Journal of Contaminant Hydrology 82 (2006) 183 194

www.elsevier.com/locate/jconhyd

Transport and degradation of ethanol in groundwater

Yi Zhang a, Imtiyaz A. Khan a, Xun-Hong Chen b, Roy F. Spalding a,*
a
Department of Agronomy and Horticulture, Keim Hall, University of Nebraska-Lincoln, Lincoln, NE 68583, USA
b
School of Natural Resources, Nebraska Hall, University of Nebraska-Lincoln, Lincoln, NE 68588, USA
Received 2 June 2005; received in revised form 8 September 2005; accepted 19 September 2005
Available online 5 December 2005

Abstract

Ethanol is rapidly replacing methyl tert-butyl ether (MtBE), the primary fuel oxygenate in the US, and
ethanol releases from spills and leaky underground storage tanks (LUSTs) are anticipated. Ethanol has
received little attention as a potential groundwater contaminant. This study investigates the fate and
transport of ethanol under transient conditions in a sand and gravel aquifer. A pulse containing
approximately 220 mg L 1 ethanol and 16 mg L 1 bromide was injected into the shallow sand and gravel
aquifer and monitored to estimate its persistence and transport. The plume was monitored for 2.5 months
using downgradient multilevel samplers (MLSs). Values for ethanol retardation were measured from
ethanol and bromide breakthrough data and compared to estimates using published K oc values for low
carbon aquifer sediments ( f oc = 10 Ag C g 1 sediment). Ethanol transport was not retarded (R = 0.99). A 3-
dimensional model reasonably simulated bromide and ethanol breakthrough curves. An average first-order
decay constant was estimated to be 0.32 d 1 (t 1 / 2 = 2.2 d). At the second fence, 75% of the injected
bromide and less than 3% of ethanol remained in the plume. Monitored terminal electron acceptor
concentrations demonstrated that the majority of the ethanol was transformed by anaerobic processes other
than denitrification and sulfate reduction.
D 2005 Elsevier B.V. All rights reserved.

Keywords: Solute transport; Ethanol degradation rate; Dispersion; Half life; Nitrate reduction

1. Introduction

Since 2000 when the US EPA announced actions to increase ethanol and reduce methyl
tertiary butyl ether (MtBE) as a fuel oxygenate, ethanol output in the USA has doubled. In that
period 11 new ethanol plants have been constructed in Nebraska, and 7 more should be operating

* Corresponding author. Tel.: +1 402 472 8214; fax: +1 402 472 2906.
E-mail address: rspalding1@unl.edu (R.F. Spalding).

0169-7722/$ - see front matter D 2005 Elsevier B.V. All rights reserved.
doi:10.1016/j.jconhyd.2005.09.007
184 Y. Zhang et al. / Journal of Contaminant Hydrology 82 (2006) 183194

in 2005. Approximately 9.4 billion liters of ethanol were produced in Nebraska in 2003, which
utilized about 30% of the corn crop (NEB, 2004). The predicted increase in demand is not only
based on expanded use as a fuel-oxygenate but also on use as a fuel to lessen dependence on
Mideast petroleum. An added incentive is the fate of these compounds in groundwater. MtBE
degradation is slow and the transformation products are potentially more toxic than MtBE
(Schmidt et al., 2004). In laboratory studies ethanol was rapidly mineralized to nontoxic
products in both aerobic and anaerobic conditions (Powers et al., 2001).
Since 1990 ethanol releases at fermentation plants, distribution facilities, and from both truck
and rail tankers in Nebraska have ranged from a few to 30,000 gal (NDEQ, 2005). Ethanol diesel
and ethanol gasoline releases also have also been documented. Gasohol releases from leaky
underground storage tanks (LUSTs) have been reported in Kansas groundwater (Ruiz-Aguilar
et al., 2003). Unlike the adverse groundwater impacts that Lahvis and Rehmann (2000) reported
for small liquid and vapor releases of MtBE, Dakhel et al. (2003) have shown in field
experiments that small volume ethanol releases are not expected to impact groundwater unless
the releases occur in very shallow water table areas (b 100 cm to ground water). However, the
fate of larger releases reaching the groundwater has not been studied.
Groundwater concentrations of ethanol typically are not reported at LUST sites because
ethanol is an unregulated contaminant, has low toxicity and persistence and is difficult to
measure in trace amounts. Hunt et al. (1997) and Corseuil et al. (1998) reported that ethanol in
petroleum spills may consume available dissolved oxygen needed for aerobic biodegradation of
BTEX (benzene, toluene, ethylbenzene, and xylenes) and could initially limit BTEX
biodegradation and elongate BTEX plumes. A field study in Brazil indicated that high ethanol
concentrations in groundwater (greater than 2%) enhanced the dissolution and transport of
BTEX (Corseuil and Alvarez, 1996). Model simulations of gasohol plumes have produced
contradictory results with one suggesting that biomass clogging will result in slower longitudinal
and increased transverse transport of BTEX in groundwater (Kildsgaard and Engesgaard, 2002)
while another suggests increased longitudinal transport of BTEX due to extended persistence in
anaerobic plumes (Barker et al., 1991).
Our objectives were to evaluate the fate and transport of ethanol under transient conditions in
a sand and gravel aquifer and to simulate the results with a commonly used 3-dimensional (3-D)
solute transport model.

