Environ Sci Pollut Res
DOI 10.1007/s11356-013-2171-2
NEW APPROACHES FOR LOW-INVASIVE CONTAMINATED SITE CHARACTERIZATION, MONITORING AND MODELLING
Integration of geophysical, geochemical and microbiological
data for a comprehensive small-scale characterization
of an aged LNAPL-contaminated site
Alessandro Arato & Markus Wehrer &
Borbala Biró & Alberto Godio
Received: 15 April 2013 / Accepted: 16 September 2013
# Springer-Verlag Berlin Heidelberg 2013
Abstract Characterization of aged hydrocarbon-contaminated
sites is often a challenge due to the heterogeneity of subsurface
conditions. Geoelectrical methods can aid in the characterization of such sites due to their non-invasive nature, but need to be
supported by geochemical and microbiological data. In this
study, a combination of respective methods was used to characterize an aged light non-aqueous phase liquid-contaminated
site, which was the scene of a crude oil blow-out in 1994. As a
consequence, a significant amount of crude oil was released
into the subsurface. Complex resistivity has been acquired, both
Responsible editor: Michael Matthies
Electronic supplementary material The online version of this article
(doi:10.1007/s11356-013-2171-2) contains supplementary material,
which is available to authorized users.
A. Arato (*) : A. Godio
Dipartimento di Ingegneria per l’Ambiente, il Territorio e le
Infrastrutture, Politecnico di Torino, Corso Duca degli Abruzzi 24,
10129 Torino, Italy
e-mail: alessandro.arato@polito.it
M. Wehrer
Lehrstuhl für Hydrogeologie, Institut für Geowissenschaften,
Burgweg 11, 07749 Jena, Deutschland
M. Wehrer
Department of Earth and Environmental Sciences, Rutgers, the State
University of New Jersey, 101 Warren Street, Newark,
NJ 07102, USA
along single boreholes and in cross-hole configuration, in a
two-borehole test site addressed with electrodes, to observe
the electrical behaviour at the site over a two-year period
(2010–2011). Geoelectrical response has been compared to
results of the analysis of hydrocarbon contamination in soil
and groundwater samples. Geochemical parameters of groundwater have been observed by collecting samples in a continuous
multi-channel tubing (CMT) piezometer system. We have also
performed a biological characterization on soil samples by
drilling new boreholes close to the monitoring wells. Particular
attention has been given to the characterization of the smear
zone that is the sub-soil zone affected by the seasonal groundwater fluctuations. In the smear zone, trapped hydrocarbons
were present, serving as organic substrate for chemical and
biological degradation, as was indicated by an increase of
microbial biomass and activity as well as ferrogenicsulfidogenic conditions in the smear zone. The results show a
good agreement between the intense electrical anomaly and the
peaks of total organic matter and degradation by-products,
particularly enhanced in the smear zone.
Keywords Multi-component NAPL . Aquifer . Borehole
geophysics . Induced polarization . Complex resistivity .
Geochemical monitoring . Multilevel piezometer . Biological
assays . Crude oil . Bio-degradation
Introduction
B. Biró
Institute for Soil Sciences and Agricultural Chemistry, Centre for
Agricultural Research, Hungarian Academy of Sciences,
Herman Otto ut 15, 1022 Budapest, Hungary
Present Address:
B. Biró
Faculty of Horticulture Department of Soil Science and Water
Management, CORVINUS University of Budapest,
Villányi ut 29-43, H-1118 Budapest, Hungary
Porous media can be considered as multi-phase systems,
consisting of solid grains and pores, partially or totally filled
by various fluids. Each phase is responsible, through its individual electrical properties, for the bulk resistivity of the
formation which is going to be measured in the field. An
important contribution also comes from the geometric arrangement of the solid matrix and the (non-)aqueous solution
Environ Sci Pollut Res
filling the inter-connected pores. The flow of electric charges
through the soil depends on electrolytic conduction (e.g.
Archie 1942; Kemna 2000), interfacial conduction and polarization mechanisms (e.g. Seigel 1959; Börner et al. 1996). We
can distinguish between these different phenomena by measuring, in time domain, direct current (DC) conduction properties (i.e. electrical conductivity or its reciprocal resistivity)
and polarization mechanisms (i.e. chargeability, defined as the
integral of a residual voltage decay after current switch-off).
Since resistivity can be considered as a complex quantity
and frequency dependent, a complete observation of electrical
phenomena requires its measurement in frequency domain.
Conduction and polarization properties of the soil are thus
assessed by measuring the module and phase of a sinusoidal
wave injected in the ground, related to the real and imaginary
part of the complex resistivity, respectively.
At hydrocarbon-contaminated sites, the presence of a third
phase (fourth within the vadose zone) increases the complexity
of the problem to be assessed. Past studies (e.g. Sauck 2000;
Marcak and Golebiowski 2008) showed that the distribution of
light non-aqueous phase liquid (LNAPL) contamination into
the subsurface can be divided into four main fractions:
–
–
–
–
dissolved phase within the saturated zone, corresponding
to a contaminant plume that undergoes flow and transport
processes into the aquifer;
free phase within the saturated layer of the so-called
smear zone (that corresponds to the zone of fluctuation
of the piezometric surface);
residual phase within the upper smearing zone (unsaturated);
vapour phase within the vadose zone.
The effect of the vapour phase can be neglected from a
merely geoelectrical point of view, since it tends to migrate
towards the surface and to be removed quite rapidly in the
atmosphere. Whilst, all the other three phases of the contamination have to be considered. The presence of hydrocarbon
contaminants in a unconsolidated porous medium produces a
distinctive geoelectrical anomaly, since the LNAPL components are characterized by extremely high resistivity values, in
a broad range between 106 and 109 Ωm (Lucius et al. 1992;
Olhoeft 1985). Therefore, the presence of a hydrocarbon
phase should result in resistive anomalies at NAPLcontaminated sites, especially in the early periods after the
contamination, as found by several authors (e.g. Mazác et al
1990; DeRyck et al. 1993).
