Our organization (and my employer), the Wyoming Outdoor Council, helped to fund an analysis by Dr. Tom Myers (Ph.D. in hydrogeology/hydrology from the University of Nevada/Reno) of the EPA’s draft study of groundwater contamination in rural Pavillion, Wyoming. Myers’ study fully supports the EPA’s preliminary finding that fluids and chemicals commonly associated with hydraulic fracturing have contaminated the groundwater resource in the area near Pavillion, Wyoming.
Dr. Myers also concluded that the well design was poor because the surface casing does not extend below the level of the water wells, a practice not permitted in many states but not disallowed in Wyoming.
In his report, Myers says that the situation in rural Pavillion is not an analogue for other ‘gas plays’ where geology and hydrology might be different. The report should nevertheless serve as a strong warning to those that would argue that hydraulic fracturing has never contaminated groundwater and that it never will. It is crucial that industry work with regulators to more effectively protect the environment and a resource that, in Wyoming, belongs not only to ours but to every generation.
Dr. Myers report is as follows —
April 30, 2012
Review of DRAFT: Investigation of Ground Water Contamination near Pavillion Wyoming
Prepared by the Environmental Protection Agency, Ada OK
Prepared by: Tom Myers, Ph.D.
SUMMARY AND RECOMMENDATIONS
After consideration of the evidence presented in the EPA report and in URS (2009 and 2010), it is clear
that hydraulic fracturing (fracking (Kramer 2011)) has caused pollution of the Wind River formation and
aquifer. The EPA documents that pollution with up to four sample events in the domestic water wells
and two sample events in two monitoring well constructed by the EPA between the level of the
domestic water wells and the gas production zone. The EPA’s conclusion is sound.
Three factors combine to make Pavillion‐area aquifers especially vulnerable to vertical contaminant
transport from the gas production zone or the gas wells – the geology, the well design, and the well
construction. Natural flow barriers are not prevalent in this area, so there are likely many pathways for
gas and contaminants to move to the surface, regardless of the source. There is also a vertical gradient,
evidenced by flowing water wells, although its magnitude and extend are undefined, to drive advective
vertical transport. The entire formation is considered an underground source of drinking water, but 169
gas wells have been constructed into it; this is fracking fluid injection directly into an underground
source of drinking water.
The well design is poor because the surface casing does not extend below the level of the water wells, as
is required in many other states, and because the wells contain substantial borehole lengths without
surface casing or cement between the production casing and the edge of the borehole. This allows
vertical transport of gas and fluids and decreases the protection against leakage during fracking or gas
production. Third, the EPA documented many instances of sporadic bonding, which simply means the
cement does not completely seal the annulus between the production casing and the edge of the
borehole. This provides pathways which could allow gas and contaminant transport along the well bore.
The EPA also appropriately accounted for the potential that their monitoring well construction could
have explained the contamination. “Since inorganic and organic concentration patterns measured in the
drilling additives do not match patterns observed in the deep monitoring wells and because large
volumes of ground water were extracted from the wells during development and prior to sampling, it is
unlikely that ground‐water chemistry was at all impacted by drilling additives.”(EPA, 2011, p 7).
Myers Review of DRAFT: Investigation of Ground Water Contamination near Pavillion Wyoming 2
The EPA also demonstrated that the inorganic geochemistry in the monitoring wells is substantially
different than that which would occur naturally in the area, and that the enrichment of numerous
constituents is most likely due to the interaction of fracking fluid with the groundwater near the
sampled well. This is particularly true for the elevated levels of potassium, chloride, and pH.
Any of the three contaminant transport pathways suggested by the EPA could be responsible for the
contamination moving from the fracking zone to the drinking water wells. The EPA has also presented
evidence that contamination in surface ponds has not caused the contamination in the water wells or
their monitoring wells.
The situation at Pavillion is not an analogue for other gas plays because the geology and regulatory
framework may be different. The vertical distance between water wells and fracking wells is much less
at Pavillion than in other areas, so the transport time through the pathways may also be low compared
to other gas plays. It is important, however, to consider that the pathways identified at Pavillion could
be applicable elsewhere (Myers, 2012; Osborn et al, 2011). In addition to improving and enforcing the
relevant regulations, monitoring the pathways between the target formation and aquifers should be
standard at all gas plays with fracking.
The following recommendations would improve the analysis and continue the study into the future
made throughout this review.
1. The EPA should continue data collection to better verify the sources and map the potential
2. EPA should map the gas production wells according to their construction date. The EPA should
also compare the locations of observed contamination with the nearby well construction dates
to estimate the travel times from the sources to the well receptors.
3. The EPA should map the depth to water prior to sampling in the water wells. Using this, they
should map vertical gradients and correlate these gradients to areas with contaminants most
likely sourced to deep aquifers.
4. The EPA should install deeper monitoring wells near the shallow pits to better map the depth of
the plume emanating from those pits.
5. Data collection should continue so the results can be replicated. An additional, deeper
monitoring well should be constructed in the gas production zone between the existing
monitoring wells to determine the vertical gradient and estimate the rate of vertical flow.