2. Materials and methods

2.1. Site description

The field site is located in Merrick County, in the central Platte Valley of Nebraska. It is
surrounded by furrow-irrigated farmland used for corn production and previously was used for
an amended denitrification study using ethanol or acetate as carbon sources. The unconfined
primary aquifer is a 12.2 m thick fluvial sand and gravel. The water table is about 3.1 m
below the land surface. The groundwater pH is 6.5, and its temperature ranges from 11 8C in
winter to 14 8C in summer. The primary aquifer has vertically uniform concentrations of
~40 mg NO3N L 1, ~60 mg SO42 L 1 and ~6 mg L 1dissolved oxygen. Background
concentration of bromide and ethanol were 0.30.5 mg L 1 and less than the method
reporting limit (MRL) of 15 Ag L 1, respectively. Microcosm studies at the site indicated that
the aquifer sediments contain indigenous microbes capable of degrading ethanol (Bates and
Spalding, 1998).
 Y. Zhang et al. / Journal of Contaminant Hydrology 82 (2006) 183194 185

Fig. 1. Plan view of ethanol study site.

The experimental design consists of a 10 cm diameter injection well (IW) located 12.2 m
from a 15 cm diameter extraction well (Fig. 1). The injection well and the extraction well are
screened in the bottom quarter of the aquifer. Five multilevel samplers (MLSs) comprise the first
fence downgradient from the injection well and three MLSs form the second fence. ML-6 and
ML-1 are located 3.1 and 9.1 m, respectively, downgradient of the injection well and are in the
regional groundwater flow path. ML-5 and ML-7 are northeast and southeast, respectively, of the
IW at a radial distance of 5.5 m. Each MLSs is constructed of 0.94 cm high density polyethylene
(HDPE) tubing with a stainless steel screens held in place with a HDPE ferrule fastened to a
piezometer with plastic bundle ties at appropriate intervals. The piezometer is 2.5 cm diameter
schedule 40 PVC pipe with a 0.6 m slotted interval capped at the bottom. The MLSs facilitate
the collection and analysis of spatial and temporal data that characterize the fate and transport of
the solutes.

2.2. Spiking solution

A 605 L concentrated spiking solution was prepared by adding 15.64 L of neat ethanol
(absolute 200 proof, McCormick Corp., Kansas, USA) and 0.68 kg of bromide (Great Lakes
Chemical Corp., West Lafayette, Indiana, USA) to a utility tank containing extracted
groundwater. During 27 h of injection, the spiking solution was fed through a diaphragm
pump (Cole-Parmer, Vernon Hills, IL, USA) at a flow rate of 0.02 m3 h 1 and mixed with water
extracted at a rate of 1.85 m3 h 1 from a nearby 35.3 m deep well which tapped the secondary
186 Y. Zhang et al. / Journal of Contaminant Hydrology 82 (2006) 183194

aquifer. Water from the deep well was anaerobic and contained no detectable nitrate, 20 mg
SO4 2 L 1 and 340 mg L 1 as CaCO3 alkalinity. The resulting injectate contained 220 F 7.4 mg
ethanol L 1 and 16 F 1.8 mg bromide L 1.