However, in the case of long-term contamination, the matured hydrocarbon mixture can undergo bio-degradation, generating a conductive plume where an enhanced concentration
of ionic species and degradation by-products can be found
(e.g. Atekwana et al. 2000; Atekwana et al. 2004; Abdel Aal
et al. 2004, 2006). Most of literature works agree that aged
hydrocarbon undergoing degradation results in a conductive
plume with enhanced polarization response (e.g. Atekwana
et al. 2000, 2002; Abdel Aal et al. 2004, 2006; Werkema et al.
2003), but there are some studies that associate strong NAPL
contamination to resistive feature into the subsoil (e.g. Benson
et al. 1997; Tezkan et al. 2005; Frohlich et al. 2008).
Quantitative information on the role of biomass degrading
activity is difficult to obtain, especially if it needs to be
converted in a useful parameter for geophysical data inversion. Studies on the effect of microbial activity at
hydrocarbon-contaminated sites have shown a decrease in
bulk resistivity and an increase in induced polarization (IP)
response in correspondence to the contaminated locations and
to the conductive plumes detected downstream. Degradation
process is responsible for surface alteration and increase in
surface conduction due to the interaction of contaminants with
grain surfaces(e.g. Abdel Aal et al. 2004, 2006), affecting the
resistivity and the IP response (both in time-domain and in
frequency-domain). Therefore, due to their non-invasive nature, geoelectrical methods offer an excellent opportunity to
aid characterization of NAPL-contaminated sites. On the other
hand, a robust model to explain the geophysical response
accounting for the presence of NAPL (and for biodegradation
activity) has yet to be developed because of the difficulty in
distinguishing surface and electrolytic effects and because the
smearing process increases the complexity of fluid distribution and saturation into the pore spaces, creating isolated blobs
(or pendular rings in water non-wetting grain surfaces) of
NAPL and enhancing the spatial heterogeneity at the contaminated locations.
This work is aimed to provide a comprehensive small-scale
characterization of hydrocarbon-degradation processes by
means of both geoelectrical cross-hole time- and frequencydomain methods, together with monitoring campaigns of
physico-chemical parameters and laboratory biological analysis at a hydrocarbon-contaminated site in Northwest Italy.
The integrated approach has been composed of geophysical
techniques, geochemical monitoring and biological laboratory
assessments. Indirect information coming from vertical profiles and 2D section of electrical resistivity and induced polarization measurements are validated by direct measurements
of the most relevant geochemical parameters of soil and water
samples, and by laboratory counts of (an-)aerobic bacterial
colonies. Biological degrading activity is estimated by carrying out enzymatic assays.
Material and methods
The experimental activity has been carried out at Trecate site
over a 2-year time period, from 2009 to 2011. After a preliminary geophysical characterization of the site (2008–2009),
presented in Godio et al. (2010), a test site for cross-hole
monitoring has been set-up in December 2008 within the most
contaminated area of the site.
Environ Sci Pollut Res
The test site has been specifically designed to monitor
time-lapse changes in electromagnetic properties both in the
vadose and in the saturated zone. More specifically, it has been
designed for cross-hole electrical resistivity tomography
(ERT) and ground-penetrating radar (GPR), and completed
with two PVC-screened boreholes (2″ diameter), named B-S3
and B-S4 as they were installed under the framework of EU
project SoilCAM. The two boreholes are 6 m spaced and
reach a depth of 18 m below ground surface (b.g.s.), and they
allow to perform cross-hole GPR surveys and geochemical
monitoring of groundwater quality, with multi-electrode
probes. Each borehole consists of 24 electrodes, 0.7 m spaced,
in the depth range between 1 m to 17.1 m b.g.s..
In December 2009, a multi-level sampling system has
been installed in the vicinity of B-S3 and B-S4. Solinst
continuous multi-channel tubing (CMT) system has been
installed in situ with a direct push Geoprobe device.
The CMT system consists of four multi-channel boreholes (named B-S5, B-S6, B-S7and B-S8, from WestEast direction), with seven separate micro-tubes in each.
A transect of 28 sampling ports is then available for
mapping the most important geochemical parameters of
interstitial filling fluid vertically and horizontally in the
saturated zone.
Figure 1 shows the geographical location of the Trecate
site, including the positioning of the test site for geophysical
cross-hole monitoring (blue dots), the CMT system (red dots)
and the sampling location for the biological characterization
(black dots). The test site for geophysical and geochemical
monitoring is described in Fig. 2.
From the drilling of two boreholes of the CMT system, soil
columns were extracted in transparent PVC liners, for laboratory characterization of lithological features, presence of contaminant and laboratory measurements of the major geochemical parameters.
In addition, microbial count analysis and in vitro growth
experiments were performed to assess the presence of active
colonies of hydrocarbon-degrading bacteria.
Geophysical characterization
Several geophysical surveys have been carried out at the
SoilCAM test site. More specifically, 1D and cross-hole 2D
complex electrical resistivity measurements have been
performed on a monthly basis from 2009 to 2011. In-hole
resistivity and IP logs were acquired by shifting the quadripole
with minimum inter-electrode separation along the total length
of the boreholes. We used a Wenner configuration, with a =
0.7 m (i.e. the electrode separation). 2D resistivity tomography has been performed with a pole–dipole array, using the 48
electrodes along the two boreholes and a 49th electrode as the
Fig. 1 Geolocalization of the Trecate test site, with indication of boreholes for geophysical small-scale monitoring (in blue), multi-level sampling
system (in red) and the location of boreholes used for microbiological analysis (in black)
Environ Sci Pollut Res
Fig. 2 Description of the test site for geophysical monitoring and characterization of chemico-physical condition of the groundwater
remote pole, which was placed 300 m farther away in the
eastern direction.
