Natural Remediation of Arsenic-Contaminated Groundwater: Solute-Trajiispuil Model Prediction!1 - frrpci iK!n RECORDS CTR «,

- t f l .«g"t "* ' i^Kenneth G. Stollenwerk1 (303-236-4994; kgstolle@usgs.gov) ' // ***

^'caK *John A. Colman2 (508-490-5027; jacolman@usgs.gov)

JU.S. Geological Survey MS 413 Box 25046, Federal Center, Denver, CO 80225, United States2U.S. Geological Survey 28 Lord Rd Suite 280, Marlborough, MA 01752, United States

Leachate from a municipal landfill in Saco, Maine has resulted in a plume of anoxic groundwatercontaining high concentrations of arsenic(III), ferrous iron, manganese, and dissolved organiccarbon. The source of arsenic appears to be reductive dissolution by dissolved organic carbon ofarsenic-containing iron oxides in the bedrock and alluvium downgradient from the landfill. Thelandfill was covered with an impermeable membrane in 1997 to eliminate the source of anoxicgroundwater and allow natural flushing of the aquifer to decrease arsenic concentrations toacceptable levels. A one-dimensional reaction-transport model was used to predict the evolutionof the plume for the first 60 years after the landfill was covered. The modeled flow path extendsfrom the landfill to a stream where the plume discharges. Groundwater analyses from wells alongthe 1 12 meter flow path were used to initialize the model. Selection of chemical reactions that areproposed to describe the evolution of the plume are based on interpretation of chemical data fromsamples of water and sediments in the aquifer, and laboratory experiments with contaminatedcores leached with uncontaminated groundwater. The dominant biogeochemical reactions in themodel were oxidation of organic carbon by dissolved oxygen, manganese oxides, and iron oxyhy-droxides. These reactions were simulated using a modified form of Monod kinetics. Transport ofarsenic was controlled by equilibrium sorption. Model parameters for these reactions wereadjusted to obtain the best fit between the model and observed breakthrough curves of constitu-ents in the laboratory experiments. Simulation results indicated that concentrations of most con-stituents in the landfill plume would rapidy decrease to near background levels within the nextfew years; however, the sorbed organic carbon in the aquifer was predicted to consume oxygenand maintain anoxic conditions for at least 60 years. Arsenic concentrations were predicted toslowly decrease from a high of 650 micrograms per liter to the current drinking water standard of50 micrograms per liter in about 30 years. After 60 years, arsenic concentrations were predicted tobe greater than 10 micrograms per liter, which is the drinking water standard currently under con-sideration by the Environmental Protection Agency.

-i'

1 INTRODUCTION

Leachate migrating from a municipal landfill in Saco, Maine (Figure 1) has resulted in a plume of contaminated groundwater extending more than 200 m downgradient. The plume is anaerobic and con­tains elevated concentrations of several constituents. Arseniccon­centrations are as high as 650 |ig/L, 13 times greater than the US Environmental Protection Agencies maximum contaminant level of >o iig/L. The landfill was capped in early 1998, stopping the source of contamination. Flow of uncontaminated groundwater through the aquifer is expected to change conditions from reducing to oxidizing, immobilizing the arsenic.

I LINE OF GEOHYDROLOGIC SECTION MW96-OR 0 MONITORING WELL AND NUMBLR

TB97-2 • TEST BORING AND NUMBER

S55 • SURFACE SAMPLE AND NUMBER < GROUND-WATER FLOW DIRECTION

43S37­

Location of ST~ Study Arc.

BASemOUUS.MMYCOfiPSOFENOIHEERS o , go ,^ f CONTOUR NTEWM. S FEET SACOMUNICIFHLLAHOnu. 124SO IMS NATIONAL GEODETIC VERTICAL OtTMl OF ISIS

0 SO 100 HFERS

Figure 1. Location of study area. The arsenic contamination is as­sociated with area 4.

f CtJ CJ CO -2? I 5

03 4^ o CD "CJ CD CO O CO $ -S T3

o %S c: §CtJ CtJ ^ E8 5 S> c O> Q) •9 §-*3 "Sco m ^m r\ >J •*- r*«

=i • §vo vu s

iQ> §

S ^ 1 ?cc g *- Cj CD

s s §° n C CD

•S S E S o aie o> g o

CD CD C CD CD CO - -o ­E CtJ **'-« ->i ^ « too CO 1^ CD -cll ss o g ^S

o § -ffi « 2^ cc •C3 CO ^ «CD **-. **­0. CD Si|li S; O OCr co oQ. CtJ 8 8 ^ c- •*—

