Groundwater Contamination (Proceedings of the Symposium held during the Third IAHS
Scientific Assembly, Baltimore, MD, May 1989), IAHS Publ. no. 185, 1989
Aquifer contamination and restoration at the Gloucester
Landfill, Ontario, Canada
R. E. Jackson, A. S. Crowe, S. Lesage and M. W. Priddle
National Water Research Institute, Canada Centre for Inland Water, Burlington, Ontario, Canada
L7R 4A6
ABSTRACT
A glacial outwash aquifer near Ottawa, Canada, has
become polluted with toxic organic chemicals following the disposal
of laboratory solvents in shallow trenches of the Gloucester
Landfill immediately above the aquifer. Several remedial options
considered by the environmental authorities for decontaminating the
aquifer are discussed. Impermeable barrier walls were rejected as
unsuitable, given the permeable nature of the underlying bedrock.
Groundwater quality monitoring data for the period 1980/88 suggests
that the existence of pools of liquid organic chemicals (DNAPLs)
within the aquifer is improbable. Therefore, much of the plume can
be removed hydraulically over a period of four years by the
operation of four purge wells pumping to an on-site treatment plant
from which the purified water is returned to the aquifer by
recharge wells. A final phase of decontamination might involve
in-situ biorestoration techniques presently under development.
CONTAMINATION ET RESTAURATION DE L'QUIFERE AU SITE D'ENFOUISSEMENT
DE GLOUCESTER, ONTARIO, CANADA
i
r
RESUME
Un aquifère d'alluvion glaciaire près d'Ottawa, Canada, a
été contaminé par des produits organiques toxiques suite au rejet
de solvants résidus de laboratoires dans des tranchées peu
profondes surmontant 1'aquifère. Les autorités environnementales
ont considéré plusieurs alternatives pour la dépollution du site.
Elles ont rejeté la possibilité d'utiliser des barrières
imperméables à cause de la perméabilité du soubassement rocheux.
Les résultats des analyses de la qualité des eaux souterraines pour
la période de 1980 à 1988 nous laissent supposer qu'il est
improbable qu'il y ait des nappes de solvants denses (DNAPLs) dans
1'aquifère. La nappe d'eau contaminée pourra donc être retirée
hydrauliquement sur une période de quatre ans en pompant à partir
de quatre puits vers une usine de traitement sur place, retournant
l'eau purifiée à l'aquifère par des puits de recharge. En phase
finale, pour l'élimination des derniers résidus, il sera possible
de considérer l'utilisation de techniques de bio-restauration
in-situ, lesquelles sont en cours de développement.
INTRODUCTION
The Gloucester Landfill is located on the property of Ottawa
International Airport (75°38"W, 45°18"N), approximately 10 km
south of downtown Ottawa, Ontario, Canada. Between 1969 and 1980,
various government agencies disposed of liquid wastes in a "Special
Waste Compound" (SWC) at one corner of the Gloucester municipal
landfill (Fig. 1 ) . The wastes were predominantly organic solvents
in glass containers from various laboratories in Ottawa. The
181
182
R. E. Jackson et al.
bottles were placed in trenches and combusted by detonation of
explosives set within the wastes.
By 1982, it was apparent that this method of waste disposal had
caused the pollution of the outwash aquifer, shown as Unit C in
Fig. 2, underlying the SWC and threatening local wells. This
aquifer is a thick (20-25 m) sequence of interstratified silts,
sands and gravel. It is confined above by marine silts which form
a discontinuous aquitard and which in turn is overlain by an
unconfined sand aquifer. The aquifer is underlain by fractured
Paleozoic limestone bedrock (Jackson et al.. 1985).
Fig. 1 The Gloucester Landfill and Special Waste Compound near
Ottawa, Ontario, Canada.
LEGEND
\
GLOUCESTER LANDFILL
i
DELZOTTO
/
\ AVENUE
^
STRATK3HAPHC
UNÎT
<«"~\c'
87
JUL„
33
— L I N E OF SECTION C-C'
• PIEZOMETER O f l WELL
•
Fig. 2
E R e g r e s t k r a l a a n a i and
Hydrostratigraphy beneath the Gloucester Landfill.