Measurements have been carried out with a complex resistivity device (Polares), developed by PASI Ltd., jointly with
geophysics research group of Politecnico di Torino, which is
designed to measure complex impedance data in the frequency range between 0.5 and 100 Hz. Analogue measurements
were carried out with a DC-current device (IRIS Syscal),
through which DC resistivity and chargeability data were
acquired.
Resistance and phase data from 1D borehole profiles were
transformed to apparent resistivity and phase logs (the same
was done with time-domain resistivity and chargeability),
while the 2D complex impedance data were inverted using a
novel code for modelling and inversion of 2D and 3D complex resistivity data, which is based on the theoretical framework developed by Borsic and Adler (2012).
A least-squares minimization formulation is herein used,
and then the L2-norm is minimized on both the data and the
regularization term of the Occam’s objective function. The
reconstructed resistivity (ρ rec) distribution comes from the
minimization of the following equation:
ρrec ¼ arg minkWd ðZðρÞ−Zmeas Þk2 þ αkWm ðρ−ρ0 Þk2
where ρ and ρ 0 are the complex resistivity distribution at the
current and previous iteration respectively (ρ 0 can incorporate
different sources of a priori information), W d is the data
weighting matrix (a covariance matrix), Z is the vector of
simulated complex impedances coming from the solution of
the forward problem on the resistivity model Z, Z meas is the
vector of the measured impedances, and, finally, α and W m
represent the Tikhonov factor and the regularization matrix
(which can control horizontal/vertical flatness of the model or
can constrain the inversion to the prior resistivity model).
The solution of this equation is carried out with a iterative
Gauss–Newton approach, and the search for the model update
Δρ k at the kth iteration is performed as
JTk WTd Wd Jk þ αWTm Wm Δρk ¼ JTk WTd Wd ½Zmeas −Zðρk Þ−αWTm Wm ðρk −ρ0 Þ
where J is the first-order differential matrix (the Jacobian),
which accounts for the sensitivity of the data to variations of
model parameters.
The results of the inversion of complex resistivity data are a
map of the amplitude (i.e. resistivity) and phase of the complex resistivity. In this context, the inversion is carried out on
complex resistivity data collected at a frequency of 1.79 Hz.
The data error values were estimated using the relationship
proposed by LaBrecque et al. (1996), who defined a linear
Environ Sci Pollut Res
model for the errors in the data with the form: ε(R )=a +bR.
The use of a pole–dipole sequence did not permit to perform a
statistical analysis of the errors based on the discrepancy
between direct and reciprocal measurements (LaBrecque
et al. 1996). We based our error model on over-estimating
the average repeatability instrumental error. Conscious that
this choice, based on empirical knowledge, is widely arbitrary,
we used an error estimation model having a =1 mΩ and
b =0.02 and 0.05 for dataset of October 2010 and May
2011, respectively. As far as the phase errors are concerned,
we applied the same linear model, using a=0.1° and b=0.05
and 0.1 for dataset of October 2010 and May 2011, respectively. Unrealistic data, such as positive phase shifts, were
ignored.
with a 0.45-μm filter with supor membrane and samples for
inductive coupled plasma-optical emission spectroscopy (ICP
OES), TOC and DOC were additionally acidified immediately
after sampling. Sulfate(SO42−), nitrate (NO3−) and chloride
(Cl−) were also analysed in the laboratory with ion chromatography. Total organic carbon (TOC) and dissolved organic
carbon (DOC) was determined as non-purgeable organic carbon. TIC was determined as purgeable carbon after acidification. Total calcium, magnesium, sodium, potassium, iron and
manganese were determined by ICP OES. TPH was extracted
according to EN ISO 9377-2:2000 and measured with GCFID (see Supplemental Material).
Microbial counts and enzymatic activity in below-ground soil
layers
Geochemical parameters in soil and groundwater samples
To better understand the small-scale geophysical response at
the test site, three measurement campaigns of the major geochemical parameters were carried out in August 2010, October 2010 and May 2011.
The soil samples were analysed for grain size characterization, water content measurements (and derivation of
related quantities such as bulk electrical conductivity and
bulk dielectric permittivity), pH, cation exchange capacity
(CEC) and measurements of total carbon (TC), total organic (TOC), total inorganic carbon (TIC) and total petroleum
hydrocarbons (TPH). Analysis of dithionite and oxalate
dissolvable minerals, containing Fe, Mn and Al, were also
performed. Details of the methods can be found in the
supplemental material.
Groundwater sampling was carried out in correspondence
of the sampling ports of the multi-channel system at very low
flow rate with submersible double-valve pumps. Each port
was previously depurated from stagnant water by pumping
several liters of water with a peristaltic pump.
To avoid oxygen exchange through air and the sample, the
double-valve pumps were operated with nitrogen gas. To
avoid sample cross-contamination, pumps and tubes were
cleaned thoroughly between two sampling events. In addition
to the multilevel piezometers, a well-located outside the contaminated zone was sampled in order to derive the background
values.
