Harbour
85
FIG. 6a. Current roses at the four depths for the July to September 1999 deployment. The circles
represent calms (less than 0.005 m/s); the other velocity ranges are shown in the lower right panel of
Figure 6b.
tion as well as did the progressive vector diagrams.
The currents measured in 1991 and 1993 at interme-
diate depths were compared with the ADCP data at
the same depths. The earlier data and the ADCP
data were consistent in terms of variability and
dominant flow direction, adding confidence to the
interpretation of the 1999 and 2000 data being rep-
resentative of long term patterns.

86
Coakley et al.
FIG. 6b. Current roses at the three depths for the May to October 2000 deployment. The circles
represent calms (less than 0.005 m/s); the other velocity ranges are shown in the lower right panel.
Analysis of the current records, especially those
that the cumulative surface flow was influenced by
the prevailing wind (toward the east and northeast).
However, such onshore flows would soon be de-
flected by the nearshore bottom and shoreline
topography along the shore. Because of the south-
collected in 1999 and 2000 containing information
on vertical structure, showed that there was some
flow in all directions at all depths. However, pro-
gressive vector plots of the various depths showed

Sewage Dispersal in Hamilton Harbour
87
easterly orientation of the downwind shoreline, it
Temporal Trends in
Effluent Transport Patterns
was expected that the flow would diverge and move
parallel to the shore either northwesterly or south-
easterly, depending on the initial direction. Judging
from visual observation of the area after a period of
onshore winds and heavy rainfall, it was expected
that the southeasterly direction would predominate.
It is commonly observed that high-turbidity plumes
from streams entering the northeastern corner of the
harbor tend to follow the eastern shoreline south-
ward before reaching the Burlington Ship Canal.
The progressive vector plot for the bottom cur-
rents showed an overall pattern that was approxi-
mately in opposition to the surface flow (i.e., to the
west-northwest). This was the expected pattern of
return from onshore-wind-driven flow. At interme-
diate depths, the flows were generally northward
with an increasingly westerly component with
depth. It would appear that the least movement
would be expected toward the south.
Mixing of the STP effluent with the harbor water
is influenced by factors such as the characteristics
of the diffuser’s discharge jets, the nature of the
local hydrodynamic climate, and the relative den-
sity of the effluent with respect to the receiving wa-
ters. Comparison of the tracer patterns for the two
surveys separated by 4 years showed little overall
change. Although the data do not allow a rigorous
statistical analysis of trends over this time, there ap-
pears to be an increase in the areas far removed
from the outfall to the southwest where above-
background values (~ 3 µg/g) extended past the lim-
its of the grid. Discounting the anomalous value in
the 1996 survey, the average coprostanol concentra-
tion increased very little over the 4-year interval,
i.e., 49 µg/g in 1996, versus 43 µg/g in 1992
(Student’s t = –0.14). The 1996 mean is slightly
higher, but sample populations for both years
are statistically identical. This suggested that, de-
spite an increase in STP discharges from less than
80,000 m3/day in 1992 to more than 90,000 m3/day
in 1996, the area of maximum impact remained vir-
tually unchanged, except for those areas mentioned
above. The impact of this volumetric increase could
have been offset by an overall decline in the sus-
pended solids content of the treated effluent from
around 10 mg/L in the early 1980s to less than 6
mg/L in 1996 due to plant improvements. Further-
more, the number of major STP bypass events re-
sponsible for much of the high-coprostanol
discharges fell from 119 in 1992 to 63 in 1996. The
high coprostanol value near the islands, though sta-
tistically discounted, raised the possibility of occa-
sional entrapment of contaminated sediment in the
area behind the artificial islands where a warm-
water fish habitat is planned. The rapid concentra-
tion reduction in the coprostanol levels near the
outfall, and the maintenance thereafter of well-de-
fined plumes, suggested that much of the effluent
dilution took place in the vicinity of the outfall. The
net effect of the diffuser jets was to mix the effluent
with surrounding bay water and to propel it toward
the surface. As a result, the effluent thereafter
tended to follow the surface flows for at least one-
third of the year, i.e., the roughly 4-month period
when the plume was buoyant as well as other times
when the density difference was slight and could
have been overcome by diffuser momentum. The
trajectory plots of the surface flows (Fig. 5) showed
that most of these flows were toward the east and
Plume Density
In reconciling the current patterns with the con-
crete evidence of the long-term tracer patterns, it
appeared that there was a correspondence with the
bi-directional dispersal pattern inferred. However,
the strong systematic differences in the vertical
structure of the currents at the outfall suggested that
