Case Study of Subsurface Vapor Intrusion at a Dry Cleaner Site

1
Case Study of Subsurface Vapor Intrusion at a
Dry Cleaner Site
Eric M. Nichols, PEEric.Nichols_at_lfr.com

• Amy Goldberg Day Amy.Goldberg.Day_at_lfr.com

AEHS Annual East Coast Conference on Soils,
Sediments and Water October 2004
2
Outline

• Background
• Conceptual Site Model
• Data Collection
• Groundwater
• Soil gas
• Indoor air
• Comparison of Attenuation Factors
• Variance from EPA Default Attenuation Factors
• Observations and Conclusions

3
Background

• Shopping center in Central California with 3 dry
  cleaners
• Routine disposal of dry cleaning fluids into
  sanitary sewer
• Sewer line leaks resulted in PCE releases
• PCE identified in downgradient municipal water
  well
• Dry cleaners implicated and ordered to perform
  RI/FS type investigation

4
Background, Continued

• Interbedded fine-grained sediments to 25 ft bgs
• Discontinuous coarse-grained sediments from 25
  to 50 feet bgs
• Depth to groundwater 50 feet bgs
• Human health risk assessment performed using
  applicable data considering source and non-source
  areas

5
Background, Continued

• Existing buildings slab-on-grade
• Some buildings had historical use of PCE
• All buildings have commercial use
• Expected transport mechanisms
• Diffusion from source zones
• Advection and diffusion across foundation

6

• Source Area

Former Dry Cleaner
Sewer Line
Subject Building
7
Groundwater Data Summary

• 3 yrs of quarterly monitoring from 18 A-zone
  wells-EPA (Level IV Data Validation)
• Analyzed using EPA Method 8260A
• Source-area PCE detected in 13 of 13 samples
• 5,000 to 85,000 ?g/l
• 95 UCL 48,300 ?g/l
• Non-source-area PCE detected in 118 of 124
  samples
• 1.5 to 12,000 ?g/l
• 95 UCL 1,800 ?g/l

8
Soil Gas Data Summary

• Soil gas samples collected from March 1997
  through June 1998
• Analyzed via on-site mobile lab using EPA Method
  8010 (Level III DV)
• 381 samples collected from 0 to 10 feet bgs
• 77 source-area PCE samples
• maximum detected 39,490,000 ?g/m3
• 95 UCL 25,485,000 ?g/m3
• 304 non-source area PCE samples
• 100 to 9,060,000 ?g/m3
• 95 UCL 605,000 ?g/m3

9
Flux Chamber Data Summary

• 13 indoor sample locations on observed floor
  seams and cracks
• 4 outdoor locations in planter boxes
• TO-14 SIM
• PCE detected in all indoor samples
• Flux range 0.29 to 26 ?g/min-ft

10
Air Data Summary

• Indoor air samples collected in 6 buildings, 1
  located close to source area 3 outdoor sample
  locations
• 15 samples collected over source area in 5
  separate sampling events over 14 months
• 1 sample collected in each of the other
  buildings
• Level III Data Validation

11
Air Data Summary

• Subject building vacant duringfirst air sampling
  event
• Doors closed HVAC on
• Cracks and seams were sealed before third
  sampling event
• Similar results
• Building was reoccupied and floor covering added
  before fourth sampling event
• Fourth and fifth sampling events were during
  normal business hours, with doors opening and
  closing throughout day

12
Vapor Intrusion Modeling

• Estimated indoor air concentration using Johnson
  Ettinger model with site-specific soil and
  building parameters
• Used JE for both soil gas and groundwater
  results (95 UCLs)
• Compared estimated indoor air concentration to
  measured indoor air concentration

13
Results of VI Modeling from Crack Flux Data

• Assumes cracks are only significant route of
  vapor entry (BIG assumption!)
• Applied box mixing model with building volume
  and air exchange rate
• Estimated indoor PCE concentration 14 ?g/m3

14
Results Comparisonsoil gas and air in ?g/m3
groundwater in ?g/l
Bold indicates higher value
15
Attenuation Factors

• Following the guidance in Appendix F
• ? indoor air/soil gas (used direct
  measured and JE estimated indoor air
  concentrations)
• ? indoor air/groundwaterHc (used direct
  measured and JE estimated indoor air
  concentrations)

16
Attenuation Factor Comparison
1.2 x10-5
2.8 x10-6
4.0 x10-3
2.8 x10-6
4.0 x10-3
4.0 x10-6
Crack flux data not useful for estimating
attenuation factor
17
Figure 3 Vapor Attenuation Factors Groundwater
to Indoor Air (Sandy Loam)
18
Figure 3 Vapor Attenuation Factors Soil Gas to
Indoor Air (Sandy Loam)
19
Observations

• Estimated attenuation factors ranged from 1x10-5
  to 4x10-6
• Figure 3 attenuation factors range from 2x10-3 to
  4x10-3
• Johnson Ettinger model with site-specific
  parameters was reasonable predictor of indoor air
  concentrations and attenuation factors using soil
  gas data

20
Observations, Continued

• Sealing floor cracks and seams did not
  significantly reduce indoor air concentrations or
  apparent attenuation factor
• Flux chamber data was least accurate predictor of
  indoor air concentrations (possibly used
  incorrect assumption)
• HVAC on or off did not significantly reduce
  indoor air concentrations or apparent attenuation
  factor
• Installation and operation of SVE system reduced
  measured indoor air concentrations to below
  reporting limits

21
Observations, Continued

• EPA Figure 3 attenuation factors are
  significantly more conservative than attenuation
  factors estimated at this site
• Indoor air concentrations likely not influenced
  by background concentrations
• Other cases with very high PCE soil gas
  concentrations had ?s in the 10-5 range

22
Attenuation Variance Possible Reasons for
Variance from EPA Figure 3

• Complex geologic subsurface conditions shallow
  fine-grained material may have restricted vapor
  intrusion
• Sampling biased towards areas of higher
  concentrations possible biases in data set

23
Attenuation Variance Possible Reasons for
Variance from EPA Figure 3

• Highest detected concentrations of PCE in both
  soil gas and groundwater were in the parking
  lot--- no indoor air samples were collected
  directly over this hottest area
• Extremely high source media concentrations

Sub-slab soil gas data could have resolved some
of these issues
24
Conclusions

• Reduction following SVE confirms origin of impact
  was from subsurface
• Measured groundwater-indoor air or soil
  gas-indoor air attenuation factors were within
  one order of magnitude of modeled attenuation
  factors

25
Conclusions

• For this well-characterized site, use of soil gas
  or groundwater data were appropriate to predict
  attenuation factors
• Site-specific subsurface and building conditions
  and extremely high source concentrations likely
  influenced differences between measured and EPA
  Figure 3 attenuation factors