6. The EPA presents no evidence regarding the extent that fracturing extends above targeted
formations. It may not be possible to prove whether this occurred at this site, but the EPA
should at least discuss the possibility. It would be useful to perform some simple testing to map
the extent of fractures, as described by Fisher and Warpinski (2010).
Myers Review of DRAFT: Investigation of Ground Water Contamination near Pavillion Wyoming 3
The Environmental Protection Agency (EPA) has released a study of groundwater contamination in the
Pavillion gas play in west‐central Wyoming. Their preliminary conclusion is that gas well development
and hydraulic fracturing (fracking (Kramer, 2011)) has caused the contamination. The EPA report is in
draft form and is open for comment until March 12, 2012. This technical memorandum reviews the EPA
report. This review was prepared with support from the Natural Resources Defense Council, Wyoming
Outdoor Council, Earthworks, Oil and Gas Accountability Project and Sierra Club.
This review discusses in detail the appropriateness of the study design, methodology, execution, results,
and interpretation and the reasonableness of the conclusions. It specifically follows and considers the
EPA’s “lines of reasoning” approach used to reach its conclusion.
The study area is in the Pavillion gas field in west‐central Wyoming. It lies northeast of the Wind River
Range. The general geology for uppermost 1000 meters (m) is the Eocene‐aged ((56 to 34 million years
before present) Wind River Formation, which is interbedded sandstone and shale with coarse‐grained
meandering stream channel deposits. The presence of stream channel deposits indicates that the
formation has been carved by river beds which left fluvial deposits interspersed among formation layers
These fluvial deposits often provide connectivity among formation layers and can fragment otherwise
continuous sedimentary layers.
The area has experienced gas development since the 1960s, with 169 gas wells constructed in the study
area. EPA Figure 2 shows the gas well construction chronology. There were three main periods of
construction – 1963‐65, 1975‐83, and 1998 – 2006, with each subsequent period having more new wells
constructed than the previous period. EPA does not specify when fracking first occurred, however.
Recommendation: Add a map of gas production wells coded for the year or time period during which the
well was completed (or fracking occurred if substantially different). This would allow an assessment of
travel time for contaminants to flow from production zones to the monitoring wells and domestic wells.
The US Geological Survey studied the water resources on the Wind River Reservation (Daddow 1996),
which surround this study area (but does not include it). The Wind River Formation is the primary
source of drinking water on the reservation. Daddow’s (1996) description of the formation indicates
that the formation consists of interbedded shale and sandstone with extremely variable permeability
that could lead to highly variable contaminant loads throughout the formation (Osiensky et al 1984).
Recommendation: A more detailed description of the geology and hydrogeology of the area, perhaps
based on the relevant Geological Survey reports would provide more insight regarding geochemical
trends as found by the USGS.
Myers Review of DRAFT: Investigation of Ground Water Contamination near Pavillion Wyoming 4
STUDY LAYOUT AND DESIGN
EPA started this study in response to citizen complaints regarding contamination in their water wells.
EPA established dedicated monitoring wells after two rounds of sampling various water wells rather
than prior to construction of the gas wells. For much of their study data, the EPA had to use sample
data collected from existing water wells. Water wells are not the best tool for monitoring groundwater
quality because, even if the well construction is of similar quality to a dedicated monitoring well, water
wells have much longer screens, or open intervals, than do monitoring wells. They screen the most
productive formation layers, usually based on observations made during drilling, to maximize the
pumping rate while minimizing the drawdown. Wells drilled specifically for monitoring wells also screen
productive zones, but target the screen to a specific zone, usually 20 feet or less thick, so that the
sample represents a given aquifer level.
Samples from water wells are therefore a mixture of water from all productive zones of the entire open
interval, weighted according to the transmissivity of each zone. A domestic water well sample is useful
for determining whether a contaminant exists at some point in the aquifer, but a dedicated monitoring
well is necessary to determine which layer is contaminated.
EPA established two dedicated monitoring wells to supplement the data obtained from the water wells.
The new monitoring wells were primarily screened below the level of the water wells (Figure 1) and
above the gas production wells to “differentiate potential deep (e.g., gas production related) versus
shallow (e.g., pits) sources of groundwater contamination” (EPA p 5). The EPA established just two
monitoring wells due to a limited budget (Id.). EPA placed the monitoring wells’ screened interval along
the conceptualized vertical pathway between the potential contaminant source (i.e. the production
wells and/or zone) and the water wells. The monitoring wells were designed appropriately to detect
and monitor contaminant movement upward from the production zone to the water wells; if the
monitoring wells had been constructed at the same depth as the water wells, they would not have
added substantial useful information.
Figure 1 (EPA Figure 3) shows that domestic water wells in the regions are screened at all levels down to
about 250 m, or more than 800 feet, with half of the wells being deeper than 300 feet, similar to the
depths found by Daddow (1996) in other areas of the aquifer. However, the EPA states the information
source was from the State Engineer and homeowner interviews (EPA p 2). It is unclear whether both
were used for each well. It is my experience that homeowners have a poor concept of the depth of their
well unless they have paperwork that documents it.
Recommendation: The EPA should provide more information about the source of its water well
construction data, showing it in EPA Table A1.