the location of the outfall plume in the water col-
umn would play an important role in its dispersal
direction. Figure 3 shows that the relative density of
the effluent plume varied considerably with time
during the year. During the period January to Au-
gust (roughly 8 months), the effluent was denser
than the receiving waters and would have a ten-
dency to sink and follow the generally northward
and westward direction of the bottom and interme-
diate flows. Such a model would explain the tracer
plume trend toward the north, then west, then
southwest as shown in Figure 4. On the other hand,
during the September to December period (4
months), the STP effluent was less dense than the
receiving waters and would thus be expected to rise
to the surface layer with the additional upward push
of the outlet diffuser ports. As indicated by the sur-
face progressive vector plots, the effluent plume
would have been strongly advected onshore (east-
ward), then deflected generally southeastward along
the shore before deposition. It can be expected that
this eastward advection had been muted somewhat
by the advection westward in the lowest part of the
water column, but that the net effect is eastward.

88
Coakley et al.
northeast, and were later deflected southeastward
toward the Burlington Ship Canal. The fate of the
plume in the canal area was difficult to predict. Bar-
ica et al. (1988) found that although flows in both
directions took place in the canal, the net outflow
from the harbor to Lake Ontario was 646,000
m3/day, or 0.5% of the total harbor volume. How-
ever, due to the complexity of the exchange flows
within the canal, it was difficult to predict the efflu-
ent volume entering the lake versus that diverted
back toward the central basin of the harbor. The
tracer plume that curved north through south-west
showed the most rapid dilution, although above-
background values in areas far removed from the
outfall could have been influenced by other effluent
inputs, such as CSOs, the other 3 STPs (Fig. 1), or
by dumping from ships in the harbor. The final sink
of these contaminated sediments was apparently the
central basin of the harbor.
covered only the summer months (July to Septem-
ber) and omitted fall and winter periods, they nev-
ertheless indicated a complex vertical structure in
the water column. Surface currents were predomi-
nantly in the downwind (toward the east and north-
east) direction, while bottom flow was
approximately in opposition to the surface flow (to-
ward the west-northwest). In addition to the current
regime, the relative density of the effluent, which
was seen to vary seasonally, played a significant
role in the vertical location within the water column
with which it tended to be advected. Thus, during
the spring and summer months when the effluent
was denser than the ambient harbor water, it tended
to be advected by flows lower in the water column,
thus moving the plume toward the north-to-west
sector. Such a flow could explain the tracer plume
along the north shore. During the rest of the year,
the effluent was lighter and more buoyant and
would have been more influenced by the surface
current, which would have advected the plume
more strongly onshore, where it would have been
deflected alongshore toward the south. This model
explained the bi-directional nature of the tracer pat-
tern in the sediment. It also argues strongly for the
incorporation of a 3-D perspective in hydrodynamic
models used even in shallow-water nearshore areas.
From a Hamilton Harbour management and re-
mediation perspective, the spatial impact of the
STP outfall showed only a slight increase over the
period studied, primarily due to plant improve-
ments. Furthermore, the tracer patterns indicated
the effluent was reduced fairly rapidly to near back-
ground levels (< 3 µg/g) within 2 kilometres of the
outfall, although between the 1992 and 1996 sur-
veys, the mixing zone appeared to have increased
slightly toward the southwest. The large area of
non-deposition of coprostanol offshore from the
outfall increased in size toward the north. The re-
sults also suggested the potential for entrapment
and accumulation of STP-source contaminants and
nutrients behind the small artificial islands con-
structed in 1995 to the east and southeast of the out-
fall. Nevertheless, the dominant trend was for the
STP effluent and its contaminant/nutrient burden to
be transported as a plume southeastward to the area
of the Burlington Ship Canal, where it was ad-
vected inward or outward according to undefined,
but complex, processes. The other trend toward the
north and west eventually was reduced to back-
ground levels near the central basin of the harbor. A
particularly noteworthy result of the study is the
conclusion that the STP plume did not move di-
Comparison with Hydrodynamic Models
Tracer results were compared with circulation
model simulations for the northeastern area of
Hamilton Harbour reported by Tsanis and Wu
(1995). Using a 2-D (depth-averaged) nested-grid
model, they ran simulations for several sections of
the harbor. In the northeastern area, they ran the
model for winds of 10 m/s from the north and west
with varying topography including the presence of
islands, no islands, and different bottom roughness.
