Multi-Scale Monitoring and Prediction of System Responses to BiostimulationSusan Hubbard*1

Co-Investigators and Collaborators: Jill Banfield*2, Jinsong Chen*1, Mark Conrad*1, Jenny Druhan*2, Andreas Englert *1, Andreas Kemna*5, LiLi*1, Phil Long*3,Michael O’Brien*4, Dimtrios Ntarlagiannis*4, Yves Personna*4, Steve Pride *1, Lee Slater*4, Carl Steefel*1, and Ken Williams *1

I. INTRODUCTION AND MOTIVIATION

Prior

Distribution

Posterior

Distribution

Prior

Information

Data

1—Define priors

2—Define likelihood functions

3—Explore posterior distribution

Prior

Distribution

Posterior

Distribution

Prior

Information

Data

1—Define priors

2—Define likelihood functions

3—Explore posterior distribution

BAYES:

P(m|d)=[P(d|m)P(m)]/P(d)

m=model parameters,

d=data

Laboratory Experiments

Petrophysical Relationships

Estimation Framework Reactive Transport Modeling

KE

Y C

OM

PO

NE

NTS

Our research includes theoretical, numerical, and experimental investigations performed at the laboratory and the field scale, which involve the remote monitoring and prediction of biogeochemicalprocesses using geophysical methods and reactive transport modeling, respectively. Linkage between the laboratory-scale and field-scale investigations will be ensured by using the same (native)sediments (and in some cases, groundwater), and by using the same geophysical measurement techniques and reactive transport code at both scales. By investigating the geophysical response tocoupled processes and by performing calibrated and validated reactive transport modeling at both scales under the same environmental conditions, we are beginning to investigate how novel monitoringand modeling approaches scale with space and time.

The experiments illustrated in Section II showed that correlationsexist between the changes in geophysical responses and thechanges in biogeochemical and physical properties. We previouslyinterpreted these changes qualitatively, based on comparing thechanges in geophysical signatures with biogeochemicalmeasurements. In this component of the project, we areinvestigating the controls on complex electrical, radar and seismicamplitude responses in saturated, porous media using theoreticaland mixing models, and testing the models using the laboratorydata.

This modeling will enable a more quantitative interpretation ofthe geophysical signatures than has been heretofore possible

Our previous work has illustrated that time-lapse geophysical signatures track the onsetand evolution of end-products associatedwith bioremediation, such as the productionof iron sulfides (Williams et al., 2005). A keycomponent of this project is thedevelopment of a stochastic estimationframework that permits the quantitativeestimation of properties associated withbiogeochemical transformations. Themethodology is performed using a Markovchain Monte Carlo approach, and by usingpetrophysical models that link thegeophysical attributes to biogeochemicalproperties. The model will be tested usingdata collected at the laboratory scale andthen applied to time-lapse field geophysicaldata collected during biostimulationexperiments.

This component focuses development andvalidation of reactive transport models atthe laboratory and field scales. We areusing the pore-water chemistry determinedover the course of lab experiments and thesolid-phase mineralogy determined via postmortem analysis to develop a defensibledescription of the reaction network (pathwaysand rates). The reactive transport modeling iscarried out using CRUNCH (Steefel et al.,2003) and TOUGHREACT codes. Thereactive transport model is beingcalibrated to the geophysical data, but onlyby using the independent constraints providedby the microbiological, chemical, and physicaldata. This is a key step, since the geophysicaldata will be crucial in developing a high-resolution data set at the field scale, wherecomplete microbiological, chemical, andphysical characterization of the subsurfacematerial will not be feasible.

Base of Column (4-6 cm)

-2

0

2

4

6

8

10

12

14

16

18

0 10 20 30 40 50 60 70

Days after Lactate Injection

Ma

x -

Ph

as

e (

mra

d)

-5.0E-05

0.0E+00

5.0E-05

1.0E-04

1.5E-04

2.0E-04

2.5E-04

3.0E-04

3.5E-04

Sim

ula

ted

Fe

S a

nd

Zn

S V

olu

me

Fra

cti

on

s

-Phase

FeS

ZnS

Near Middle of Column (13-17 cm)

0

0.5

1

1.5

2

2.5

3

3.5

4

0 10 20 30 40 50 60 70

Days after Lactate Injection

Ma

x -

Ph

as

e (

mra

d)