The following table summarizes in general terms the wells that were sampled during each sampling
phase (other media were also sampled but not included in this table). It is apparent that the wells
sampled in phases subsequent to the first phase depended in part on the results of the prior phases.
Myers Review of DRAFT: Investigation of Ground Water Contamination near Pavillion Wyoming 5
Phase Date Domestic
Wells Stock Wells Monitoring
I 3/09 35 2 0 0
II 1/10 17 (10
2 4 0 This phase came about
because EPA had detected
methane and dissolved
hydrocarbons during Phase I.
III 10/10 3 (2
0 0 2 Gas samples also collected
from the well casing of EPA’s
two deep monitoring wells.
IV 4/11 8 previously
sampled 0 3 previously
sampled 2 Added glycols, alcohols, low
molecular weight acids
Figure 1: Snapshot from EPA (2011) Figure 3 showing frequency of depth for gas wells (top), surface casing for gas wells, and
base of domestic wells.
EPA Table A1 lists the wells and the phase during which they were sampled, broken into eight data
1. anions and alkalinity
3. alcohols and VOCs
4. low molecular weight acids and glycols
5. semi‐volatile organic compounds (SVOCs), pesticides, PCBs, and tentatively identified
Myers Review of DRAFT: Investigation of Ground Water Contamination near Pavillion Wyoming 6
6. gas/diesel related compounds, and hydrocarbons
8. fixed gases, heavy hydrocarbons, dissolved carbon, and gas and water isotopic ratios
EPA Table A2a presents the geochemical results – anions, cations, and alkalinity. Unfortunately, this
table does not consistently state in which phase the initial sample was taken. Additional samples are
identified with a suffix on the sample number. The other data tables in Appendix A provide results by
phase, but some results are found only in other reports, including URS (2009 and 2010).
URS (2009) reports the Phase 1 sampling (water wells only) in their Table 9, which shows concentration
of SVOC contaminants, including caprolactam at 1.4 ug/l at PGDW20, dimethylphthalate detected at
nine wells, and Bis (2‐ethylhexyl)phthata at 9.8, 6.4 and 12 ug/l in PGDW25, ‐20 and ‐141
Recommendation: The EPA should present and discuss the correlation of contaminant detects in the
domestic wells with depth.
and detect levels at ten other wells. Total purgeable hydrocarbons were 26 and 25 ug/l in wells
PGDW05 and PGDW30, respectively. Measurable methane concentrations were found in 8 wells. Total
purgeable organics are generally gasoline and diesel range organics. PGDW25 is one of the deeper wells
at 243.8 m below ground surface (bgs) and PGDW05 and ‐30 are at 64.0 and 79.2 m bgs, respectively.
URS (2010) reports the Phase 2 sampling in more detail. It shows more than 20 wells with detectable
levels of a variety of semi‐volatile organics (URS 2010, Table 9). The report does not assess these
detects with the depth of the well, but a quick glance suggests that most of them are on the deeper half
of the domestic wells. An exception is PGDW39, reported to be just 6.1 m deep, although the EPA
should consider whether “6.1” is correct because if so it would be tens of meters shallower than any
other water well in the aquifer.
EPA based this study on four sample events including various subsets of domestic, municipal, and stock
wells and two sample events in the monitoring wells. A reasonable question is whether the number of
samples is sufficient for developing an opinion? A time series would help to identify a trend, but is not
necessary to establish presence/absence. Objections to this data on the basis of there being just two
samples are without merit – simple presence of a substance that would not naturally occur in the
aquifer, if other causes can be eliminated, is sufficient to reach a preliminary conclusion that fracking
fluid has affected the aquifer. However, the EPA should continue the sampling to determine whether
the concentrations are trending higher, or not, and determine how or whether the plume expands.
The EPA identifies three potential pathways for contaminants to reach the water wells from the fracking
(EPA, p 32).
! Fluid and gas movement up compromised gas wells.
The table did not highlight the values at PGDW14 and ‐20 as being exceedences.
Myers Review of DRAFT: Investigation of Ground Water Contamination near Pavillion Wyoming 7
! Fluid excursion from thin discontinuous tight sandstone units into sandstone units of greater
! Out‐of‐formation fracking, whereby new fractures are created or existing fractures are enlarged
above the target formation, increasing the connectivity of the fracture system.
The EPA does not conclude which or whether any of these pathways actually facilitated the
contamination at Pavillion, although arguments throughout the document (and reviewed in this report)
support the potential for any of them. EPA correctly notes that for all three pathways there would be a
correlation between the concentration of gas in the water wells and the proximity to gas well, as found
by Osborn et al (2011) in the Marcellus shale in Pennsylvania. They also note that for all three
pathways, “advective/dispersive transport would be accompanied by degradation causing a vertical
chemical gradient” (EPA, p 32) as discussed in other portions of the report. In other words, with
increasing distance from the source, both vertical and horizontal, the contaminant concentration would
decrease. This would be due in part to chemical degradation, dispersion of a finite mass over a larger
volume, attenuation due to chemicals adsorbing to soil particles, and dilution by mixing with
The following sections consider evidence from various aspects of the EPA report in context of the
Very low permeability layers can prevent or impede the upward movement of fluid or gas from depth to
the water well zone, which in the Wind River Formation is the upper 250 meters (based on the reported
water well depth). Extensive layers of shale are often sources of gas and/or capstones, which prevent
gas in underlying sandstone from escaping to the surface. However, the shale must be horizontally
extensive and not fractured to be an effective seal, which is not the situation in the Pavillion field as
quoted above. The formation is most productive (for gas) at its base with gas trapping occurring in
“localized stratigraphic sandstone pinchouts on the crest and along flanks of a broad dome” (EPA p 2).