Aquifer contamination and restoration
183
The outwash aquifer has an average hydraulic conductivity of 5
x 10
m/s, a mean transmissivity of 10
m /s and a storativity on
the order of 2 x 10 . The average groundwater velocity in the
aquifer is approximately 7 cm/day as measured by borehole dilution
techniques (Jackson et al., 1985). The organic carbon content of
the outwash aquifer materials has a mean of 0.06% (Jackson et al.,
1985) . The mineralogy of the outwash aquifer is principally quartz
and feldspar with minor amounts (i.e. < 10% each) of mica, calcite,
dolomite, hornblende and garnet.
CONTAMINANT MIGRATION AND FATE
The groundwater quality of the flow system beneath the Gloucester
Landfill has been monitored since 1976. By 1982, it was apparent
that widespread contamination of both the unconfined and the
outwash aquifers had occurred (see Fig. 1 ) . Monitoring results
from May 1988 indicate that the leading edge of the contamination,
as shown by the chloride plume in Fig. 3, is beyond the property
limits of the Airport. According to Provincial law, this fact
requires that decontamination or containment measures be undertaken
to prevent further off site seepage.
Fig. 3 Plan view of chloride plume in confined, outwash aquifer,
May, 1988
Patterson et al. (1985) showed that the organic contaminants
had been transported distances from the SWC that were inversely
proportional to their respective octanol-water partition
coefficients (K ) . It was concluded that each contaminant had
ow
been differently retarded due to sorption by organic materials
within the aquifer sediments. This process, known as hydrophobic
partitioning, causes the chromatographic dispersion of the various
contaminants and yields retardation factors (see Jackson et al.,
1985, Table 17) that are dependent on the compound's K
value.
184
R. E. Jackson et al.
In the most heavily contaminated zones of the aquifer, such as
the zone of greater than 100 /sg total organo-halogen compounds
(TOH)/L shown in Fig. 4, the pH varies between 7 and 8, the E.,
H
is less than 200 mV and dissolved oxygen is not detectable {< 0.3
mg 0 ? /L).
rat
c
SPECIAL WASTE COMPOUND
rt
w
E
o «o
VERTICAL EXAGGERATION 2.4X
,
1 C%M£ê
07
,50*»..
LEGEND
\ DELZOTTO"/
\AVENUE
^
GLOUCESTER LANDFILL
~Tc*
33
— LINE OF SECTION C-C'
• PIEZOMETER 0 3 WELL
SAMPLE LOCATION
STRATIQRAPHIC UNIT
A
B
C
D
E
LIMESTONE
TILL
OUTWASH
SILT
BANDS
-25-
CONCENTRATION CONTOUR
BUI
>260mflCI7L (1988)
%3M
>100ngTOH/L(1987)
Fig. 4 Cross-sectional view of chloride and TOH plumes in the
outwash aquifer, May 1988.
Recent sampling (1987/88) of the contaminated groundwater shows
that degradation products from the transformation of chlorinated
aliphatic hydrocarbons have appeared in the groundwater. These
products include vinyl chloride and 1,1-dichloroethene (see Table
1), which are produced by abiotic and biotic dechlorination
reactions as shown in Fig. 5.
CCI 2 CI 2
I POE h
CHjCCi-,
CH 3 CHCI 2
I 1,1-DCE f]
J 1.1-OCA P
I
CHCiCCIj
ITCEH
8
CH 2 CHCI
j 1,1-DCE []
]
CIS
[)
CH3COOH
CH3CHPH
Ethanot
[ TFIANS [)
jl
Acetic
acid
Fig. 5 (a) Sequential transformation of 1,1,1-trichloroethane to
chloroethene (VC), chloroethane (CA) and other degradation products
by abiotic (labelled a) and biotic (labelled b) mechanisms (from
Vogel and McCarty, 1987); (b) sequential transformation of
tetrachloroethene (PCE), to trichloroethene (TCE) and to other
degradation products (from Barrio-Lage et al., 198 6) .