-1.0E-05

0.0E+00

1.0E-05

2.0E-05

3.0E-05

4.0E-05

5.0E-05

6.0E-05

7.0E-05

8.0E-05

Sim

ula

ted

Fe

S a

nd

Zn

S V

olu

me

Fra

cti

on

s

-Phase

FeS

ZnS

To advance solutions needed for remediation of DOE contaminated sites, approaches are needed that can elucidate and predict reactions associated with coupled biological,geochemical, and hydrological processes over a variety of spatial scales and in heterogeneous environments. Our previous laboratory experimental experiments, which wereconducted under controlled and homogeneous conditions, suggest that geophysical methods have the potential for elucidating system transformations that often occur duringremediation. Examples include tracking the onset and aggregation of precipitates associated with sulfate reduction using seismic and complex resistivity methods (Williams et al.,2005; Ntarlagiannis et al., 2005) as well as estimating the volume of evolved gas associated with denitrification using radar velocity. These exciting studies illustrated that geophysicalresponses correlated with biogeochemical changes, but also that multiple factors could impact the geophysical signature and thus a better understanding as well as integration toolswere needed to advance the techniques to the point where they can be used to provide quantitative estimates of system transformations.

Our current research includes theoretical, numerical, and experimental investigations, performed at the laboratory and the field scales, to determine if geophysical methods can beused to uniquely monitor system transformations. Our work is geared toward the Uranium Mill Tailings site at Rifle, CO, site (Figure 1), where ERSP-sponsored investigations areexploring the efficacy of electron-donor amendments for facilitating sustainable microbial reduction of U(VI) to U(IV) through a series of local-scale field experiments conducted in2002-2005 (Anderson et al., 2003; Vrionis et al, 2005) Early experiments at the site showed that U(VI) loss from groundwater occurred synchronously with growth of Geobacter afteracetate amendment, and illustrated the importance of maintaining iron-reducing conditions for optimal U(VI) removal. Since the interplay between iron and sulfate reduction isbelieved to be of critical importance to the sustainable reduction of U(VI) at this site, quantitative interpretation of geophysical data in terms of redox state, exhaustion of bioavailableiron mineral phases, or onset of sulfate reduction is expected to greatly benefit the understanding and sustained remediation of uranium at the site.

The sections below describe the key components of our work, including laboratory and field characterization and monitoring during biostimulation experiments that were conducted in2006, reactive transport modeling, and biogeochemical property estimation using geophysical datasets.

II. RESEARCH METHODS

III LABORATORY EXPERIMENTAL RESULTS

Laboratory studies are being performed tomimic experimental conditions at the OldRifle, CO, field site using native groundwaterand sediments. Radar, seismic, SP, andcomplex resistivity measurements are beingcollected during the experiment concomitantwith geochemical, hydrological, andmicrobiological measurements. Experimentsare investigating the responses ofgeophysical attributes to:•Ferric Oxide Reduction duringBiostimulation;

•Stimulation of natural microbialcommunity in native Sediments undersulfate reducing conditions;•Stimulation of natural microbialcommunity in native sediments undersustained iron reducing conditions•FeS Oxidation.These experiments will indicate the

potential of using geophysicalmeasurements to remotely indicate when

a system is undergoing iron or sulfatereduction.

Field Characterization andMonitoring

Figure 1. Rifle, Colorado UMTRA Site

2005 Mini -gallery: More sensitive to FeS; FeRB signal overwhelmed?

2004 Mini -gallery: FeS signal NOT observed; constrained by injection zone

Dep

th (

m)

Dep

th (

m)

0 5 10 15 20 25 30

0

-2

-4-6

-8

-10

col

row

-4.0 -2.0 0.0 2.0 4.0

phadiffary3f1

ary3_(t1-t0)_f1, iso5

0 5 10 15 20 25 30

0

-2

-4-6

-8

-10

col

row

-4.0 -2.0 0.0 2.0 4.0

phadiff2ary3f1

ary3_(t2-t0)_f1, iso5

3-Weeks after Injection Began

10-Weeks after Injection Began

Distance Along Array (m)

Phase Difference (mrad)