Dissolved oxygen, electrical conductivity, redox potential,
pH and CO2 were analysed online in flow through-cells with
ion selective electrodes (WTW, Weilheim) during pumping on
site. Nitrate (NO3−), nitrite (NO2−), sulfate(SO42−), sulfide
(S2−), iron (Fe2+), manganese (Mn2+; for the second campaign
only) and turbidity were determined photometrically on site
immediately after sampling. Samples for laboratory analysis
were kept refrigerated immediately after sampling during all
time. The samples for laboratory analyses (carried out in Jena,
Germany) were filtered in the field (except TOC samples)
Microbial counts and growth experiments have been
performed, in the laboratories of the RISSAC HAS (Budapest,
Hungary), on contaminated and uncontaminated samples
from nine different sampling locations at Trecate site (indicated by black dots in Fig. 1, with progressive ID number S1 to
S9). Samples at different depths were assessed in order to
obtain a vertical distribution of bacterial abundance and
degrading activity.
Isolation and enumeration of soil microorganisms were
performed using the dilution plate-counting technique. For
each depth, two series of diluted specimens were used, for
the estimation of both aerobic and anaerobic microbial components in the soil samples.
The catabolic enzymatic activity, through which it is
possible to infer the presence of extra-cellular enzymes and
thus to verify the presence of additional sources of energy,
was characterized by fluorescent diacetate hydrolysis (FDA).
FDA consists of a colourless transformation to fluorescein
by both free and membrane-bound bacterial enzymes. FDA
hydrolysis was measured to account for the general microbiological activity, as stated by Villanyi et al. (2006). Furthermore, the measure of FDA activity provides a description of the microbiological state of the soils, and can be
compared to the amount of the contaminants (Biró et al.
2012, 2013).
Results
We hereby present the results of the integrated geochemical,
biological and geophysical characterization at the contaminated site of Trecate. First, the results of direct (in situ and in
laboratory) measurement of the geochemical evidences of the
contamination and its natural degradation are shown. Finally,
the results of the geoelectrical surveys are presented to spread
the characterization over the whole section delimited by the
boreholes B-S3 and B-S4.
Environ Sci Pollut Res
Fig. 3 Soil texture profiles from
boreholes B-S5 and B-S7
Geological stratification
The geological stratification has been obtained by performing
a grain size analysis on soil liners from drilling of wells B-S5
and B-S7. As shown in Fig. 3, it has been pointed out a
prevalence of gravel and sand along the entire column. Seventy percent of silty material (and 5 to 8 % of clay) has been
measured in the top soil layer, in the depth range between 0
and 3 m. Downwards, no significant amount of clay was
detected. Silt is present in low percentage (2–5 %) below a
depth of 3 m, with an increase in B-S5 (up to 10 %) around a
depth of 8 m.
Distribution of geochemical parameters in soil
and groundwater samples
We mainly focus on the TOC, total petroleum hydrocarbon
(TPH) and DOC as the main indicators of the vertical distribution of contaminants in groundwater and in soil. The TOC,
DOC, Fetot and Mntot concentrations in the groundwater during three sampling campaigns in the four piezometer wells
BS5 to BS8 are shown in Fig. 4 and, while in Fig. 5 the Total
Carbon (TC), TOC and TPH in the solid phase are presented.
TOC concentrations in soil samples below the depth of 3 m
and in groundwater are due to the hydrocarbon content and
can be used as representative for TPH, which was only measured during the third campaign (correlation coefficient TOC-
TPH soil: 0.94, p =0.01; correlation coefficient TOC-TPH
groundwater: 0.87, p =0.01). Figure 4a shows that in particular the depths from −8.5 to −10 m are a contaminant hot spot
with large hydrocarbon concentrations. DOC concentrations
do not show such a pronounced peak in these depths (Fig. 4b).
The difference between TOC and DOC comprises oil droplets
of size 0.45 μm or larger while DOC either represents smaller
droplets, truly dissolved hydrocarbons or metabolic products
of the hydrocarbon degradation. Depths between −7 and −8.5
and also between −10 and −12 m are affected by this fraction
only, but not by the larger size droplets. There is remarkable
little difference in the hydrocarbon concentrations at the three
different sampling dates despite the groundwater fluctuations,
which happened between these dates. The depths with large
hydrocarbon concentrations in the solid phase (Fig. 5) correspond with the depth of the most contaminated groundwater.
Figure 4c, d shows concentrations of total iron and specific
conductivity of the pore water. Both iron and manganese (not
shown, but reflective of the values for iron species) are largely
comprised of the mobile species Fe2+ and Mn2+, respectively.
Particularly the depths from −7 to −11 m are affected by large
concentrations. Like TOC/DOC concentrations, the three
sampling dates have very similar concentrations. Nitrate and
nitrites were mostly absent (data not shown). Due to the
sampling strategy, the detection limit for oxygen was quite
high. A complete removal of oxygen out of the sampling
system would have taken impractical long pumping intervals,
Environ Sci Pollut Res
Fig. 4 Plots of a TOC, b DOC, c
FeTOT and d pore-water
conductivity for the three
sampling campaigns (blue, first;
red, second and green, third
campaign) in the four CMT wells.
The background values (ref.) are
indicated by the well in the
uncontaminated location
so samples were derived as soon as oxygen concentrations
were <1 mg/l. We assume the oxygen is only present in trace
amounts in the contaminated zone. Sulfate was below detection limit over the entire profile of the aquifer, compared to
background concentrations of about 30 mg/l (data not
shown). Only small amounts of sulfide were detected,
which cannot explain the missing sulfate (data not
shown). It seems that other reduced sulfur species,
which were not analysed, are more abundant, for example
HS− or H2S, which is to be expected at a circumneutral pH.
Overall, the changes in ionic composition of the groundwater in the most contaminated zone result in an increase of
the pore fluid electrical conductivity, as it is evident from
Fig. 4d.
The vertical distribution of oxalate dissolvable iron, which
corresponds to amorphous and poorly crystallized oxides
(Schwertmann 1964), is shown in Fig. 6 (left). The dithionite
dissolvable fraction, which represents the sum of all iron
oxides, differs hardly from the oxalate dissolvable fraction in
greater depths below 6 m (data not shown). An estimate of
magnetite content was obtained from empirical relations (e.g.
