Dating of young ground water withCFCs, SF6, 3H, and 3H/3He:

Examples from the Blue Ridge and Valleyand Ridge of parts of Virginia and Pennsylvania

L.N. Plummer, USGS, Reston, VA

“Owing to dispersion, the “age” of a groundwater sample corresponds generally to a time distributionof many elementary flows. Thus, except in the theoretical case of a pure piston flow system, or ofstationary waters entrapped in a geological formation, the concept of groundwater age has little significance”

J.-Ch. Fontes (1983) in Guidebook on Nuclear Techniques in Hydrology, IAEA

Selected Environmental TracersSelected Environmental Tracers00--50 Year Timescale50 Year Timescale

• 3H , 3H/3He• 85Kr• CFC-11, CFC-12,

CFC-113• SF6

• Event Markers: 3H, 36Cl, 14C

• Age: time elapsed since recharge

tracer1b.grf (JK Böhlke) 4/22/03

CFC-12

3H°

14C

36Cl

85Kr

1940 1950 1960 1970 1980 1990 2000Year

0

50

100

150

200

250

300

350

400

14C

(pm

c) a

nd 3

6 Cl (

atom

s m

-2 y

r-1

x10-

9 )

0

200

400

600

800

1000

1200

1400

3 H (

TU

) a

nd 8

5 Kr

(mB

q m

-3)

and

CFC

's (p

ptv)

and

SF

6 (1

0-2

ppt

v)

-11

-113

3He*

3H(1995)

SF6

Cook and Bohlke, 2000

Why measure environmental tracers in Why measure environmental tracers in ground water?ground water?

• Estimate fractions of young water and the mean age of the young fraction in mixtures.

• Evaluate vulnerability to contamination.

• Estimate recharge rates.• Calibrate models of groundwater

flow.• Estimate rates of geochemical

and microbiological processes. • Retrieve historical records of

contaminant loading to aquifers.• Estimate remediation times.

Approach to “Dating” Young GW

• Collect water samples without contacting air.

• Minimize mixing effects by sampling monitoring wells with narrow screens.

• Analyze with high precision for CFCs, SF6, 3H, 3H/3He, and others (multi-tracer approach).

• Age interpretation. Evaluate multiple tracer data in context of models of groundwater flow.

• Age is model dependent.

Age InterpretationAge Interpretation

• Comparison of simulated and observed tracer concentrations. Lumped-parameter models. (1) multiple tracers from the source, (2) a time series from the source, or (3) multiple tracers from multiple sources in the system. Choose a model based on hydrogeology.

• Tracer plots. Method of comparing simulated and observed multiple tracer data; recognizing cases of possible piston flow and binary mixing (dilution); elimination of some mixing models.

• Flow-model calibration and simulation of age.

• All “ages”, regardless of method, are model dependent.

Tritium/HeliumTritium/Helium--33

• Half-life 12.3 years; decays to 3He.

• Atmospheric thermonuclear weapons testing from the 1950’s, and especially in the period 1962-1963

• Initial 3H measured• Terrigenic He sources• Dispersion around bomb peak

• Confinement of 3He

ChlorofluorocarbonsChlorofluorocarbons

• CFC-12 (CF2Cl2), 1930• CFC-11 (CFCl3), 1936• CFC-113 (C2F3Cl3), 1944• Input smooth, increasing until

1990s (dating range ~1950 to early 1990s).

• Stable in aerobic environments.• In future- dual ages.• Use of ratios.• Can detect post 1940’s water.• Collection/analysis not overly

labor intensive.• Problems with contamination,

degradation (anoxic), sorption?

New bottle method of collection.See http://water.usgs.gov/lab/cfc

Sulfur HexafluorideSulfur Hexafluoride

• Electrical insulator in high voltage switches

• First produced 1953• Very low solubility in water.• Does not degrade• Terrigenic source• Dating range: 1970-modern.• Smooth input, increasing in

air at 6%/yr; 5 pptv today.