Injection

2004 ⊥ Flow

-2 0 2 4-4

0 5 10 15 20 25 30

0-2

-4

-6

-8

-10

col

row

-8 -4 0 4 8

ary3_Ct1_t0_f0_pha

C: ary3_(t1-t0)_f0, iso10

0 5 10 15 20 25 30

0-2

-4

-6

-8

-10

col

row

-8 -4 0 4 8

ary3_Ct2_t0_f0_pha

C: ary3_(t2-t0)_f0, iso10

10-Weeks after Injection Began

2-Weeks after Injection Began

Phase Difference (mrad)-4 0 4 8-8

Dep

th (

m)

Dep

th (

m)

Injection

2005 ⊥ Flow

2-weeks

10-weeks

3-weeks

10-weeks

Distance Along Array (m)

Crosshole seismic, radar and complexresistivity, and SP data were collected thisyear between several well pairs and withinall three flow cells at the Rifle Site.Comparison of the characterization data withborehole electrical, geological, andflowmeter logs and tracer data is beingperformed as a first step in estimatinghydrogeological heterogeneity at the fieldscale. The time-lapse geophysical datasets,collected in association with the field-scalebiostimulation experiments, are beinginterpreted in terms of biogeochemicaltransformations. Comparison of thecharacterization and monitoring datasetspermits assessment of the role ofheterogeneity on systemtransformations.

REFERENCESAnderson, R.T., Helen A. Vrionis, Irene Ortiz-Bernad, Charles T. Resch, Philip E. Long, Richard Dayvault, Ken Karp, Sam Marutzky, Donald R. Metzler, Aaron Peacock, David C. White, Mary Lowe, and Derek R. Lovley. 2003. Stimulated in situ removal of U(VI) from groundwater of a uranium-contaminated aquifer, Applied And Environmental Microbiology, p. 5884-5891, Vol. 69, No. 10Pride, S., J.G. Berryman, and J.M. Harris, Seismic attenuation due to wave-induced flow, J. Geophysics. Res., 109, 2004.Slater, L., D Ntarlagiannis, D.B. Wishart, On the relationship between induced polarization and surface area in metal-sand and clay-sand mixtures, Geophysics, Volume 71, Issue 2, pp. A1-A5, 2006Steefel C., User’s Manual for CRUNCH, Software for Multicomponent Reactive Transport, LBNL, 2005.Vrionis, Helen A., Robert T. Anderson, Irene Ortiz-Bernad, Kathleen R. O'Neill1, Charles, T. Resch, Aaron D. Peacock, Richard Dayvault, David C. White, Philip E. Long, and Derek R. Lovley, Microbiological and Geochemical Heterogeneity in an In Situ Uranium Bioremediation Field Site. Appl Environ Microbiol., 2005Williams, K.H., D. Ntarlagiannis, L. Slater, A. Dohnalkova, 2005a, S. Hubbard and J. Banfield, Geophysical imaging of stimulated microbial biomineralization, Env. Science and Tech., 39, 7592-7600.

Phase impacts of oxidation. Comparison of thelaboratory phase response of oxidized and reducedRifle, CO sediments over the indicated frequency range.The sediments were obtained from an un-amendedportion of the Rifle site from a depth approximately 4 to4.5-m below ground surface, sieved to include the < 2-mm fraction, and saturated with site groundwater. Forthe reduced sediments, an extended 70-day in situincubation was performed in the measurement cell usinggroundwater amended with 50-mM acetate. Over thistime interval, extensive precipitation of FeS wasdocumented having a acid volatile sulfide (AVS)concentration of 18-µmol S2- g-1 sediment, in comparisonto the oxidized sediment, which had a AVSconcentration of 0.05-µmol S2- g-1 sediment.

Previous laboratory experiments in sandy materials with a single microbial strain illustrated that complexresistivity data can be used to detect the onset of sulfide production associated with sulfate reduction(Williams et al., 2005). Here, we expand upon this work by repeating the experiments using site-derivedsediments and groundwater (and thus microbial community), and by also performing additionalexperiments aimed at investigating the geophysical transformations to biogeochemical transformations,such as:

•clay mineral alteration accompanying the reduction of structural ferric iron;•oxidation end-products accompanying an increase in groundwater oxygen concentration, and•reduction of structural iron in bulk sediments through exposure to sulfide-rich groundwater.

These studies, which are briefly described below, indicate that several of these processes should bedetectable at the field scale.

