/ £% "t,
U.S. ENVIRONMENTAL PROTECTION AGENCY
OFFICE OF INSPECTOR GENERAL
Improving air quality
EPA Needs to Improve
Its Emergency Planning
to Better Address
Air Quality Concerns
During Future Disasters
Report No. 20-P-0062
December 16, 2019

-------
Report Contributors:	James Hatfield
Gabrielle Fekete
Seth Gerhart
Julie Narimatsu
Abbreviations
AEGL
Acute Exposure Guideline Level
AMCV
Air Monitoring Comparison Value
ASPECT
Airborne Spectral Photometric Environmental Collection Technology
CCP
Crisis Communication Plan
EPA
U.S. Environmental Protection Agency
ESF
Emergency Support Function
NAAQS
National Ambient Air Quality Standards
OIG
Office of Inspector General
ppm
Parts Per Million
SLAMS
State and Local Air Monitoring System
SSM
Startup, Shutdown, Malfunction
TAGA
Trace Atmospheric Gas Analyzer
TCEQ
Texas Commission on Environmental Quality
Cover Photo: Residential neighborhood in Houston, Texas, with industrial facilities in the
background. (OIG photo)
Are you aware of fraud, waste or abuse in an
EPA program?
EPA Inspector General Hotline
1200 Pennsylvania Avenue, NW (2431T)
Washington, D.C. 20460
(888) 546-8740
(202) 566-2599 (fax)
OIG Hotline@epa.gov
Learn more about our OIG Hotline.
EPA Office of Inspector General
1200 Pennsylvania Avenue, NW (2410T)
Washington, D.C. 20460
(202) 566-2391
www.epa.gov/oiq
Subscribe to our Email Updates
Follow us on Twitter @EPAoig
Send us your Project Suggestions

-------
x-^to sr4j.
*. U.S. Environmental Protection Agency	20-P-0062
^ ¦¦ \ Office of Inspector General	December 16,2019
mZ'
\»
At a Glance
Why We Did This Project
We conducted this audit to
determine whether the air
quality monitoring and related
activities conducted in the
greater Houston area by the
U.S. Environmental Protection
Agency (EPA) and the state of
Texas:
•	Addressed potential
high-risk areas.
•	Indicated any potential
health concerns.
•	Accurately communicated
air monitoring results and
potential health concerns to
the public.
On August 25, 2017, Hurricane
Harvey made landfall on the
U.S. Gulf Coast as a
Category 4 storm. Many of the
Houston area's air monitors
were shut down and secured
prior to the storm's landfall to
prevent damage. The EPA and
state and local agencies
subsequently conducted mobile
monitoring to assess air quality
conditions, including the levels
of hazardous air pollutants,
which are also called air toxics.
This report addresses the
following:
•	Improving air quality.
Address inquiries to our public
affairs office at (202) 566-2391 or
OIG WEBCOMMENTS@epa.gov.
List of OIG reports.
EPA Needs to Improve Its Emergency Planning
to Better Address Air Quality Concerns
During Future Disasters
Developing EPA guidance for
collecting and communicating air
quality data could improve public
confidence in the agency during
future disaster responses.
What We Found
Most air toxic emission incidents during
Hurricane Harvey occurred within a 5-day
period of the storm's landfall. The majority
of these emissions were due to industrial
facilities shutting down and restarting
operations in response to the storm and
storage tank failures. However, state, local and EPA mobile air monitoring
activities were not initiated in time to assess the impact of these emissions.
Additionally, once started, monitoring efforts did not always generate data
considered suitable for making health-based assessments, in part because there
was no guidance outlining how to monitor air quality following an emergency.
The air monitoring data collected did not indicate that the levels of individual air
toxics after Hurricane Harvey exceeded the health-based thresholds established
by the state of Texas and the EPA. However, these thresholds do not consider
the cumulative impact of exposure to multiple air pollutants at one time. Further,
the EPA's thresholds are based on short-term exposure to a single air pollutant
and do not consider lifetime exposures. Consequently, the thresholds may not be
sufficiently protective of residents in communities that neighbor industrial facilities
and experience repeated or ongoing exposures to air toxics.
We did not identify instances of inaccurate communication from the EPA to the
public regarding air quality after Hurricane Harvey. However, public
communication of air monitoring results was limited. As a result, communities
were unaware of the agency's activities and data collection efforts. This lack of
awareness can diminish public trust and confidence in the EPA.
Recommendations and Planned Agency Corrective Actions
We recommend that the Assistant Administrator for Land and Emergency
Management develop guidance for emergency air monitoring in heavily
industrialized areas, develop a plan to provide public access to air monitoring
data, and assess the availability and use of remote and portable monitoring
methods. We also recommend that the Region 6 Regional Administrator develop
a plan to inform communities near industrial areas of adverse health risks and to
limit exposure to air toxics in these communities, and conduct environmental
justice training. We further recommend that the Associate Administrator for Public
Affairs establish a process to communicate the resolution of public concerns.
Two of our six recommendations are resolved with corrective actions pending.
The remaining four recommendations, which we revised after we issued our draft
report, are unresolved pending receipt of corrective action plans from the EPA.

-------
UNITED STATES ENVIRONMENTAL PROTECTION AGENCY
WASHINGTON, D.C. 20460
OFFICE OF
INSPECTOR GENERAL
December 16, 2019
MEMORANDUM
SUBJECT: EPA Needs to Improve Its Emergency Planning to Better Address
Air Quality Concerns During Future Disasters
Report No. 20-P-0062
[yj/XfaMA
FROM: Charles J. Sheehan, Acting Inspector General
TO:
See Attached List
This is our report on the subject audit conducted by the Office of Inspector General (OIG) of the
U.S. Environmental Protection Agency (EPA). The project number for this audit was OA&E-FY18-0266.
This report contains findings that describe the problems the OIG has identified and corrective actions the
OIG recommends. This report represents the opinion of the OIG and does not necessarily represent the
final EPA position. Final determinations on matters in this report will be made by EPA managers in
accordance with established audit resolution procedures.
The EPA provided acceptable corrective actions and milestone dates for two recommendations:
Recommendation 5, which is addressed to the Associate Administrator of Public Affairs, and
Recommendation 6, which is addressed to the Regional Administrator for Region 6. In accordance with
EPA Manual 2750, both recommendations are resolved, and no further response to these
recommendations is required.
Action Required
We consider four recommendations to be unresolved: Recommendations 1 through 3, which are addressed
to the Assistant Administrator for Land and Emergency Management, and Recommendation 4, which is
addressed to the Regional Administrator for Region 6. In accordance with EPA Manual 2750, you are
required to provide a written response to this report within 60 calendar days. You should include planned
corrective actions and completion dates for the four recommendations that need additional information for
resolution. Your response will be posted on the OIG's website, along with our memorandum commenting
on your response. Your response should be provided as an Adobe PDF file that complies with the
accessibility requirements of Section 508 of the Rehabilitation Act of 1973, as amended. The final response
should not contain data that you do not want to be released to the public; if your response contains such
data, you should identify the data for redaction or removal along with corresponding justification.
We will post this report to our website at www.epa.gov/oig.
Addressees
Peter Wright, Assistant Administrator for Land and Emergency Management
Ken McQueen, Regional Administrator for Region 6
Corry Schiermeyer, Associate Administrator for Public Affairs

-------
EPA Needs to Improve Its Emergency
Planning to Better Address Air Quality
Concerns During Future Disasters
20-P-0062
Table of C
Chapters
1	Introduction		1
Purpose		1
Background		1
Responsible Offices		12
Scope and Methodology		12
Prior OIG Reports		14
2	Better Planning Was Needed to Coordinate Air Quality
Monitoring Efforts		16
Monitoring Not Conducted During Most Air Toxic Emission Incidents		16
Some Data Considered Unusable for Health Assessments
Due to Monitoring Duration		18
EPA Lacked Guidance for Emergency Air Quality Monitoring Efforts		20
Conclusion		21
Recommendations		21
Agency Response and OIG Evaluation		22
3	Data Did Not Indicate That Air Toxic Levels Were Exceeded,
but Health Risks to Fenceline Communities from
Emission Spikes Are Unknown		23
EPA Used State Thresholds to Assess Houston's Air Quality		23
Monitoring Thresholds Do Not Consider Exposure
to Multiple Pollutants		24
Conclusion		26
Recommendation		27
Agency Response and OIG Evaluation		27
4	Lack of Communication Left Communities Unaware of Risks		28
Guidance Outlines Community Engagement During an Incident		28
EPA Deployed Community Liaisons		29
Residents Were Not Informed How EPA Resolved Their Concerns		30
Environmental Justice Not Adequately Addressed in Emergency
Response Implementation		32
Conclusion		33
Recommendations		33
Agency Response and OIG Evaluation		34
Status of Recommendations and Potential Monetary Benefits		35
- continued -

-------
EPA Needs to Improve Its Emergency	20-P-0062
Planning to Better Address Air Quality
Concerns During Future Disasters
Appendices
A Agency Response to Draft Report	 36
B Distribution	 43

-------
Chapter 1
Introduction
Purpose
The Office of Inspector General (OIG) for the U.S. Environmental Protection
Agency (EPA) conducted this audit to determine whether the EPA's and the state
of Texas' air quality monitoring and related activities after Hurricane Harvey
(1) addressed potential high-risk areas, (2) indicated any potential health
concerns, and (3) were accurately communicated to the public with respect to
monitoring results and potential health concerns.
Background
On August 25, 2017, Hurricane Harvey made landfall on the U.S. Gulf Coast as a
Category 4 storm, dropping over 19 trillion gallons of rain across the region
(Figure 1). During this unprecedented weather event, the highest total rainfall in
the nation's history—60.58 inches—was recorded near Nederland, Texas, about
90 miles east of Houston. According to state officials, more than 270,000 homes
were impacted, with approximately 80,000 homes inundated with at least
18 inches of water. Hurricane Harvey was the most expensive natural disaster in
more than a decade and the second costliest in U.S. histoiy, causing an estimated
$125 billion in damage.
Figure 1: Region impacted by Hurricane Harvey
City of
/Houston

Source: OIG analysis using Esri's ArcMap, a mapping and location analytics platform.
20-P-0062
1