1950 1960 1970 1980 1990 2000Year

0

200

400

600

CFC

, SF

6 C

onc

entr

atio

n (p

ptv)

1

10

100

1000

10000

Triti

um (

TU)

CFC-1

1

CFC

-12

CFC-113SF

6X10

0

3H

Dating with CFCs and SF6

• Henry’s law solubility Ci = KH x pi

• Requirements:

ØGas-water equilibrium at rechargeØRecharge temperatureØBarometric pressure at rechargeØKnowledge of atmospheric history of the gas

1950 1960 1970 1980 1990 2000Year

0

200

400

600

CFC

, SF 6

Con

cen

trat

ion

(ppt

v)

CFC-1

1

CFC

-12

CFC-113

SF6X

100

Limitations

• Cost• Terrigenic He• Bubbles, Gas-stripping,

confinement• Mixing

• Contamination• Degradation • Mixing

3H/3He

SF6

•Terrigenic SF6

•Mixing

CFCs

85Kr

•Very difficult collectionand analysis

•No labs available to us

There is no perfect tracer. Multi-tracerapproach recommended.

Tracer Method of dating the young fraction in binary mixtures of young and old

• CFC pptv ratio defines age of young fraction.• % young water (pptv(measured) /pptv(ratio year)) x 100

• Cannot date outside range for ratio.• Cannot use if one of the CFCs in the selected ratio is “contaminated”, even

if ratio is “in range”.• Ratio-based age must be less than (younger than) apparent (piston flow)

model ages for both CFCs in the ratio.

1950 1960 1970 1980 1990 2000Year

0

0.2

0.4

0.6

0.8

1

Atm

osp

her

ic R

atio

(p

ptv

/pp

tv)

CFC-11/CFC-12

CFC-113/CFC-12

SF6X100/CFC-12

3H/3He Age Applies to that of the Young Fraction in Simple Binary Mixtures

τ =1λ

⋅ ln 1+3Hetri3Hm

λ = decay constant = ln 2/t1/2 = 0.05635 year-1

t1/2 = 12.3 years

3H/3He Age is based on an isotope ratio.

Blue Ridge Mountains of Virginia

Plummer, L.N., Busenberg, E., Böhlke, J.K., Nelms, D.L., Michel, R.L.,and Schlosser, P., 2001, Groundwater residence times in Shenandoah National Park, Blue Ridge Mountains, Virginia, USA: A multi-tracerapproach. Chemical Geology, v. 179/1-4, p. 93-111.

• 800 km2;elev. 170-1230 m• Annual precipitation averaged

114 cm; mean annual temperature,7.8 oC in high altitude, central part.

• Precambrian to Cambrian fracturedcrystalline rocks, metabasalts withthin cover of colluvium and residuum.

• Two largest springs have maximumdischarges of 20 and 11 l/sec. Mostwith max. discharge of < 2 l/sec.

• In drought, sprs. about 10% of max.• wells produce < 6 l/sec, and

typically < 1/sec.• 34 springs, 15 wells: 1996

(wet season), 1997 (dry season).• Shallow recharge through residuum

and colluvium recharge fracturesystem which has low storage.

• CFCs, 3H/3He, SF6, 35S, stable isotopes, dissolved N2, Ar.

Well14

LewisSpring

SHENANDOAH NATIONAL PARKBIG MEADOWS AREA

Ground-Water-Flow System

EL=2800 EL=2800

EL=3520

NWSE

Lewis Mountain Spring

Sampling from a TypicalSpring Box and Weir

0 200 400 600CFC-12 (pptv)

0

100

200

300

CF

C-1

1 (p

ptv

)

Sample 1997.0Piston FlowExponential MixingBinary MixtureSprings, '96Springs, '97Wells, '96Wells, '97

300

100

50

3040

20

10

5

5

10

20

30

40

Hudson Spring

0 200 400 600CFC-12 (pptv)

0

20

40

60

80

100

CF

C-1

13 (p

ptv

)

Sample 1997.0Piston FlowExponential MixingBinary MixingSprings in 1996Springs in 1997Wells in 1996Wells in 1997

5

10

20

3040

10

20

30

40

50

100

300

52

2

Hudson Spring

0 200 400 600 800CFC-12 (pptv)

0

2

4

6

SF 6

(pp

tv)

Sample 1997.0Piston FlowExponential MixingBinary MixtureSprings, '96Springs, '97Wells, '96Wells, '97

5

10

20

3040

50100

300

5

10

20

3040

2

L

L

F

F

F

F

L

F Funace Spr.L Lewis Spr.