2.3. In situ injection

The injection well was equipped with a 8.9 cm diameter BaskiR inflatable packer (Baski Inc.,
Denver, CO, USA) which was set at 11.6 m below the land surface (bls) to pressure inject
(2.1  105 Pa) 50,000 L ethanol/bromide solution into the bottom 3.1 m (11.614.6 m bls) of the
primary aquifer. The pressure injection forced the injectate to move immediately into the aquifer
formation through the 3.1 m section of well screen. A hydraulic gradient to the extraction well
was established by pumping the extraction well at the rate of 9.1 m3 h 1 for 2 days prior to the
injection. During the 27 h injection the extraction well pumped at a rate of 4.5 m3 h 1 to initially
direct plume towards the multilevel fence (458 eastnortheast). The extraction well was shut off
at the conclusion of the injection. Initial pumping of the extraction well combined with
pressurized injection created a hydraulic condition in which the plume was transported under
forced gradient flow; however, this gradient dissipated quickly when injection and extraction
ceased and natural flow conditions resumed.

2.4. Sampling and analysis

Groundwater samples were collected from four depths (10.6, 12.2, 13.4, and 14.6 m) at five
downgradient MLSs using a peristaltic pump (Cole-Parmer, Niles, IL, USA). Three tube
volumes were purged prior to sampling. The plume was monitored for 2.5 months. Samples
were collected daily for the first 3 weeks and twice a week thereafter. Samples were chilled with
ice in a cooler and transported to the laboratory where they were stored at 4 8C until analysis.
Ethanol samples were collected in 40 ml VOC vials and preserved by the addition of 4 drops
of concentrated H2SO4 (pH b 2) and analyzed by solid-phase microextraction and gas
chromatography/mass spectrometry (SPME-GC/MS). The method was developed by Cassada
et al. (2000) to measure trace levels of ethanol and has a MRL of 15 Ag L 1. Briefly, the aqueous
extraction is enhanced by the addition of ~25% NaCl; the ethanol is sorbed to a preconditioned 2
cm 50/30 Am DVB/Carboxen/PDMS fiber; desorbed at 220 8C in the GC injection port;
separated; on a fused-silica capillary column and quantified with an HP 5970 mass spectrometer.
Samples collected for bromide, nitrite, nitrate, and sulfate analysis were collected in 40 mL
polyethylene bottles, filtered through 0.2-Am Whatman membrane filters and analyzed by ion-
chromatography (IC) using Dionex DX-100 (Dionex Corporation, Sunnyvale, CA, USA) with a
50 AL sample loop, ASRS-11 self-regenerant suppressor, Ion-Pac AS14 column, and AG14
guard column (APHA, 1998). The chromatograms were processed using a Dionex 4400
integrator. Method detection limits for anions were 0.1 mg L 1, and the analytical precision
was ~F5%.

2.5. Transport model

MT3D (Zheng, 1990) was used in conjunction with MODFLOW (McDonald and Harbaugh,
1988) to simulate the transport of the bromide/ethanol plume under transient conditions. MT3D/
MODFLOW is user friendly and provides the needed flexibility for three-dimensional transport
simulation in a groundwater domain where both pumping and injection occur. Groundwater head
 Y. Zhang et al. / Journal of Contaminant Hydrology 82 (2006) 183194 187