Aquifer contamination and restoration
185
Table 1. Groundwater quality data (in /ig/L) at selected multilevel
samplers and piezometers within the outwash aquifer, September
1987. See Figs. 1 & 4 for locations. n.d. = not detectable, i.e.,
< 1 yUg/L. Abbreviations refer to Fig. 5.
Parameter
Chloroethene
Chloroethane
1,1-Dichloroethene
trans 1,2-Dichloroethene
cis 1,2-Dichloroethene
1,1-Dichloroethane
Trichloroethene
1,1,1-Trichloroethane
Tetrachloroethene
Abbreviation
vc
CA
11-DCE
TRANS
CIS
11-DCA
TCE
TCA
PCE
67M-9
15
n.d.
49
n.d.
1. 5
81
490
82
2. 9
54M-15
n.d.
n.d.
66
n.d.
n.d.
9.5
n.d.
93
n.d.
128P
0.6
n.d.
2.1
n.d.
0.5
1.9
n.d.
n.d.
n.d.
AQUIFER DECONTAMINATION
Following the assessment of groundwater contamination in the
outwash aquifer beneath the Gloucester Landfill, attention turned
to the options available for containing the contamination or
decontaminating the aquifer. One proposed solution, which is
commonly employed at similar sites, was to install an impermeable
slurry wall or grout curtain around the zone of highest
contamination. In order for such a method to be effective, there
must be an impermeable floor beneath the contamination zone. In
this situation, however, the limestone bedrock underlying the plume
is very permeable and is hydraulically connected to the overlying
outwash aquifer.
Consequently, after groundwater quality monitoring data
indicated the probable absence of pools of liquid-phase organics
(i.e., DNAPLs) within the aquifer, it was decided to decontaminate
the aquifer by a "pump and treat" scheme (A. J. Graham Engineering
Consultants, 1985). This scheme involves four purge wells situated
along the length of the plume, which pump into a treatment plant,
comprising lime addition, air stripping and granular activated
carbon. The treated groundwater is then returned to the aquifer by
four recharge wells. The purge wells are to operate at a total
rate of 1050 m /day beginning in 1990 and continuing for five
years. The cost of this operation is estimated to be' of the order
of Cdn $6 million (US $5 million).
In order to prove the efficacy of decontaminating the outwash
aquifer by purge wells (i.e., the "pump and treat" option), a test
was conducted (Whiffin & Bahr, 1985) . This involved the injection
of uncontaminated groundwater and two, non-reactive tracers into
the contaminated aquifer via an injection well and the withdrawal
of a similar volume of contaminated groundwater from another well 5
meters away. A multilevel sampler, Ml, was situated between the
two wells from which samples were analyzed for the tracers and for
three organic solvents present in that part of the aquifer
—1,4-dioxane, tetrahydrofuran and diethyl ether. The behavior of
the organic contaminants during desorption is shown in Fig. 6, in
which the time axis is replaced by the number of pore volumes
pumped through the test section of the aquifer.
186
R. E. Jackson et al,
\
-10
0
1
»!
o
1j
o
20
i
WITHDRAWAL
W-2
INJECTION
W-1
°
10 h
S
Ml
MONITOR
A
o
30
°
o
Oo
*
z
40
o
t-
Q_
CE
O
CO
50
60
70
-
DEE
THF
u
DIOXANE
o°
o
•
°
°o
0
•
80
90
Q
o
G
D
L
0
2
4
1
o
°
O
D
a
>
6
8
10
12
14
16
NUMBER OF PORE VOLUMES AT MULTILEVEL Mi -12
Fig. 6 Behavior of diethyl ether (DEE), tetrahydrofuran (THF) and
1,4-dioxane during the purge well test.
Analysis of these disappearance curves indicates that there is
a log-linear relationship (Fig. 7) between the number of pore
volumes withdrawn from the aquifer to obtain 90% apparent
desorption of each of the organics and their respective K
values.
This site specific relationship can be extrapolated to obtain an
estimate o'f the number of pore volumes required to decontaminate
the aquifer of other contaminants not present in that particular
part of the contaminated aquifer.