0 2 4 6SF6 (pptv)

0

4

8

12

16

203 H

(T

U)

Sample 1997.0Piston FlowExponential MixingBinary MixtureSprings, '96Springs, '97Wells, '96Wells, '97Wash. D.C. , Exp. Model

Two 3H inputfunctions:

• Wash. DC(upper)

• Wash. DCscaled to SNP(lower)

Also 0-2 cc/Lof excess air

Precipitation

Ground-Water Reservoir Spring

Discharge

Average -8.2 per milAmplitude 9.6 per milStd. Dev. 2.8 per mil

Average -8.2 per milAmplitude1.1 per milStd. Dev. 0.3 per mil

ResidenceTime2 Years

-16

-12

-8

-4

δO

perm

il

-16

-12

-8

-4

-16 -12 -8 -4

δ18O per mil

-120

-100

-80

-60

-40

-20

0δD

per

mil Precipitation

d Excess 16.1 per mil

Springs and Wellsd Excess 16.4 per mil

1999.0 1999.2 1999.4 1999.6 1999.8Date

-12

-10

-8

-6

-4

δ18 O

(per

mil)

SpringBrowntown ValleyByrd's Nest 3Furnace HudsonLewisPrecipitation

1 Std. Dev. Model Max. - M

in. Model

0.00 0.01 0.10 1.00 10.00 100.00Residence Time, years

0.1

1.0

10.0

δ18 O

Sea

son

al V

aria

tio

n, p

er m

il

3H/3He (Springs)SF6 (Springs)CFC-12 (Springs)3H/3He (Wells)SF5 (Wells)CFC-12 (Wells)

Average Seasonal∆ δ18O = 0.12 per milAverage Age ∼ 5 years

"Average" Spring

Conclusions from Blue Ridge (SNP)• δ 18O data: Mean residence time of 5 years. (Range 0-10

yrs.) Age includes uz travel time.

• Ages based on CFCs and SF6 young; generally consistent with stable isotope data; do not include uz travel time.

• Ages based on 3H/3He biased young (0-2 yrs, most 0).

• Water from wells have ages of 0 to 25 yrs. 3H/3He dating works well for these (and applies to the young fraction).

• Some wells discharge mixtures; ratios of CFCs can define age of young fraction and percent young water in mixture.

• Excess CFCs-- Anthropogenic sources? Evidence of shallow recharge?

Chesapeake Bay Watershed

Region of 165,000 km2 over parts ofNY, PA, MD, DE, VA, and WV.

Evaluate residence times and nitratetransport in groundwater dischargingto streams in the watershed.

Assess lag time between changes at theland surface and the response in thebase-flow component of groundwaterdischarge to the bay.

Washington D.C.

Lindsey and others, 2003, USGS WRIR 03-4035, http://pa.water.usgs.gov/reports/wrir03-4035.pdf

E. Mahantango Cr.

Layered fracture density model(Gburek, Folmar, and Urban (1998)Depth (ft)

0

20

40

80

60

Layered fracture density model

Burton et al. 2002

CFC ratio plot

0 200 400 600CFC-12 Concentration in pptv

0

40

80

120

CFC

-113

Con

cen

trat

ion

in p

ptv

Piston FlowBinary MixingEast TransectWest Transect

19902000

1980

1985

1994

1975

Bedding

Cleavage

NE SW

NW SE

drainage

East Piezometer Transect

West Piezometer Transect

..although these ages are really mixture ages!

--Preferred CFC-12 ages NE of stream are older...

19.942.5

47.2

41.4

49.647.6

42.8

31.5

13.413.9

25.4

16.9

15.9

12.9

8.9

13.4

5.4

5.5

..than those underneath and SW of stream

22/94

--Ages are even more mixed, but % of young water (2nd #) shows older waters generally prevailing to NW 28.1

28/959/23

17/4117/17

21/24…and younger to SE

12.511/45

6/38

14/63

11.9

0.0

Piezometerage animation

Burton et al. 2002

USGS Chesapeake Bay watershed nutrient study

• SF6 vs 3H, CFCs• Monitoring

wells, springs from watersheds Valley and

Ridge

Piedmont/ Blue Ridge

Coastal Plain

Muddy Cr.

