e6079
Remediation of a Groundwater Nitrogen
(N) Plume Utilising Enhanced Microbial
Degradation and Phyto-Remediation
Processes
Mark Stuckey, Environmental & Earth Sciences Pty Ltd, [email protected]
EXECUTIVE SUMMARY
A remediation program was undertaken to reduce dissolved and free-phase
concentrations of total petroleum hydrocarbons (TPH) and benzene, toluene, ethylbenzene and xylene (BTEX), derived from leakage and spills as a result of past petroleum
storage at a site located within the Port Phillip Bay inshore segment on the Mornington
Peninsula in Victoria. Remediation included injection of nitrogen (N) in the form of urea
[CO(NH2)2] to the shallow unconfined water-table aquifer, which promoted microbial
degradation of hydrocarbon compounds and led to a rapid decline in dissolved TPH and
BTEX levels in groundwater.
Hybrid willow trees (Salix matsudana x alba) were planted along the northern site
boundary for a twofold purpose: to retard N-rich groundwater movement away from the
site towards Port Phillip Bay; and to increase the rate of de-nitrification and/or plant uptake
of nitrate (NO3-) dissolved in the aquifer.
On-site boreholes positioned in the centre of the N-plume reported increased Ca-HCO3
levels, interpreted to be due to microbial activity from hydrocarbon compound consumption
respiring carbon dioxide (CO2), seen as bicarbonate (HCO3-) alkalinity in the pH range
observed (6.1-7.7).
Redox levels became significantly more negative or reducing over time in the area of the
willow trees, which is inferred to be a result of an increased microbial population in the
rhizosphere of the tree roots. The microbial population consumed molecular oxygen in the
form of NO3- and sulfate (SO42-), resulting in significant decreases in the concentration of
these two anions. A trend of decreasing NO3- and SO42- with increasing HCO3- and
alkalinity was apparent over time.
The breakdown of N dissolved in the aquifer from ammonium (NH4+) to NO3- and finally
nitrogen gas and nitrous oxide gases under natural processes of ammonification,
nitrification, and de-nitrification was observed.
The final stage of works involved sampling soil and groundwater in the vicinity of the willow
tree root systems, and collecting samples of willow tree leaves and stems. This work was
performed to provide further evidence that nitrification and denitrification occurred to
reduce dissolved N levels in the aquifer, and that the willow trees assisted in this process;
thus demonstrating that phyto-remediation had also occurred.
The breakdown of N-compounds by an active microbial population was demonstrated
under both aerobic and anaerobic conditions. Aerobic conditions were enhanced by air-
sparging, and anaerobic conditions through stimulation of the indigenous microbial
population to consume dissolved and molecular oxygen.
INTRODUCTION
The purpose of this work was to confirm that the site was in a suitable condition for
medium density residential use, and to record and substantiate that groundwater
remediation works were conducted in accordance with statutory requirements. In
particular, it was to be determined that no potential for impact on any potential beneficial
uses, either ecological or human, existed due to migration of groundwater from the site.
This paper particularly aims to demonstrate that the remedial technologies adopted, and
methodologies employed, created processes that actually occurred and resulted in the
‘remediation’ of soil and groundwater at the site to an appropriate level.
METHODOLOGY
Eleven on-site and three off-site down-gradient piezometers were installed into the aquifer
to confirm hydrocarbon compound degradation and N-plume movement and behaviour, as
well as indicate background chemistry and geo-chemical evolution in the aquifer. In
addition, one down-gradient piezometer was installed through the base of the shallow
water-table aquifer into the underlying saline-wedge beneath the beach associated with
sea-water in Port Phillip Bay at a depth of 7-9 metres.
Groundwater monitoring events (GMEs) were undertaken in July, September and
November 2002, May 2003, May and October 2004, and February 2005.