-------
According to the United States Global Research Program's most recent climate
assessment,1 "heavy precipitation events in most parts of the United States have
increased in both intensity and frequency since 1901 and are projected to continue
to increase over this century." Further, "the heaviest rainfall amounts from intense
storms, including hurricanes, have increased by 6% to 7%, on average, compared
to what they would have been a century ago." Similarly, a study published in the
Proceedings of the National Academy of Sciences of the United States of America
indicates that the annual probability of rainfall in excess of 19 inches has
increased sixfold since the late 20th century.2 Thus, the likelihood that the EPA,
states and local governments will have to continue to respond to disasters similar
to Hurricane Harvey has also increased.
Air Quality Impacts of Hurricane Harvey
Before Hurricane Harvey made landfall, many industrial sources of air pollution—
such as oil and gas production facilities—shut down their operations in anticipation
of heavy rainfall and flooding. When industrial facilities shut down or restart their
plant operations, significant spikes in air pollutants—including hazardous air
pollutants, which are also referred to as air toxics—can result. These spikes are
often referred to as startup, shutdown, malfunction (SSM) emissions.
Many industrial facilities affected by Hurricane Harvey were forced to make last-
minute decisions regarding whether to shut down because of the uncertain course
of the storm. Facilities in Corpus Christi, Texas, which is located southwest of
Houston, were forecasted to be in the storm's path and were able to coordinate
shutdown activities early, thus reducing SSM emissions. However, the hurricane's
course toward Houston was not as clear. When the storm did make landfall, it
stalled over southeastern Texas, leading to massive flooding. Many facilities in
Houston, therefore, were shutting down within 24 hours of when the heavy
rainfall began. After the storm passed and flooding subsided, all the facilities that
shut down resumed normal operations.
According to excess emissions reports voluntarily submitted to the Texas
Commission on Environmental Quality (TCEQ) by impacted facilities in Harris
and Jefferson counties, Hurricane Harvey resulted in industrial facilities releasing
an extra 340 tons of air toxics.3 These emissions were from accidents, facility
1	USGCRP, 2017: Climate Science Special Report: Fourth National Climate Assessment, Volume I [Wuebbles, D.J.,
D.W. Fahey, K.A. Hibbard, D.J. Dokken, B.C. Stewart, and T.K. Maycock (eds.)]. U.S. Global Change Research
Program, Washington, DC, USA, 470 pp, doi: 10.7930/J0J964J6.
2	Kerry Emanuel, "Assessing the present and future probability of Hurricane Harvey's rainfall," Proceedings of the
National Academy of Sciences of the United States of America 114, no. 48 (November 28, 2017): 12681-84.
3	Excess emissions are self-reported by facilities to the TCEQ. The reporting rule requiring these submissions was
suspended during and for 7 months after Hurricane Harvey. Thus, the total emissions reported likely underrepresent
the total excess emissions due to Hurricane Harvey. For example, only 13 of nearly 400 major industrial facilities
operating in Harris and Jefferson counties reported excess emissions due to facility shutdowns or startups during the
hurricane. Of these 13 facilities, six reported only emissions related to a shutdown event.
20-P-0062
2

-------
shutdowns during the hurricane and facility startups after the hurricane. For
example:
•	A gasoline spill at Magellan in the Galena Park Terminal released an
estimated 282 tons of combined air toxics, including over 6 tons of benzene.
•	A floating roof tank failure at Valero released an estimated 12.5 tons of
combined air toxics.
•	During a startup event, the Flint Hills Resources Port Arthur Facility
released 0.89 tons of air toxics.
•	During a shutdown event, the ExxonMobil Beaumont Refinery released
0.07 tons of air toxics.
The impact on air quality concerned community members and health officials.
Short-term exposure to air toxics such as benzene can cause drowsiness;
dizziness; headaches; irritation to eyes; skin and respiratory tract problems; and,
at very high levels, unconsciousness and death. In addition, residents who live
near Houston-area industrial facilities already experience chronic exposure to
high levels of air pollution.
Health Impacts in Fenceline Communities
According to a study published in Environmental Science and Technology, the
health impacts of direct and indirect particulate matter emissions from SSM
events in Texas were estimated to cost $148 million in 2015.4 An analysis of air
pollution risks in the greater Houston area conducted for the Houston Mayor's
Task Force on the Health Effects of Air Pollution reached the following
conclusion:
East Houston neighborhoods that face a number of vulnerabilities
based on their marginal social and economic standing also carry a
heavier burden of health risks from breathing pollutants in their air.
They tend to be located closer to major point sources than most
other neighborhoods in the Greater Houston area and to be nearer
to major transportation corridors.
Air pollution can lead to health effects that often go unaddressed in communities
where residents have limited financial and health care resources. Further,
residents of fenceline communities—neighborhoods that are directly next to a
facility and are directly impacted by the facility's operations, including air
emissions—are often unable to relocate because of low home values. The lack of
resources and the disproportionate layering of intersecting social factors create
additional challenges in these communities when faced with a weather event like
Hurricane Harvey.
4 Zirogiannis, Nikolaos, Alex J. Hollingsworth and David M. Konisky, "Understanding Excess Emissions from
Industrial Facilities: Evidence from Texas," Environmental Science and Technology 52, no. 5 (2018): 2482-90.
20-P-0062
3

-------
Industrial Makeup and Demographics of Greater Houston Area
The greater Houston area encompasses nine counties along the Gulf Coast in
southeastern Texas and is the fifth-most populous metropolitan statistical area in
the United States, with a population of over 6 million people as of 2014 (Figure 2).
The Houston area is also a major industrial center and is home to hundreds of
petrochemical facilities, including two of the four largest petroleum refineries in
the United States. According to the Mayor's Task Force on Health Effects of Air
Pollution, the massive petrochemical complex along the Houston Ship Channel is
the largest in the country, and the Port of Houston is the sixth largest port in the
world and is the second largest in the country in terms of total tonnage. These
facilities release several types of air pollutants, including air toxics that can cause
cancer or other serious health problems.
Figure 2: Houston population estimates and industrial air polluter locations
(as of 2017 and 2019, respectively)
Greater
Houston Area
o
Miles
0 115230 460 690 920
O
Austin
Waller
Chambers
Legend
O Major Industrial Air Polluters
• Other Industrial Air Polluters
Population Estimate (People)
< 50,000
50,001 -100,000
| 100,001 -350,000
| 350,001 - 760,000
¦ > 4,500,000
| Surrounding Counties



Miles
40
0 5 10
20
30

Gulf of Mexico
Service Layer Credits: Sources: Esri, GEBCO,
NOAA, National Geographic, Garmin, HERE,
Geonames.org, and other contributors
Esri, Garmin, GEBCO, NOAANGDC, and
other contributors
Source: OIG analysis using Esri's ArcMap.
20-P-0062
4

-------
Houston Ship Channel. (OIG video)
The National Air Toxics Assessment is the EPA's periodic estimate of the
public's cancer and noncancer health risks from long-term exposure to air toxics
in the United States. The most recent estimate of national average cancer risk—
the 2014 National Air Toxics Assessment5—was estimated as 30 in 1 million.
This estimate has not historically accounted for SSM emissions, however. As
noted on the EPA's National Air Toxics Assessment website, the assessment "may
not accurately capture sources that emit only at certain times (e.g., ... startups,
shutdowns, malfunctions and
upsets)." Still, for 2014, this
screening tool estimated elevated
risk levels for all census tracts in
the Houston area,6 with a
county wide average cancer risk of
45.89 in 1 million but with some
cancer risks estimated as high as
348 in 1 million. Most of the
Houston area's highest risk census
tracts were in East and Southeast
Houston.
The Houston area is unusual in
Houston community center playground neighboring an industrial	thai due to a lack of zoning
facility, with smokestacks in the background. (OIG video)	. ' '	. . . .
requirements—many residential
communities are located next to or near industrial sources of air pollution. The
number and density of industrial sources and their proximity to residents
contribute to the elevated health risks in the Houston area. The area's fenceline
5	The EPA released the 2014 National Toxics Assessment on August 22, 2018. The assessment is based on air toxics
emissions for calendar year 2014.
6	Per the U.S. Census Bureau, a census tract is a small, relatively permanent statistical subdivision of a county for
the purpose of presenting data. Census tracts nest within counties, and their boundaries normally follow visible
features but may follow legal geography boundaries and other nonvisible features in some instances. Census tracts
ideally contain about 4,000 people and 1,600 housing units.
20-P-0062
5

-------
communities are also often environmental justice communities,7 which are
communities predominantly comprising minority and low-income residents.
For example, as shown in Figure 3 below, the Harrisburg/Manchester
neighborhood in Harris County in East Flouston sits along the Houston Ship
Channel, home to several industrial emitters of wastewater, air contaminants and
hazardous waste. According to the Mayor's Task Force on Health Effects of Air
Polluti on, this neighborhood routinely exceeded safe levels for seven of the 12 air
pollutants that the task force deemed "definite risks." Furthermore, the
Harrisburg/Manchester neighborhood is surrounded by major transportation
corridors. Both the Sidney Sherman Bridge, which services Interstate 610 over the
Houston Ship Channel, and multiple rail tracks run through the community.
Figure 3: Houston's Harrisburg/Manchester neighborhood
Houston*
Houston
Legend
Service Layer Credits: Sources: Esri. GEBCO, NOAA.
National Geographic. Gsrmin. HERE. Geonames.org.
ana other contributors
Esri, Garmin, GEBCO. NOAA NGDC. and other
contributors
•	Major Roads
The City of Houston
| Harrisburg/Manchester Neighborhood
•	Houston Ship Channel
Source: OIG analysis using Esri's ArcMap.
Union Pacific rail tracks, Houston, (OIG photo)
? Environmental justice is defined by the EPA as "the fair treatment and meaningful involvement of all people
regardless of race, color, national origin, or income, with respect to the development, implementation, and
enforcement of environmental laws, regulations, and policies."
20-P-0062
6

-------
Ship Channel Bridge, Houston. (OIG video)
In addition to the inherent vulnerability of the community's location,
Harrisburg/Manchester residents face several socioeconomic challenges.
According to the U.S. Census Bureau American Community Survey, 2013-2017,
more than 25 percent of the neighborhood's residents live at or below the poverty
line. Approximately 37 percent of Harrisburg/Manchester residents, ages 16 to 64,
were either unemployed or worked less than 6 months in 2017. More than
one-third (36 percent) of Harrisburg/Manchester residents ages 25 to 64 reported
that they had not graduated from high school. Finally, in 2017, about 22.5 percent
of the population age 5 and above speak English "not well" or "not at all."
EPA Assisted Texas' Response to Hurricane Harvey under the
Stafford Act
On August 25, 2017, the President declared a major disaster in Texas at the
request of the Texas Governor. This declaration allowed the federal government
to assist local emergency responders under the authority of the Robert T. Stafford
Disaster Relief and Emergency Assistance Act and under the direction of the
U.S. Federal Emergency Management Agency. The federal government supports
state and local entities during an emergency response, consequently, the TCEQ
served as the lead agency for the Hurricane Harvey environmental response.
To coordinate Flurricane Harvey response activities, a unified command was
established among the EPA, the TCEQ, the General Land Office of Texas and the
U.S. Coast Guard to oversee the evaluation and cleanup of spills, releases and
orphan containers. This command was supported by three operational branches in
Corpus Christi, Houston and Port Arthur. In addition, the EPA's Emergency
Operations Center serves as the agency's emergency response operational focal
point for all its emergency response efforts, as well as a communication hub to
increase data management and coordination capabilities. The EPA also staffed
on-scene coordinators to monitor or direct responses to all oil spills and hazardous
substance releases reported to the federal government. The on-scene coordinators
20-P-0062
7

-------
worked with, provided support to and disseminated information to local, state and
regional response communities regarding all federal efforts.
National Incident Management System and Response Framework
The federal government's response to a national disaster is guided by the National
Incident Management System and the National Response Framework, which work
together to provide a comprehensive approach to domestic incidents (Figure 4).
The National Incident Management System provides management and
organizational structures—such as the Incident Command System—to assist
operations across jurisdictions and disciplines. The Incident Command System is
a management structure that assists in managing resources, making decisions and
assigning responsibilities. It also establishes a chain of command detailing how
authority and information flow during an incident. Under the Incident Command
System, the Incident Commander has overall responsibility for the incident; for
determining incident objectives; and for establishing priorities based on the nature
of the incident, the resources available and agency policy.
Figure 4: National Incident Management System and National Response Framework
National Incident
Management System
• Incident Command
System
o Command
o Planning
o Operations
o Logistics
o Finance
National Response Framework
Emergency Support Functions
Mechanisms to provide federal
resources and capabilities to
support state and local
responders
Support Annexes
Essential supporting aspects of
the federal response common to
all incidents

Incident Annexes
Incident-specific applications of
the framework
Partner Guides
Next level of detail in response
actions tailored to the actionable
entity