Mullins 1977) on the basis of magnetic susceptibility measurements. A susceptibility value of ca. 100×10−6 cm3/g
roughly corresponds to 0.2 mg/g of magnetite content. A peak
Environ Sci Pollut Res
Fig. 5 TC, TOC and TPH
concentration on soil samples
collected during the drilling
operations for boreholes B-S5
(left) and B-S7 (right)
in magnetic susceptibility, around 130×10−6 cm3/g, is observed in borehole B-S7 at the depth of 8 m but apart from
that only in depths above −3 m such large values were
recorded. In all other depths, the magnetic susceptibility is
usually less than 50×10−6 cm3/g.
Fig. 6 Concentration of iron
oxides (left) and magnetic
susceptibility (right) on soil
samples from boreholes B-S5 and
B-S7
Microbial counts and estimation of catabolic enzymatic
activity
Results of microbial counts are presented for the sampling
position S1 and S4, respectively located outside the
Environ Sci Pollut Res
Fig. 7 Colony-forming units for aerobic microbes, measured at different depths in the sampling point S1 (left) and S4 (right)
contaminated area and in the vicinity of the test site used for
geophysical and geochemical monitoring. Figures 7 and 8
show the microbial counts for the aerobic and anaerobic
microbes, respectively. Regarding the aerobic microbes, the
abundance was found to be rather variable as a function of
different soil layers. In the non-contaminated samples, the
highest amount of bacterial colonies was found in the sample
originating from the 4.3 m b.g.s. layer (Fig. 7, left).
Fig. 8 Colony-forming units for anaerobic microbes, measured at different depths in the sampling point S1 (left) and S4 (right)
Environ Sci Pollut Res
Fig. 9 Fluorescein diacetate analysis, measured at different depths in the sampling point S1 (left) and S4 (right)
Approximately the same quantity was estimated in the borehole S4 at depth of 6.1 m (Fig. 7, right), while lower counts
were estimated in the saturated zone, where lower oxygen
concentration was measured. The anaerobic microbes were
discovered to be more abundant in the contaminated locations
(here presented for borehole S4, in Fig. 8, right) than in the
control non-contaminated samples (Fig. 8, left). In particular,
higher counts of anaerobic microbes were obtained within the
soil samples at depths around 6 and 11 m b.g.s..
Results of FDA hydrolysis are presented in Fig. 9. It is clear
how a greater amount of fluorescein dye was estimated in the
contaminated samples from borehole S4 (Fig. 9, right), compared to samples from borehole S1 (Fig. 9, left). No great
differences were discovered in the amount of fluorescein
produced in S4, within the depth range between 6.1 and
11 m b.g.s.
Geoelectrical characterization
Results of 1D apparent resistivity and IP logs are shown for
the borehole B-S3 (Fig. 10a, b) and B-S4 (Fig. 10c, d),
respectively.
The electrical behaviour in the two boreholes, in particular
the IP response, is dramatically different.
The resistivity profiles point out four main layers:
–
a conductive superficial layer at the top (around 50–
100 Ω m), corresponding to a silty layer (0–≈3 m);
–
–
–
a medium resistive layer (several hundred ohms metre,
with some peaks to 1,000 Ω m in B-S3; ≈3–≈8 m) above
the piezometric surface, which presents oscillations due
to various fluid saturation, and probable hydrocarbon
smearing;
a conductive layer below the water table (decreasing from
several hundred to 100–200 Ω m; ≈8–≈12 m), corresponding to the contaminated saturated zone.
a highly conductive layer corresponding to the slightlycontaminated saturated zone (100 Ω m; ≈12–16 m)
IP logs show a complex behaviour, in particular along the
borehole B-S3. The IP data appear to be consistent with the
oscillations of the water table, with a peak (up to 100 mV/V)
detected 1–2 m below the piezometric level.
The main significant apparent IP anomaly is in the depth
range between 8 and 12 m b.g.s. It is interesting to notice the
accordance between the position of the IP peak and the water
table level. On the contrary, very low IP values, mostly less
than 5 mV/V, were found in B-S4.
We observe from complex resistivity measurements a good
agreement between apparent chargeability and phase values.
We apply the method proposed by Kemna et al. (1997) to
convert time domain into frequency domain data for vertical
logs in borehole B-S3; an example on dataset acquired
in February 2011 is shown in Fig. 11. A log of phase
values obtained by linear conversion from DC apparent
chargeability data measured with a DC georesistivimeter
Environ Sci Pollut Res
Fig. 10 Logs of apparent
electrical resistivity and apparent
chargeability from borehole B-S3,
(a) and (b), and from borehole BS4, (c) and (d), respectively.
Colored bars on the right side of
a and c indicate water table level
at each time-step
(IRIS Syscal Pro) is compared to logs of apparent phase angle,
measured in the frequency range between 0.5 and 100 Hz with
PASI Polares.
Results of inversion of 2D resistivity data shown in Figs. 12
and 13 point out a lateral variability in both amplitude and
phase of the complex resistivity, both for the October 2010 data
Environ Sci Pollut Res
Fig. 11 Comparative plot of
vertical phase logs (data of
February 2011) along B-S3, and
phase log from time domain
chargeability data (TDIP),
obtained by using the relation
proposed by Kemna et al. (1997)
and for the May 2011 data. The inversion process of the 2D data
greatly reduces the overall IP values, and the anomalies in B-S3
are in the order of 10-20 mradians, in the depth range between 6
to 10 m. The resulting map of complex resistivity for the dataset
of May 2011 shows a relatively more prominent phase response,
with a phase anomaly extended at a depth of 12 m b.g.s..