BACKGROUND
Based on thermodynamic reactions, NO3- reduction (to NH4+ or N gas) is highly favoured in
the soil environment (Sposito 1994 pp49-52). Considering chemical equilibria in soils, it
can be estimated that NH4+ will be the dominant N species below a pe + pH relationship of
9.5 (at pH 7.0). As the soil profile, particularly in the saturated zone (beneath the watertable), is likely to be an oxygen limited or reducing (i.e. low pe) environment, NH4+ is
therefore likely to be the dominant form. Under such conditions N2(g) from the air can also
reduce to NH4+ and accumulate in soils. The NH4+ level in soil is also strongly pH
dependent, with increasing acidity causing more soil (and potentially atmospheric) N to
form NH4+. For example, at a pH of 5.0, the pe + pH relationship increases to about 11.2
(Lindsay, 1979).
Ammonium will however be oxidised to NO3- (nitrification) under favourable oxygenated
conditions, a process which is catalysed by bacteria. Once nitrification has occurred, the
major behavioural mechanisms of N are leaching, plant uptake, or respiration to form N
gas (denitrification). Conditions which favour microbial activity encourage nitrification and
denitrification. Denitrification occurs when microbial demand for oxygen exceeds its
availability (seen in water as increased biological oxygen demand [BOD]).
In highly reducing groundwater environments with a high N input, NH4+ is expected to be
the dominant form of dissolved N in a NO3- reducing environment (Lawrence et. al. 2000).
Robertson and Cherry (1992) report a process of NH4+ oxidation to NO3- (nitrification) in
the unsaturated zone of a sand profile, followed by denitrification in the anaerobic aquifer
sediments of the saturated zone beneath the water-table. The driving force for
denitrification was determined to be elevated solid-phase organic carbon (OC) content
(2.5%) in the aquifer sediments, a conclusion which appeared to be confirmed by similar
studies of low OC aquifers that did not report the same level of denitrification. Soil aquifer
zone OC levels on this site ranged between 0.03 and 0.6%.
DISCUSSION
Hydro-geochemistry
The hydraulic conductivity (K) of the water-table aquifer was determined to be 2.45 m/day
from field single borehole falling head and recovery tests. The hydraulic gradient (i) was
determined to be 0.005 m/m in a northerly direction. A maximum groundwater velocity of
0.06 m/day (22 m/year) was calculated in a northerly direction towards Port Phillip Bay
(100 metres north of the northern site boundary). See Figure 1.
Figure 1 Borehole locations and inferred potentiometric surface contours, October 2004
Groundwater located within the aquifer at depths between 0.8 and 1.54 mBGL is neutral to
slightly alkaline, with a pH range in all boreholes at the times of sampling of 6.1 to 7.7
(average 7.1). The salinity (EC) of the water-table aquifer ranged from 610 to 10920
µS/cm (av. 4790 µS/cm), and the deeper aquifer at borehole BH8D from 16900 to 19710
µS/cm (av. 17900 µS/cm) over the period of sampling (field determined values).
The laboratory determined TDS results show that for 54 samples collected from the
shallow uppermost-unconfined water-table aquifer, an average salinity of 2740 mg/L was
observed. Off-site boreholes BH8S, BH9 and BH10 had an average TDS of 1100 mg/L,
with on-site boreholes having an average 3370 mg/L TDS. The TDS relationship across
the area of investigation shows that salinity increases significantly to the south, south-east
during all sampling events, indicating that despite a groundwater gradient towards the bay
to the north, recharge of fresh rain-water at the site appears to be predominantly along the
curb-side drainage trench of Point Nepean Road.
The shallow unconfined groundwater across the site is dominated by the ions sodium
(Na+) and chloride (Cl-), with magnesium (Mg2+), calcium (Ca2+), HCO3- and SO42- being
sub-dominant (Na-Cl >Ca-HCO3/Mg-SO4).
Sixty hybrid willow trees were planted over a 180 m2 area in August and September 2003
across the northern property boundary (denoted as the shaded area on Figure 1 above).
The hybrid variety of willows was selected on the basis of their rapid growth and ability to
consume large volumes of water, to assist in retarding off-site groundwater movement and
increase denitrification and/or transpiration of NO3- in the aquifer.
Increased reducing conditions were observed over time at boreholes BH1, BH2, BH4,
BH13, BH14, BH16, and BH17 (130 to –155 mV; pe 2.2 to –2.62). The redox relationship
is an indication of an active microbial population, consuming all dissolved oxygen from the
area of the aquifer beneath and to the north of the site. It is expected that under these
conditions OC and dissolved N will be food sources, and O2 >NO3 >SO4 will be electron
acceptors and hence be rapidly consumed in the order indicated.
