MEMORANDUM
Subject:

Evaluation of the Perimeter Air Monitoring Strategy and Identification of
Corrective Actions at the TVA Kingston Fly Ash Release Time-Critical Removal
Action

From:

Leo Francendese
EPA On-Scene Coordinator

To:

Steve McCracken
TVA Kingston Project Manager

Date:

January 25, 2010

The following memorandum discusses the results of the recent perimeter air monitoring
evaluation at the TVA Kingston Fly Ash Release site (the Site) and identifies corrective actions
that are required as part of these findings.
During the week of January the 11, 2010, an audit was conducted by the USEPA Region 4
Science and Ecosystem Support Division (SESD) that included a review of soil, sediment, biota
and air monitoring data, as well as a field observation of water sampling methods, being
performed by TVA. Included in the review of the air monitoring data were metals, silica, filterbased low volume PM2.5 and PM10 particulate matter, high volume particulate PM10 particulate
matter, meteorological, and continuous tapered element oscillating microbalance (TEOM) data.
No significant issues were found with any of these monitoring parameters with the exception of
the PM2.5 and PM10 filter based particulate data generated during the period from midSeptember 2009 to mid-January 2010. The draft audit findings have been documented in a
separate bulleted summary. Please see Attachment 1. These deficiencies necessitated an
evaluation of the protectiveness of the perimeter air monitoring strategy during the period in
question, as well as identification of corrective actions to be implemented to ensure similar
deficiencies are not repeated.
The findings of the quality assurance audit document that low-volume, filter media, airborne
particulate sampling at the five stationary monitoring stations failed to meet quality assurance
procedures as specified in the Dust Control and Air Monitoring Plan and accompanying Quality
Assurance Project Plan required by the Administrative Order and Agreement on Consent (AOC)
between our respective agencies. Both plans may be found at www.epakingstontva.com.
Essentially, low-volume analyses during this period were not consistent with required EPA
protocol, and will therefore not be utilized. Other data collected as part of the overall air
monitoring strategy, however, indicate that public health was protected during the time period in
question.
Effectiveness of Perimeter Air Monitoring Strategy
Redundant air monitoring systems exist on the Site and there is quality assured data
available for the September to January time period. These systems, when viewed in total and
considering the environmental conditions, indicate that the public has not been exposed to
unacceptable risk from fugitive dust migrating from the Site.
The overall perimeter air monitoring strategy comprises multiple elements categorized as
follows:
1

 •
•
•
•

Surrounding the adjacent perimeter of the Site are five fixed monitoring stations.
Mobile monitoring is performed on a daily basis.
Industrial hygiene sampling within the exclusion zone to monitor working conditions.
Work management practices that include visual on-site dust monitoring with
accompanying engineering controls.

The five fixed air monitoring stations are described and itemized in Attachment 2. Filter-based
particulate media sampling, which may include PM2.5, PM10 or both, is performed at all these
locations, and is described as the set of data that failed to meet EPA quality assurance
requirements. Station PS07 is viewed as the most critical station as it is located downwind of
the prevailing wind direction from the Site and is thus positioned to monitor worst case air
quality conditions. In addition to filter-based sampling, Station PS07 is also equipped with
Tapered Element Oscillating Microbalance (TEOM) real-time particulate monitors. TEOMs are
considered federal reference equivalent methods. PM2.5 and PM10 TEOM monitors are
maintained at Station PS07 by TVA and the Tennessee Department of Environment and
Conservation (TDEC), respectively. Data from the TEOM monitors have been determined to
meet quality assurance standards.
In addition to TEOM monitors, and because it is immediately adjacent and downwind of the Site,
Station PS07 is equipped with additional monitoring capabilities not present at the other four
locations, including equipment to sample for airborne silica and metals. Silica and metals data
generated from this station have been determined to meet quality assurance standards.
To augment monitoring provided by the five fixed stations, mobile air quality monitoring of the
Site perimeter is conducted using PM10 handheld instrumentation on a near daily basis.
Frequency of such monitoring ranges from 200 to 400 individual samples per day. The typical
spatial distribution of real-time PM10 monitoring is given in Attachment 3.
Industrial hygiene sampling is conducted on a frequent basis to determine whether workers are
being adequately protected and augments data generated through perimeter air sampling and
monitoring. As industrial hygiene sampling is performed within the actual work zone, the
generated data are viewed as representing worst case air quality conditions. Please see
Attachment 4 for the spatial distribution of these samples during this time period.
Site management includes a full-time staff whose responsibility includes frequent visual
observations to view, document, and control fugitive air emissions. Flexterra® temporary
ground cover covers 30% to 75% of the on-site extent of coal ash at any given time. Such
response actions have significantly reduced the extent of the coal ash that potentially
contributes to fugitive air emissions. Coal ash that remains uncovered is managed by
maintaining appropriate moisture to prevent migration or is graded to allow a hard crust to form
over its surface, which effectively seals the coal ash in place.
Evaluation of the effectiveness of the redundant air monitoring systems must consider
environmental conditions that existed during the timeframe in question. The risk of fugitive dust
migration is reduced to near zero during rain events and for a period of time thereafter.
Precipitation occurred on-site 58 out of 142 days from mid-September until mid-January, or
approximately 40% of the total number of days. The frequency and volume of rainfall during this
period is such that high ambient soil moisture conditions were maintained during much of the
time.
2