Discussion
The results of the integrated characterization at the Trecate test
site indicate that a strong contamination at the site and residual
NAPL is present mostly in the aquifer affected by the seasonal
water table fluctuations (from 6 to 11 m b.g.s.). An oil emulsion, measured as difference between TOC and DOC, is
present within a more confined region between 8.5 and
10 m. Remarkably, 10 m is also the depth of the groundwater
level at the beginning of February, when the blow out occurred in 1994. Despite the groundwater fluctuations, the
measured concentrations are rather time invariant at the three
sampling dates. All arguments hint on a strong retention of
NAPL in the contaminated zone, even emulsified NAPL
remains in place. The buoyancy of the droplets is not strong
enough that all oil floats on top and the horizontal hydrodynamic forces due to groundwater flow are not large enough for
a down gradient transport. Potentially, horizontal groundwater
flow in the smear zone is restricted by blockage of pore throats
in the smear zone. Blockage could be an effect of the increased
biomass and/or NAPL filled pore throats. The parameters
indicate also a horizontal heterogeneity of the contamination.
However, the horizontal coverage of data is too sparse to allow
further conclusions. Microbial degradation of hydrocarbons
has multiple effects on the hydrochemistry. Hydrocarbons are
degraded and converted to CO2 or metabolic byproducts.
Dissolution of CO2 results in increase of electrical conductivity of the pore water and decrease of the pH. Microbial
degradation of crude oil hydrocarbons requires terminal electron acceptors, which are used preferentially depending on the
energy gain of the reaction. First, oxygen and nitrate are
depleted. Then, manganese and iron oxides are reduced,
Environ Sci Pollut Res
Fig. 12 October 2010. Module
(left) and phase (right) of the
complex resistivity (at 1.79 Hz)
distribution at Trecate site. The
white line indicates the level of
the piezometric surface
which results in the production of Mn2+ and Fe2+ ions. The
next lower energy gain is derived through the reduction of
sulfate to sulfur or hydrogen sulfide (Christensen et al. 2000).
The presence of dissolved Mn and Fe indicates manganese
and iron oxide-reducing conditions. Because sulfate is missing, it must have been used as electron acceptor also. In
summary, the conditions can be characterized as ferrogenic
on the limit to sulfidogenic.
This is supported by the anaerobic CFU counts in the
contaminated area (Fig. 8), which are more than 2 orders of
magnitude greater than the anaerobic CFUs in the
uncontaminated control location. In contrast, no considerable
differences on aerobic microbial CFUs can be found in the
contaminated region versus uncontaminated location (Fig. 7).
The results of FDA hydrolysis, presented in Fig. 9, are important to check whether the soil condition are favourable or not
for a conspicuous bacterial growth. The amount of fluorescein
produced in the analysed samples, which is related to the
presence of additional energy sources (in this case, the contaminant), is 10 times greater in the deeper samples of S4 (in
the contaminated region). The depth range between 6 and
11 m corresponds in fact to the hydrocarbon-smearing zone.
It is interesting to note that sulfate is missing in all sampled
depth of the profile, while Mn and Fe only occur in the most
contaminated zone. The reason could be that sulfate is reduced
upstream of the investigated area, which is located in the
centre of the contaminated zone, and transverse mixing results
in a homogeneous distribution of the depleted groundwater
over the whole water column. In contrast, Mn and Fe are less
mobile due to precipitation outside the iron and manganese
reduction zone or due to the precipitation of iron and manganese sulfides (Baedecker et al. 1993; Bennett et al. 1993).
Consequently, they are not affected by the transverse mixing
as much. This finding contradicts the traditional assumption of
the geometry of redox zonation (Lyngkilde and Christensen
1992), where zones of low redox potential, as indicated by the
redox sensitive species, should be surrounded by zones of
higher redox potential. Instead, our data supports growing
evidence that reaction and transport characteristics of the
individual species and site heterogeneities are important for
the characterization of plumes at aged NAPL sites (Wilson
et al. 2004).
As far as the geophysical results are concerned, we attribute
the low resistivity of the top 3 m to the high silt and clay
content, resulting in a larger water-holding capacity than the
layers below. The high resistivity in the second layer
(≈3–≈8 m) is most likely due to the low water content of the
unsaturated coarse material. Detection of contaminated horizons in the unsaturated zone is complicated, since the NAPL
is present as a residual phase and adds to the bulk resistivity
Environ Sci Pollut Res
Fig. 13 May 2011. Module
(right) and phase (right) of the
complex resistivity (at 1.79 Hz)
distribution at Trecate site. The
white line indicates the level of
the piezometric surface
increase due to the unsaturated conditions. We do not assume
that the resistivity changes at depths >3 m can be attributed to
texture changes. Although sand and gravel alternate strongly
below the depth of 3 m, this alternation is not reflected in the
resistivity profile. The reason for this is that division into the
two fractions, sand (<2 mm) and gravel (>2 mm) probably
exaggerates the differences in texture and pore size distribution between layers. The strong NAPL contamination in the
third layer (≈8–≈12 m) below the water level can contribute to
partially increased resistivity, reducing the contrast between
saturated and unsaturated zone. The beforehand discussed
blockage of pore throats seems to contribute to an increase
of the resistivity in the contaminated zone. This effect seems
to outbalance the conductivity increase of the pore water due
to the microbial activity in this layer. Consequently in the
fourth layer ≈12–16 m, a decrease of resistivity is observed
due to the absence of NAPL.
Some of the time series show considerable IP values above
the water table. Because of capacitive effects from bad
electrode-ground contact, anomalous polarization responses
can be originated, and we believe they represent an
overestimation of the real polarizability of the soil at those
depths. In addition, inductive coupling effects from the measurement system have to be considered into the interpretation.