Source: EPA analysis of Federal Emergency Management Agency information.
The National Response Framework is composed of 15 Emergency Support
Functions (ESFs) that detail how agencies implement their capabilities and
coordinate the resources required in a national response. For Hurricane Harvey,
the Federal Emergency Management Agency activated EPA Region 6 under
ESF #10, Oil and Hazardous Materials Response, on August 28, 2017. ESF #10
"includes the appropriate actions to prepare for and respond to a threat to public
health, welfare, or the environment caused by actual or potential oil and
hazardous materials incidents."
20-P-0062
8

-------
Each ESF contains a range of possible mission assignments for federal agencies
activated to respond to a national disaster. ESF #10 actions can include those to
"prevent, minimize, or mitigate a release"; "detect and assess the extent of
environmental contamination, including environmental monitoring"; and
"stabilize the release and prevent the spread of contamination." Under EPA
Order 2071, National Approach to Response, which documents agency policy for
the National Incident Management System, the EPA's role under ESF #10 may
include air quality sampling and monitoring.
In addition to EPA Order 2071, the EPA's response to national emergencies is
governed by EPA Order 2010, Crisis Communication Plan (CCP). The CCP
outlines the process for the EPA to coordinate and communicate environmental
information to the public. The EPA initiated its CCP under ESF #15—External
Affairs—on August 28, 2017, "to ensure rapid response to providing coordinated,
accurate, up-to-date information regarding its field activities." Figure 5 shows the
EPA's roles under the National Response Framework.
Figure 5: EPA's roles under the National Response Framework
The EPA's role under ESF #10 and EPA
Order 2071 may include:
•	Detect, identify, contain, clean up, or dispose of released
oil or hazardous materials	
•	Removal of drums, barrels, tanks, or other bulk
containers that contain oil or hazardous materials	
•	Household hazardous waste collection
•	Monitoring of debris disposal
» Water quality monitoring and protection	
•	Air quality sampling and monitoring
•	Protection of natural resources
Source: National Response Framework and EPA Order 2071.
Note: Yellow highlighted text indicates the EPA's roles.
Through ESF #15, the National Response Framework delivers "coordinated,
prompt, reliable, and actionable information" on threats and hazards to the entire
affected community to "expedite the delivery of emergency services and aid the
public in taking protective actions." Per EPA Order 2071, the EPA's role under
ESF #15 "integrates Public Affairs and the Joint Information Center,
Congressional Affairs, Intergovernmental Affairs (state, local, tribal and
territorial), Planning and Products and the Private Sector under the coordinating
auspices of external affairs." The order also says that the Joint Information Center
"ensures the coordinated release of information," while the "Planning and
Products component of external affairs develops all external and internal
communications strategies and products."
NRF Emergency Support
Functions
#1
Transportation
#2
Communications
#3
Public Works and Engineering
#4
Firefighting
#5
Information and Planning
#6
Mass Care, Emergency Assistance,
Temporary Housing, and Human Services
#7
Logistics
#8
Public Health and Medical Services
#9
Search and Rescue
#10
Oil and Hazardous Materials Response
#11
Agriculture and Natural Resources
#12
Energy
#13
Public Safety and Security
#14
Superseded
#15
External Affairs
20-P-0062
9

-------
¦¦	¦ . ¦ ¦	In addition to measuring criteria air pollutants, the TCEQ's
TCEQ air monitor in Houston. _T .,	.	,, , ^	,-rr	¦
(OIG photo)	SI.AVIS routinely collects data tor over 100 ditterent air toxics to
determine whether their levels exceed Air Monitoring Comparison
Value (AMCV) thresholds established by the TCEQ. If a TCEQ SLAMS monitor
detects a chemical concentration that exceeds its associated AMCV threshold,
adverse health effects in the public are not necessarily anticipated. However, the
TCEQ considers these data during any future permitting process.
Air Monitoring Conducted after Hurricane Harvey
Managed by the TCEQ, the state and local air monitoring system
(SLAMS) network in Texas collects data about six criteria air
pollutants to determine whether air quality meets the National
Ambient Air Quality Standards (NAAQS)
established by the EPA.8
There are adverse health effects associated with each of the six
criteria air pollutants. For example, short term exposure to ozone is
associated with deaths from respiratory causes, while long-term
exposure to ozone is linked to asthma aggravation and
development, as well as permanent lung damage.
Starting on August 23, 2017, before
Hurricane Harvey made landfall, the
TCEQ began preparations to shut down
its SLAMS sites and monitors in the
Houston area to protect the network from
storm damage. Once the storm was over,
the TCEQ began taking steps to restore its
air monitoring operations. By
September 13, 2017, most of the air
monitoring network in the Houston area
was once again operational. By
September 29, 2017, Houston's network
was 100 percent operational.
8 The Clean Air Act. as amended, requires the EPA to set NAAQS for pollutants considered harmful to public health
and the enviromnent. The EPA established NAAQS for six principal pollutants, which are called criteria air
pollutants', carbon monoxide, lead, ground-level ozone, nitrogen dioxide, particulate matter and sulfur dioxide. Per
the act, states are responsible for maintaining an air quality monitoring network to provide "timely air quality data
upon which to base national assessments and policy decisions." The Clean Air Act also requires each state to have a
state implementation plan to attain and maintain the NAAQS. Many of these state implementation plans (such as
Texas') included provisions that govern SSM events and provided automatic exemptions from enforcement for
facilities whose SSM emissions violate the Clean Air Act standards. In 2015, the EPA found that the SSM provisions
included in the state implementation plans for Texas and 35 other states were "substantially inadequate" to meet
Clean Air Act requirements {State Implementation Plans: Response to Petition for Rulemaking, 80 Fed. Reg. 33840,
33845 (June 12, 2015)). However, in April 2019, EPA Region 6 proposed to deviate from the agency's finding and
allow Texas to maintain its existing SSM provisions. As of October 2019, the EPA was revising its SSM policy.
Top to bottom: TAGA bus. ASPECT aircraft.
( PA photos)
20-P-0062
10

-------
Although the SLAMS can provide useful air quality information during or after an
emergency, these fixed, stationary networks were not specifically designed for
that purpose and may not be able to withstand emergency conditions. An
emergency response may therefore require portable, remote sensing or other
monitoring techniques to obtain air quality data, especially for those locations and
pollutants not regularly monitored by existing networks. Existing technologies—
such as the EPA's Airborne Spectral Photometric Environmental Collection
Technology (ASPECT) and the EPA's Trace Atmospheric Gas Analyzer
(TAGA)—provide alternative solutions to this issue by either analyzing remote
infrared and photographic imagery or by directly collecting pollutant
concentrations using gas chromatography.
After Hurricane Harvey, the EPA and the city of Houston used a variety of
temporary monitoring methods to capture conditions around industrial sites.
These efforts included monitoring conditions next to industrial fencelines with
handheld instruments, such as toxic vapor analyzers, summa canisters, optical gas
imaging cameras and portable multi-gas monitors. In addition, from August 31
through September 11, 2017, the EPA conducted
flyovers of facilities with the ASPECT plane, screening
pollutant plumes for potential hazardous emissions near
high priority industrial targets. The agency also drove a
TAGA bus throughout the impacted region from
September 6 through 20, 2017. Additional air
monitoring was conducted using portable instruments
by a firm under contract to the Environmental Defense
Fund, which is a nongovernmental organization.
Although this private monitoring was not conducted at
the request of the EPA or state and local agencies, the
results were made available to the EPA and the TCEQ.
In a September 8, 2017, press release, the EPA and the
TCEQ informed Houston communities that available
data collected around the Valero facility indicated that
local residents should not be concerned about air
quality issues related to the effects of the storm. The
EPA issued six press releases related to fuel waivers,
four related to water or Superfund issues, and six that
specifically addressed air toxic exposure concerns
Top: Valero facility fencing displaying	j d tQ gn explosion and fires at the Arkema plant in
community banner. Bottom: Community park	v	1
and housing adjacent to Valero facility in the
background. (OIG photos)
20-P-0062
11

-------
Crosby, Texas.9 The six press releases related to Arkema, some of which were
issued jointly with the TCEQ, informed members of the public about the fire and
chemical release; assured them that the TCEQ and the EPA were monitoring the
smoke and air quality; and advised them to limit their exposure by staying
indoors, keeping their doors and windows closed, and continually running their
air conditioners. On September 1, 2017, an EPA press release stated that neither
aerial surveillance nor ground-level air quality monitoring "found toxic
concentration levels in areas away from the evacuated facility."
Responsible Offices
The EPA's Office of Emergency Management, within the Office of Land and
Emergency Management, develops and implements regulations related to
emergency management and is central to the EPA's emergency preparedness and
response efforts. The Office of Emergency Management also maintains valuable
air quality assets that can be used during emergencies.
EPA Region 6 worked directly with the TCEQ and other government and
nongovernmental stakeholders in the overall emergency response effort and,
specifically, the air monitoring response effort.
The EPA's Office of Public Affairs within the Office of the Administrator is
responsible for coordinating the agency's external message for emergency
response activities.
Scope and Methodology
We conducted our audit from August 2018 through July 2019. We conducted this
audit in accordance with generally accepted government auditing standards.
Those standards require that we plan and perform the audit to obtain sufficient,
appropriate evidence to provide a reasonable basis for our findings and
conclusions based on our objectives. We believe that the evidence obtained
provides a reasonable basis for our findings and conclusions based on our
objectives.
We encountered an impediment to obtaining all the desired information to
complete our audit, as described below. We were still able to obtain enough
information to answer our objectives, although this impediment impacted our
9 The Arkema plant manufactures organic peroxides. Due to extensive flooding from Hurricane Harvey, the plant
lost power, backup power and critical organic peroxide refrigeration systems. On August 31, 2017, organic peroxide
products stored inside a refrigerated trailer decomposed, causing the peroxides and the trailer to burn. After the
vapor from the decomposing products traveled across a public highway adjacent to the plant, 21 people sought
medical attention from exposure to the fumes. Over the next several days, a second fire and a controlled burn
consumed eight more trailers holding Arkema's remaining organic peroxide products. During these three fires, over
350,000 pounds of organic peroxide combusted, and more than 200 residents living within 1.5 miles of the facility
evacuated the area and could not return home for a week. A U.S. Chemical Safety and Hazard Investigation Board
report (No. 2017-08-1-TX). issued May 2018, provides more details on the Arkema explosion and fires.
20-P-0062
12