data generated by MODFLOW were used to compute groundwater flow velocities which then
were used in MT3D to compute contaminant transport and degradation.
The 150  150 m area was divided into 1.5  1.5 m grids. The aquifer depth was divided into
two layers to account for partial penetration of the injection well. The top layer extends 9.1 m
below the water table, while the bottom layer containing the wellTs screened interval is 3.1 m
thick. Constant heads were assigned to the upgradient (515.2 m) and downgradient (515.9 m)
boundaries of the model area. The top and bottom domain boundaries were oriented parallel to
the direction of regional groundwater flow. Given the small volume of solution/water injected
into and extracted from the aquifer, the boundaries had no effect on the flow condition in the
central part of the domain where the injection and extraction wells were placed.
In the transient model 220 mg L 1 of ethanol and 16 mg L 1 of bromide were injected at a
rate of 45 m3 d 1 for 27 h in the bottom layer of the aquifer. Beneath the site, the natural
groundwater flow direction is 108 eastnortheast. The regional hydraulic gradient is 0.0012 m/m
(Sniegoeki, 1955), and the average horizontal hydraulic conductivity (K h) determined from the
pump test is about 100 m d 1. A 130 m d 1 K h was derived from a pump test at the land
owners irrigation well using an effective porosity of 0.3 (Zlotnik et al., 1993). Chen et al. (2003)
have assigned 0.2 as the specific yield for sand and gravel aquifers in the region. The
longitudinal and transverse dispersivities (geometric properties of a porous medium which
determine the dispersion characteristics of the medium) and the ethanol decay constant (k) were
obtained by goodness of fit of the simulated to the observed breakthrough responses. The
dipersivity values of 0.05, 0.01, and 0.01 m were assigned to the x, y and z directions,
respectively. Sorption was not considered in the model because ethanol has a very low log K oc of
0.71 (Moyer, 2003) and the sands and gravels of the central Platte Valley are considered low
carbon aquifers with f oc b 0.01% carbon.
In low and medium carbon aquifers (Wilson, 2003), the calculated retardation factor R for
ethanol using the equation: R = 1 + q bk d / h, where q b=bulk density, h=effective porosity and
k d=soil distribution coefficient , is very close to 1. Site-measured R values from observed
breakthrough curves, where R is the velocity a conservative (bromide) divided by the velocity of
the contaminant of interest (ethanol) were also about 1. Velocities to each MLS were estimated

Fig. 2. Observed versus simulated bromide breakthrough responses at the 12.2 m depth at the first fence of MLSs.
Simulations were performed with K h = 100 m d 1.
188 Y. Zhang et al. / Journal of Contaminant Hydrology 82 (2006) 183194

Fig. 3. Observed bromide breakthrough responses at the second fence suggesting transverse and vertical aquifer
heterogeneity.

for bromide and ethanol using a central limit theorem; that is C / C o will be 0.5 whenever x = vt
where x is horizontal distance, v is velocity and t is time (Nielsen and Biggar, 1962).

3. Results and discussions

3.1. Breakthrough results

The longitudinal advancement of bromide based on observed breakthrough curves ranged

from 0.6 to 1.5 m d 1 with a mean of 1.0 m d 1 (Figs. 24). The variation in flow velocities

Fig. 4. Observed versus simulated bromide breakthrough responses at the 12.2 m depth at the second fence of MLSs.
Simulation was performed with K h = 100 m d 1.
 Y. Zhang et al. / Journal of Contaminant Hydrology 82 (2006) 183194 189

Fig. 5. Observed versus simulated ethanol concentrations at the first fence. Simulations were performed with K h = 100 m
d 1 and k = 0.32 d 1.

shown in the breakthrough curves between the first and second fences primarily resulted from
transient conditions associated with the initial pumping of the extraction well. Velocity
differences with depth at a multilevel sampler appear related to aquifer heterogeneity.
The breakthrough curves (Figs. 2 and 5) indicated that the transport times for ethanol and
bromide were essentially equivalent. The observed mean retardation value was 0.99 F 0.01.
The core of the bromide/ethanol plume was centered on an axis between inner fence ML-6
and ML-1 in the outer fence. Maximum injectate concentrations occurred at depths with the
greatest groundwater velocities. Breakthrough of bromide and ethanol appeared in MLSs 5 and 6
in 1.5 days, and both were no longer detected after day 16 (Figs. 2 and 5). Maximum bromide
and ethanol concentrations in ML-6 were 14.5 mg L 1 (C / C o for Br was 0.9) and 190 mg