This is clearer for IP logs whose measures were carried out in
frequency domain, where phase values are affected by inductive coupling effects, which increase with frequency. IP response could also be related to the presence of dispersed
ferromagnetic minerals, which were found in slightly larger
abundances in B-S7 at a depth of 8 m and near the ground
surface. However, the IP response does not appear to be
sensitive to these small amounts, otherwise a clear relationship
between magnetic susceptibility and IP response should be
visible. This is not the case. Instead, we suggest that the deeper
IP anomalies are prevalently associated to degradation byproducts and/or microbial colonization. It is interesting to note
that the groundwater level is almost the same for the two
campaigns, but the global electrical properties in the section
is different between the two time series. This is mainly due to
the fact that after a maximum depth of 11 m b.g.s. reached at
the end of February 2011, the piezometric level is being forced
to raise as a consequence of the abundant amount of water
used for irrigation of the area. This huge rise produces a
smearing of NAPL free phase, and it is responsible for the
isolated blob of trapped LNAPL (e.g. Sauck 2000; Atekwana
and Atekwana 2010). This, in addition to the recharging fresh
water from irrigation, can contribute to enhance biodegradation processes of the available organic matter (i.e.
Environ Sci Pollut Res
the LNAPL) and therefore a different IP response of the two
dates. Similar results can be found in Flores Orozco et al.
(2012), who found good agreement between complex resistivity anomalies and hydrocarbon (BTEX) contamination.
They derived a qualitative relationship between spectral IP
parameters and BTEX contamination, with an increase of
polarization effects for large BTEX concentrations but a decrease in phase values in presence of free phase products. This
can explain the variability of phase anomalies encountered at
Trecate site, which is likely to be related to the variability of
hydrocarbon distribution and microbial activity.
electrical measurements used for NAPL site characterization
and monitoring. The conclusive statement based on the results
presented in this paper is that changes in fluid composition,
due to the high amount of free phase NAPL (smeared by the
huge water level fluctuations), microbial degradation and
growth and local heterogeneities represent the main factors
influencing the complex geoelectrical response.
Acknowledgments This study was supported by the European Commission (SOILCAM Project, contract number 212663, www.soilcam.eu)
and the DFG (WE 4979/1-1). We would like to thank Helen French, Kai
Uwe Totsche and Lee Slater for valuable advice and Sarah Heck, Christian Egel and Diego Franco for field assistance.
Conclusions
The integrated characterization of the hydrocarboncontaminated site of Trecate permitted to achieve a complete
knowledge of the various and complex phenomena undergoing at the site. The crude oil contamination at the site is spread
over the groundwater fluctuation zone and large NAPL concentrations are present in the groundwater as an emulsion. The
contamination appears fairly time invariant due to trapping of
NAPL. NAPL products act as additional (continuous) source
for biological degradation. From a chemical point of view, this
has been shown through the presence or absence of the different redox-sensitive species in an anoxic and reductive
environment. The geometry of the redox zonation at the site
supports the hypothesis that the traditional concept of plume
development, where zones of low redox potential are
surrounded by zones of high redox potential, may be an
over-simplified approach. The results of micro-biological laboratory assays support the geochemical findings with respect
to hydrocarbon degradation, indicating that the smearing zone
is the richest zone of the anaerobic bacteria population. Both
CFUs and enzymatic activity tests pointed out that there is an
enhanced abundance and activity of microbial communities
and that the smearing zone is rich of additional organic matter
for the bacteria to proliferate.
The results of the geophysical surveys (both ERT and
vertical resistivity and IP logs) pointed out an anomalous
response in the upper saturated zone, in the depth range
between 8 and 11 m b.g.s. It is unlikely that this observation
can be attributed to textural differences because these do not
appear to be related to the geoelectrical anomalies. Therefore,
resistivity and phase anomalies are attributed to the presence
of the NAPL contamination. However, the analysis of both
resistivity and IP data and geochemical data partly brings
counter-intuitive observations. The presence of NAPL seems
to result in a resistivity increase, which outbalances an increase in ion concentration in the pore water caused by microbial activity. The phase response hints on a large spatial
heterogeneity, which is reflected in the geochemical data.
Such effects need to be taken into account when complex
References
Abdel Aal G, Atekwana EA, Slater LD (2004) Effects of microbial
processes on electrolytic and interfacial electrical properties of unconsolidated sediments. Geophys Res Lett 31, L12505
Abdel Aal GZ, Slater LD, Atekwana EA (2006) Induced-polarization
measurements on unconsolidated sediments from a site of active
hydrocarbon biodegradation. Geophysics 71:H13–H24
Archie GE (1942) The electrical resistivity log as an aid in determining
some reservoir characteristics. Trans Am Inst Min Metall Pet Eng
146:54–62
Atekwana EA, Atekwana EA (2010) Geophysical signatures of microbial
activity at hydrocarbon contaminated sites: a review. Surv Geophys
31:247–283
Atekwana EA, Sauck WA, Werkema DD (2000) Investigations of
geoelectrical signatures at a hydrocarbon contaminated site. J Appl
Geophys 44:167–180
Atekwana EA, Sauck WA, Abdel Aal GZ, Werkema DD (2002) Geophysical investigation of vadose zone conductivity anomalies at a
hydrocarbon contaminated site: implications for the assessment of
intrinsic bioremediation. J Environ Eng Geophys 7:103–110
Atekwana EA, Atekwana E, Rowe RS, Werkema DD, Legall FD (2004)
The relationship of total dissolved solids measurements to bulk
electrical conductivity in an aquifer contaminated with hydrocarbon.