Hypothesis: The lithology in the Pavillion area does not prevent the vertical movement of gas or
contaminants to the surface because it is either not sufficiently extensive or impervious. EPA claims
there is no “lithologic barrier … to stop upward vertical migration” (EPA p viii) and also that “there is
little lateral and vertical continuity of hydraulically fractured tight sandstones” (Id.).
Evidence: EPA presented a lithologic cross‐section (Figure 20) showing mapped shale layers, production,
water, and monitoring wells and the points where the production wells had been fracked. EPA found
that the lithology is “highly variable and difficult to correlate from borehole to borehole” (EPA p 15).
“Sandstone and shale layers appeared thin and of limited lateral extent” (Id.). Pathways could go
around the intermittent shale so that contaminants in a given monitoring well may not result from the
nearest production well. Pathways for movement through sandstone could be tortuous (EPA p 37);
vertical pathways through sandstone could be more tortuous than horizontal pathways because the
particles in sandstone tend to be elongated with the longer side being horizontal.
Myers Review of DRAFT: Investigation of Ground Water Contamination near Pavillion Wyoming 8
Fracking has occurred for up to 45 years, so there is potential for many pathways from various sources
to a receptor well. The travel time to a given point could be any time period up to 45 years.
Additionally, out‐of‐formation fracking occurring at any time could have shortened the pathway.
Conclusion: The lithology in most areas would not prevent the vertical movement of contaminants to
the water wells because of the lateral variation.
Vertical flow and gradient
In order for contaminants to move from the fracked zones or from deep well bores to surface aquifers,
there should be a vertical hydraulic gradient. Lacking such a gradient, movement could still be possible
due to lateral dispersion and upward concentration gradients, but it would be much slower.
Hypothesis: There is upward flow in the Pavillion gas field that would support advection of
contaminants associated with fracking fluids to the monitoring and water wells.
Evidence: In the Pavillion area, there are flowing wells, which would indicate an upward gradient, at
least at depth, which could drive vertical advection, or contaminant transport with the groundwater
flow . Daddow (1996) also documented flowing wells in other areas of the Wind River Range, with the
depth range from 225 to 450 feet bgs. EPA uses PGDW44 as an example (p 36). This water well lies near
the middle of the field near MW01. MW01 showed a depth to water equal to 61.2 m at the beginning of
a purge for sampling (p 11 and Figure 8). MW02 had depth to water of 80.5 m (p 12). The depth to
water in the monitoring wells does not support the idea of an upward gradient, but being the only wells
at that depth, the data is not conclusive. Table A1 reports the PGDW44 well depth is 228.6 m; PGDW25
is deeper, at 243.8 m bgs. MW01 is just 10 m deeper. There is apparently an upward gradient at that
point because the well is flowing, but the analysis could be improved, as follows.
EPA documents that the shallower monitoring well has more natural breakdown products of the organic
contaminant like BTEX or glycol that are found in the deeper monitoring well and in fracking fluids (p
36). It suggests that the contaminants in the shallow well are derived from the natural breakdown of
the contaminants found in the deeper well. This could only occur if the wells represent a vertical flow
path, which they do and therefore these findings support the hypothesis of upward movement.
The gas found in the deep Wind River Formation is chemically similar to gas in the underlying Fort
Union Formation suggesting that gas in the Wind River Formation has naturally moved upward until
captured in localized capstones, or “localized stratigraphic sandstone pinchouts” (EPA, p 2). EPA
concludes that differences in gas composition and isotopes support the hypothesis of upward migration
through the various layers in the Wind River formation (p 29). The fraction of ethane and propane in the
gas from domestic wells is mostly less than in the produced gas, but the isotopic composition is clearly
thermogenic, which suggest there is an ongoing “preferential loss of ethane and propane relative to
methane” (p 29, 38). This evidence supports the hypothesis of upward fluid and gas movement.
Vertical movement could occur in the absence of a vertical gradient, if the pressurization caused by the
fracking is sufficient and there is a poorly developed well bore nearby. Contaminants can migrate
Myers Review of DRAFT: Investigation of Ground Water Contamination near Pavillion Wyoming 9
quickly upward through a leaky borehole due to the transient pressure gradient across an aquitard
created by the fracking pressure (Lacombe et al, 1995).
Conclusion: There is evidence to support the concept of upward movement in the area, but it is not
conclusive. The EPA should complete more studies documenting the vertical hydraulic gradient
throughout the area.