Polecat Cr.

East Mahantango Cr.

Upper Pocomoke R.

0 1 2 3 4 5SF6 Concentration, pptv

0

200

400

600

CFC

-12

Con

cent

rati

on, p

ptv

Piston FlowRecharge DateBinary MixingExponential MixingResidence Time, Yrs.

50

5

10

200

100

20

1

19991995

1990

1985

1980

1975

1970

0 1 2 3 4 5SF6 Concentration, pptv

0

10

20

30

40

3 H D

ecay

ed t

o Y

ear

1999

, TU

Piston FlowRecharge DateBinary MixingExponential MixingResidence Time, Yrs.

50 510

200

100

20

1

19991995199019851980

1975

1970

Winchester

Harrisonburg

Staunton

Waynesboro Charlottesville

38º30'N

39º30'N

39ºN

38ºN

78º3

0'W

79ºW

78ºW

77º3

0'W

0 15 30 MILES

Virginia

Maryland

WestVirginia

Virginia Aquifer Susceptibility StudyValley & Ridge Carbonate

Valley & Ridge Siliciclastic

Blue Ridge

Piedmont

Shenandoah National Park StudySpring

Well

Chesapeake Bay StudyCarbonate Spring

EXPLANATION

PD-20

VR-24

VR-17

VR-16

VR-46

VR-27

VR-26

VR-20 VR-45

Stanley

Shenandoah

Waynesboro

VR-25

BR-10

PD-24

BR-07

BR-06

PD-17

PD-16

BR-01 BR-02

BR-03

PD-19

VRC9 VRC15

VRC12VRC23

VRC16

VRC20

VRC2

Soil & RegolithSoil & Regolith

AlluviumAlluvium

LimestoneLimestone

Younger Ground WaterYounger Ground Water

Older Ground WaterOlder Ground Water

Mixture of Younger and OlderMixture of Younger and OlderGround WaterGround Water

Modified from Brahana and others, 1986Modified from Brahana and others, 1986

SPRINGSSPRINGSSOLUTIONSOLUTIONCHANNELSCHANNELS

NOT TO SCALENOT TO SCALE

FRACTURESFRACTURES

Hydrogeology of Carbonate Hydrogeology of Carbonate TerranesTerranesof the Valley & Ridge Provinceof the Valley & Ridge Province

Tritium vs CFC-12

• Young, piston flow•Binary mixtures of young and old•Few CFC-12 contamination

0 200 400 600CFC-12 Concentration, pptv

0

10

20

30

40

3 H D

ecay

ed to

Yea

r 19

99, T

U

Piston FlowRecharge DateBinary MixingExponential MixingResidence Time, Yrs.C-Bay SpringsMuddy Creek Sprs.Muddy Creek Dom. wellsMuddy Creek Mon. wellsShenandoah Sprs.Shenandoah WellsVAS Sprs.VAS Wells

Valley and Ridge CarbonatesC-Bay Sprs.Muddy Creek Sprs.Muddy Creek Dom. wellsMuddy Creek Mon. wellsShenandoah Sprs.Shenandoah wellsVAS Sprs.VAS wells

Mixture

s

Bear Lithia Spring • 2 samples beyond dating range of 3H/3He.• 9 samples may be unmixed.• 13 samples look like mixtures.

0 20 40 603H/3He Age in Years

0

20

40

60

CF

C-1

2 P

isto

n-F

low

Age

in Y

ears

1:1 L

ine

Beyond 3H/3HeDating range

0 10 20 30CFC-113/CFC-12 Age of Young Fraction, Years

0

20

40

60

80

100

120P

erce

nt Y

oun

g W

ater

in M

ixtu

re Valley and Ridge CarbonatesC-Bay Sprs.Muddy Creek Sprs.Muddy Creek Dom. wellsMuddy Creek Mon. wellsShenandoah Sprs.Shenandoah wellsVAS Sprs.VAS wells

Ages and mixing fractions determined from CFC-113/CFC-12

5-22 years and 10-100%young water in mixture.