 Results of Measurements in Redundant Systems Indicate Public Health Protected
Prior to mid-September 2009, filter-based samples collected by TVA at all five perimeter
monitoring stations were analyzed using the correct analytical procedure. Data collected from
these five stations during the period prior to mid-September correlated well with one another
and exhibited no substantial variation from station to station. The filter-based data also
correlated well with concurrent TEOM PM2.5 and PM10 data.
From the mid-September through mid-January time period during which filter-based PM data
were found not to meet quality assurance standards, PS07 TEOM PM2.5 and PM10
measurements indicate that airborne particulate matter concentrations for both parameters were
below National Ambient Air Quality Standards (NAAQS) reference values. The Dust Control and
Air Monitoring Plan uses a project management action level of 75% of NAAQS, or 112
microgram per cubic meter (ug/m3) for PM10 and 26 ug/m3 for PM2.5 . Given the strong
correlation between TEOM and filter-based data established prior to mid-September, the
Agency is confident that mid-September thru mid-January PS07 TEOM measurements
are representative of conditions at other 4 locations around the perimeter of the Site.
The above observation alone gives the Agency confidence that the public was protected during
the mid-September to mid-January timeframe as these measurements were consistently below
NAAQS reference values. Please see Attachment 5 for a graphic representation of these
relationships as well as the TEOMs’ reliability in tracking station to station relationships.
Additional weights of evidence supporting this determination are as follows:
•

Thousands of mobile PM10 instantaneous readings during the time period of midSeptember to mid-January also are significantly below NAAQS reference values.

•

Approximately 150 industrial hygiene sampling events were conducted within the
exclusion (work) zone during the time period in question. The data derived from these
sampling events demonstrate that on-site dust control and air quality objectives were
being met in that these measurements were significantly below on-site health and safety
triggers. These data also support the reasonable conclusion that Site conditions during
this period would not promote fugitive emissions that would adversely impact perimeter
ambient air quality conditions.

•

Filter-based media sampling data for silica and metals met quality assurance
requirements during the subject time period and indicate that airborne concentrations for
these parameters were below reference values.

•

Precipitation events occurred 40% of the time during this period. Assuming that
moisture content was retained for a period of time after each, it is reasonable to
conclude that Site conditions were such that fugitive dust emission risks were low for
Site conditions for greater than 40% of the period in question.

•

Rigorous on-site management practices for dust suppression have significantly reduced
the coal ash that is subject to wind entrainment and dispersion.

3

 Corrective Actions Underway
While, based on quality assured data available from Station PS07 and from other data
generated through the multiple and redundant monitoring strategy, the Agency is confident that
the public was protected during the mid-September to mid-January period, it is imperative that
corrective action be taken to ensure that all data collected at the Site meets applicable federal
guidelines. The following corrective actions will be conducted:
•
•
•
•
•
•

TVA will submit an investigative report regarding this incident.
EPA audit staff will participate in TVA’s QA weekly conference calls with its analytical
laboratories.
EPA audit staff will conduct audits of TVA contract laboratories on a periodic basis.
EPA will conduct periodic independent sampling and performance audits of TVA’s air
monitoring network.
TVA will submit a Corrective Action Plan and schedule for accelerating the validation of
raw data in order that it can be released for review on a more timely basis.
TVA will submit a Corrective Action Plan for upgrading its perimeter stations to include
additional TEOM monitors.