Recommendation: The EPA report should document the depth to water in the domestic wells prior to
sampling so that they could map water levels for different well depths and determine the zones of
Contamination from shallow pits
The presence of shallow disposal pits is an alternative source of contamination. EPA notes that there
are 33 shallow pits that had been used for the “storage/disposal of drilling wastes, produced water, and
flowback fluids in the area of investigation” (EPA p 17). As part of this study, the EPA communicated
with stakeholders to further determine the location of pits. Shallow monitoring wells have found very
high concentrations of several contaminants that were also found in deeper water wells and the EPA
monitoring wells. These pits could have received the detritus of fracking operations in the past.
Hypothesis: Contaminated water seeping from these pits could be responsible for the observed
Evidence: Shallow monitoring wells that had been installed previously for reasons not associated with
this project (EPA, p 11) are reported to have very high contaminant concentrations, although this data is
not well summarized in the report. The shallow monitoring wells are only 4.6 m bgs (EPA p 17), so there
is little information about how deep the contamination extends beneath the pits. Assuming the pits are
some distance away from homes and people avoided them when constructing their water wells, it is
possible the shallow disposal pits are sources of contamination beyond the level the EPA considers
shallow, or 31 m bgs (Id.).
Irrigation could help to contain the contamination near the shallow pits because they would be located
in low recharge areas, either by design or in comparison with irrigated fields. It would be unlikely that
the pits would have been constructed within irrigated fields, so the seepage from the pits may be much
less than the seepage beneath irrigated fields because of the continuous application of water to the
field, and for a much shorter time period. Irrigation water would have seeped deeper and faster due to
the likely higher rate of application and effectively diluted or prevented the deeper circulation of
seepage from the pit.
Conclusion: The EPA concludes that these shallow pits are not the source of contaminants found in
deeper water wells. Because there is little contamination in intermediate‐depth wells, their conclusion
is sound, but the document would benefit from more analysis and discussion.
Recommendation: The EPA should document more fully the contaminant plumes near the pits.
Specifically, deeper monitoring wells near the pits should be constructed to construct a contamination
Myers Review of DRAFT: Investigation of Ground Water Contamination near Pavillion Wyoming 10
profile beneath the pits. Better investigation of the pits as a source would also facilitate the remediation
of the groundwater near those pits.
LINES OF REASONING
The EPA used a line of reasoning analysis regarding the presence of fracking fluid constituents and gas in
monitoring wells in support of their preliminary conclusion that fracking has contaminated aquifers in
Pavillion Wyoming. This is critical because the conclusion is not just that leakage from the wells or spills
caused contamination, but that the fracking process itself caused the contamination. EPA deemed the
multiple lines of reasoning approach necessary due to the complexity in detecting contaminants in
groundwater from deep sources. This section critically reviews each of the EPA’s lines of reasoning.
High pH Values
The EPA monitoring wells both have very high pH, ranging from 11.2 to 12.0, which is much higher than
the level seen in the domestic water wells in the Wind River formation. EPA concluded the high pH was
due to hydroxide (OH) which indicated the addition of a strong base to the background water (EPA p xii).
EPA’s reaction path modeling suggested that the addition of just a small amount of potassium hydroxide
to the sodium‐sulfate waters typical of deep portions of the Wind River formation would cause such a
pH change; EPA concludes from the modeling that the typical groundwater in the Pavillion aquifer “is
especially vulnerable to the addition of a strong base” (EPA p 20).
Potassium hydroxide was used as a crosslinker and solvent for fracking the production wells in the area
(EPA p 33), which could be a source of the OH to increase the pH of the water in the area of the
The use of soda ash as a drilling additive when drilling the monitoring wells, often to control the pH, is a
possible alternate explanation for the elevated pH2
EPA Figure 12 verifies these pH values are higher than in the domestic wells, but also shows they fall on
the general trend of pH with elevation of the well open interval. Based on this information, it is not
possible to conclude that the high pH is not natural, but the EPA’s conclusion appears to be justified
based cumulatively on all of the facts concerning pH. EPA should consider geophysical logging
completed by the industry if it includes pH logs to improve their analysis; such logs could provide pH
values for deeper areas that could be compared with the pH values for their monitoring wells.
. Soda ash is 100% Na2CO3. At a 1:100 mixing ratio
with water, the pH of dense soda ash was 11.2 (EPA Table 2). The recommended ratio for use in
fracking fluid is 1:100 to 1:50 (EPA Table 1). The pH of drilling mud varied between 8 and 9. The
concentrations of neither sodium nor carbonate are abnormal in the monitoring wells. If the soda ash
did separate from the drilling mud, mixing with background groundwater would further dilute it so that
the pH would be less than observed at the 1:100 mixing ratio.
MSE%3a%3a1053024648177449, visited 1/13/12
Myers Review of DRAFT: Investigation of Ground Water Contamination near Pavillion Wyoming 11
Chemistry in the shallow wells has been affected by irrigation with Wind River water. This irrigation
water has very low total dissolved solids (TDS) and neutral pH (<8) (EPA Figure 11) but the other shallow
groundwater wells show that the irrigation water picks up contaminants as it seeps.
The methods used to collect samples probably minimized contamination causing high pH in the
monitoring wells. EPA purged the monitor wells until pH stabilized, a process which would minimize the
potential that any residual contamination from well development would have been sampled.