Fig. 6. Observed versus simulated ethanol concentrations in the second fence. Simulation was performed with K h = 100 m
d 1 and k = 0.35 d 1.
190 Y. Zhang et al. / Journal of Contaminant Hydrology 82 (2006) 183194

L 1, respectively. At ML-1 in the second fence, the plume centroid arrived at a depth of 12.2 m
after 15 d. Here maximum bromide and ethanol concentrations were 12 mg L 1 (C / C o for Br
was 0.75) and 4.35 mg L 1, respectively (Figs. 4 and 6). Khan and Spalding (2003) previously
examined continuous sediment cores from the primary aquifer and noted the presence of vertical
variations. Preferential flow resulted in differently shaped bromide and ethanol breakthrough
curves and variable peak break-through transport times at the second MLS fence (Fig. 3). At the
13.4 m depth at ML-1 peak concentrations of 10.38 mg bromide L 1 and 6.56 mg ethanol L 1,
occurred after 10 d, while at the 10.7 m depth peak concentrations of 6.56 mg bromide L 1 and
1.81 mg ethanol L 1 occurred 8 d later.
Peak bromide concentrations occurred in ML-2 15 d after appearing at ML-1. Unlike ML-1,
ML-2 did not appear to be in as immediate hydraulic connection with transport from the
injection well. Bromide arrived on day 25 at the 12.2 m depth and on day 16 at a 13.4 m depth.
Peak ethanol concentrations at these depths were 4.28 and 1.2 mg L 1, respectively. Although
bromide persisted, ethanol was rapidly degraded throughout transport from the injection site and
about 97% of the initial ethanol was depleted when the plume left the second fence of MLSs.
Ethanol-C can be an electron donor in several microbiological processes (Chapelle, 1993),
and it can metabolized in several different pathways such as aerobic respiration, anaerobic
denitrification, sulfate reduction, and methanogenesis. McCarty et al. (1969) have shown that
reaction stoichiometry can be used to estimate carbon consumed in both respiration and cell
growth during deoxygenation and denitrification.
The relative-core bromide concentration (C / C o) at ML-6 was 0.9 and while at ML-1 farther
downgradient it was 0.75. The observed decrease was due to dispersive mixing of the injectate
with 10% ambient groundwater at the first fence and with 25% ambient groundwater at the
second fence. Mixing of injectate plume with the ambient groundwater containing 40 mg NO3
N L 1, 60 mg SO4 2 L 1, and 6 mg dissolved oxygen L 1 resulted in total deoxygenation,
denitrification and sulfate reduction within 20 days at ML-1 (Fig. 7). In order to estimate the
amount of ethanol metabolized in the three reduction processes concentrations of dissolved
oxygen, nitrate and sulfate in the plume were input into stoichiometric equations. Bromide

Fig. 7. Changes in bromide, nitrate, and sulfate concentration in the ethanol/bromide plume at 12.2 m at ML-1.
 Y. Zhang et al. / Journal of Contaminant Hydrology 82 (2006) 183194 191

concentrations at ML-6 suggest that that dispersion resulted in a ~10% reduction in ethanol
concentration in the plumes core. About 13% of the ethanol was consumed during
deoxygenation (0.2%), denitrification (8.3%) and sulfate reduction (4.5%). The remaining
ethanol (77%) appears to have been transformed in anaerobic processes such as fermentation,
methanogenesis, and cell growth. Powers et al. (2001) report that the anaerobic degradation of
ethanol to acetic acid (acetogenesis) and methanogenesis are the major processes for ethanol
degradation in anaerobic plume cores. After the ethanol was depleted nitrate and sulfate
concentrations rebounded to background levels in ML-1.
Khan and Spalding (2003) and Nuttall (2002) have shown that ethanol promotes growth of
microbes that secrete excessive biofouling and reduce aquifer permeability. Slow moving-broad
plumes would result from excessive biofouling in sand and gravel aquifers receiving continuous
ethanol releases (Kildsgaard and Engesgaard, 2002). From the site data it appears that the release
was not long enough or concentrated enough to cause significant clogging.

