SEPA
            United States
            Environmental Protection
            Agency
           Office of Air Quality
           Planning and Standards
           Research Triangie Park NC 27711
EPA-450/4-86-007
July 1986
            Air
Guideline on the
Identification and
Use of Air Quality
Data Affected by
Exceptional Events

-------

-------
                            EPA-450/4-86-007
 Guideline on the Identification
  and Use of Air Quality Data
Affected by  Exceptional Events
        U.S. ENVIRONMENTAL PROTECTION AGENCY
             Office of Air and Radiation
         Office of Air Quality Planning and Standards
          Monitoring and Data Analysis Division
           Research Triangle Park, NC 27711

                 July 1986

-------
                                   DISCLAIMER





     This report has been reviewed by the Office of Air Quality Planning and



Standards, Environmental  Protection Agency, and approved for publication.



Mention of trade names or commercial  products is not intended to constitute



endorsement or recommendation for use.

-------
                                  CONTENTS



    1.  INTRODUCTION  .....'	      1


    2.  PURPOSE AND MANAGEMENT OF FLAGGING SYSTEM,	      5

        2.1  Pre-Airs	      5

        2.2  Post-Airs	       6


    3.  USES OF FLAGGED DATA	      8

        3,1  General Guidelines  on the Use of  Flagged Data   	      3

        3.2  NAAQS  Status  	      9

        3.3  Trends Analysis  ......  	      10

        3.4  SIP  Regulatory Activities	      11


    4.  DEFINITION  AND  APPLICATION OF  EXCEPTIONAL  EVENTS  	     12

      . " 4.1  Definitions  of Exceptional  Events	     12

             4.1.1  General Exceptional  Event  Criteria  .......      13"

             4.1.2  Specific  Definitions	  .  .  ._     14

        4.2  Demonstration of Causal  Relation  ,...,.. 	     24

        4.3  Application  of Definitions  and/or Criteria 	  .     25

        4.4  Documentation	     26


APPENDIX	'. .  ,  .  .  .     28

     Development of  Criteria for Possible Exceptional  Events

        I.  Identification of  Events  .....  	  .      28

       II.  Events Not Considered Exceptional for Data Flagging
              Purposes	      31


REFERENCES , .'	      34

-------

-------
                                  SECTION 1

                                 INTRODUCTION

     Federal, State, and local  air pollution control  officials have expressed
a great deal  of concern regarding the handling of air quality data that are
influenced by both natural  and  man-made events that are considered exceptional.
These events are considered exceptional for two reasons;  they are not expected
to recur routinely at a given location, or they are possibly uncontrollable
or unrealistic to control through the State Implementation Plan (SIP) process.
In some cases in the past, air  quality data collected during these "exceptional"
events have not been submitted  to the National Air Data Bank (NADB) because
State or local agencies were concerned about the potential misuse of such
data.  This guideline document  was prepared as a response to this concern and
is intended to provide national guidance for Identifying ("flagging") and using
ambient air quality data influenced by exceptional events.
     The guidance includes definitions of 18 acceptable exceptional events
and describes the procedures for submitting flagged data influenced by these
events to EPA's Aerometric Information Retrieval System (AIRS),  The Appendix-
provides information on other events which were also initially proposed for
consideration as  "exceptional"  but are not  included in the final flagging
system.
     The need for a  flagging (or "identification") system was  implied  in
previous Agency guidelines and regulations.   The  first example is  Office of
Air  Quality  Planning and Standards' Guideline  No.  1.2-008 (revised February
1977)  entitled, "Guidelines  for the  Interpretation of  Air Quality  Standards."!