On-site boreholes BH13, BH16 and BH17 (positioned in the centre of the N-plume)
reported increased Ca-HCO3 levels. This was interpreted to be due to microbial activity
from hydrocarbon compound consumption respiring CO2, which in turn is converted to
HCO3- alkalinity in the pH range 6.1-7.7 observed at these locations. Microbial respiration
is also consuming molecular oxygen (seen as SO42- and NO3-), resulting in significant
decreases in the concentration of these two anions. The trend of decreasing NO3- and
SO42- with increasing HCO3- and alkalinity is apparent over time at boreholes BH1, BH2,
BH11, BH12, BH14, BH16 and BH17. Chart 1 shows the change in chemistry of
groundwater at borehole BH13 over time.
10000
1000
100
meq/L
10
1
TDS
Na
Ca
Mg
K
NH4
Cl
SO4
HCO3
NO3
PO4
F
0.1
0.01
0.001
0.0001
Ion
BH13 July 2002
BH13 Sept 2002
BH13 May 2003
BH13 May 2004
BH13 Oct 2004
Chart 1 Borehole BH13 groundwater chemistry over time
The processes occurring since urea injection in August 2002 can be seen to be
ammonification, as CO(NH2)2 was converted to NH4+ between August and December
2002, followed by nitrification, as NH4+ was converted to NO3- between December 2002
and May 2003. Between May 2003 and May 2004 de-nitrification of NO3- resulted in the
final stage of conversion of applied N to N gas and gaseous oxides of N by anaerobic
bacteria. Nitrogen gases dissolved in the aquifer are readily available for plant uptake,
particularly by the willow trees, and de-nitrification catalysed by bacteria such as
Pseudomonas, Bacillus, Micrococcus and Achromobacter (The University of Sydney
1992). Charts 2 and 3 also show this process.
100
Ammonium as N (mg N/L)
10
1
0.1
0.01
May-02
Sep-02
Dec-02
Mar-03
Jun-03
Oct-03
Jan-04
Apr-04
Aug-04
Nov-04
Feb-05
Nov-04
Feb-05
Date
BH1
BH2
BH4
BH11
BH12
BH13
BH14
Chart 2 NH4 (as N) levels in groundwater, August 2002 to October 2004 (note log scale on y-axis)
100
Nitrate as N (mg N/L)
10
1
0.1
0.01
May-02
Sep-02
Dec-02
Mar-03
Jun-03
Oct-03
Jan-04
Apr-04
Aug-04
Date
BH1
BH2
BH4
BH11
BH12
BH13
BH14
BH15
BH16
BH17
Chart 3 NO3 (as N) levels in groundwater, August 2002 to October 2004 (note log scale on y-axis)
Fluoride (F-) levels on-site ranged from 0.01 to 0.67 mg/L, and concentration trends mirror
that of TDS/EC (increasing to the south of the site). Gupta et. al. (2005) state that a
consistent cause of elevated F- in groundwater is evaporative enrichment co-incident with
elevated salinity (TDS). As F- levels across the site exceed that of rainwater (zero to 0.089
mg/L; Gupta et. al. 2005, Hem 1992 and Jankowski 1999) but are less than sea-water
(mean 1.3 mg/L), it can be hypothesised that rainfall recharge, followed by evaporative
enrichment, influenced F- (and other ion and TDS) distribution across the site.
The Na/Cl ratios presented suggest all groundwater has been residing in the aquifer for
some time due to the dominance of Na over the less reactive Cl, despite recharge of
storm-water run-off along Point Nepean Road (Jankowski 1999). Further, the consistency
of the non-reactive solutes (Na/Cl) ratio over time at all locations indicates that minimal
dilution has occurred (Robertson and Cherry 1992). Na is also generally dominant over
Ca and Mg in this Na-Cl aquifer, however increasing dissolved Ca (as seen by Na/Ca <2,
Mg/Ca <1 and Ca/K >10) is apparent along with increasing alkalinity (Cl/HCO3 <1) at a
number of locations with increasing microbial populations, including boreholes BH13,
BH16 and BH17.