3.2. Simulation results

Breakthrough curves (Figs. 2 and 4) for bromide were simulated by using the longitudinal
dispersivity values from 0.02 m to 1.0 m and transverse and vertical dispersivity values equal to
one fifth of the longitudinal dispersivity. Simulations were calibrated with the field data until the
modeled-solute travel times were within 10% of observed times. Only the most direct transport
data from the injection well to the MLSs was used. A longitudinal dispersivity of 0.05 m,
transverse dispersivity of 0.01 m, and vertical dispersivity of 0.01 m provided a reasonable fit to
the data. Bromide solute tailing in the field data (Figs. 2 and 4) may be attributed to physical
nonequilibrium (solute exchange with immobile water) and/or differential biomass accumula-
tion. Simulations were based a uniform hydraulic conductivity and uniform dispersivities
throughout the domain. Therefore, the model lacked the ability to reproduce the demonstrated
tailing in the field data. Tailing appeared dependent on transport time and was considerably more
pronounced at the second fence. In order to more closely match the simulated bromide
concentrations with the field data, adjustments would be necessary for heterogeneity within the
3m thick injection zone. However, calibration with field data would be impossible since geologic
samples for K h estimates within the injection zone were not collected and would be extremely
difficult to collect.
Ethanol breakthrough curves (Figs. 5 and 6) were simulated by varying the ethanol k from
0.2 to 0.5 per day and were calibrated to fit ethanol breakthrough responses observed at various
depths at the first and second fences. The ethanol k varied vertically as well as transversely
and ranged from 0.25 to 0.4 d 1 and averaged 0.32 d 1. By assuming a k of 0.32 d 1 (half life,
t 1 / 2 = 2.2 d) and varying K h, a K h between 95 and 100 m d 1 produced minimum root mean
square errors of 12 and 0.9 mg l 1 at ML-6 and ML-1, respectively, when the simulated results
were compared to field data. This k is within the wide range of measured ethanol decay constants
(0.1 to 0.53 d 1) reported by Powers et al. (2001) for laboratory microcosm studies in which
they measured process specific ethanol consumption rates for deoxygenation, denitrification,
iron reduction, sulfate reduction and methanogenesis.
Sensitivity analysis is provided to evaluate model uncertainty associated with heterogeneity.
The sensitivity of the simulated ethanol k to change in dispersivity was assessed by simulating
the degradation constants for a set of varying longitudinal and transverse dispersivities (Table 1).
The k varied by a factor of approximately two and was more sensitive to changes in longitudinal
dispersivity (Table 1). Extreme cases of high and low longitudinal and transverse dispersivity
192 Y. Zhang et al. / Journal of Contaminant Hydrology 82 (2006) 183194

Table 1
1
Sensitivity analysis of ethanol degradation rate constant (d ) to varied dispersivity values
Longitudinal dispersivity (m) Transverse dispersivity (m)
0.002 0.01 0.05
1
Simulated degradation rate constant, k (d )
0.02 0.28 0.25 0.22
0.05 0.35 0.32 0.27
0.15 0.38 0.34 0.30
0.30 0.45 0.40 0.36

used in the sensitivity analysis that resulted in the two-fold increase in k should be viewed as a
worst case scenario. Stenback et al. (2004) reported a three-fold variation in the estimated
toluene k when both longitudinal and transverse dispersivity values were varied by one order of
magnitude.
The sensitivity of the simulated ethanol k to horizontal aquifer hydraulic conductivity (K h)
was assessed in a comparison of simulated ks using a low K h value of 25 m d 1 (fine sand) and a
higher K h value of 200 m d 1 (very course sand and gravel). Longitudinal and transverse
dispersivity values were held constant. Using the simulated k of 0.32 d 1 ethanol completely
degraded within 3 m from the source in the low K h zone; however, in the higher K h zone 25% of
ethanol plume was transported beyond the monitored area. In order to match the field ethanol
concentrations to those generated by the model, the k needed to be four times faster in the higher
K h zone than in the lower K h zone. The analysis indicated that the simulated fit with the
monitored ethanol concentrations was sensitive to both K h and k; however, variability in average
K h in the investigated zone was considerably less than illustrated in the above exercise and thus
would have less impact on the simulated decay constant. Although sensitivity analysis indicates
that a large uncertainty in the simulated ethanol ks can be introduced from using non-
representative values for K h, pump test data at the study site and a nearby suggest that the
average K h should not vary by more than F30 m d 1 from the selected value of 100 m d 1.
Analysis of sensitivity of the simulated ethanol k to changes in porosity indicated that a 5%
change in porosity did not significantly impact transport.