All the simulations indicated transport toward the
south for both wind directions. There was no simu-
lation that reflected the tracer’s transport toward the
north. This discrepancy revealed a potential short-
coming in the hydrodynamic model used, most
likely the result of using a 2-D (depth averaged) ap-
proach and thus ignoring the important vertical
changes in flow characteristics revealed in the cur-
rent data described here. The above circulation
model also was not set up to take into account sea-
sonal changes in buoyancy of inputs.
SUMMARY AND CONCLUSIONS
The distribution patterns identified by the co-
prostanol tracer around the Burlington Skyway STP
outfall reflected net transport and dispersal trends
of STP-related contaminants in the northeastern
sector of Hamilton Harbour. They could also be a
potential proxy for the transport of other particle-re-
active pollutants, such as POPs and trace metals.
Although the current measurements near the outfall

Sewage Dispersal in Hamilton Harbour
89
Dutka, B.J., Chau, A.S.Y., and Coburn, J. 1974. Rela-
tionship between bacterial indicators of water pollu-
tion and faecal sterols. Wat. Res. 8:1047–1055.
Eganhouse, R.P., 1997. Molecular markers and environ-
mental organic geochemistry: an overview. In Molec-
ular Markers in Environmental Geochemistry, R.P.
Eganhouse (Ed.), pp. 1–20. Am. Chem. Soc. Sympo-
sium Series 671, Washington, D.C.
rectly offshore where minimum values were consis-
tently found, but inshore and alongshore. This
raises questions regarding the potential impact of
the plume distribution on other Hamilton Harbour
issues such as aquatic recreation activities,
nearshore macrophyte growth, and uncontaminated
habitats for fish and birdlife in these nearshore
zones.
Emery, W.J., and Thomson, R.E. 2001. Data analysis
methods in physical oceanography. Amsterdam: Else-
vier Science B.V.
ACKNOWLEDGMENTS
Grimalt, J.O., Fernandez, P., Bayona, J.M., and
Albaiges, J. 1990. Assessment of faecal sterols and
ketones as indicators of urban sewage inputs to
coastal waters. Environ. Sci. Technol. 24:357–363.
Hatcher, P.G., and McGillivary, P.A. 1979. Sewage con-
tamination in the New York Bight: Coprostanol as an
indicator. Environ. Sci. Technol. 13:1225–1229.
The authors thank Lina Thiessen, Doug Doede,
and Fausto Chiocchio, and the NWRI Technical
Operations Services for technical and field support.
Data on operation and history of the Burlington
Skyway STP were provided by Benny Seminario
(Plant Supervisor), Jeffrey Large (previous Supervi-
sor), and Len Yust (Plant Manager, Region of Hal-
ton Public Works Department). The authors thank
B.G. Krishnappan for data on effluent temperature
profiles, and Murray N. Charlton for information
related to impacts of STP effluent on the harbour
ecology. We also thank the two reviewers, D. Boyd
and B.M. Lesht, for their thorough and constructive
review of this paper.
, Keister, L.E., and McGillivary, P.A. 1977.
———
Steroids as sewage specific indicators in New York
Bight sediments. Bull. Envir. Contam. Toxicol. 17:
491–498.
Holm, S.E., and Windsor, J.G. 1990. Exposure assess-
ment of sewage treatment plant effluent by a selected
chemical marker method. Arch. Envir. Contam. Toxi-
col. 19:674–679.
International Joint Commission. 1978. Great Lakes
Water Quality Agreement of 1978, as amended
November 18, 1987. IJC, Windsor, Ontario.