Impact of Iron Reduction. For the first test,the clay-sized fraction (< 2-um) was isolatedfrom the bulk sediments and reducedenzymatically using a site-isolated strain ofGeobacter. The complex resistivity datarevealed a slight decrease in both the phaseresponse and the imaginary component ofresistivity following bioreduction, suggestingthat the alteration of clay minerals bymicroorganisms is capable of decreasing thepolarization response of sediments containingiron-bearing phyllosilicates.

We also explored the impact of oxidative and abioticprocesses on the real component of resistivity through theprecipitation of low conductivity phases, such as elementalsulfur created through the oxidation of bisulfide by ferric ironpresent in crystalline iron (oxy)hydroxides. Oxidized aquifersediments were added to sulfide-rich groundwater lacking areadily available source of acetate and the complex resistivityresponse was measured over the course of 2 weeks. Duringthis time, the real component of conductivity (i.e. bulkconductivity) decreased by up to 16%, presumably as a result ofthe precipitation of S0 and/or disseminated, disordered FeS.Over the same time period, the polarization response increasedsignificantly (44% at 10-Hz), most likely as a result of theprecipitation of FeS and in agreement with our previousobservations monitoring metal sulfide precipitation.

0.06

0.062

0.064

0.066

0.068

0.07

0.072

0.074

0.076

0.078

0.08

1 10 100 1000

Frequency (Hz)

Rea

l Con

d. (

S/m

)

0-hr 19-hr 24-hr42-hr 66-hr 115-hr142-hr 306-hr 388-hr400-hr 522-hr

HS--oxidation; S0 precipitation

FeS precipitation

σreal decrease = 16%

!" # 3-4 mrad

VI. FIELD SCALE CHARACTERIZATION AND MONITORINGThrough field characterization and monitoring using seismic, radar, complex electrical and SP measurements, and through coupling the field characterization with the lab-scale experiments and reactive transport modeling described above, we willinvestigate at the Rifle Site the following questions:

•Can geophysical data be used to distinguish between iron and sulfate reduction processes and track these process over space and time?•What is the role of heterogeneity on the system transformations?

Characterization of subsurface hydrogeological heterogeneity isbeing performed across the Rifle Flow Cells using a combination ofgeologic logs, flowmeter data, tracer test, and electrical logs fromwellbores together with radar, seismic, and electrical tomographicdatasets. As characterization data were collected at different times inrelationship to the manipulation experiments, a first step in theanalysis is determining the value of the data for characterizing initial(i.e., unperturbed) hydrogeological conditions and for developingsite-specific petrophysical relationships between the hydrological andgeophysical attributes, as is shown by the figure to the right.

Preliminary analysis of the geological and geophysical datasetssuggests that there is consistent layering across the Rifle study siteand that this layering may impact transport and transformations.

Monitoring is also being performed using time-lapse complexresistivity, radar, seismic, and SP to image spatiotemporal changesin the system in response to biostimulation.

The figures to the right illustrate the complex resistivity responsesassociated with the 2006 biostimulation experiments, which weredesigned to facilitate iron reduction in the ‘2004’ flow cell andsulfate reduction in the ‘2005’ flow cell. Figure A illustrates wherethe surface datasets were collected relative to the two injectiongalleries. Figures B-C illustrate the inversion of the complexresistivity data relative to baseline conditions, while Figure Dillustrates illustrates geochemical responses. The expectedtransformations associated with the 2004 & the 2005 experimentsare shown below. Based on the 2005 and 2006 complex resistivityresponses, which indicate the formation of iron sulfides, weinterpret that a zone of mixing formed between the two injectionareas. 2004 Injection Area: FeOOH(s) + 3H++ e- → Fe2+(aq) + 2H20 2005 Injection Area: SO42- + 9H+ + 8e- → HS-(aq) + 4H20 2006 Zone of mixing: Fe2+(aq) + HS-(aq) → FeS(s) + H+The flow diversion from the 2005 to the 2004 flow cell may havebeen impacted by pumping activity as well as local heterogeneity.These field-scale geophysical responses:

•Are in agreement with the laboratory responses;•Provide different information at different frequencies,which suggests that the geochemical estimation frameworkdescribed in IV should be applicable to field datasets:•Suggest that complex resistivity approaches are useful foridentifying important changes in mineralogy associatedwith sulfate reduction at the field scale.