-------
ability to analyze all air quality data and to definitively determine the rationale for
certain decisions. We believe that the evidence obtained provides a reasonable
basis for our findings and conclusions based on our audit objectives.
To understand the EPA's responsibilities during emergency situations, we
reviewed the following statutes, policies, guidance and documents:
•	The Clean Air Act, as amended.
•	The Stafford Act.
•	The Emergency Planning and Community Right-to-Know Act.
•	The National Response Framework.
•	The National Incident Management System.
•	EPA Order 2071, National Approach to Response.
•	EPA Order 2010, Crisis Communication Plan.
•	EPA press releases.
•	EPA internal documents related to emergency response.
We also conducted interviews with staff from EPA Region 6, the Office of Land
and Emergency Management, the Office of Air and Radiation, and the Office of
Research and Development. We discussed emergency response activities at the
county and city levels with officials representing Harris County and the city of
Houston. Finally, we discussed the EPA, state and local emergency responses
with nongovernmental organizations and community members.
To understand how and when air monitoring occurred, we collected and analyzed
air toxic data from several sources, including the EPA's Air Quality System,
TAGA bus and ASPECT aircraft; the TCEQ's Air Emission Event Report
database; the city of Houston; and Entanglement Technologies, a private company
under contract with the Environmental Defense Fund. We compared these data to
the TCEQ's short-term AMCVs and the EPA's Acute Exposure Guideline Levels
(AEGLs) to identify any potential health impacts of Harvey-related air
emissions.10 We also compared the location, timing and duration of the
monitoring with reported excess emissions incidents to identify any potential data
gaps in areas of elevated air emissions.
After the hurricane, the EPA's Office of Emergency Management and Region 6
developed after-action reports based on online surveys, written questionnaires and
interviews with EPA response personnel. These reports identified areas of
strength, lessons learned and recommendations to be used in future EPA
responses. We reviewed these documents and developed an OIG survey to assess
10 The TCEQ maintains two sets of AMCVs: short-term comparison values and long-term comparison values. Short-
term AMCVs are based on acute (short-term) health effects data and are used to evaluate air quality averaged over
short time frames (e.g., 30 minutes to 1 hour), while long-term AMCVs are based on chronic health effects data and
are used to evaluate air quality averaged over a year or more. The EPA's AEGLs describe the human health effects
from once-in-a-lifetime, or rare, exposure to airborne chemicals. The AEGLs are generally used by emergency
responders when dealing with chemical spills or other catastrophic exposures.
20-P-0062
13

-------
the effectiveness of the EPA's communications regarding air quality in response
to Hurricane Harvey. This survey was designed to determine whether the EPA's
on-the-ground response and Harvey-related EPA communications were effective.
We distributed the survey to 59 EPA staff who served as community liaisons
during the response. We received 44 responses and analyzed the data.
Impediment to Obtaining Information
TCEQ staff, managers and officials declined to meet with us to discuss their
response to the hurricane and their reasoning for various decisions or actions
described in this report. We provided the TCEQ with an initial list of questions
before scheduling a meeting at TCEQ offices in September 2018. The TCEQ
cancelled the meeting the day before we were scheduled to meet due to an
impending tropical storm. Also, the TCEQ declined to meet with us during a
subsequent week when we visited the Houston area to meet with city officials and
community representatives from impacted areas. Further, despite several
conversations to arrange for written answers to our initial list of questions, we
never received a response from the TCEQ. Subsequent to our unsuccessful
attempts to arrange meetings and obtain information from the TCEQ, we learned
that the TCEQ collected air monitoring data from helicopter flyovers following
Hurricane Harvey. We were unable to review those data as a part of this audit.
However, we believe that the evidence we obtained provides a reasonable basis
for our findings and conclusions based on our audit objectives.
Prior OIG Reports
EPA OIG Report No. 2006-P-00033. Lessons Learned: EPA's Response to
Hurricane Katrina, issued September 14, 2006, identified deficiencies in the EPA's
coordination with state and local officials, as well as in the EPA's use of its
floodwater database. The OIG recommended, among other things, interagency
meetings and training for EPA Region 6 and state staff on the Incident Command
System and the ESFs. The agency agreed with the OIG's recommendations and
implemented appropriate corrective actions.
EPA OIG Report No. 19-P-0236. Region 6 Quickly Assessed Water Infrastructure
after Hurricane Harvey but Can Improve Emergency Outreach to Disadvantaged
Communities, issued July 16, 2019, found that EPA Region 6 conducted extensive
preparation activities and forged close working relationships with state emergency
response partners well before Hurricane Harvey made landfall. This preparation
enabled Region 6 to protect human health and water sector resources as part of its
Hurricane Harvey mission assignment. The OIG identified one area for
improvement—staff outreach to residents of vulnerable communities—that would
further enhance the region's emergency response capabilities. The OIG
recommended, among other things, that the EPA Region 6 Regional Administrator
include environmental justice outreach in planning and pre-landfall preparation
exercises by gathering data to determine the population, unique needs and
20-P-0062
14

-------
challenges of vulnerable communities. The agency agreed with the OIG's
recommendations and, as of October 2019, was in the process of implementing
appropriate corrective actions.
EPA OIG Report No. 20-P-0010. EPA Adequately Managed Hurricane Harvey
Funding Receivedfrom FEMA, issued on October 23, 2019, found that the EPA
effectively managed its Hurricane Harvey Disaster Relief Funding. The OIG did
not identify any significant issues in the EPA's contracting, logistics or resource
acquisition processes. The OIG noted that the agency had already identified
strengths and areas for improvement and had implemented corrective actions in
response to the OIG's recommendations in its 2006, 2008 and 2014 reports
regarding its emergency responses. The OIG made no recommendations to the
agency in this audit.
20-P-0062
15

-------
Chapter 2
Better Planning Was Needed to Coordinate
Air Quality Monitoring Efforts
The EPA, the state of Texas and the city of Houston lacked guidance and
procedures for conducting air quality monitoring in response to an emergency. As
a result, their ability to assess and communicate air quality-related health risks to
the public during and after the Hurricane Harvey emergency response was
limited. The nature of an emergency response requires flexibility and cannot be
predetermined. However, EPA guidance would help future efforts address when,
where and how long to monitor air quality; the minimum quality assurance
needed to obtain data that can be used to assess health risks; and other issues
related to air monitoring. Although the data from Hurricane Harvey monitoring
efforts did not exceed the health-based thresholds used during the response (e.g.,
the TCEQ's AMCVs), pre-emergency planning and coordination by the EPA and
the TCEQ could lead to more effective monitoring and communication during
future emergency responses.
Monitoring Not Conducted During Most Air Toxic Emission Incidents
In response to the Hurricane Harvey disaster, a nongovernmental organization,
local governmental entities and the EPA collected air monitoring data with
four distinct mobile monitoring efforts over a span of 21 days (August 3 I
September 20, 2017). Despite the broad range of monitoring efforts, this
monitoring:
•	Did not coincide with most industry-reported air toxic emission incidents
occurring during the disaster.
•	Sometimes used ineffective techniques to collect data. For example, a
nongovernmental organization collected samples over a duration too short to
analyze whether the concentrations were harmful to
human health.
Over half of all known air toxic emission incidents
began when no monitors were operating.
Companies in the Houston area reported over
319 tons of air toxic emissions due to Harvey-
related SSM activities. However, when these
facilities were shutting down and when the first
malfunctions and air toxic emissions occurred, most
of the TCEQ's monitors in Houston's air
monitoring network had been turned off and
secured to protect them from storm damage.
Known Air Toxic Releases Over Time

(8/20/2017 through 9/20/2017)
I Total Air Toxics Emitted:
0 tons

Total Incidents
0 incidents
Legend

Air Monitor Method
® ¦« Ml


*~1
Air Toxic Releaus
(tons)

•	D1 1
•	'
# ....

	—
Video showing air toxic releases and monitoring
methods used overtime. (OIG video)
20-P-0062
16

-------
Figure 6 illustrates the different air monitoring efforts during the Hurricane
Harvey emergency response,11 as well as the asset owner/operator. Our
comparison of these monitoring timelines to the TCEQ's repository of self-
reported SSM emission data revealed that most air toxic emission incidents
occurred from August 26 through 31. 2017—after the TCEQ disabled its SLAMS
in the Houston area and before the EPA began collecting data with its ASPECT
flight response. Many of the air toxic emissions during the peak incident period
were from storage tank leaks due to excessive rainfall. However, since these
reported emissions occurred before temporary monitoring had begun or the
SLAMS was redeployed, we were unable to assess their impact on air quality.
Figure 6: Monitoring efforts and air toxic emission incidents during the Hurricane Harvey response
EPATAGA
# of SLAMS
Operational
(Out of 5)
EPA ASPECT
# of SLAMS Operational (Out of 5)
Hurricane
City of Houston Mobile Ambient Air Monitoring Laboratory
• • • •
Entanglement
Air Toxic Release
Incidents
I
oo co
(N m ^
I tH
en oi cn CT)
lo io oo
r-r—^
Ol Ol (J1 Ol Q (J)
Source: OIG analysis.
Notes: This chart includes only SLAMS monitors capable of detecting air toxics, not NAAQS monitors.
ASPECT operation dates are based on actual data submitted to the OIG.
An example of an air toxic emission incident during the peak incident period was
Valero Partners' roof tank failure. This incident—which released an estimated
12.5 tons of air toxics, including benzene, hexane and toluene—began on
August 27, 2017, when all SLAMS monitors were offline and before emergency
monitoring had begun. The Arkema Crosby Plant explosion, another widely
publicized event, occurred on August 31, 2017, before the EPA's TAGA bus or
the city of Houston's Mobile Ambient Air Monitoring Laboratory had been
deployed. At the time of the explosion, only the ASPECT was operational.
11 Although NAAQS monitors were also offline during this time, given our audit focus on air toxics, we did not
extensively assess criteria air pollutants. However, according to an Environmental Integrity Project report, based on
self-reported data in the State of Texas Environmental Electronic Reporting System ozone precursor emissions were
high along the Gulf Coast following Hurricane Harvey. Per the report, from August 23 through September 1, 2017,
approximately 3.9 million pounds of volatile organic compounds were released into the Houston region by surrounding
industries, and "[n]itrogen oxides totaled about 154,000 pounds during the same period in the Houston region."
20-P-0062
17

-------
As demonstrated in Table 1, communities located close to industries faced an
increased likelihood of exposure to SSM emissions during the emergency response
period. For example, 38 percent of all known air toxic emission incidents due to
Hurricane Harvey that were reported by Houston-area industries occurred fewer
than 4 miles from the Harrisburg/Manchester neighborhood in East Houston.
These incidents accounted for over 93 percent (a total of nearly 300 tons) of all
known air toxic emissions occurring in Harris County during the disaster, despite
this geographical region accounting for only 4.5 percent of the county.
Table 1: Proximity of air toxic emissions to Harrisburg/Manchester, August 20-
September 20, 2017
Category
Value
Tons released in a 4-mile radius of Harrisburg/Manchester
298.71
Total tons released in Harris County
319.97
Percent of Harris County emissions released in a 4-mile radius of
Harrisburg/Manchester
93%
Source: OIG analysis of industry data reported to TCEQ.
In 2018, the EPA's Office of Air and Radiation amended the National Emission
Standard for Hazardous Air Pollutants for petroleum refineries to require that,
starting in January 2019, these facilities report their monitoring data for benzene
concentrations at the perimeters of their facilities.12 The monitoring and reporting
requirements were not yet in place when Hurricane Harvey hit Houston.
However, the monitors used to collect the benzene data could provide useful
information for assessing air quality impacts related to future emergency
responses in the Houston area and other industrialized locations. These monitors
are also relatively cost-effective and replaceable if damaged, unlike the TCEQ's
SLAMS monitors. These low-cost sensors could therefore be used in fenceline
communities during emergency situations.
Some Data Considered Unusable for Health Assessments Due to
Monitoring Duration
Governmental and nongovernmental organizations collected data to evaluate the
region's air quality after the hurricane by comparing these data to existing health-
based air quality thresholds. The results of these comparisons were used to assess
whether the air quality was likely to result in adverse human health effects.
However, due to quality control-related reasons, the TCEQ did not use much of
the data collected to make health-based assessments. Table 2 shows which data
collected could not be used to make health assessments related to local air quality.
12 83 Fed. Reg. 60696, November 26, 2018.
20-P-0062
18

-------
Table 2: Usability of data collected during Hurricane Harvey for health-based assessments
Monitoring asset
Asset owner
Monitoring type
Data usable?