The next step is the optimization of the placement of the purge
wells so that a minimum amount of uncontaminated groundwater is
pumped and sent for treatment over the specified duration of
purging, by the end of which time the aquifer should have been
decontaminated to a prescribed level. This is being accomplished
by the development of computer models of groundwater flow and
solute transport within the Gloucester flow system. The 2-D
groundwater flow code is embedded in the USGS code AQMAN (Lefkoff &
Gorelick, 1987), which combines the flow simulation with
mathematical optimization to develop and evaluate aquifer
management strategies. AQMAN prepares a data file that defines the
objective and all constraint functions plus other information
required; the file is then used in a mathematical programming code
such as MINOS (Murtagh & Saunders, 1983) . The optimal locations
and pumping strategy determined with AQMAN are entered into a 2-D
solute transport code (Kent et al., 1986) to test that the proposed
purge-well system will decontaminate the aquifer. The simulation
of contaminant transport subject to sorption in a 2-D aquifer
undergoing pumping, uses the method of characteristics to solve the
solute transport equation (Konikow & Bredehoeft, 1978).
Aquifer contamination and restoration
187
Because of the high cost of "pump and treat" operations, and
because in-situ biorestoration is still under development, the
latter was not considered as a remedial option for the outwash
aquifer at Gloucester. However, it is anticipated that, at the end
of the five years of purging the contaminated aquifer, in-situ
biorestoration will be required for the cost effective clean up of
the residual contamination that will undoubtedly remain within the
aquifer.
2.0
3.0
L O G Koyy
Fig. 7 Estimation of the number of pore volumes that must be
purged from the outwash aquifer to achieve 90% decontamination.
Pore volume values for diethyl ether, tetrahydrofuran and dioxane
are obtained from Fig. 6. Other values are obtained by
extrapolation.
REFERENCES
Barrio-Lage, G., Parsons, F. Z., Nassar, R. S., & Lorrenzo, P. A.
(1986) Sequential dehalogenation of chlorinated ethenes.
Environ. Sci. & Tech. 20(1), 96-99.
Graham, A. J., Engineering Ltd. (1985) Gloucester Landfill Waste
Site. Problem Definition and Remedial Alternatives.
Gloucester, Ontario.
188
R. E. Jackson et al.
Jackson, R. E, Patterson, R. J., Graham, B. W., Bahr, J., Bélanger,
D., Lockwood, J. & Priddle, M. W. (1985) Contaminant
hydrogeology of toxic organic chemicals at a disposal site,
Gloucester, Ontario, I.W.D. Scientific Series No. 141, Ottawa,
Canada.
Kent, D. C , LeMaster, L. & Wagner, J. (1986) Modified N . R e version of the U.S.G.S. Solute Transport Model. Vol. 1:
Modifications. U.S. Environmental Protection Agency, Ada,
Oklahoma.
Konikow, L. F. S Bredehoeft, J. D. (1978) Computer solution of
two-dimensional solute transport and dispersion in groundwater.
Techniques of the U.S. Geological Survey, Chapter C2, Book 7,
U.S. Geological Survey, Reston, Virginia.
Lefkoff, L. J. & Gorelick, S. M. (1987) AQMAN: Linear and
quadratic programming matrix generator using two-dimensional
groundwater flow simulation for aquifer management modeling.
U.S. Geological Survey, WRI Report 87-4061. Menlo Park,
California.
Murtagh, B. A. & Saunders, M. A. (1983) MINOS 5.1 User's guide.
Tech. Reot. SOL83-20R, Stanford University, DePt. of Operations
Research, Stanford, California.
Patterson, R. J., Jackson, R. E., Graham, B. W., Chaput, D. &
Priddle, M. W. (1985) Retardation of toxic chemicals in a
contaminated outwash aquifer. Water Sci. and Tech., 17, 57-69.
Vogel, T. M. & McCarty, P. L. (1987) Abiotic and biotic
transformations of 1,1,1-trichloroethane under methanogenic
conditions. Environ. Sci. Tech. 21(12), 1208-1213.
Whiffin, R. B. & Bahr, J. (1985) Assessment of purge well
effectiveness for aquifer decontamination. In: Proc. of 4th
National Symp. on Aguifer Restoration and Groundwater
Monitoring. 75-81. NWWA, Dublin, Ohio.