The figures to the right illustrate the use of borehole SP data tomonitor transformations associated with the biostimulationexperiment that was performed in 2006 at the Rifle Site. Thefigures illustrate that there is an excellent correlation betweenthe SP response and the geochemical parameters, suggestingthe usefulness for SP logging as a low cost means forestimating persistence of low redox conditions associatedwith sulfate-reduction.

This ongoing research suggests that field IP methods tend tobe extremely sensitive to mineralogical transformations, whileSP data are most sensitive to aqueous geochemical changes

associated with remedial treatments.

IV. GEOCHEMICAL ESTIMATION FRAMEWORK V. ISOTOPE STUDIES & REACTIVE TRANSPORT MODELINGWe have developed a stochastic framework that permits estimation of the evolution of end-products associated withbiogeochemical transformations using time-lapse geophysical data and the petrophysical models described below. Bothmodels include an error term (ε).

Xt-1 Xt Xt+1

Yt-1 Yt Yt+1

Geochemical

parameters

Geophysical

data

Xt-1 Xt Xt+1

Yt-1 Yt Yt+1

Xt-1Xt-1 XtXt Xt+1Xt+1

Yt-1Yt-1 YtYt Yt+1Yt+1

Geochemical

parameters

Geophysical

data

Seismic P-wave velocity(v) and attenuation(1/Q) as a function ofprecipitate radius (r) andvolume (p) arerepresented using apatchy model (Pride etal., 2004).

Chargeability (m) and time constant (τ) are related to volume and mean grain size ofevolved precipitates using a modified Cole-Cole model (Slater et al., 2006).

1 1

1 2 2

( ) ( ( ), ( ))

( ) ( ( ), ( ))

p i i i

i i i

v t g p t r t

Q t g p t r t

!

!

= +"#

= +$

1 3

2

2 4

( )( )

( )

( ) ( )

b

in i

i

i i

p tm t a

r t

t a r t

!

" !

# $ %& = +' () * +&

= +,

We have used a Markov chain monte carlo method for the estimation problem, which has been expressed in a Bayesianformulation. This approach permits updating of the prior information about geochemical parameters using the time-lapsegeophysical data and the petrophysical models, which are expressed in terms of likelihood functions. The developed stochasticframework is general and flexible and can be used to combined various types of geophysical data for geochemical parameter estimation. Here, we have applied the method to the time-lapse seismic and IP datasets collected inassociation with the laboratory column experiments described by Williams et al. (2005) to estimate the evolution of the volumeand mean grain size of sulfide precipitates.

{ }

{ }

1 2

1

n

1

=

1 2

i 1

1{ ( ), ( ), ( ), ( ), 1, }

( |

)

( ( ) | ( ), ( ), , ) ( ( ) | ( ), )

( ( ) | ( ), ( )) ( ( ) | ( ), ( ))

{ ( ), ( ), 1, }, , ,

n

n i i i i i

i

p i

i

i i i i

n i i p i

i

i

i

f

f m t p t

m t t v t Q t i n

r t a b f t r t a

f v t p t r t f Q t p t r

p t r t i n a a b

t

!

!=

"

=

=

#

#

1 2

1

( ) ( ) ( ) ( ( ), ( ))n

i i

i

f a f a f b f p t r t=#

Prior distribution

Likelihood function of complex resistivity data

Likelihood function of seismic data

In this first attempt to use geophysical data to quantitatively estimate end-products associated with remediation, we findthat geophysical data provides useful information about the onset and evolution of sulfide precipitates. As is describe in inSection V, iteration of the advanced modeling and monitoring is being performed to improve both approaches.

Hydraulic Conductivities Along the Center-Line (Perpendicular to

Injection Gallery)

-14

-13

-12

-11

-10

-9

-8

-7

-6

-5 0 5 10 15 20

Distance to Injection Gallery [m]

Ln

of

Hyd

rau

lic C

on

du

cti

vit

y [

m/s

]

Statistics (By Well)

Statistics (All Wells)

Mean of Upper Zone

Mean of Lower Zone

B-02 M-03 M-08 M-13

Injection Gallery

ln(EC) and ln(K i) at M 03 and M 13

1.5

1.7

1.9

2.1

2.3

2.5

2.7

2.9

3.1

3.3

3.5

-14 -13 -12 -11 -10 -9 -8 -7

ln Hydraulic Conductivity [m/s]

ln E

lectr

ical C

on

du

cti

vit

y [

mS

/m]

ln(K) m/s

ln (

EC

) m

S/m

Data from Frank Spane and Phil Long (PNNL)

K impacted by

stimulation??