-------
 This  guideline  addressed  the  submission  and  use  of  all  valid  air quality  data
 for determining an  area's overall  compliance status with  respect to  National
 Ambient  Air Quality Standards (NAAQS).   Issue 9  in  the  guideline posed  the
 following question: "How  should  participate  matter  (PM),  carbon  monoxide
 (CO), and other pollutant concentrations resulting  from severe recurring  dust
 storms,  forest  fires,  volcanic activity, and other  natural  sources  be taken
 into  account in determining compliance with  NAAQS?"  The  guideline  recommended,
 "Regardless of  the  source, ambient pollutant concentrations exceeding a NAAQS
 constitute a violation."   The guideline, however,  implied a need for data
 'flagging by further stating,  "Detailed  information  establishing  that viola-
 tions are due to uncontrollable natural  sources  may be  used in determining
 the feasibility of  modifying  control  strategies."
      The second example is in 40 CFR 51.12 (d),  the regulations  for the
 development, adoption, and submittal  of  SIP's.2   The regulations stated:  "For
 purposes of developing a  control strategy, data  derived from  measurements of
 existing ambient levels of a  pollutant may be adjusted  to reflect the extent
•to which occasional natural or- accidental  phenomena, e.g., dust  storms,
 forest fires, industrial  accidents, demonstrably affected such ambient levels
 during the measurement period."
      A third example is provided in the  March 20,  1984, 40 CFR Part 50
 Feder_a| Regi_ste_r proposed revisions to  the national ambient air quality stan-
 dards for particulate matter.3  Appendix K of Part 50 would allow consideration
 of the influence of rare or unusual events on PMjQ data by various techniques.

-------
       All three  examples,  OAQPS  1.2-008, 40 CFR  51.12(d),  and the  proposed
  40  CFR  50  Appendix  K  reflect  concern that some  air quality data associated
  with  the occurrence of  certain  types of events  may require special
  consideration in  order  to avoid misuse.
       The guideline's  general  policy  is to allow consideration  of  excluding
  flagged data from use in  regulatory  actions.  The actual  exclusion of the use
  of  flagged data would only be allowed  if, as  a  result  of  a public review
  process, the responsible  government  agency e.g., the  State Air Agency during
  the State  regulatory  process, and  the  U.  S.  EPA during the Federal  review/
•  approval process, determines  that  the  data are  inappropriate for  use  in a
  specific  regulatory activity.  This  consideration for-exclusion of  flagged
  data carries with it  no prior presumption towards use  or  non-use  of flagged
  data.
       By establishing  uniform  procedures  and  criteria  for  flagging and
  determining the use of  data associated with  exceptional events, EPA expects
  data collectors to  submit to  the  NADB  all valid ambient air  quality data,
  i.e., data collected  in accordance with  40  CFR  58.   Having  a complete national
  air quality data file will provide a data base  adequate to  evaluate and
  substantiate the impact of exceptional  events on  air  quality and  to assist
  users in interpreting the data.
       The guideline provides criteria and  procedures  by which  potential  users
  of air quality  data can be informed of "exceptional  events"  which may have
  influenced the  data.  The guideline has  no  regulatory or  legal significance
  regarding use of any air quality data.  Use or  non-use of air  quality data,
  whether flagged or not, must be subjected to full  public  disclosure and
  rulemaking procedures.

-------
     The criteria for identification of "exceptional  events"  are designed to
be expansive enough to encompass most good faith claims by State and local
agencies of when data should be considered for.special  treatment.  It is not
intended to reflect EPA's views on the validity of these claims.  The flagging
of data is merely a way for a State or local  agency to  state  that it regards
the data as influenced by exceptional events, and may later claim that the
data should be discounted for certain purposes.
     As experience with this guideline is gained, periodic revisions may be
made.  The guideline is to be implemented on a trial  basis for approximately
2 years, after which the Standing Air Monitoring Work Group (SAMWG)  will
assess its effectiveness and make any appropriate recommendations for revision.
The SAMWG is composed of Federal, State, and local air  pollution control
officials who constitute a forum for discussion and resolution of ambient air
quality monitoring issues.