EPA acknowledges that TVA has already taken immediate action to terminate use of the
laboratory that generated the non-assured data and looks forward to implementation of the
corrective actions specified above. EPA will also consider any additional steps that TVA might
offer to improve the Site air monitoring strategy.

4

 Attachment 1

Audit Summary for the TVA Kingston, Tennessee Fossil Plant
EPA Science and Ecosystem Support Division
January 11-14, 2010
Auditors:
Ray Terhune, Quality Assurance Section
Greg Noah, Air and Superfund Section
Tim Slagle, Air and Superfund Section
Linda George, Air and Superfund Section
Jeff Hendel, Quality Assurance Section

Introduction
On January 11-14, 2010, the EPA Science and Ecosystem Support Division (SESD)
conducted an audit of the TVA Fossil Plant located in Kingston, Tennessee at the request
of Leo Francendese, On-Scene Coordinator for the TVA Kingston Fly Ash Spill
response. The audit included a review of water, sediment, ash, tissue and air monitoring
data as well as field observations of water sampling methods. Included in the review of
the air monitoring data were metals, silica, filter based low volume PM2.5 and PM10
particulate matter, high volume PM10 particulate matter, meteorological, and continuous
TEOM data. Data collected from April 2009 to December 2009 were included in this
audit.

Summary of Audit Observations and Recommendations
No significant issues were found during the audit of sampling activities for the surface
water, sediment, ash or air. There were also no significant issues found with the
laboratory data for the surface water, sediment, tissue or ash samples. Nor were there any
problems found with the laboratory data for the metals in air. However, there was a
notable exception with a portion of the laboratory analysis of the PM2.5 and PM10 filter
based air particulate data. The low-volume PM2.5 and PM10 filter based sampling at the 5
fixed particulate monitoring stations that occurred from September 16, 2009 to January
13, 2010 analyzed by the Bureau Veritas contract laboratory failed to meet required
quality assurance requirements and analysis methodology required by the Air Monitoring
Plan and accompanying quality assurance project plan. All low-volume PM2.5 and PM10
filter analysis conducted before September 16, 2009 by the Inter-Mountain Laboratory
showed acceptable quality assurance and conformance to analysis methodology. EPA
strongly recommends that the data analyzed by the Bureau Veritas contract laboratory for
the September 16 to January 13 timeframe be “R” flagged.
The auditors also noted that the amount of time that elapses between the air sampling
events and the data being released from TVA’s Quality Assurance may lead to future
potential excessive data loss if other similar laboratory errors occur. The auditors

1

 recommend TVA reduce the time it takes to complete the quality assurance reviews from
3-months to 30-days if possible.

Audit Observations
Water, Sediment, Ash, and Tissue
Sixteen random data packages were reviewed of various matrixes (water, sediment, ash,
and tissue (several species). Data packages from several laboratories were reviewed
including: TestAmerica - Nashville, TN; TestAmerica - Knoxville, TN; TestAmerica,
Pittsburgh, PA; and Pace Analytical - Green Bay, WI. The primary analyses of interest
were metals and nutrients (ammonia, TKN, etc) and physical properties (TSS, TDS,
turbidity, pH, etc).
•
•
•
•
•

•

The data was acceptable for nutrients in the reviewed packages in all matrices
tested (water).
The data was acceptable for physical analyses in the reviewed packages in all
matrices tested (water).
The Quality Assurance requirements for laboratory analyses were more stringent
in the data analysis plan than the analytical method requirements for each analyte.
There were numerous checks and balances between the laboratories and the data
review team with excellent communication between all parties.
Metal analysis method/procedure blanks for tissue samples had numerous low
level hits of analytes of concern. This indicates a minor contamination issue with
the procedures. Although this is quite common in tissue preparation it does cast
some doubt on low level concentration values of metals in samples.
The electronic data and deliverables are stored on two redundant servers and are
backed up four times daily.

Field Sampling Methods
•
•
•
•
•
•
•

The sampling procedures were followed as written with no deviations noted.
The SOPs have been updated and reviewed as needed when project operations
change.
The field logs reviewed were detailed and complete.
One sample in ten is collected for quality assurance (matrix spikes and/or
duplicates) which surpasses the EPA requirement of one sample in twenty.
The calibration room was set up very well and the procedure for performing the
day’s calibration was very efficient.
Hydrolab probes were adequately rinsed between standards. The Hydrolab data
was stored in the memory of the instrument at each of the sampling locations.
In addition to storing the information in the memory of the instrument, the field
measurements were also recorded in the field notebook. This redundancy is a
good practice in the event of a memory failure within the hydrolab software.