TAGA
EPA
Temporary mobile monitor
No

ASPECT
EPA
Temporary mobile monitor
No

SLAMS
TCEQ
Permanent stationary monitors
Yes

Mobile Ambient Air
Monitoring Laboratory
City of Houston
Temporary mobile monitor
No a

Portable monitoring
Entanglement
Technologies
Temporary mobile monitor
No b
Source: OIG analysis.
Note: The colors differentiating the assets in this table correlate with the colors used in Figure 6, which illustrates
when the monitoring efforts using these assets were conducted.
a The laboratory's monitoring time frames were sufficiently long enough to produce data useable for health-based
assessments; however, the TCEQ disqualified the data because the onboard global positioning system failed.
b Some monitoring time frames were sufficiently long enough to produce data useable for health-based
assessments; however, the TCEQ determined that most time frames were too short. The monitor must be
active for at least 30 minutes to 1 hour to be usable for health-based assessments.
Although the EPA's TAGA operation was primarily intended to screen for
elevated air toxic concentrations, the data collected by this method were also
compared against the TCEQ's short-term AMCV thresholds (described in
Chapter 1) to make health-based assessments. Although the EPA, the TCEQ and
the city of Houston assessed that the data indicated there was no concern—and
subsequently issued a press release communicating this assessment to the
public—we found that the TAGA's sampling time frame was too short to generate
data that could accurately assess airborne toxin concentrations for making health-
based assessments.13 In addition, we found that the data collected by the TAGA
operation were not timely. Before the TAGA buses were activated by the EPA's
Emergency Operations Center for Hurricane Harvey, they were parked in
Las Vegas, Nevada, and Research Triangle Park, North Carolina. Their transit to
the Houston area after they were activated impaired the timeliness of the data
collection.
Entanglement Technologies supported the assessment of air quality following
Hurricane Harvey's landfall by using a portable monitor from September 4
through 9, 2017. These data were submitted to the TCEQ and the city of Houston
for review. However, the TCEQ concluded that the data were unsuitable for
making health-based assessments because most air samples were collected over a
period lasting fewer than 5 minutes. The EPA also conducted handheld
monitoring in Manchester from September 3 through 8, 2017. However, this
handheld monitoring collected data on only one air toxic (benzene), and no
readings exceeded the method detection limit (effectively 0 parts per million
[ppm]).
ASPECT data are intended only for screening purposes, as this monitoring
method (i.e., remote sensing) does not provide sufficiently reliable data for health-
13 Short-term AMCVs require monitoring data to be averaged for a 30-minute to 1-hour period prior to comparing
the data to the air quality thresholds. The TAGA monitoring method averages data for only 1-2 seconds.
20-P-0062
19

-------
based assessments. As a part of this screening process, the EPA dispatched
follow-up ground monitoring units and established evacuation zones as
necessary to protect human health when elevated pollutant concentrations were
detected. This follow-up occurred, for example, on September 2, 2017, when the
ASPECT detected benzoyl peroxide concentrations above the ASPECT'S method
detection limit.
EPA Lacked Guidance for Emergency Air Quality Monitoring Efforts
Emergency air monitoring efforts were initiated without a plan to help guide and
coordinate governmental and nongovernmental efforts, including the minimum
level of quality assurance needed to obtain data suitable for health-based
assessments and how to effectively share data among all interested parties.
While many entities collected air monitoring data in the weeks following
Hurricane Harvey's landfall, the data acquisition itself was not performed in a
manner that would provide a holistic picture of air quality in the Houston region:
1.	Despite efforts by Entanglement Technologies and the city of Houston to
share information with the TCEQ, the TCEQ did not forward these raw
datasets to the EPA. We also found no evidence that the EPA requested
access to these data or that these data were shared with the public.
2.	The raw data collected by the EPA via the TAGA were stored in the
agency's Environmental Response Team Information Management
System, a data repository that can only be accessed by the EPA team
members.
3.	The EPA's ASPECT flight data were retained in the Environmental Unit
of the EPA's Office of Emergency Management, with the air toxic
concentration values stripped from the dataset.
4.	Although the EPA presented via press releases that some preliminary
analyses of data were received, the raw data were never publicly
distributed.
Ultimately, this isolation of raw data limited analysts' and the public's ability to
perform monitoring data comparisons and make informed and comprehensive
conclusions regarding the region's overall air quality.
Even if these monitoring datasets were housed in a central database that was
accessible to all interested parties, the unique formatting of each dataset would
have presented substantial challenges in terms of data interpretation. For example,
the ASPECT'S concentration values were split into 97 separate Excel
spreadsheets. Furthermore, we found that concentration values were
20-P-0062
20

-------
inconsistently expressed using a range of units, such as parts per million, parts per
billion, milligrams per cubic meter and micrograms per cubic meter.
Although EPA Region 6 and the TCEQ collaborate annually to plan and train for
hurricanes, the EPA lacked both internal and external guidance on how to
appropriately collaborate with others to collect, assess and store air quality data
during extreme weather events or other emergency situations. A focus on air
quality monitoring when planning for disasters in industrial cities like Houston
would facilitate the timely, proper and collaborative use of alternative monitoring
devices.
Conclusion
Overall, the EPA's lack of monitoring guidance and various technological
limitations prevented nongovernmental organizations, local governmental entities
and the EPA itself from monitoring air quality during the peak period of excess
emissions due to Hurricane Harvey. Further, the monitoring data that were
collected were not always useful for assessing potential impacts on human health.
Additionally, inconsistent formatting and isolated storage of air monitoring data
prevented the EPA, the public and other stakeholders from gaining a holistic
understanding of air quality.
The EPA could better plan and coordinate future emergency response efforts with
governmental and nongovernmental organizations to help ensure that the air
quality in potentially high-risk areas is monitored during periods of elevated air
toxic emissions. During the Hurricane Harvey response, high-risk areas were
predominantly located adjacent to or near large industrial facilities. Increased
planning and coordination could provide these communities with timely
information about their air quality during an emergency, enabling them to take
precautions to reduce their exposure to air toxics.
Recommendations
We recommend that the Assistant Administrator for Land and Emergency
Management:
1.	Develop general guidance to help state and local agencies and external
stakeholders develop air monitoring plans for emergency situations in
heavily industrialized areas so that usable data are collected in targeted
areas of concern.
2.	Develop, in coordination with the Associate Administrator for Public
Affairs, a plan for providing public access to air monitoring data collected
during an emergency response.
20-P-0062
21

-------
3. Coordinate with the Office of Research and Development and the Office
of Air Quality Planning and Standards within the Office of Air to assess
the availability and use of remote and portable monitoring methods to
monitor air toxics when stationary monitoring methods are not available.
Agency Response and OIG Evaluation
The agency disagreed with our draft report recommendations for this chapter. The
agency noted that each emergency is unique and that developing guidance that
would cover all scenarios would be challenging. Further, per the agency's response,
state and local governments are primarily responsible for emergency response
efforts, with the EPA regions assisting when requested. The agency said that the
EPA has developed a variety of tools and procedures for emergency assistance.
Based on discussions with the agency and its response to our draft report, we
revised our recommendations for the final report to better clarify the
recommendations. Recommendations 1 through 3 are unresolved pending the
OIG's receipt of acceptable corrective action plans and proposed completion dates
from the EPA in response to the final report.
The agency's response to our draft report and our additional comments are in
Appendix A. The agency provided specific suggestions for our consideration, and
we revised the report as appropriate.
20-P-0062
22

-------
Chapter 3
Data Did Not Indicate That Air Toxic Levels Were
Exceeded, but Health Risks to Fenceline
Communities from Emission Spikes Are Unknown
Although available monitoring data did not indicate that the levels of air toxics in
the Houston area during the Hurricane Harvey disaster exceeded Texas' short-
term AMCVs or the EPA's AEGLs,14 these thresholds do not consider the
cumulative impact of being exposed to multiple pollutants. Instead, the thresholds
are based on an individual exposed to one specific pollutant (e.g., benzene).
Further, the EPA's thresholds do not consider chronic exposure that some
populations, such as those residing near industrial facilities, may have already
experienced. Consequently, emergency exposure thresholds may not be
sufficiently protective of populations already experiencing chronic exposure to
multiple air toxics.
EPA Used State Thresholds to Assess Houston's Air Quality
According to EPA staff, the agency coordinates with the relevant state when an
incident occurs to determine which health-based thresholds to use when analyzing
air monitoring results. A review of internal agency documents from September 5
and 6, 2017, showed that there was confusion among EPA staff regarding whether
to use the TCEQ's short-term AMCVs or other TCEQ thresholds. Ultimately, the
TCEQ decided that the EPA should use the AMCVs after discussing the issue
with the federal agency. The TCEQ and the EPA subsequently compared air
monitoring data collected from various handheld monitors, summa canisters,
ASPECT and the TAGA bus to the AMCVs. The TCEQ also compared data
collected by the city of Houston to the AMCVs. None of the data were found to
exceed the AMCVs.
Relative to the EPA's Level 2 and Level 3 AEGLs,15 the EPA's Level 1 AEGL
thresholds most closely match the short-term AMCV thresholds, although the
differences between these threshold categories are substantial and their underlying
purposes are different. Short-term AMCVs were developed by the TCEQ to
14	As described earlier in the "Scope and Methodology" section the TCEQ's short-term AMCVs are used to
evaluate air quality averaged over short time frames (e.g., 30 minutes to 1 hour). The EPA's AEGLs describe the
human health effects from rare exposure to airborne chemicals and are generally used by emergency responders
when dealing with chemical spills or other catastrophic exposures.
15	AEGLs, which address the acute (or short-term) effects of air toxics, are established at three levels, with each level
representing the severity of health impacts. Level 1 is the lowest impact level and represents the airborne
concentration above which notable discomfort or irritation could be experienced, but the effects are not disabling and
are reversible once exposure stops. Level 2 is the middle impact level and represents the exposure level at which
irreversible harm; other serious, long-lasting adverse health effects; or an impaired ability to escape are caused.
Finally, a Level 3 exposure causes life-threatening health effects or death.
20-P-0062
23

-------
screen air quality in more general, day-to-day situations, while AEGLs were
developed by the EPA to screen situations involving a once-in-a-lifetime,
accidental exposure. As an example of the difference between these thresholds,
the AEGL Level 1 short-term (30 minutes and 60 minutes) thresholds for
1,3-butadiene is 670 ppm versus the short-term AMCV threshold of 1.7 ppm.
Thus, the use of short-term AMCVs as health-based thresholds for assessing air
quality data after Harvey was more conservative—in other words, protective of
health—than if AEGLs were used.
However, the use of state thresholds to assess adequate margins of safety could
lead the EPA to endorse different conclusions regarding public safety when air
quality conditions are similar. For example, Figure 7 shows the differences in
common air quality thresholds issued by Texas and California.
Figure 7: Comparison of Texas and California air quality threshold levels
? 12
a.
Texas AMCV	California Reference Exposure Level
Source: OIG analysis.
This lack of standardization in state air toxic thresholds could cause the EPA to
provide inconsistent advice as it supports local entities in disasters. For example,
using California's air quality thresholds, the EPA could advise local governments
in that state to issue a shelter-in-place order if monitoring results showed a
benzene concentration of 0.1 ppm. That same concentration, however, would not
have triggered any health advisories during the Hurricane Harvey response, since
the Texas' short-term AMCVs have a higher threshold for benzene.
Monitoring Thresholds Do Not Consider Exposure to Multiple Pollutants
Studies have shown that fenceline communities are exposed to a heavy daily load
of multiple pollutants beyond SSM emissions. For example, the Houston Mayor's
Task Force on Health Effects of Air Pollution found that the communities in East
Houston, which includes the Harrisburg/Manchester neighborhood, are exposed to
more high-risk pollutants than other Houston communities. In East Houston,
20-P-0062
24