Some are inconsistent with 3H (CFC contam.)

Valley and Ridge CarbonatesC-Bay Sprs.Muddy Creek Sprs.Muddy Creek Dom. wellsMuddy Creek Mon. wellsShenandoah Sprs.Shenandoah wellsVAS Sprs.VAS wells

Numbers are % ofYoung fraction inMixture.

3H + 3He(tritiogenic)as function of 3H/3Heage.

Bear LithiaSpr.

1940 1960 1980 2000Date

0

10

20

30

40

50T

riti

um

in P

reci

pita

tio

n in

TU

18

31

5

11

34

839

32

42

78

45

64

1

47

6

Bear Lithia Spring (9/2/99)

Piston-Flow AgesCFC-11 27.2 yrsCFC-12 27.2 yrsCFC-113 >Modern (100 pptv)3H/3He 30.0 yrs

Agreement in ages suggestsPiston flow (30 yrs in pipe flow).

Major Contradiction:Tritium = 1.2 ± 0.2 TU1969.6 water would containedabout 130 TU; decays to 25 TU

Piston FlowRecharge DateBinary MixingExponential MixingResidence Time, Yrs.C-Bay SpringsMuddy Creek Sprs.Muddy Creek Dom. wellsMuddy Creek Mon. wellsShenandoah Sprs.Shenandoah WellsVAS Sprs.VAS Wells

0 200 400 600CFC-12 Concentration, pptv

0

10

20

30

403 H

Dec

ayed

to

Yea

r 19

99, T

U

1990

19801960

19951975

1970

15

100

200

1050

20

1985

1999

Bear Lithia

Without multipleTracers, we would not know this is aMixture.

Bear Lithia Spring (9/2/99)

Correcting for assumed pre-bomb tritiogenic 3He in the old fraction, young fraction varies from 5% at 30 yrs to 12% at 0 yrs age.

From CFC-11: 33.0 % 0 yrs water.From CFC-12: 32.8 % 0 yrs water.

CFCs Imply 3H should have been around 3 TU instead of 1.2 TU.

CONCLUSION: Discharge 67-88 %pre-bomb water mixed with approx.12-33% modern water. 3H/3He and CFC age of young fraction is very uncertain (0-30 yrs). 1950 1960 1970 1980 1990 2000

Year

0.1

1

10

100

1000

10000

Tri

tiu

m (T

U)

1953

1955

1960

1965

1970

19751980

Washington D.C.

Pre-Bomb 3H5 - 8 TU? < 0.3-0.5 TU today

GC-ECD Chromatograms

• Analysis by GC-ECD: Purge and trap gas chromatography with electron-capture detector.

• Detects halogenated VOCs (examples: CFCs, CCl4, Halons, TCE, TCA, etc.

• Traditional analysis: GC-MS < 0.1 ug/L (= 100 ng/L = 100,000 pg/L)

• GC-ECD: < 1 pg/L (= 5 orders of magnitude below normal reporting levels)

Shenandoah National ParkSpringDrinking WaterModern Clean Water

Spring Yorktown, VADrinking Water

Spring, Clean Modern Water 10 Year Old Water

CFC-12 CFC-12

CFC-11 CFC-11

CFC-113 CFC-113

CCl4 120 pptv in air

Chloroform?Low ppb range

CCl4

A Town well in VADrinking Water WellDepth: 190 metersValley and Ridge

A Town well in VADrinking Water WellDepth: 115 metersBlue Ridge

CFC-12 CFC-12

Vinyl Chloride?, Methyl chloride?

CFC-11 CFC-11DCA?, TCA?, CH3Br?CFC-113?

CCl4 ppb rangeCCl4 natural range

???

Chloroform ? Others?, ppb? Chloroform ?

PCE Tetrachloroethene ??

???

?