-------
                                   SECTION 2
                PURPOSE AND MANAGEMENT OF THE'FLAGGING SYSTEM

     The basic purpose of the flagging system is to identify those air quality
measurements that are influenced by exceptional  events.   These are events
which, if unidentified, could lead to possible misinterpretation or misuse of
the data.  Because the flagging system relies heavily on the identification
and understanding of events that may have influenced a particular air quality
measurement, its major thrust is information exchange.  If a particular air
quality measurement is influenced by an exceptional event, it is important
for all those who may review and ultimately use the data to be aware of this
influence and to take care that such data are not misinterpreted or misused.
Knowledge and understanding of what the data represent are critical in the
overall air quality planning process.
     Under the flagging system, State and local air pollution control agencies
will be responsible for initially identifying and documenting data influenced
by exceptional events.  These agencies also must develop the appropriate back-
ground information used to support the decision to flag an individual piece
of data; they must submit the information to EPA for concurrence and make it
available for the public's review upon request.  Because of the potential
implications on  the use of flagged data, the agency flagging the data must
(as discussed in Section 4.2) clearly demonstrate  a causality between the
exceptional event and  the  flagged air quality data.

2.1   PRE-AIRS 'PROCEDURES
      Until  the Aerometric  Information Retrieval System  (AIRS) is operational,
the  flagged data should be specifically  identified and discussed in the   "

-------
annual State and Local Air Monitoring System (SLAMS) Report.  The EPA Regional
Offices will be responsible for review and concurrence or non-concurrence with
the flag, except for data related to stratospheric ozone intrusions which
will be reviewed and concurred or non-concurred with by OAQPS.  States should
initiate discussions with EPA Regional Offices regarding data that may be
flagged as soon as possible after data collection, and should not wait until
the annual SLAMS Report is submitted.

2.2  POST-AIRS PROCEDURES
     After AIRS is operational, all flagged data will be entered and stored
in the AIRS; and as-data are retrieved, a user will be able to identify those
data that have been flagged.  Each exceptional event will be assigned a
unique flag code by the AIRS for the exceptional events included in this
guideline.  The procedure for submitting, reviewing, and assigning appropriate
flags for data identified by State or local agencies to be associated with an
exceptional event are as follows:
     (1) The State or local agency should submit their flagged data with the
proper unique flag code to AIRS as part of their routine data submissions to
EPA's data bank.
     (2) The State or local agency should provide to the Regional Office,
within 30 days, appropriate documentation and demonstration of causality as
discussed in Section 4.4.
     (3) The Regional Office should concur or non-concur with the data flag
and notify the State or local agency within 30 days of receipt of documenta-
tion from the agency.   If the Regional Office concurs with the flagging of
the data, they will change the unique flag code associated with the data to

-------
designate EPA Regional  Office concurrence with the flag.  If the Region non-



concurs with the data flag, the State or local agency flag will remain in



AIRS with the particular data point.  The non-concurrence by the Regional



Office could be revised upon subsequent submission of adequate justification



by the State and local  agency.  The exception to the procedure is stratospheric



ozone intrusion, which is to be reviewed and acted upon by OAQPS rather than



the Regional Office.



     (4)  Those States electing to submit to NADB only an annual SLAMS report



rather than raw data should include a section in their annual report which



lists all flagged data, SARQAD I.D.., date of occurrence, and type of acceptable



exceptional event.



     Notification must be sent to AIRS and the Regional Office by the



responsible State or local agency when flagged data are to be changed to a



nonflagged  status for any reason.



     Two data records will be maintained in AIRS.  One record will contain



all the data including the flagged data, and the other will  contain only



data that have  not- been flagged.  All users of the data will have access to



both records, including the rationale for the flag and .the EPA Regional



Office  concurrence or non-concurrence with the flag.