-------
90 percent of the census tracts face four or more "definite-risk" pollutants,16 while
one tract in the Harrisburg/Manchester neighborhood faces seven definite-risk
pollutants. Of the greater Houston census tracts exposed to six or more definite-
risk pollutants, half of them are in East Houston. These figures suggest that these
communities—given their cumulative exposure to multiple definite-risk
pollutants—face a higher lifetime risk of cancer and chronic disease than other
Houston communities exposed to only one or two definite-risk pollutants.
During Hurricane Harvey, these East Houston communities faced exposures to
many pollutants at one time. Within a 3-hour period, the city of Houston's Mobile
Ambient Air Monitoring Laboratory identified
46 pollutant concentrations greater than 0 ppm
occuring in Manchester Park on September 4,
2017, including benzene (0.008 ppm), n-hexane
(0.096 ppm) and n-heptane (0.072 ppm). While
none of these concentrations exceeded their
respective short-term AMCVs, this example
illustrates the large number of distinct pollutants u	A 4
°	r	Houston s Mobile Ambient Air Monitoring
in the air at that time.	Laboratory. (City of Houston photo)
One limitation to using the AMCVs or AEGLs to assess health risks during an
emergency response is that neither one accounts for the following situations that
could potentially impact health:
•	Concurrent exposure to multiple air pollutants (i.e., cumulative exposure).
•	Accumulation of consecutive distinct exposures to a pollutant over time
(i.e., aggregated exposure).
As Figure 8 shows, when compared to the rest of Harris County, a
disproportionate amount of air toxic emissions reported for Hurricane Harvey
were within 4 miles of the Harrisburg/Manchester neighborhood. These residents
were potentially exposed to a variety of air toxics, such as xylene, toluene, hexane
and ethylbenzene. However, the TCEQ only tracks these incidents and assesses
the air toxics' health effects at certain exposure levels on a pollutant-by-pollutant
basis; there is no way of quantifying potential effects across the AMCV or AEGL
standards.
16 The task force defined definite-risk pollutants as "those substances for which there was compelling and
convincing evidence of significant risk to the general population or vulnerable subgroups at current ambient
concentrations." The following 12 air pollutants were classified as definite risks: ozone; fine particulate matter
(PM 2.5); diesel particular matter; 1,3-butadiene; chromium VI; benzene; ethylene dibromide; acrylonitrile;
formaldehyde; acrolein; chlorine; and hexamethylene diisocyanate.
20-P-0062
25

-------
Figure 8: Known emissions near Harrisburg/Manchester during Hurricane Harvey
N
s
Harris County •/
Legend
Air Toxic Releases
(tons)
•	< 0.37
•	0 37 - 1 32
•	1.32-3.17
•	3.17-16.54
F 282.08
| Harrisburg/Manchester
1	Mile
2	Miles
3	Miles
4	Miles
Houston Ship Channel
Service Layer Credits: Sources: Esri. GEBCO.
NOAA, National Geographic, Garmin, HERE.
Geonames.org. and other contributors
Source: OIG analysis using Esri's ArcMap.
The EPA's guidance on the development of AEGLs only relies on multiple
exposure studies when single exposure data are lacking. AEGLs may therefore
not be protective enough of disproportionately burdened communities like
Harrisburg/Manchester, given their proximity to large industrial facilities and the
number of air toxics they could be exposed to during large-scale SSM incidents
before, during and after an emergency or disaster situation. Although AEGLs
were not used to make public health assessments after Hurricane Harvey, with the
exception of California, no other states have developed acute air toxic thresholds
like Texas. The other states may therefore opt to use AEGLs to assess air quality
Based on a review of TCEQ guidance, we determined that cumulative risks from
multiple pollutant exposures are not addressed in AMCVs. While short-term
AMCVs are more protective of health than AEGLS when assessing exposure to a
single air toxic, whether these values were sufficiently protective of health is
unknown, considering the multiple pollutant exposures experienced after
Hurricane Harvey.
Conclusion
The available monitoring data did not indicate that air toxic levels during the
Hurricane Harvey disaster exceeded Texas or EPA thresholds. It is unclear,
however, whether or how SSM emissions compound the health risks of residents in
fenceline communities. Short-term AMCVs and other risk-based thresholds used
by the EPA and the TCEQ to assess the risk of emissions during Hurricane Harvey
20-P-0062
26

-------
do not account for communities that are exposed—daily and/or over the long-
term—to multiple pollutants and chronic daily exposures in addition to spikes from
large-scale SSM events.
Recommendation
We recommend that the Region 6 Regional Administrator:
4. Develop and implement, in coordination with the states, a plan to inform
residents in fenceline and nearby communities about adverse health risks
resulting from multiple facility startups and shutdowns during
emergencies and to limit these residents' exposure to air toxics.
Agency Response and OIG Evaluation
In this chapter in our draft report, we included one recommendation addressing the
use of acute exposure thresholds to assess air quality during an emergency. The
agency disagreed with this recommendation and noted that there are existing air
quality standards that the EPA uses to estimate the risks to communities for criteria
air pollutants. The agency further explained that the EPA uses its AEGLs to assess
public risk from air toxics exposure.
Our draft report also included two additional recommendations in this chapter
addressing how to limit the potential health impact of multiple shutdowns and
startups on nearby residents during an emergency. The agency noted that neither
the EPA nor the states have authority over facilities' SSM schedules. The agency
stated that the EPA coordinates with local officials, states and tribes regarding
shelter in place, evacuations or other protective measures for fenceline and nearby
communities.
Based on discussions with the agency, its response to our draft report, and internal
management discussions, we developed one recommendation for this chapter in our
final report (Recommendation 4).
Recommendation 4 is unresolved pending the OIG's receipt of an acceptable
corrective action plan and proposed completion date from the agency in response to
our final report. The agency's response to our draft report and our additional
comments are in Appendix A. The agency provided specific suggestions for our
consideration, and we revised the report as appropriate.
20-P-0062
27

-------
Chapter 4
Lack of Communication Left Communities
Unaware of Risks
We did not identify any instances of inaccurate communication regarding air
quality during the Hurricane Harvey response effort. However, we found that
official communication from the EPA regarding air quality was limited. For
example, a lack of guidance regarding how the EPA should disseminate air
quality data meant air monitoring results and air quality risks did not always reach
residents of impacted communities. In addition, the lack of a feedback mechanism
meant field staff did not communicate how the EPA resolved residents' concerns.
As a result, some communities were left unaware of important issues, which can
lead to a lack of trust and confidence in the EPA's actions and findings.
Guidance Outlines Community Engagement During an Incident
Pursuant to EPA Order 2010, Crisis Communication Plan (CCP), the agency's
Public Information Officers must consider five factors when communicating with
the public during an emergency:
1.	Community engagement.
2.	Language access.
3.	Environmental justice.
4.	Environmental data.
5.	EPA authority.
In addition, the EPA's CCP states that information provided to the public during
an incident must be understandable, timely, accurate and consistent. Further, the
CCP stresses the following points:
•	The agency will widely disseminate information concerning EPA
activities to the public.
•	Information should be developed in languages other than English under
the Commitment to Language Access Obligations in Executive
Order 13166.
•	The agency will develop information to address environmental justice as
prescribed by EPA Memorandum, Incorporating Environmental Justice
Considerations into EPA Disaster Preparedness and Response
Procedures, issued November 2, 2006.
Some EPA offices have incorporated environmental justice into their office-
specific guidance about risk communication, which the EPA defines as the
20-P-0062
28

-------
"process of informing people about potential hazards to their person, property, or
community." For example, the EPA's Office of Research and Development
produced the Risk Communication Workbook, which explains that risk
communication must "transcend barriers of literacy, language, and ethnicity to
ensure acceptance or understanding." An Office of Research and Development
document regarding risk communication during water security emergencies warns
that poor risk communication "can ... undermine public trust and confidence" and
that the goal should be to "enhance knowledge and understanding [and] build trust
and credibility." The EPA Superfund program's risk communication guidance
emphasizes that individuals perceive risk differently depending on different
factors of the risk,1 including voluntariness, controllability, familiarity, fairness,
catastrophic potential, reversibility, equity and effects on children.
EPA Deployed Community Liaisons
EPA Region 6 deployed more than 80 community liaisons to the region impacted
by Hurricane Harvey—the first instance in which so many liaisons were used by
the agency to respond to a disaster, according to an EPA staff person. These
liaisons, who were coordinated by three leaders, provided information to the
public regarding how to best protect themselves from environmental risks,
collected citizen concerns, and forwarded these concerns to EPA management.
The liaisons were not tasked with resolving environmental issues.
During the Hurricane Harvey
response effort, the EPA's
community liaisons communicated
with the public by distributing
preapproved flyers, which were
available in English, Spanish and
Vietnamese. The community
liaisons held daily meetings with
the community liaison lead18 and
maintained a dedicated
environmental justice email address that the community could use. During our
audit, we received feedback from the community that the liaisons in the Port
Arthur/Beaumont area were present and played an active role.
EPA community liaison providing information in
Houston. (EPA photo)
17	The EPA's Superfund program addresses the nation's most contaminated sites and responds to environmental
emergencies and natural disasters.
18	This individual was located in the Region 6 office in Dallas and provided updates to senior management regarding
the work of the liaisons on the ground.
20-P-0062
29

-------
Si se siguen estas pautas especificas a! poner
en la acera los escombros que
deja un huracan se acelerara
el proceso de recoleccion
Following these speofc $*Sres
When Mating rwmcane-reiaied
debns to the curt Ml make for .
a speedw removaf	4
HAULING HURRICANE-RELATED DEBRIS TO THE CURB IN P
Following these specific guidelines will make for a speed ie!
Check with your local government on what is avaiia
PICKING UP THE PIECES
PONER EN LA ACERA LOS ESCOMBROS QUE DEJA UN HURACAN EN LAS AREAS
Si sigoe estas pautas espedficss se acelerara el proceso de recoleccidn.
Confirm* con las autoridades del goblemo local lo que Ie corresponds a usf
RECOGER ESCOMBROS
EPA HOA KY
LO LlfiU CHO CAC OdNG 06 NAT SAU THlCN TAI
KEO CAC 00 DAC DO NAT SAU TRAN BAO RA IE Dl/ONG TAI (AC VUNG CO THAM GIA
lam theo ck huong dan sau day se lam cho quy trinh lay do bi di difoc nhanh Hon.
Ki^m lai v6i chinh quySn tai da pfiuong oia quy vj ve nhCrng gl hi4n cd cho mlnh.
Lam theo cic hiking dan sau day
khi keo do dac do nat sau tran
bao ra ngoai ledtKmg se lam 4
cho quy trinh lay dobodi duec
From left: English, Spanish and Vietnamese versions of EPA flyers regarding debris management. (EPA photo)
Residents Were Not Informed How EPA Resolved Their Concerns
Despite concerns about air quality and other issues in the Houston area after
Hurricane Harvey, the EPA did not adequately communicate important
information so that all impacted communities received it. A lack of information
hindered residents' ability to make informed and independent decisions to protect
their health.
Residents Expressed Concerns about Health Impacts of
Hurricane Harvey
The public expressed concern about
the health effects related to the
hurricane's impact on the community,
including drinking water quality and
air quality issues. As shown in
Figure 9, over half of the 59 EPA
staff who served as community
liaisons and responded to an OIG
survey stated that outdoor air quality
was a concern to the community.
These staff cited odors, safety, fires or
hazardous air emissions from
facilities as community concerns.
An aerial view of the flooding caused by Hurricane
Harvey in Houston on August 31, 2017.
(U.S. Department of Defense photo)
20-P-0062
30