4. Conclusions

At the first and second fences relative bromide concentrations in the plume centroid were
90% and 75% of the injectate, respectively. Ethanol degraded rapidly with a first-order rate
constant of 0.32 d 1 (t 1 / 2 = 2.2 d). Most of the ethanol appeared to be transformed by anaerobic
processes such as fermentation and methanogenesis. Ethanol transport was not retarded (R = 1).
Field data and simulations indicated that ethanol did not migrate a substantial distance beyond
the injection well and 97% was biodegraded after one month. Simulated ethanol degradation
rates are dependent upon using representative hydraulic conductivity and longitudinal and
transverse dispersivity in the model.

Acknowledgements

This project was funded by the Nebraska Ethanol Board and the Central Platte Natural
Resources District. We thank David Cassada, and Jeff Toavs for their field and laboratory
 Y. Zhang et al. / Journal of Contaminant Hydrology 82 (2006) 183194 193

assistance. We also thank M. Exner for editing the manuscript. Mr. John Jefferson, a local
farmer, provided land for the project. This manuscript has been assigned Journal Ser. No.13841,
Agric. Res. Div., University of Nebraska.

References

American Public Health Association (APHA), 1998. Standard Methods for the Examination of Water and Wastewater,
20th edition. American Public Health Association, Washington, DC.
Barker, J.F., Gillham, R.W., Lemon, L., Mayfield, C.I., Poulsen, M., Sudicky, E.A., 1991. Chemical Fate and Impact of
Oxygenates in Groundwater-Solubility of BTEX From Gasoline-Oxygenate Compounds. American Petroleum
Institute Publication, Washington, DC, p. 4531.
Bates, H.K., Spalding, R.F., 1998. Aquifer denitrification as interpreted from in-situ microcosm experiments. J. Environ.
Qual. 27 (1), 174 182.
Cassada, D.A., Zhang, Y., Snow, D.D., Spalding, R.F., 2000. Trace analysis of ethanol, MTBE, and related oxygenated
in water using solid-phase microextraction and gas chromatography/mass spectrometry. Anal. Chem. 72, 4654 4658.
Chapelle, F.H., 1993. Groundwater Microbiology and Geochemistry. John Wiley and Sons, Inc., New York.
Chen, X.H., Goeke, J., Ayers, J.F., Summerside, S., 2003. Observation well network design for pumping tests in
unconfined aquifers. J. Am. Water Resour. Assoc. 39 (1), 17 32.
Corseuil, H.X., Alvarez, P.J.J., 1996. Natural bioremediation perspective for BTX-contaminated groundwater in Brazil.
Water Sci. Technol. 35, 9 16.
Corseuil, H.X., Hunt, C.S., Dos-Santos, R.C.F., Alvarez, P.J.J., 1998. The influence of the gasoline oxygenate ethanol on
aerobic and anaerobic BTX biodegradation. Water Resour. Res. 32, 2065 2072.
Dakhel, N., Pasteris, G., Werner, D., Hohener, P., 2003. Small-volume releases of gasoline in the vadose zone: impact of
the additives MTBE and ethanol on ground water quality. Environ. Sci. Technol. 10, 2127 2133.
Hunt, C.S., Dos-Santos, R.C.F., Corseuil, H.X., Alvarez, P.J.J., 1997. Effect of ethanol on aerobic BTX degradation. In:
Alleman, B.C., Leeson, A.L (Eds.), In Situ and OnSite Bioremediation, vol. 4 (1). Battelle Press, pp. 49 54.
Khan, I.A., Spalding, R.F., 2003. Development of a procedure for sustainable in-situ aquifer denitrification. Remediation:
the journal of environmental cleanup costs. Technol. Tech. 13 (2), 53 69.