M16

3.60

4.10

4.60

5.10

5.60

6.10

0 0.1 0.2 0.3 0.4

The high K peak is often at the base of the ‘clayey gravel unit ’, where there are fewer large

pebbles

> 20% of total flow

going through

this zone

20% of flowtravelsthrough

this zone

M16

3

3.5

4

4.5

5

5.5

6

15 25Electrical Conductivity (mS/m)

m T

OC

Geologic Log Flowmeter Radar Velocity EC Log (mS/m)M16 M16 M 16-M17 M17

Comparison of Different DatasetsDowngradient of Injection Wells, 2004 Gallery

May 18, 2006Aug 13, 2006

ERT4 M16 M17 X1 M21 X3

May 18, 2006 May 18, 2006 May 18, 2006

2004 Gallery 2005 Gallery

Use indirect and direct measurements to

provide characterization data

for flowmodel

Radar Tomography Velocity ‘Fence Diagram’

May 18, 2006 May 18, 2006 May 18, 2006

OtherGeologic

Data

GeophysicalData

HydrogeologicData

StatisticalModels

EstimatedSpatial Distribution of

Hydrogeologic Parameters

OtherGeologic

Data

GeophysicalData

HydrogeologicData

StatisticalModels

EstimatedSpatial Distribution of

Hydrogeologic Parameters

X1 M21 X3

0

20

40

60

80

100

120

140

0 5 10 15 20 25 30

Fe-Experiment

Fe-calculation

Fe-Extractable (umol/g sediment)

Location (cm)

Modeling Studies. We are using the pore-water chemistrydetermined over the course of lab experiments and the solid-phasemineralogy determined via post mortem analysis to develop adefensible description of the reaction network (pathways andrates). The reactive transport modeling is carried out usingCRUNCH and TOUGHREACT codes. The figure to the rightshows that predictions of FeS and ZnS accumulation rates agreeclosely with those determined by 0.5 N HCl extractionmeasurements from column studies of Williams et al. (2005). Thestudies also suggest that the bacteria are highly motile andchemotaxis must be considered in the transport model.

Comparing Geophysical and Modeling Estimates. The estimatesof evolved precipitates obtained using the time-lapse geophysical data werecompared with those estimates obtained using the numerical modelingapproach. We found that the time evolution profiles were very similar inshape, but different in mass balance by a factor of 1000. Our workinghypothesis is that both methods give different information about the evolvedprecipitates: the geochemical modeling predicts the absolute volume ofprecipitates, whereas the geophysical methods are likely to sense thesurface area that is impacted by FeS and ZnS secondary phases growing ongrains or biomass. This observation suggests that a combination ofgeophysical monitoring and geochemical modeling may lead to new insightsabout the dynamic system.

Blue-Calculated from the direct

measurements of geochemical

parameters

Red-Estimated from geophysical data

X1200

Isolated precipitates

(3-5nm)

Sand grains(600-800_ m)

Aggregatecells (10 -60 um)

Encrustedsingle cell (1 -2 um)

Seismicmethod

InducedPolarization (IP)

Geochemicalsimulation

Isolated precipitates

(3-5nm)

Sand grains(600-800_ m)

Aggregatecells (10 -60 um)

Encrustedsingle cell (1 -2 um)

Seismicmethod

InducedPolarization (IP)

Geochemicalsimulation

Isolated precipitates

(3-5nm)

Sand grains(600-800_ m)

Aggregatecells (10 -60 um)

Encrustedsingle cell (1 -2 um)

Seismicmethod

InducedPolarization (IP)

Geochemicalsimulation

Isotope Studies. Sulfur isotopes of SO42- and HS- were analyzed in a

background well and three down-gradient monitoring wells beginning near the onsetof SO4

2- reduction associated with the 2006 biostimulation experiment at the RifleSite. Analysis of δ34S of SO42- and HS- during in situ U(VI) bioremediation hasindicated that variation in isotopic composition of sulfur species can be detectedunder these conditions. The data indicate a high level of fractionation betweenSO42- and HS- at both onset and recovery and minimal fractionation at the heightof SO42- reduction. These methods can therefore be applied as an indicator of thestage of SO42- reduction, and may provide a more sensitive means of obtaining atimeline of these processes needed for the modeling effort. In addition, the data areuseful determining the composition of incoming SO42-, and the causes of aqueousHS- depletion.