This occurrence is an indicator of microbial respiration occurring as a result of their
breakdown of dissolved organic compounds in the aquifer and adsorbed organic
compounds on aquifer sediments and organic matter.
Locations where the Na/Cl ratio exceeds that of sea-water and rainwater (0.55) include
boreholes BH13, BH16 and BH17 in May 2004, and boreholes BH1 and BH13 in October
2004. As Ca-HCO3 is also elevated at these locations at these times, this suggests that
Na is being exchanged from the solid-phase in the plume centre during elevated microbial
activity from de-nitrification that results in increased Ca-HCO3 dissolved in the aquifer and
subsequent ion exchange of Ca for Na (Daessle et. al., 2005).
Nitrogen Inputs
To assist the hydrocarbon compound bioremediation process, 20 kg of urea (CO[NH2]2)
was added to the excavation pit, with 40 g of urea added to air sparging boreholes. These
amounts were based upon existing ratios of nutrients within the aquifer so as not to disturb
nutrient concentrations following completion of remediation works. Table 1 has been
provided to demonstrate the calculations made following the May 2003 groundwater
monitoring event to ensure all added N could be attenuated by the aquifer, without
considering de-nitrification.
Table 1 Estimated mass of N added to aquifer
Concentration Contour
mg NO3/L
128
64
32
16
8
4
2
1
0.5
Total
Area
m2
20
180
400
250
600
3500
3500
2100
2700
13250
Groundwater Volume
m3
12
103
250
143
356
2150
2081
1244
1623
7962
Nitrate
kg
1.47
9.29
11.94
3.44
4.28
12.92
6.24
1.87
1.14
52.58
Total N
kg
0.33
2.11
2.71
0.78
0.97
2.94
1.42
0.42
0.26
11.95
Table 1 shows that close to 12 kg N was present in the aquifer on 30 May 2003. This
corresponds to 25 kg urea (which has 46.7% N), thus the prediction of N mass added
(based on dissolved levels in groundwater) compares relatively well to that added to the
aquifer as urea. Note that urea is completely soluble in water. The concentrations
detected in groundwater will be readily attenuated to 6.6 mg NO3-/L or 1.5 mg N/L within a
13000 m2 area.
Willow Trees Assessment
The final stage of works (undertaken in February 2005) involved sampling soil and
groundwater in the vicinity of the willow tree root systems, in addition to collecting samples
of willow tree leaves and stems. These samples (which were collected from both plumecentre and background locations) were analysed for N content, and N-fixing fungi and
bacteria. An incubation experiment was also undertaken on soil and groundwater
collected from the aquifer in the willow trees rhizosphere to reproduce the N-utilising
microbial population and associated nitrification/denitrification processes.
This work was performed with the aim of providing further evidence that nitrification and
denitrification occurred to reduce dissolved N levels in the aquifer, and that the willow trees
assisted in this process; thus demonstrating that phyto-remediation had taken place.
Results are summarized in Table 2.
Table 2 Willow Tree Assessment Results Summary
Plume N-leaf Soil Rhizosphere Groundwater
Active
Total
Active
Total
Water
Location tips
(mg/kg)
(mg/L)
Biomass Biomass Biomass Biomass Biomass
Units
%
N
NO3 NH4 N NO3 NH4
µg/g
µg/g
µg/mL
µg/mL
Ratio
Centre
1.46 340 0.13 0.71 0.94 0.05 0.05
1.7
96
0.06
5.12
0.01
Periphery 1.14
1.99
152
Outside
0.91 420 0.96 1.84 1.18 2.5 0.07
1.7
78.6
0.19
1.05
0.18
The results of leaf analysis from samples collected demonstrate that trees contained more
N in their leaf tips closer to the plume centre. Soil analysis from the plant rhizosphere
determined that total N, NO3 and NH4 levels were lowest in the plume centre and highest
outside the plume. This is considered to be a reflection of microbial activity and hence N
consumption and degradation being higher in the plume centre. Likewise, a groundwater
sample collected from the plume centre (from borehole BH13) contained less total N, NO3
and NH4 than a sample from outside the plume (borehole BH14). This is seen to support
the findings of the soil assessment, that nitrification and de-nitrification processes are
greater in the plume centre.