-------
Figure 9: Community liaison survey results—outdoor air quality concerns*
Did the community have outdoor air quality concerns?
Not answered
Number of respondents = 44
Source: OIG survey analysis.
* This chart is based on the perspectives of EPA's community liaisons.
The city of Houston also received public expressions of outdoor air quality
concerns after Hurricane Harvey via the city's 311 hotline. For example, the city
received 33 odor complaints from August 27 through September 17, 2017. Many
of these complaints pertained specifically to odors emanating from refineries in
the Ship Channel area.
In addition, a few nongovernmental organizations requested air quality data from
the EPA. One of these nongovernmental organizations had contacts living in
affected communities who could reach the impacted constituency. However, the
EPA was not responsive to requests from nongovernmental organizations for air
quality data.
Resolution of Concerns Not Communicated to Affected Residents
The EPA lacked a process for providing feedback to the community after
residents' concerns were considered resolved or addressed. While response
activities were communicated daily to EPA headquarters via written reports,
community liaisons and field staff reported in the EPA's post-hurricane surveys
that this information was not being relayed to field teams and that they were not
informed whether problems were resolved. For example, one community liaison
who communicated an incident at a local refinery up the established chain of
command subsequently asked about the health risks from that incident and about
the resolution status. That community liaison told the OIG that the only response
received from the chain of command was that the TCEQ was taking care of the
situation. The community liaison expressed concern about the community and
whether it was exposed to health risks from the incident.
Over half of the community liaisons who responded to our survey reported
hearing about air quality concerns in communities, but about half also said that
20-P-0062
31

-------
the EPA did not address or only sometimes addressed the concerns they submitted
in their daily reports (Figure 10).
Figure 10: Community liaison survey results—EPA management responsiveness*
Was EPA management responsive to community concerns?
Sometimes Not answered
Number of respondents = 44
Source: OIG survey analysis.
* This chart is based on the perspectives of EPA's community liaisons. Numbers do not add to
100 percent due to rounding.
According to the survey respondents and EPA staff we interviewed, community
concerns were passed up the EPA's chain of command and were then forwarded
to the governmental party responsible for resolving the issue (e.g., air quality
concerns were forwarded to the TCEQ). Once the relevant party was notified, the
EPA considered the matter "closed." Region 6 staff from the Office of
Environmental Justice and Tribal Affairs did conduct outreach with local
government officials and community organizations;19 however, some community
liaisons reported that information about how issues were resolved was lacking. In
addition, after the EPA referred an issue, the EPA's process did not include
following up to confirm resolution of the issue and communicating that resolution
to the concerned party.
Environmental Justice Not Adequately Addressed in Emergency
Response Implementation
According to the Office of Emergency Management's 2017 Hurricane and
Wildfire Response After-Action Report, environmental justice considerations were
not adequately integrated into the Incident Command System structure. The report
recommended integrating environmental justice considerations, "such as through
coordination with nongovernmental organizations to maintain awareness of their
concerns," into the CCP.
19 In the March 2019 Region 6 realignment, this office became the Office of Communities, Tribes and
Enviromnental Assessment.
20-P-0062
32

-------
EPA-conducted questionnaires, our survey and our interviews with community
members indicated a lack of knowledge on behalf of the EPA about the needs of
the Houston region's various communities and how best to reach them. This
knowledge is especially critical for community liaisons to effectively
communicate with environmental justice communities. For example, community
liaisons should have experience with these communities so that the liaisons can
address the cultural differences, communication barriers and geographical
challenges that make some of these communities hard to reach. Knowing when
and where communities gather is also important to effectively communicate and
distribute essential information.
We also identified some concern among regional staff and managers that
information did not reach all environmental justice communities. Some residents
were not aware of the EPA's presence in these communities. Although
community liaisons were deployed into affected communities, we confirmed with
some community members that they never saw a community liaison in their
neighborhoods after Hurricane Harvey. In addition, many community liaisons and
organizations expressed concern about the lack of printed materials in languages
other than English that are spoken prevalently in the Houston area.
Conclusion
Based on the results of our review, some residents impacted by Hurricane Harvey
were unaware of air monitoring results and air quality risks during and immediately
after the hurricane. The EPA has limited guidance on how to disseminate air
quality data and lacks a feedback mechanism allowing EPA field staff to
communicate the status of concerns to affected communities.
These challenges led to limited public awareness of potential air quality issues,
which in turn could reduce public trust and confidence in the government's actions
in response to an emergency. Given the number of impacts of the hurricane—
including flooding, loss of power and the fear naturally instigated by a natural
disaster—unaddressed concerns regarding air quality likely compounded the public
perception of risks.
Recommendations
We recommend that the Associate Administrator for Public Affairs:
5. Revise the EPA's Crisis Communication Plan to include a communication
process to inform affected communities about the resolution of community
concerns raised during an emergency.
20-P-0062
33

-------
We recommend that the Region 6 Regional Administrator:
6. Conduct environmental justice training for community liaisons and
Incident Command System staff, thereby fulfilling that element of the
EPA's Crisis Communication Plan.
Agency Response and OIG Evaluation
The agency concurred with Recommendations 5 and 6 and provided acceptable
planned corrective actions and completion dates. To address Recommendation 5
(Recommendation 7 in our draft report), the EPA's Office of Public Affairs plans
to update the agency's CCP. In an email to the OIG dated December 3, 2019, the
agency clarified that its update to the CCP will include a communication process
to inform affected communities about the resolution of community concerns
raised during an emergency. To address Recommendation 6 (Recommendation 9
in our draft report), Region 6 will provide annual environmental justice training to
all EPA Region 6 employees, including emergency response personnel. The EPA
will also provide training to community involvement core team, Incident
Command staff and other appropriate community liaisons consistent with the
EPA's CCP. Recommendations 5 and 6 are considered resolved with corrective
actions pending.
The agency's response to our draft report and our additional comments are in
Appendix A. The agency provided specific suggestions for our consideration, and
we revised the report as appropriate.
20-P-0062
34

-------
Status of Recommendations and
Potential Monetary Benefits
RECOMMENDATIONS
Rec.
No.
Page
No.
Subject
Status1
Action Official
Planned
Completion
Date
Potential
Monetary
Benefits
(in $000s)
1
21
Develop general guidance to help state and local agencies and
external stakeholders develop air monitoring plans for
emergency situations in heavily industrialized areas so that
usable data are collected in targeted areas of concern.
U
Assistant Administrator for
Land and Emergency
Management


2
21
Develop, in coordination with the Associate Administrator for
Public Affairs, a plan for providing public access to air monitoring
data collected during an emergency response.
U
Assistant Administrator for
Land and Emergency
Management


3
22
Coordinate with the Office of Research and Development and
the Office of Air Quality Planning and Standards within the Office
of Air to assess the availability and use of remote and portable
monitoring methods to monitor air toxics when stationary
monitoring methods are not available.
u
Assistant Administrator for
Land and Emergency
Management


4
27
Develop and implement, in coordination with the states, a plan to
inform residents in fenceline and nearby communities about
adverse health risks resulting from multiple facility startups and
shutdowns during emergencies and to limit these residents'
exposure to air toxics.
u
Region 6 Regional
Administrator


5
33
Revise the EPA's Crisis Communication Plan to include a
communication process to inform affected communities about
the resolution of community concerns raised during an
emergency.
R
Associate Administrator for
Public Affairs
12/30/20

6
34
Conduct environmental justice training for community liaisons
and Incident Command System staff, thereby fulfilling that
R
Region 6 Regional
Administrator
9/20/20
and annually

element of the EPA's Crisis Communication Plan.	thereafter
1 C = Corrective action completed.
R = Recommendation resolved with corrective action pending.
U = Recommendation unresolved with resolution efforts in progress.
20-P-0062
35

-------
Appendix A
Agency Response to Draft Report
49k
UNITED STATES ENVIRONMENTAL PROTECTION AGENCY

O
T
WASHINGTON, D C 20460
SEP 6 ' 2019
OFFICE OF
LAND AND EMERGENCY
MANAGEMENT
MEMORANDUM
SUBJECT: Response to Office of Inspector General Draft Report No. OA&E F Y18 0266 "EPA
Needs to Improve Its Emergency Planning to Better Address Air Quality Concerns
During Future Disasters" '	>19
Thank you for the opportunity to respond to the issues and recommendations in the subject audit
report. Following is a summary of the agency's overall position, along with its position on each
of the report recommendations. For those report recommendations with which the agency agrees,
we have provided high-level intended corrective actions and estimated completion dates to the
extent we can. For those report recommendations with which the agency does not agree, we have
explained our position, provided the legal basis, and proposed alternatives to recommendations.
For your consideration, we have included a Technical Comments attachment to supplement this
response.
¦\GEN"€Y" S QVERALi. POSITION
The report seems to make broad conclusions applicable to the Agency and several national
programs based on the limited review of one event in which flooding was the primary focus of
the response. Instead, a review of the Region 9 response to the Kilauea volcanic activi ty in
Hawaii, as an extended response intensively focused on air monitoring, would provide a better
overall picture of EPA's existing processes, capabilities, and thorough coordination with state
and local agencies.
In general, the Agency does not agree with nor advise developing overarching monitoring
guidance for emergency responses - beyond what already exists. First, states and local
governments are responsible for their emergency response efforts. If federal assistance is
requested, or EPA receives a mission assignment from FEMA, the response is handled by the
particular EPA Region. Each emergency is unique, as are the associated responses. Overarching
FROM: Peter C. Wright
Assistant Administrator
TO:
Charles J. Sheehan, Acting Inspector General
Office of Inspector General
20-P-0062
36

-------
guidance for monitoring that would encompass the myriad emergency scenarios that could
possibly occur would be challenging. Decisions regarding monitoring are made based on an
evaluation of the specific incident. Further, in Region 6, all states have State Implementation
Plans (SlP)-approved authority to permit planned startup, shutdown and maintenance (SSM)
emissions for most facilities. EPA only approves state permitting regulations. EPA can, and
does, also enforce these permits, SIP required conditions, and National Emission Standards for
Hazardous Air Pollutants. Current regulations do not allow the state or EPA to dictate SSM
schedules.
We understand from our discussions that we can propose alternatives for the draft
recommendations, and we have provided suggestions in the "Disagreements" table below. In
general, we propose to:
1.	remove the Office of Air and Radiation as an action official from the
recommendations,
2.	combine recommendations 2 and 8,
3.	combine recommendations 5 and 6,
4.	exclude the term "implement" in the revised recommendations,
5.	assume that "develop guidance" includes the option to incorporate the requested
provisions into existing guidance or other appropriate document(s), and
6.	focus the revisions on monitoring related to permitted and non-permitted air toxic
releases during an emergency event rather than hazardous air pollutants covered
by SIPs.
This response and these revisions have been coordinated with the Office of Air and Radiation,
the Office of Public Affairs, and EPA Region 6.
20-P-0062
37