J Appl Geophys 56:281–294
Baedecker MJ et al (1993) Crude-oil in a shallow sand and gravel
aquifer.III. Biogeochemical reactions and mass-balance modeling
in anoxic groundwater. Appl Geochem 8(6):569–586
Bennett PC et al (1993) Crude-oil in a shallow sand and gravel aquifer. 1.
Hydrogeology and inorganic geochemistry. Appl Geochem 8(6):
529–549
Benson AK, Payne KL, Stubben MA (1997) Mapping groundwater contamination using dc resistivity and VLF geophysical methods—a case
study. Geophysics 62:80–86
Biró B, Domonkos M, Kiss E (2012) Catabolic FDA microbiological
activity as site-dependent monitoring tool in soils of an industrial
town. Int Rev Appl Sci Eng 3(1):41–46. doi:10.1556/IRASE.3.
2012.1.5
Biró B, Horváth N, Matics H, Domonkos M, Malov X (2013) Enhanced
degradation of deicing fluids in soils and soil-plant systems by improving soil nutrient status and quality. Növénytermelés (Plantbreeding), 62:
393–396. DOI:10.12666/Novenyterm.62.2013.suppl
Börner FD, Schopper JR, Weller A (1996) Evaluation of transport and
storage properties in the soil and groundwater zone from induced
polarisation measurements. Geophys Prospect 44:583–601
Borsic A, Adler A (2012) A primal dual - interior point framework for
using the L1-norm or the L2-norm on the data and regularization
Environ Sci Pollut Res
terms of inverse problems: inverse problems, 28, no. 9, 095011,
DOI:10.1088/0266-5611/28/9/095011
Christensen TH, Bjerg PL, Banwart SA, Jacobsen R, Heron G, Albrechtsen
HJ (2000) Characterization of redox conditions in groundwater contaminant plumes. J Contam Hydrol 45(3–4):165–241
DeRyck SM, Redman JD, Annan AP (1993) Geophysical monitoring of
controlled kerosene spill. In: Proceedings of the symposium on the
application of geophysics to engineering and environmental problems (SAGEEP), San Diego, pp 5–19
Flores Orozco A, Kemna A, Oberdörster C, Zschornack L, Leven C,
Dietrich P, Weiss H (2012) Delineation of subsurface hydrocarbon
contamination at a former hydrogenation plant using spectral induced polarization imaging. J Contam Hydrol 136–137:131–144
Frohlich RK, Barosh PJ, Boving T (2008) Investigating changes of
electrical characteristics of the saturated zone affected by hazardous
organic waste. J Appl Geophys 64:25–36
Godio A, Arato A, Stocco S (2010) Geophysical characterization of a
nonaqueous-phase liquid–contaminated site. Environ Geosci 17(4):
141–161
Kemna A (2000) Tomographic inversion of complex Resistivity. Theory
and application. PhD Thesis. Der Andere Verlag
Kemna A, Räkers E, Binley AM (1997) Application of complex resistivity tomography to field data from a kerosene-contaminated site.
Proc. 3rd Mtg. Environmental and Engineering Geophysics, Environ. Eng. Geophys. Soc., Eur. Section, 151–154
LaBrecque DJ, Miletto M, Daily W, Ramirez A, Owen E (1996) The
effect of noise on Occam’s inversion of resistivity tomography data.
Geophysics 61:538–548
Lucius J, Olhoeft GR, Hill PL, Duke SK (1992) Properties and hazards of
108 selected substances, 1992 edn. United Staes Geological Survey
Open File Report, 92–527, 560 pp
Lyngkilde J, Christensen TH (1992) Redox zones of a landfill leachate
pollution plume (Vejen, Denmark). J Contam Hydrol 10(4):273–289
Marcak H, Golebiowski T (2008) Changes of GPR spectra due to the
presence of hydrocarbon contamination in the ground. Acta
Geophys 56:485–504
Mazác O, Benes L, Landa I, Maskova A (1990) Determination of
the extent of oil contamination in groundwater by
geoelectrical methods. In: Ward, S. H. (ed.) Geotechnical
and environmental geophysics II: Society of Exploration
Geophysicists, 107–112
Mullins CE (1977) Magnetic susceptibility of the soil and its
significance in soil science - a review. J Soil Science 28, 2,
223–246
Olhoeft GR (1985) Low frequency electrical properties. Geophysics 50:
2492–2503
Sauck WA (2000) A model for the resistivity structure of LNAPL
plumes and their environs in sandy sedments. J Appl
Geophys 44:151–165
Schwertmann U (1964) Differenzierung der Eisenoxide des
Bodens durch Extraktion mit Ammoniumoxalat-Lösung.
Zeitschrift Fur Pflanzenernahrung Und Bodenkunde 105:
194–202
Seigel HO (1959) Mathematical formulation and type curves for induced
polarization. Geophysics 24:547–565
Tezkan B, Georgescu P, Fauzi U (2005) A radiomagnetotelluric survey on
an oil-contaminated area near the Brazi Refinery, Romania.
Geophys Prospect 53:311–323
Villanyi I, Füzy A, Angerer I, Biró B (2006) Total catabolic
enzyme activity of microbial communities. Fluorescein
diacetate analysis (FDA). In: Understanding and modelling
plant-soil interactions in the rhizosphere environment. Handbook of methods used in rhizosphere research. Chapter 4.3.
Microbiology, Biochemistry and Molecular Biology. p. 441–
442
Werkema DD, Atekwana EA, Endres AL, Sauck WA, Cassidy DP (2003)
Investigating the geoelectrical response of hydrocarbon contamination undergoing biodegradation. Geophys Res Lett 30:1647. doi:10.
1029/2003GL017346
Wilson RD, Thornton SF, Mackay DM (2004) Challenges in monitoring
the natural attenuation of spatially variable plumes. Biodegradation
15(6):359–369