EPA’s analysis associated with Figures 11 and 12, explaining the shallow water geochemistry, is accurate
and useful. It utilizes data from all of the wells in the area and surface waters to show water chemistry
trends through the study area. It also shows how EPA’s monitoring wells differ substantially from the
general trends, supporting the conclusion that elevated pH in water samples from EPA’s deep
monitoring wells was likely caused by contamination with hydraulic fracturing chemicals.
Elevated potassium and chloride
The monitoring wells both have concentrations of K and Cl much higher, 14 to 18 times, than the
domestic water wells (EPA p 34). Potassium concentration ranged from 43.6 to 53.9 mg/l and Cl
concentration averaged 466 mg/l (Id.). The drilling additives reported by EPA to have been used at
Pavillion had a much lower concentration for both anions. The fracking fluid contained several
compounds with high concentrations of both ions (Id.). Therefore, the high concentrations of K and Cl
suggest contamination with fracking fluid.
The chloride concentration data plotted in EPA Figure 12 shows clearly that Cl concentration in two of
the three samples from EPA’s deep monitoring wells are much higher than those in domestic wells, and
EPA correctly assesses there must be a cause other than natural variation for the high concentrations.
However, in this case I disagree with EPA’s assessment that “regional anion trends tend to show
decreasing Cl concentrations with depth” (EPA p 19) because EPA Figure 12 shows little variation with
depth although there are a couple of high concentration outliers near the surface. Regardless of the
interpretation of trend, concentrations from the EPA monitoring wells plot far higher than the Cl data
from domestic wells.
The chloride concentrations reported from the EPA monitoring wells are also much higher than reported
by the USGS in their Wind River study (Daddow 1996). He describes the formation water as having TDS
concentration as high as 5000 mg/l, but Cl is a small proportion of that. He also reported that the
highest Cl concentration on surface water sites was less than about 30 mg/l, so assuming the river
recharges the alluvial aquifer, the source of the groundwater is relatively clean with respect to chloride.
Cl concentrations at EPA’s monitoring wells are much higher than the regional values reported by USGS
in either ground or surface water on the Wind River Reservation, and are unlikely to be properly
considered “naturally occurring”.
For potassium, it is much clearer that the monitoring well concentrations exceed the domestic water
well concentrations by many times (EPA Figure 12, p 20).
Myers Review of DRAFT: Investigation of Ground Water Contamination near Pavillion Wyoming 12
There is too little of either K or Cl in drilling mud or additives for it to have been the source or cause of
the enrichment in the monitoring wells. Also, purging prior to sampling occurred until the specific
conductivity (SC) of the purged water reached a relative steady state (EPA Figure 9). K and Cl both
contribute to the SC of the water being sampled. Any potential contamination due to well construction
or development has most likely been purged from the system.
The high K and Cl concentrations are clearly present in the formation water near the monitoring wells.
Without a natural source as explanation, the mostly likely source is the fracking fluid which used
compounds that have high concentrations of both anions. EPA has reasonably concluded the most likely
source of elevated K and Cl is fracking fluid.
Detection of synthetic organic compounds
The EPA found in the monitoring wells significant concentrations of isopropanol, diethylene glycol,
triethylene glycol, and tert‐butyl alcohol (TBA) (in MW02). TBA was not directly used as a fracking fluid,
but “is a known breakdown product of methyl tert‐butyl ether and tert‐butyl hydroperoxide”. The first
three products are found in fracking fluid based on the material safety data sheets (MSDSs) analyzed by
EPA, but the parent compounds of TBA have not been reported as such; importantly, MSDSs, which are
the source of the fracking fluid additives lists in the report, do not list all chemicals because the formulas
are proprietary. That a chemical is missing from the list of additives is not evidence they were never in
Isopropanol was found in “concentrated solutions of drilling additives” at concentrations much lower
than detected in the monitoring wells (EPA p 35) and the others, glycols and alcohols, were not used for
None of these compounds naturally occur in groundwater. The EPA is correct in its conclusion that
there is no acceptable alternative explanation and the most likely source of these contaminants is
Detection of petroleum hydrocarbons
EPA detected benzene, toluene, ethylbenzene, and xylenes (BTEX), trimethylbenzenes, and naphthalene
at MW02 (EPA, p 35). They detected gasoline and diesel range organics at both monitoring wells (Id.).
These are not found in drilling additives, but the MSDSs showed a long list of additives in the fracking
fluid that could be the source of the contamination just cited (EPA p 35, 36). For example, a BTEX
mixture had been used in the fracking fluid as a breaker and a diesel oil mixture was used in guar
polymer slurry (Id.).
EPA rejects alternative explanations that claim that substances, used on the well or pump, caused these
contaminant detections. Specifically, the agency points out that the contact time for water with the well
or pump during purging and sampling would be so low that contamination would be unlikely, especially
after purging. This would be especially true for the Phase 4 sampling which would have occurred after
Myers Review of DRAFT: Investigation of Ground Water Contamination near Pavillion Wyoming 13
the well had been purged for sampling twice and had several months of natural groundwater flow
An alternate explanation considered by EPA is that the constituents are due to the groundwater being
above a natural gas field. In fact, the EPA has noted that historically some wells encountered gas at
levels shallower than the monitoring wells. EPA encountered methane while logging MW01 (EPA p 11).