2

 •
•
•
•

Safety glasses were not utilized during calibrations which involve the handling of
standards, reagents, and/or sample bottle preparation in the calibration and sample
preparation rooms.
Eyewash stations were not present at the instrument calibration and sample
preparation rooms.
Three equipment volumes should be purged prior to sample collection.
The filter cartridge should be purged prior to sample collection.

3

 Air Monitoring Data
Several random data packages of several monitoring parameters were reviewed including
low volume particulate data (PM2.5 and PM10), Tapered Element Oscillating
Microbalance (TEOM) continuous particulate data (PM2.5 and PM10), meteorological,
metals, and silica data. Data packages from several laboratories were reviewed
including: Inter-Mountain Laboratory (IML) – Sheridan, WY; Bureau Veritas, Novi, MI;
and Galson Laboratories – Syracuse, NY. Observations for each monitoring parameter
are below:
Low Volume Particulate Data
Two contract laboratories have been used to weigh PM10 and PM2.5 low volume
samples during the time of this project. IML was contracted for filter weighing
from the start of the ambient perimeter monitoring network through midSeptember, and Bureau Veritas was used from mid-September through the time of
the audit. The reviews for each laboratory are below.
Inter-Mountain Laboratory (IML)
•

•

Several samples were traced from the pre-weighing session, through field
handling, and post weighing to determine if essential documentation was
present and acceptable for validation. All records and quality assurance
records were present for validating these samples.
The records were not adequately organized by the record keeper and the
review took more time than expected for this exercise. The data packages
and field data should be catalogued and organized in a logical manner to
speed review.

Bureau Veritas
•

Several samples were traced from the preweighing session, through field
handling, and post weighing to determine if essential documentation was
present and acceptable for validation.
• No laboratory blank data or duplicate data was reported, and the working
standard weights did not effectively bracket the expected weight of the
filter media.
• The data packages were not clearly organized.
• TVA reported that several filter weighings showed significantly high
negative numbers that are impossible to achieve in routine monitoring.
• TVA observed and showed EPA a wider range of scatter than previously
observed among measurements at the TVA ambient air monitoring sites.
The above observations prompted a laboratory audit of the Bureau Veritas
laboratory by TVA. The following findings are the critical elements of
deficiency recorded by TVA during the Bureau Veritas Laboratory audit.

4

 1. Inadequate static control in the laboratory was observed.
2. Temperature and humidity requirements as specified by 40 CFR Part
50, App.L Sec 8.2 were not followed.
3. Pre- and post-weighing humidity criteria as specified by 40 CFR Part
50, App.L Sec 8.3.3 were not met.
4. No lab blanks or replicate weighings were conducted as specified in 40
CFR Part 50, App.L Sec 8.3 and 7.11.
5. The balance check standards were inappropriate to properly bracket
the filter media weight and confirm the repeatability of the balance.
6. The balance was configured to only weigh to 5 places (0.00000g)
instead of the required 6 places (0.000000g).
7. From interviewing the TVA auditors, their opinion was that the
laboratory was using a “NIOSH 0600-type” weighing procedure
instead of the EPA Method Specified in the Dust Control and Air
Monitoring Plan. The TVA auditors concluded that the laboratory
Bureau Veritas was not familiar with EPA filter weighing
methodology. However, further review of the data by the EPA
auditors concluded that the laboratory made significant deviations
from the NIOSH 0600 Method.
EPA strongly recommends that the low volume PM2.5 and PM10 data
analyzed by the Bureau Veritas contract laboratory for the September 16
to January 13 timeframe be “R” flagged.
Continuous PM2.5 Tapered Element Oscillating Microbalance (TEOM) Data
•
•
•
•

•

The PM2.5 TEOM data is automatically transmitted electronically to the
TVA Environmental Compliance & Modeling Support Group located in
Muscle Shoals, AL.
The data is reported in real-time and is posted on the TVA Kingston
website.
The TEOM is audited approximately every two months, and the quality
assurance checks (flow, temperature, barometric pressure) are acceptable.
However, EPA recommends monthly flow checks for the TEOM.
Data collected on September 10 through 14 show several consecutive zero
measurements indicating a possible instrument problem. The review
process should be augmented to include daily reviews and investigation
and communication of any anomalous measurements.
The data was found to be acceptable.