M-21

-30

-25

-20

-15

-10

-5

0

7/7/06 7/27/06 8/16/06 9/5/06 9/25/06 10/15/06 11/4/06

date

34S

0

2

4

6

8

10

12

SO

4 m

M

d34S Sulfide

Sulfate

Sulfide x 10

Fe(II) x 10

_

Start StopM-21

-30

-25

-20

-15

-10

-5

0

7/7/06 7/27/06 8/16/06 9/5/06 9/25/06 10/15/06 11/4/06

date

34S

0

2

4

6

8

10

12

SO

4 m

M

d34S Sulfide

Sulfate

Sulfide x 10

Fe(II) x 10

_

Start Stop

*1 Lawrence Berkeley National Laboratory*2 UC Berkeley*3 Pacific Northwest National Laboratory*4 Rutgers University*5 Forschungszentrum Jülich

[A] Plan map of the acetate injection gallery andmonitoring well locations used during the 2006Old Rifle, CO bioremediation experiment.Borehole SP data were acquired along wells SP-1and SP-2, with copper measurement electrodes(anodes) installed on the outside of the wellcasing and the Cu/CuSO4 reference electrode(cathode) at the ground surface.[B] Graphical representation of the cathodic andanodic reactions occurring at the electrodesurfaces under iron- and sulfate-reducingconditions. The cell is completed upon connectionthrough a high-impedence voltmeter

SP data acquired along well SP-1. The magnitude of the cellpotential can be estimated usingthe difference between thecathodic and anodic half-cellpotentials, with a given anodicreaction (e.g. sulfide-mediatedcopper-oxidation) yielding avalue reflective of the dominantmetabolic process. Here, themeasured potentials arecharacteristic of sulfate-reduction dominating theresponse downgradient from thezone of acetate injection.

Temporal relationshipsbetween three keygeochemical parametersand the SP response of adiscrete electrode located5.0-m below ground surfacein well SP-1.

Figure A. Plan map of the injection galleries andmonitoring well locations used during the 2006 Rifle,CO bioremediation experiment. Injection in the gallerylabeled ‘2005’ focused on long-term (~70-day) acetateamendment targeting sulfate-reduction, whereas thegallery labeled ‘2004’ was used for a short-duration(~21-day) acetate amendment targeting iron-reduction;both amendment experiments ended on the samedate. Surface-based complex resistivity data wereacquired at three time-intervals during the amendmentexperiments along two lines, one located parallel andperpendicular to flow; the later was locatedapproximately 1.5-m downgradient from the injectiongalleries. Data were acquired between 30 Cu/CuSO4electrodes equally spaced over a distance of 30-m.

Figure B. Phase inversion results for the surface complex resistivity dataacquired at using two frequencies and three time intervals with orientationas indicated; both the injection gallery boreholes and the monitoring welllocations are indicated and refer to the locations shown in Fig. A. Perconvention, the phase values shown are plotted as their negative value inmilliradians, with the scale expanded for the data collected parallel togroundwater flow to enhance temporal changes in the phase response.The red boxes illustrate our hypothesis that mixing between the twogalleries occurred, possibly due to the impact of preferential flowpaths onthe transformations.

Figure C. Example of the 2005 phaseinversion results for 0.125 Hz surface complexresistivity data acquired perpendicular to flowfor comparison with Figure B. Acetateamendment during the 2005 experiment waslimited to 25-days and no parallel injectionwas performed in the ‘2004’ gallery location.Stimulation of iron-reduction in the ‘2004’gallery during the 2006 experiment may havecontributed to the elevated phase responserelative to the phase shifts observed in 2005.

Figure F. Geochemicaldata obtained from threeof the monitoring wellsshown in Fig. A. Incontrast to the upgradientcontrol well (B-05), eachof the two downgradientwells (M-21 and M-24)exhibit evidence ofstimulated microbialactivity, with an initialincrease in Fe2+ over thefirst 15-days, followed byan extended period ofsulfate-reduction. Of note,removal of Fe2+ during theonset and duration ofsulfate-reduction is likelyreflective of precipitationand sequestration ofinsoluble FeS.

The surface complex resistivity data shown in Figure Bwere collected at 30- and 57-days.