REFERENCES
Kirchmer, C.J. 1971. 5β-cholestan-3β-ol: an indicator of
faecal pollution. Ph.D. thesis, University of Florida,
Gainesville, FL.
LeBlanc, L.A., Latimer, J.S., Ellis, J.T., and Quinn , J.G.
1992. The geochemistry of coprostanol in waters and
surface sediments from Narragansett Bay. Estuarine,
Coastal, and Shelf Science 34:439–458.
Leeming, R., Latham, V., Rayner, M., and Nichols, P.
1997. Detecting and distinguishing sources of sewage
pollution in Australian inland and coastal waters and
sediments. In Molecular Markers in Environmental
Geochemistry, R.P. Eganhouse (Ed.), pp. 306–319.
Am. Chem. Soc. Symposium Series 671, Washington,
D.C.
Nishimura, M., and Koyama, T. 1977. The occurrence of
stanols in living organisms and the behaviour of
sterols in contemporary sediments. Geochim. Cos-
mochim. Acta 41:379–385.
Purdom, C.E., Hardiman, P.A., Bye, V.J., Eno, N.C.,
Tyler, C.R., and Sumpter, J.P. 1994. Estrogenic
effects of effluents from sewage treatment works.
Chemistry and Ecology 8:275–285.
Remedial Action Plan (RAP) 1992. Remedial Action
Plan for Hamilton Harbour, Stage 2A. Prepared for
Ontario Ministry of the Environment and Environ-
ment Canada. Toronto, ON, Canada.
Bachtiar, T., Coakley, J.P., and Risk, M.J. 1996. Tracing
sewage-contaminated sediments in Hamilton Harbour
using selected geochemical indicators. Sci. Total Env-
iron. Special Volume on Transport and Accumulation
Processes of Contaminants 179:3–16.
Barica J., Poulton, D.J., Kohli, B., and Charlton, M.N.
1988. Water exchange between Lake Ontario and
Hamilton Harbour: water quality implications. Water
Poll. Res. J. Canada 23:213–226.
Brown, R.C., and Wade, T.L. 1984. Sedimentary
coprostanol and hydrocarbon distribution adjacent to
a sewage outfall. Wat. Res. 18: 621–632.
Charlton, M., Chiocchio, F., and Ventresca, A. 1998.
Comparison of physical data from a centre station
with data from survey stations in Hamilton Harbour
for the period 1987–92. NWRI Unpublished Report.
Coakley, J.P., and Poulton, D.J. 1991. Tracers for fine
sediment transport in Humber Bay, Lake Ontario. J.
Great Lakes Res. 17:289–303.
, Carey, J.H., and Eadie, B.J. 1992. Specific
———
organic components as tracers of contaminated fine
sediment dispersal in Lake Ontario near Toronto.
Hydrobiologia 235/236:85–96.
, Smith, D.S., Bourbonniere, R.A., and Fox, M.E.
———
1996. Stability of coprostanol and its isomers in
freshwater sedimentary environments, Lake Ontario,
Canada. NWRI Contribution 96–76, Environment
Canada.
Sherblom, P.M., Henry, M.S., and Kelly, D. 1997. Ques-
tions remain in the use of Coprostanol and Epico-

90
Coakley et al.
prostanol as domestic waste markers: examples from
ocean sediments: novel tracers to examine habitats of
cetaceans, pinnipeds, penguins, and humans. Marine
Biology 102: 431–437.
coastal Florida. In Molecular Markers in Environmen-
tal Geochemistry, R.P. Eganhouse (Ed.), pp. 320–331.
Am. Chem. Soc. Symposium Series 671, Washington,
D.C.
Wu, J., Tsanis, I.K., and Chiocchio, F. 1996. Observed
currents and water levels in Hamilton Harbour. J.
Great Lakes Res. 22(2): 224–240.
Tsanis, I.K., and Wu, J. 1995. A nested grid hydrody-
namic / pollutant transport model for nearshore areas
of Hamilton Harbour. Water Qual. Res. Jour. Canada
30:205–229.
Submitted: 14 December 2000
Accepted: 4 December 2001
Venkatesan, M.I., and Santiago, C.A. 1989. Sterols in
Editorial handling: Paul V. Doskey