EPA notes that the gas from the Wind River formation is “dry and unlikely to yield liquid condensates”
(EPA p 36). They also argue that the monitoring wells have substantially different compositions of liquid
condensates, which would not result if they came from a common source of gas. The explanation is
reasonable, unless there is a variation with depth. Because these contaminants occur only at low
concentrations in the deepest domestic wells, the data does not rule out a natural gradient from the gas
sources at depth to the shallower zones of the formation. However, the EPA explanation is supported
by the fact that the monitoring wells are far enough apart, more than a mile, that they must have
different gas well sources and represent different pathways..
Recommendation: To further decrease the uncertainty, the EPA should complete an additional sampling
event with more domestic wells sampled. It would also be desirable to have another monitor well
screened at the level of the gas wells. The EPA could then develop a concentration profile as a function
of depth and formation layer.
Breakdown products of organic compounds
EPA verified a vertical pathway by showing that organic compounds in the shallower monitoring wells
are daughter products of the organic compounds found in the deeper monitoring wells. This supports
the concept of upward migration with ongoing biologic transformation or natural degradation. It
supports the concept of an upward flow gradient. It cannot be asserted that the EPA monitoring wells
are on the same flow pathway, as they are more than a mile apart, therefore, the presence of
contaminants in the monitoring wells is evidence that there are multiple sources of contaminants at the
level of the gas production wells.
As part of this line of reasoning, the EPA presents the “hypothetical conceptual model” that “highly
concentrated contaminant plumes exist within the zone of injection with dispersed lower concentration
areas vertically and laterally distant from the injection points”. This refers to how the fracking fluids,
once injected, simply disperse in all directions because there are no confinements, similar to how they
disperse from coal seam fracking. It is consistent with the lower concentrations found further from the
EPA’s hypothesis is reasonable and explains the vertical movement of contaminants from a broad zone
of production wells. Its simplicity indicates that fracking in such a formation will eventually lead to
contamination moving vertically from the gas wells – it is only a matter of time (Myers, 2012).
Sporadic bonding outside of production casing and hydraulic fracturing in thin discontinuous
Myers Review of DRAFT: Investigation of Ground Water Contamination near Pavillion Wyoming 14
The last two lines of reasoning are considered together because they describe two pathways for fracking
fluid to get into the aquifer. The fracking that occurs in the Pavillion gas field directly injects fracking
fluid into an underground source of drinking water. Fracking occurs as little as 150 m below the bottom
of the deeper water wells. The sandstone and intervening shale zones are discontinuous, which
suggests there are no significant continuous barriers to a vertical component of flow and contaminant
movement. Fracking has also occurred for up to 40 years, so the pathways could have required up to 40
years for transport. Sporadic bonding above the zone being fracked basically means the annulus
between the production zone and surface casing may not be fully sealed with cement which may allow
gas or fluids to move vertically among formation layers. During fracking, the high pressure could force
some of the fracking fluid through improperly sealed well bores to contaminate formations nearer the
Both of these lines of reasoning correctly describe potential pathways and sources of fluids in the
aquifer. The EPA’s conclusions in this regard are reasonable and appropriate and conform to the
available facts and data.
Gas in Monitoring and Shallow Wells
Many shallow water wells have gas concentrations that exceed expected background levels. EPA also
uses several lines of reasoning to conclude that gas has migrated to domestic wells from the fracked
zones, in addition to or instead of it occurring naturally in those wells.
Isotopic composition of gas samples from shallow wells, deeper monitoring wells and produced gas are
all similar in that all have a thermogenic origin. However, the shallower domestic water wells have very
little higher chain carbon‐based gas, which suggests some dispersion and decomposition with vertical
movement (ethane and propane degrade faster). The isotopic composition of most wells is thermogenic
and indicative of a deep source; URS (2010) noted that methane in one domestic well of eight sampled
with measurable methane had biogenic origins.
EPA also found that the concentration of methane in domestic water wells was generally higher in areas
of higher gas production, as counted by the number of gas wells. Although it could be coincidental
because more gas wells are constructed where more gas naturally occurs, this seems unlikely because
the presence of gas in domestic water wells shows that gas is occurring outside of the production zones
deep in the Wind River Formation or high in the underlying Fort Union Formation. Gas would only move
naturally from depth to areas near the surface if there is a lack of containment which would have
depleted the gas source at some point in the last 40,000,000 years. Thus, the gas wells have apparently
provided a migration pathway for gas released by fracking into overlying formations; this migration
occurred at a rate sufficient to allow gas to accumulate to a concentration capable of causing a blowout
at 159 m bgs near well PDGW05.