Meteorological Data
•

The meteorological data is automatically transmitted electronically to the
TVA Environmental Compliance & Modeling Support Group located in
Muscle Shoals, AL.

5

 •
•

The Environmental Compliance & Modeling Support Group has audited
the meteorological equipment quarterly, and the monitor has passed each
audit.
No issues were found with data integrity or quality assurance.

Metals Data
•

Several data packages were reviewed from both IML and Bureau Veritas
and found no issues with data integrity or quality assurance.

Silica Data
•

Several data packages were reviewed from Galson Laboratories and no
issues were found with data integrity or quality assurance.

6

 Attachment 2

Particulate Monitoring Stations

e

CTEH

®

Harriman High School - FRM PM10

PS06
FRM PM2.5

!
(
!
(

PS07
FRM PM2.5 (2)
FRM PM10 (2)
TISCH Hi Vol PM10 (2)
TISCH TSP (2)
SKC PCXR8
PM2.5 TEOM
TDEC Operated
TISCH TSP (1)
PM10 TEOM
MetOne SASS

!
(
PS08
FRM PM2.5

PS05
FRM PM2.5

!
(

!
(
PS09
FRM PM2.5
FRM PM10

!
(

Analytical Sampling Stations - Instruments
Ash Extent 12/23/2008

Stationary
Particulate Monitoring
Stations

I

FIGURE 1
0

0.125

0.25

CTEH
Project No.

0.5 Miles

4735

®

Tennessee Valley Authority
Kingston, TN
Roane County
Date Printed: 07/28/2008

 Attachment 3

PM10 Mobile Monitoring

Real Time PM10 Mobile Monitoring Samples
(10/01/09 - 10/08/09)

0

0.5

1

2
Miles

Date of Imagery:
Plant Area: 12/01/2009
Other Area: 2008

Tennessee Valley Authority
E&T - Environmental Resources
Geographic Information & Engineering

 Attachment 4

Industrial Hygiene Monitoring

Industrial Hygiene Monitors

0

0.125

0.25

0.5
Miles

Date of Imagery: 12/01/2009

Tennessee Valley Authority
E&T - Environmental Resources
Geographic Information & Engineering

 Attachment 5

40
TVA PS07 PM 2.5
35

TVA TEOM PM 2.5

PM 2.5 Limit (35 ug/M3)

35

30
26
Values in ug/M3

25

20

15

10

5

0

Week Ending

Station 7 PM 2.5 FRM vs TEOM

Graph 1

75% of the NAAQS

 160
150
140
TVA PS07 PM 10

TDEC PS07 TEOM (PM 10)

PM 10 Limit (150 ug/M3)

120
112

Values in ug/M3

100

80

60

40

20

0

Week Ending

Station 7 PM 10 ‐ FRM vs TEOM

Graph 2

 

75% of the NAAQS

 View Magnified 533%
30

Values in ug/M3

TVA PS07 PM 10

TDEC PS07 TEOM (PM 10)

PM 10 Limit (150 ug/M3)

15

0

Week Ending

Station 7 PM 10 ‐ FRM vs TEOM
Graph 2a

Action Level (112 ug/M3)

 40

35

35

TVA PS05 PM 2.5

TVA PS06 PM 2.5

TVA PS07 PM 2.5

TVA PS08 PM 2.5

TVA PS09 PM 2.5

TVA TEOM PM 2.5

PM 2.5 Limit (35 ug/M3)

75% of the NAAQS

30
26
Values in ug/M3

25

20

15

10

5

0

Week Ending

Proportional Relationship Between Stations for PM 2.5
Graph 3

 160
150
140

120

TVA PS09 PM 10

TVA PS07 PM 10

PM 10 Limit (150 ug/M3)

75% of the NAAQS

TDEC PS07 TEOM (PM 10)

112

Values in ug/M3

100

80

60

40

20

0

Week Ending

Proportional Relationship Between Stations for PM 10

Graph 4

 View Magnified 533%

Values in ug/M3

30
TVA PS09 PM 10

TVA PS07 PM 10

PM 10 Limit (150 ug/M3)

Action Level (112 ug/M3)

TDEC PS07 TEOM (PM 10)

15

0

Week Ending

Proportional Relationship Between Stations for PM 10

Graph 4a