Kildsgaard, J., Engesgaard, P., 2002. Tracer tests and image analysis of biological clogging in a two-dimensional sandbox
experiment. Ground Water Monit. Remediat. 22, 60 67.
Lahvis, M.A., Rehmann, L.C., 2000. Simulation of transport of methyl tert-butyl-ether (MTBE) to groundwater from
small volume releases of gasoline in the vadose zone. American Petroleum Institute Soil and Groundwater Research
Bulletin, vol. 10. American Petroleum Institute, Washington, D.C. 8 pp.
McCarty, P.L., Beck, L., Amant, P. St., 1969. Biological denitrification of wastewater by the addition of organic
materials. Proceedings of the 24th Industrial Waste Conference, May 6, 7, and 8. Indiana-Purdue University,
Lafayette.
McDonald, M.G., Harbaugh, A.W., 1988. Modular three-dimensional finite-difference ground water flow model. U.S.
Geol. Surv. Tech. Water-Resour. Invest. 06-A1. 586 pp.
Moyer, E.E., 2003. Chemical and physical properties. In: Moyer, E.E., Kostecki, P.K. (Eds.), MTBE Remediation
Handbook. American Scientific Publishers, Amherst, MA, pp. 11 18.
Nebraska Department of Environmental Quality (NDEQ), 2005. Reported ethanol spills to NDEQ, Lincoln, Nebraska.
Personal communication with Dave Chambers.
Nebraska Ethanol Board (NEB), 2004. Ethanol production in United States. Energy Information Division, Department of
Energy, Washington, DC. http://wmaster@eia.doe.gov.
Nielsen, D.R., Biggar, J.W., 1962. Miscible displacement: III. Theoretical considerations. Soil Sci. Soc. Am. Proc. 26,
216 221.
Nuttall, H.E., 2002. A case study of denitrification in a New Mexico aquifer. In: Faris, B. (Ed.), A Systematic Approach
to In-situ Bioremediation for Groundwater. Interstate Regulatory and Technology Council, Washington, DC, p. 158.
August 2002.
Powers, S.E., Hunt, C.S., Heermann, S.E., Corseuil, H.X., Rice, D., Alvarez, P.J.J., 2001. The transport and fate of
ethanol and BTEX in ground water contaminated by gasohol. Crit. Rev. Environ. Sci. Technol. 31 (1), 79 123.
Ruiz-Aguilar, G.M.L., OReilly, K., Alvarez, P.J.J., 2003. A comparison of benzene and toluene plume lengths for sites
contaminated with regular vs. ethanol-amended gasoline. Ground Water Monit. Remediat. 23 (1), 48 53.
Schmidt, T.C., Schirmer, M., Weih, H., Haderlein, S.B., 2004. Microbial degradation of methyl tert-butyl ether and tert-
butyl alcohol in the subsurface. J. Contam. Hydrol. 70 (3-4), 173 203.
194 Y. Zhang et al. / Journal of Contaminant Hydrology 82 (2006) 183194

Sniegoeki, R.T., 1955. Groundwater resources of the Prairie Creek unit of the Lower Platte River Basin, Nebraska. Geol.
Surv. Water-Supply Pap., 1327.
Stenback, G.A., Ong, S.K., Rogers, S.W., Kjartanson, B.H., 2004. Impact of transverse and longitudinal dispersion on
first-order degradation rate constant estimation. J. Contam. Hydol. 73, 3 14.
Wilson, J.T., 2003. Fate and transport of MTBE and other gasoline components. In: Moyer, E.E., Kostecki, P.K. (Eds.),
MTBE Remediation Handbook. American Scientific Publishers, Amherst, MA, pp. 19 61.
Zheng, C., 1990. MT3D, a modular three-dimensional transport model for simulation of advection, dispersion and
chemical reactions of contaminants in groundwater systems. Report to the U.S. Environmental Protection Agency.
170 pp.
Zlotnik, V.A., Spalding, R.F, Exner, M.E., Burbach, M.E., 1993. Sampling of non-point source contamination in high-
capacity wells. Water Sci. Technol. 28 (3-5), 409 413.