The area also generally has gas well designs that are below current industry standards in some states,
with surface casing not extending below the maximum depth of water wells and with a “lack of cement
or sporadic bonding of cement outside of production casing” (EPA p 38). This would provide a pathway
from depth to at least the bottom of the surface casing, and allow gas leakage to higher levels in the
Myers Review of DRAFT: Investigation of Ground Water Contamination near Pavillion Wyoming 15
aquifer. Many states and areas require surface casing to extend below the maximum depth of USDWs
(a USDW must generally have TDS less than10,000 mg/l). The gas well design in Pavillion appears to be
below industry standards because the surface casing does not extend even below the bottom of the
zone of domestic wells. The pathways discussed above for fluid movement would also facilitate gas
The EPA acknowledges that poorly sealed domestic wells could also be a pathway (EPA p 38‐39). This is
true but not a relevant argument because the gas wells are much deeper and actually tap formation
layers with gas. Once gas reaches a domestic well, it is possible that the well provides an additional
pathway, but it is not the source of the contamination or the primary pathway from the gas source zone
to the aquifers.
The EPA also references the fact of citizen’s complaints (EPA p 39) as an indicator that gas
contamination started after fracking. Citizens do not complain until a problem occurs. Assuming their
water well was initially acceptable, they would complain when they noticed a change.
DISCUSSION OF CONTAMINANT TRANSPORT PATHWAYS
The general dispersion of contaminants upward from the fracking zone would result from either well
bore transport or transport through overlying higher permeability sandstone. Transport through
wellbores that cross multiple aquifer layers, as the gas wells do near Pavillion, would allow contaminants
to reach the different levels. However, the concentration reaching shallower formations would be much
less because the contaminants bleed off to the deeper aquifer zones (Nordbotten et al 2004). Fracking
could also create the vertical gradient to temporarily cause contaminants to move vertically upward
through wellbores to contaminate shallower aquifer layers (Lacombe et al 1995).
Because there are not any significant horizontal confining units within the Pavillion Field, the upward
vertical contaminant transport is partially due to dispersion through relatively porous media. In areas
with extensive horizontal confining layers, such as the Marcellus shale areas, transport through vertical
fractures, similar to that through wellbores, could transport substantial contaminant mass through the
impervious zones (Myers, 2012). If the bulk media bounding the fractures have conductivity less than
one hundredth that in the fracture, the contaminants will transport with little dispersion, or loss, into
the bulk media (Zheng and Gorelick, 2003).
This appears to be the case in the Pavillion Field, given the existing geology. Thus, unless fracking is very
carefully done, and well bores are solidly (not intermittently) bonded, this result is to be expected. In
the case of the Pavillion Field, sporadic bonding is revealed and reported for 9 of the wells that EPA
examined well bore data made available to them. To the extent that this is indicative of the entire field,
it would greatly increase the likelihood that transport of contaminants from the gas wells to the water
wells of the rural Pavillion residents would occur.
Myers Review of DRAFT: Investigation of Ground Water Contamination near Pavillion Wyoming 16
Daddow, R.L. 1996. Water Resources of the Wind River Indian Reservation, Wyoming. U.S. Geological
Survey Water‐Resources Investigations Report 95‐4223.
Fisher, K, and N. Warpinski. 2010. Hydraulic fracture‐height growth: real data. Paper SPE 145949
presented at the Annual Technical Conference and Exhibition held in Denver, CO, October 30 –
November 2, 2011. Doi: 10.2118/145949‐MS
Kramer, D. 2011. Shale‐gas extraction faces growing public and regulatory challenges. Physics Today 64,
no. 7: 23‐25.
Lacombe, S., E.A. Sudicky, S.K. Frape, and A.J.A. Unger. 1995. Influence of leaky boreholes on cross‐
formational groundwater flow and contaminant transport. Water Resources Research 31(8):1871‐1882.
Myers, T. 2012. Potential contaminant pathways from hydraulically fractured shale to aquifers. Ground
Water, doi: 10.1111/j.1745-6584.2012.00933.x.
Nordbotten, J.M., M.A. Celia, and S. Bachu. 2004. Analytical solutions for leakage rates through
abandoned wells. Water Resources Research v 40, W04204.
Osborn S.G., Vengosh, A., Warner, N.R., and Jackson, R.B. (2011). Methane contamination of drinking
water accompanying gas‐well drilling and hydraulic fracturing. Proceedings of the National Academy of
Sciences, v. 108, p. 8172‐8176.
Osiensky, J.L., G.V. Winter, R.E. Williams. 1984. Monitoring and mathematical modeling of
contaminated ground‐water plumes in fluvial environments. Ground Water 22, no. 3: 298‐307.
U.S. Environmental Protection Agency (EPA). 2011. Draft, Investigation of Ground Water Contamination
near Pavillion, Wyoming. Office of Research and Development, Ada, OK.
URS Operating Services, Inc. (URS) 2010. Expanded Site Investigation – Analytical Results Report,
Pavillion Area Groundwater Investigation, Pavillion, Fremont County, Wyoming, CERCLIS ID #
WYN000802835. Denver, CO.
URS Operating Services, Inc. (URS) 2009. Site Inspection – Analytical Results Report, Pavillion Area
Groundwater Investigation Site, CERCLIS ID# WYN000802735. File
Pavillion_GWInvestigationARRTestAndMaps.pdf. Denver, CO
Zheng, C., and S. M. Gorelick 2003. Analysis of solute transport in flow fields influenced by preferential
flowpaths at the decimeter scale. Ground Water 41, no. 2: 142‐155.