Microbial analysis of soil collected from the aquifer zone for N fixing bacteria in the plant
rhizosphere determined 21600 CFU/gm outside the plume area, and 8400 CFU/gm in the
plume centre. Active bacterial biomass soil results indicate that the microbial population
may be dying off and becoming less active as all available N has been consumed from soil
and groundwater within the plume area by the time of sampling in February 2005.
This hypothesis is supported by redox levels in groundwater, which show highly reducing
conditions in the plume centre (boreholes BH13, BH14 and BH16) and reducing conditions
at the plume periphery (boreholes BH1, BH2, BH12 and BH17) by October 2004. Thus
aerobic bacteria will die off as oxygen (dissolved and molecular) is removed from the
aquifer and N fixing can no longer occur.
The fact that total biomass exceeds active biomass by a significant degree indicates soil
pH is becoming more alkaline (Lower et. al. 2002), which may be a result of increased
alkalinity in the groundwater (due to microbial respiration). Likewise in groundwater,
borehole BH14 outside the plume had a higher active bacterial biomass, whereas borehole
BH13 in the plume centre had a greater total bacterial biomass. Hence, borehole BH14
has a higher active to total biomass ratio, suggesting severely depleted, but past elevated,
N levels in the former plume centre. Elevated total bacterial biomass in the plume centre
and periphery suggests that active bacterial levels have been elevated in the past,
indicating past microbial activity. This is supported by elevated N in willow tree leaf tips in
the plume centre.
CONCLUSION
This project has demonstrated that the combined use of enhanced monitored natural
attenuation and phyto-remediation provided a solution to a groundwater pollution issue on
a site located in a sensitive ecological area (adjacent to Port Phillip Bay), and zoned for
residential development. The breakdown of N-compounds by an active microbial
population was demonstrated under both aerobic and anaerobic conditions. Aerobic
conditions were enhanced by air-sparging, and anaerobic conditions through stimulation of
the indigenous microbial population to consume dissolved and molecular oxygen.
Excess N, in the form of NO3- and NH4+ in the aquifer, has been removed through
ammonification, nitrification (assisted by air-sparging) and de-nitrification (assisted by
phyto-remediation) processes that converted NO3- to N gas and nitrous oxide gases that
are available for plant uptake.
ACKNOWLEDGEMENTS
The author wishes to acknowledge the permission of Michael Buckley to present this
information in a public forum.
REFERENCES
Daessle, L W, Sanchez, E C, Camacho-Ibar, V F, Mendoza-Espinosa, L G, Carriquiry, J D,
Macias, V A and Castro, P G (2005) — Geochemical evolution of groundwater in the
Maneadero coastal aquifer during a dry year in Baja California, Mexico. Hydrogeology
Journal (2005) 13:584-595.
Gupta, S K, Deshpande, R D, Agarwal, M and Raval, B R (2005) — Origin of high fluoride
in groundwater in the North Gujarat-Cambay region, India. Hydrogeology Journal (2005)
13:596-605.
Hem, J D (1992) ⎯ Study and interpretation of the chemical characteristics of natural
water. Third Edition. US Geological Survey Water-Supply Paper 2254.
Jankowski J (1999) — Hydrogeochemistry. The University of New South Wales,
Groundwater Centre.
Lawrence, A R, Gooddy, D C, Kanatharana, P, Meesilp, W and Ramnarong, V (2000) —
Groundwater evolution beneath Hat Yai, a rapidly developing city in Thailand.
Hydrogeology Journal (2000) 8:564-575.
Lindsay W L (1979) — Chemical equilibria in soils. The Blackburn Press.
Lower S S and Orians C M (2002) — Soil nutrients and water availability interact to
influence willow growth and chemistry but not leaf beetle performance. The Netherlands
Entomological Society 107: 69-79.
Robertson, W D and Cherry, J A (1992) — Hydrogeology of an unconfined sand aquifer
and its effect on the behaviour of nitrogen from a large-flux septic system. Hydrogeology
Journal (1992) 0:0.
Sposito G (1994) — Chemical equilibria and kinetics in soils. Oxford University Press.
The University of Sydney (1992) — Soil microbiology. Department of Microbiology, Soil
Science 2 Practical Course Manual.