O c •S> § c: Jb s _ IS **5 .if m«^ ** '-^ 1 Nl»1 tsE^- s 3 c i-a

ic CM cB CtJ g' Q) .0 " "S g> ** s*-» : •=; CtJ 0-tS -f: CD «A a•sa ctj ctj o

*ik rTT fe"SS a, ^^fc-_ fm\ +^—^> s £ O CD O

_o • O0) *i o §

ll CD To C .0 CDC

CD ^§° ). CO O

i ~ •*< «J t: o p^ 5* r i "s1*­8 ^D CD -S E Q)

1 RESULTS

Dissolved solids in effluent from the core rapidly decreased within a few pore volumes (Figure 2). Carbonate dissolution a/r/ carbonic acid from or­ganic carbon oxidation account for most of the el­evated conductance values after pore volume 5. By pore volume 30, the effluent specific conduc­tance was about equal to the influent.

1500

• Specific conductance of column effluent Specific conductance of column Influent

Z 1000I o o j£ 500

O UJ 0. v­

• • 9m'. *«^^» •••»'• •**•• -«^ "

l I 10 20 30 40 50

PORE VOLUME

Figure 2. Specific conductance of core effluent.

uo

Initial dissolved organic carbon in the pore water was 31 mg/L (Figure 3). Concentrations decreased to less than 1 mg/L by pore volume 5 and remained relatively constant for the remaining 50 pore vol­umes. About 5% of the total organic carbon on aqui­fer solids had been removed by pore volume 50.

35

"a,30

O 25 ffi DC • Dissolved organic carbon in column effluent

Dissolved organic carbon In Influent = 0

O QHI

en 5

10 20 30 40 50 PORE VOLUME

3. Dissolved organic carbon in core effluent,

^———. ^ —^W^ ^ HH^ ^ ^ ^ ^ ^ ^ ^ ^ ^BBBW^ - HH - VVIIBBBBH^ ^BBHVW^ —^ HH^ - HV -H^ ^ ^ M^ ^H^

Dissolved oxygen in the influent groundWater <j was 6 mg/L Oxidation of organic carbon in the core resulted in effluent concentrations less than 0.06 mg/L for50pore volumes (Figure 4). Reduction of oxygen by organic carbon $ks ki­netically controlled. During periods of higher flow rates, some oxygen was measured in the effluent. During periods of slower flow rates, dissolved oxygen was less than the detection limit of 0.005 mg/L

0.5

0-4 r •> Dissolved oxygen In column effluent

Dissolved oxygen in Influent = 6 mg/L

0.3

X o o g 0.2

O CO CO Q 0.1

(l* • y«* • 0 10 20 30

PORE VOLUME

Figure 4. Dissolved oxygen in core effluent.

/ 7 After depletion of oxygen, the electron accep­tors became manganese and iron oxides. The concentration of iron in the initial pore water was 33 mg/L (Figure &). Concentrations de­creased to about 5 mg/L by pore volume 10, then remained relatively constant.

40

• Iron (II) In column effluent 30

Iron In influent = 0

<tr 20 UJ

o o o

10

O DC

••„ v:» \ 10 20 30 40 50

PORE VOLUME

Figure 6, iron concentration in core effluent,

Arsenic in the pore water from this core was 480 \ig/L (Figure 6). After the initial rapid decrease in doncentration during the first few pore volumes, krsenic concentrations decreased at a slower fate. About 30 pore volumes were required for ar­senic levels to decrease to the maximum contam­inant level of 50 yig/L Arsenic was in the trivalent Oxidation state.

500

' • Arsenic (III) In column effluent 3400 Maximum contaminant level (EPA)

O

LU O

|300

oo o

200

LU eg 100

10 20 30 40 PORE VOLUME

Figure 6. Arsenic concentration in core effluent.

The most likely source of arsenic in groundwater ^thought to b£ the aqUifdf solids. The landfill is located on glacial till, derived from igneous and metasedimentary rocks. Uncontaminated till col­lided near the landfill contains about 02 mg/kg arsenic in the iron oxides coating the aquifer sol­ids (based on 0.04 M hydroxylamine hydrochlo­tide extraction). Reductive dissolution of these iron oxides by organic carbon in leachate from the landfill would result in mobilization of the associ­ated arsenic. Once in solution, transport of ar­senic is controlled by adsorption.

CONCLUSIONS \°

Accumulation of organic carbon on aquifer solids downgradient from the Saco landfill has resulted in significant biological oxygen demand. The resulting anaerobic conditions cause reductive dissolution of iron oxides and mobilize arsenic. The concentration of arsenic (III) in leachate from contaminated aqui­fer cores remained above the EPA maximum con­taminant level for 30 pore volumes. The experimental data Indicate that decades may be re­quired for natural remediation of the contaminated

^j*

areas of the aquifer.