-------
AGENCY'S RESPONSE TO REPORT RECOMMENDATIONS
Agreements
No.
Recommendation
High-Level Intended Corrective
Action(s)
Estimated Completion by
Quarter and FY
7
(OPA) Revise the EPA's
Crisis Communication Plan
to include a communication
process to inform affected
communities about the
resolution of community
concerns raised during an
emergency.
7.1 Update Crisis
Communications Plan
1st Quarter FY 2020,
December 30
OIG Response #1: The agency concurred with the recommendation and provided a planned
corrective action and completion date. In an email to the OIG dated December 3, 2019, the agency
clarified that its update to the CCP will include a communication process to inform affected
communities about the resolution of community concerns raised during an emergency. We consider
this recommendation—which is Recommendation 5 in the final report—resolved with corrective
actions pending.
9
(Region 6) Conduct
environmental justice
training for community
liaisons and Incident
Command System staff,
thereby fulfilling that
element of the EPA's Crisis
Communication Plan.
9.1 Continue to provide annual
EJ training to all EPA Region 6
employees including emergency
response personnel. EPA will
consider adding a module to
emphasize environmental justice
communications during
emergency response.
4th Quarter FY 2020
September 30, and annually
thereafter
9.2 Provide training to
community involvement core
team, Incident Command staff,
and other appropriate
community liaisons consistent
with EPA's Crisis
Communication Plan.
3rd Quarter FY 2020, June 30
OIG Response #2: The agency concurred with the recommendation and provided planned corrective
actions and completion dates. Specifically, the agency's corrective action number 9.2 addresses the
recommendation. This recommendation, which is Recommendation 6 in the final report, is resolved
with corrective actions pending.
Disagreements
No.
Recommendation
Agency Explanation/Response
Proposed Alternative
1
(OLEM) Develop and
implement ambient
air quality monitoring
guidance for
emergency responses
Air monitoring during a response is
individualized and highly dependent upon
the unique characteristics of the incident.
Overarching guidance for monitoring
encompassing the myriad emergency
(OLEM) In order to
collect useable data and
target concerns during an
emergency response
develop guidance (e.g, job
20-P-0062
38

-------

in heavily
industrialized areas.
This guidance should
address, at a
minimum, how to
select monitoring
locations, duration,
timing and methods
depending on the
intended use of the
data.
scenarios that could possible occur is not
feasible. Decisions regarding monitoring
are made based on an evaluation of the
specific incident. For example,
circumstances such as flooded streets,
power outages, accessibility to
facilitate/sites, or personnel safety, would
dictate what could/could not be done.
aid) to assist state, local
and tribal agencies;
industry; and the affected
public in developing air
monitoring plans in
heavily industrialized
areas during an
emergency.
OIG Response #3: Our report recognizes the individual nature of each emergency response. Our
intent was not to recommend that the EPA develop prescriptive guidance to cover all potential
situations; rather, our intent was to recommend that the EPA develop general guidance to help state
and local agencies, as well as nongovernmental organizations, develop their emergency monitoring
plans. In discussions with the agency, we agreed on alternative language for Recommendation 1 and
revised it for the final report. This recommendation is unresolved pending receipt of a correction
action plan and proposed completion date from the EPA.
2
(OLEM) Develop and
implement a method
for storing and
providing public
access to ambient air
monitoring data
collected during an
emergency response.
EPA has several existing tools and
procedures such as SCRIBE, Viper,
Common operation Picture, and story
maps. Emergency Response Team
Sampling guidelines can be found at:
httDs://neDis.eDa.sov/Exe/ZvPDF.csi/2000
FZYG.PDF?Dockev=2000FZYG.PDF
(OLEM/OPA) Develop a
method for storing and
providing public access to
air monitoring data during
an emergency response.
OIG Response #4: Viper, a wireless network-based communications system, was not used to
distribute raw air monitoring data to the public during the agency's response to Hurricane Harvey and
still has not been used to retroactively provide this information. This tool also lacks user-friendly
features that would allow members of the public to easily identify and extract information relevant to
their exposures or interest level. The remaining tools that the EPA mentions in its response also do not
adequately address our concerns or resolve our recommendation, because the tools only provide
summary-level information or require the installation of complex software onto the user's computer.
Based on discussions with the agency, we developed alternative language for Recommendation 2 and
revised the recommendation for the final report. This recommendation is unresolved pending receipt
of a correction action plan and proposed completion date from the EPA.
3
(OLEM) Test and
evaluate the use of
low-cost air monitors
throughout fenceline
communities to
monitor air toxics and
other air pollutants
during emergency
situations when state
and local air
If pre-event monitoring systems are
rendered non-operational by an
emergency conditions, EPA uses
screening level tools (TAGA, ASPECT)
to pinpoint areas of concern for further,
targeted air monitoring.
(OLEM) To improve the
availability of air
monitoring immediately
post-event, incorporate
into existing procedures
coordination with ORD
and OAQPS to assess the
availability and use of
remote and portable
monitoring methods to
20-P-0062
39

-------

monitoring systems
and networks are not
operational.

monitor air toxic when
stationary methods are not
available.
OIG Response #5: The screening-level tools cited by the agency are all described in our report. In
discussions with the agency, we came to an agreement on alternative language for Recommendation 3
and revised the recommendation for the final report. This recommendation is unresolved pending
receipt of a corrective action plan and proposed completion date from the EPA.
4
(OLEM/OAR)
Identify and
standardize the use of
appropriate health-
based ambient air
quality thresholds in
communities during
emergency responses.
There are already existing air quality
standards [National Ambient Air Quality
Standards (NAAQS)]. They do not change
during emergency responses. During a
response, if we detect a specific
contaminant of concern, we go to the
existing acute values for that chemical in
order to estimate risk to communities.
These values already exist (e.g., AEGLs)
and indicate the concentrations at which
public health impacts may occur for a
particular chemical hazard. In the rare
instance that there is no established value
for a particular substance, one is
developed based on existing data or by
using existing tools to estimate toxicity.
This is done in coordination with entities
such as EPA's ORD, AT SDR, and other
experts in toxicology and risk assessment.
AEGLS are expressed as specific
concentrations of airborne chemicals at
which health effects may occur. They are
designed to protect the elderly and
children, and other susceptible
populations.
Remove OAR and revise
recommendation to read:
(OLEM) In the absence of
federal acute exposure
thresholds (AEGL
standards) for air toxics
and to avoid delays in
assessing the potential
health impacts of
concentrations detected
during an emergency,
incorporate into existing
preparedness guidance the
requirement for Regions
to coordinate with states
to identify the air
pollutant standards for
making decisions about
public health impacts
from potential toxic air
emissions.
OIG Response #6: We recognize that there are existing air quality standards for criteria air pollutants,
but there are no federal air quality standards for air toxics. We also acknowledge that the EPA
developed the AEGLs for assessing public health risk from exposure to air toxics during an
emergency. However, the AEGLs do not account for cumulative or aggregated exposures to airborne
chemicals, meaning the AEGLs may not be sufficiently protective of sensitive communities.
Additionally, Texas developed its own acute exposure thresholds, and a key decision during the
Hurricane Harvey response was whether to use the state or EPA thresholds as action levels. Our report
does not question the selection of the thresholds used for the response. After further discussions with
the agency and among OIG management, we have withdrawn this recommendation.
5
(Region 6) Assess the
potential for adverse
health risks to
residents living near
industrial areas from
Current regulations do not allow the state
or EPA to dictate SSM schedules. Public
health evaluations are the responsibility of
department of Health and Human
Services, not EPA. EPA can provide air
(Region 6) SSMs are
governed by state and
federal regulations which
are already designed to
limit emissions including
20-P-0062
40

-------
increased [startup,
shutdown and
maintenance] SSM
emissions during
emergencies
monitoring data to support HHS analysis,
as needed. EPA's emergency responses
are undertaken to protect human health
and the environment from immediate
threats posed by discharges and hazardous
substance releases resulting from a natural
disaster. These responses follow statutes,
regulations, policy, guidance, which
provide for coordination with other
federal agencies and state, tribal and local
response agencies. For fenceline and
nearby communities, EPA coordinates
with local officials, states and tribes
regarding shelter in place, evacuations, or
other protective measures.
during emergencies.
During an emergency, air
quality concerns are
addressed through
monitoring using
established acute values
(e.g. AEGLs) for the
chemicals of concern, in
order to estimate risk to
communities. EPA's
enforcement program also
evaluates facility
operations and takes
enforcement actions as
needed when violations
occur.
OIG Response #7: We understand that the EPA cannot dictate when a facility should shut down or
start up in response to an emergency and that characterizing the risk from these exposures is difficult.
However, a public health concern during the Hurricane Harvey response was the potential health
impact of residents' exposure to air toxics from multiple facility SSMs during a condensed time
period. We therefore believe that Region 6 should develop a strategy, in coordination with its states,
to limit fenceline communities' exposures in heavily industrialized areas during future emergencies.
Based on discussions with the agency, we revised and combined two draft report recommendations
(Recommendations 5 and 6) into one final recommendation (Recommendation 4). The final report
recommendation is unresolved pending receipt of a corrective action plan and proposed completion
date from the EPA.
(Region 6) Develop
and implement a plan
for limiting air toxic
exposures in
fenceline and other
nearby communities
from startup,
shutdown and
malfunction
emissions during a
large-scale
emergency.	
Delete this
recommendation and
combine with #5.
OIG Response #8: See OIG Response #7
20-P-0062
41

-------
8
Develop and
implement a strategy
for public
dissemination of air
quality data.
See OLEM response to recommendation
#2.
Delete this
recommendation and
incorporate into #2
OIG Response #9: Based on our discussions with the agency, we agreed t
and 8 in the draft report were similar and could be combined into one recoi
deleted draft Recommendation 8 and made minor revisions to Recommenc
See OIG Response #4.
lat Recommendations 2
nmendation. We therefore
ation 2 for the final report.
CONTACT INFORMATION
If you have any questions regarding this response, please contact Reggie Cheatham, Director, of
the Office of Emergency Management at Cheatham.Reggie@epa.gov or (202) 564-8003 or
Becki Clark, Deputy Director, of the Office of Emergency Management at Clark.Becki@epa. gov
or (202) 564-3818.
Attachment - Technical Comments
cc: Anne Idsal, OAR
Nancy Grantham, OPA
Ken McQueen, Region 6
Reggie Cheatham, OEM
Kevin Christensen, OIG
James Hatfield, OIG
Gabrielle Fekete, OIG
20-P-0062
42

-------
Appendix B
Distribution
The Administrator
Assistant Deputy Administrator
Associate Deputy Administrator
Chief of Staff
Deputy Chief of Staff
Agency Follow-Up Official (the CFO)
Agency Follow-Up Coordinator
General Counsel
Associate Administrator for Congressional and Intergovernmental Relations
Associate Administrator for Public Affairs
Director, Office of Continuous Improvement, Office of the Administrator
Assistant Administrator for Air and Radiation
Deputy Assistant Administrator for Air and Radiation
Assistant Administrator for Land and Emergency Management
Principal Deputy Assistant Administrator for Land and Emergency Management
Deputy Assistant Administrator for Land and Emergency Management
Regional Administrator, Region 6
Deputy Regional Administrator, Region 6
Assistant Administrator and EPA Science Advisor, Office of Research and Development
Principal Deputy Assistant Administrator for Science and EPA Science Advisor, Office of
Research and Development
Director, Office of Emergency Management, Office of Land and Emergency Management
Director, Office of Air Quality Planning and Standards, Office of Air and Radiation
Director, Office of Regional Operations
Division Director, Superfund, Region 6
Audit Follow-Up Coordinator, Office of the Administrator
Audit Follow-Up Coordinator, Office of Air and Radiation
Audit Follow-Up Coordinator, Office of Land and Emergency Management
Audit Follow-Up Coordinator, Office of Research and Development
Audit Follow-Up Coordinator, Office of Air Quality Planning and Standards, Office of Air
and Radiation
Audit Follow-Up Coordinator, Region 6
20-P-0062
43

-------