Nelms, D.L. and others,2003, USGSWRIR 03-4278

http://water.usgs.gov/pubs/of/2003/ofr03-246/ofr03-246.htmShapiro, S.D., Busenberg, E., Focazio, M.J., and Plummer, L.N., 2004, Historical trends in occurrence and atmospheric inputs of halogenatedvolatile organic compounds in untreated ground water used as a source of drinking water. Science of the Total Environment, v. 321, 201-217.

Halon 1211Drilling Tracer

Minutes0 1 2 3 4 5 6 7 8 9

milli

volts

0

2

4

6

8

10

12

14

1.

65

258

1.

79

1005

0

1.87

59

31

2.0

2 5

866

2

.14

166

CF

C-1

2 2

.29

227

06

2.

52

5067

2.

90

5833

62

3.

13 9

4

CF

C-1

1 3

.75

721

07

4

.03

386

4

4

.86

756

2

CF

C-1

13

6.21

49

75

6

.43

264

46

7

.44

210

828

8

.18

2

8

.65

179

3

8.67

10

32

9.

29

1588

9

.40

987

9.

52

271

9.

61

48

Detector 2ORhinkle011603c20

NameRetention TimeArea

Minutes0 1 2 3 4 5 6 7 8 9

mill

ivol

ts

0

2

4

6

8

10

12

14

1.

60

775

1

.68

192

1

.85

682

75

2.0

1 1

1083

CFC

-12

2.3

0 8

9941

2.

54

2608

2.

65

475

2.

83

865

2

.93

133

5

CF

C-1

1 3

.52

938

3.

80

1097

263

3

.99

480

8

4

.95

129

15

5.

27

91

CF

C-1

13

6.18

273

04

6.

47

3276

5

7.

49

5212

23

8

.65

238

3

Detector 2ORhinkle011603c23

NameRetention TimeArea

Well MW-N3D. Large Halon peak.CFC-11 and CFC-12 indicate mid- tolate 1970s. Since CFCs came withHalon, water is older than 1970s, butcannot be dated further with CFCsbecause of contamination withdrilling air. Possible mixture of oldwater and water contaminated withdrilling air. Without the Halon data,we would have interpreted a CFCage that is too young.

Halon-1211

Well MW-N4D. Trace or no Halonpresent. CFC data probably unaffectedby drilling air. CFC data suggest a modernage for the water. Without the Halon data,we would not know that the CFCs are valid in this sample.

Halon-1211?

CFC-11

CFC-12 CFC-11

CFC-12

CFC-113

CONCLUSIONS• Use of multiple tracers and “tracer plots” can help to

eliminate some mixing models and refine estimates of mean tracer age.

• To a first approximation, the ages and mixing fractions of many samples from karst or fractured rock can be interpreted using a simple binary mixing model. Is this the result of the sampling process?

• Young ages in Blue Ridge. Mixtures from wells.

• About half of the 3H/3He samples from the Valley and Ridge karst have initial tritium consistent with piston flow (0-15 yrs, unmixed). Rest are mixtures of 0-25 yrs (apparent age) mixed with old (pre-bomb) water.

CONCLUSIONS (cont.)

• CFC-113/CFC-12 data demonstrate cases of piston flow and binary mixing too, but can be affected by contamination.

• Most ground water from fractured rock orkarstic aquifers is vulnerable to contamination.

• Should include tracers in well drilling.

CONCLUSIONS (cont.)

“… the concept of groundwater age has little significance” (Fontes, 1983).

Investigation of multiple environmental tracers in groundwater systems can often help to refine the interpretation of age, refine hydrologic concepts, and identify vulnerability to contamination.

•Chesapeake Bay Study. Scott Phillips, Bruce Lindsey, Gary Spieran, Mike Focazio, J.K. Bohlke, Bill Burton, Colleen Donnelly, Ed Busenberg.

•Virginia Aquifer Susceptibility Study. DaveNelms and George Harlow.

•Shenandoah National Park Study. Ed Busenberg, Dave Nelms, Jerry Casile, Julian Wayland, WandeeKirkland, Stephanie Shapiro, Brian Norton.

•Reston Chlorofluorocarbon Laboratory.

•Noble Gas Laboratory of Lamont-Doherty Earth Observatory, Columbia University.

Thanks