-------

-------
                                   SECTION 3
                              USES OF FLAGGED DATA

3.1  GENERAL GUIDELINES ON THE USE OF FLAGSED DATA
     Concern over-the potential for misuse has made some State and local
agencies reluctant  to submit air quality data that were influenced by what
they consider to be an exceptional event, i.e., an event that 1s not expected
to recur routinely at a given location, or that is possibly uncontrollable
or unrealistic to control through the State Implementation Plan (SIP) process.
The views of Federal, State, and local agency officials have varied concerning
the use of air quality data collected during an exceptional event.  Some
believed that air quality data collected during an exceptional event should
not be used under any circumstances; some" believed that all valid (i.e.,  col-~
lected in accordance with 40 CFR 58) data should be used (without exception);
some believed that the data should be used only to determine the status of
the area with respect to the NAAQS; and still others believed that the data
not only should be used to determine  the compliance-status of the area, but
also to develop trends analyses and control strategies (with some qualifica-
tions).  This document addresses these differences by requiring the States
(and EPA) to identify and explain the use or non-use of data influenced by
exceptional events during a  public review process.  Furthermore, nhe guidance
provided does not dictate any  prior presumption toward use or non-use for any
specific purpose.
     In  general, decisions on  the use  or  non-use  of flagged data will be made
on a case-by-case basis  for  specific  purposes  (attainment designations, control
strategies, etc.), and the  public must be  informed tnat the data  exist, whether

-------
the data are used or not.  The main,concern lies with-understanding what the
flagged data represent.-  Three steps usually should be taken in order to
decide whether questionable data should be flagged, and to assist potential
users In determining how the data should be used.  The first step is to deter-
mine the portion of the measured air quality level attributable to the single
-event, as well as the cumulative effect of several similar exceptional events,
that create substantial impacts at a monitoring site.  In some cases, modeling
(either source apportionment or dispersion) or other procedures may be used to
determine the relative contribution of the event.  The relative contribution of
the event provides a better understanding of' what the air quality level for the
day or days in question actually represents.
     The second step is to determine the area influenced by the event.  In
most cases, the impact will be limited.  Therefore, if the data are .being con-
sidered for use in or exclusion from regulatory purposes, the agency should
determine the area which would be influenced by the determination.
     The third step is to demonstrate how the flagged data relate to data
previously collected at the monitoring site.  This step is extremely critical
for trends analyses and for preparing reports to the public on air quality
levels for a given area.  Obviously, if flagged data are used or excluded in
preparing summaries of air quality data, the reader should be made aware of how
the. data were used.

3-.2  NMQS Status
     NAAQS compliance  status  involves the  use of data to determine whether the
area represented by the data  is meeting or exceeding the NAAQS for the pollutant
being monitored.   It is EPA's policy and a regulatory requirement to have valid

-------
NAMS and SLAHS data (i.e.,  data collected in  accordance with  40  CFR  58)
submitted to the NADB or included in  the  annual  SLAMS report.   All data,
flagged or unflagged, should be available to  the public for comparison  to  the
NAAQS to determine if exceedances have occurred.  All relevant  flagged  data
along with the reasons for  flagging,  and  a demonstration of causality between
the exceptional event and the flagged data, shall  be submitted  for considera-
tion of use/non-use during  any public hearing or comment period  called  under
Sections 107, 110-113, 119, 120, 122, 123, 126,  Part C or Part  D of  the Clean
Air Act.^   Consideration of ambient  air  quality data during public  reviews
called under the authority of other sections  or  titles of the Clean  Air Act,
although not prohibited, does not appear  to be relevant to the  purpose  of  those
reviews.  For example, consideration  of ambient  data at a hearing held  regarding
automotive emissions standards may or may not be appropriate.  Such  consideration
is neither mandatory nor prohibited.

3.3  Trends Analysis
     Trends analysis involves the evaluation of the  long-term trends associated
with the measured levels of a given pollutant for a given area.   These  analyses
are useful  in  evaluating the overall  progress of the air pollution control pro-
gram for the given pollutant and in understanding why the concentration levels
of a pollutant  are increasing or decreasing.
     In  some cases the data collected during an exceptional event can be
used for trends analysis (as long as the  analyst understands the limitations
associated  with the  data).  The  trends analyses should clearly state how the
flagged  data were treated  and to what extent the flagged data were or were
not  included in the  analyses and why.  The trends analyses also should consider
                                        10

-------
(to the extent possible)  what effect the. flagged data have on the overall
trend line for the area in light of other confounding variables that may also
affect the trend Tine,
3.4  SIP ReguUto_ry_ Activities
     The use of flagged a'ir quality data for SIP regulatory activities
(areawide or local control strategy development, SIP design values,  attain-
ment/non-attainment status, enforcement actions, etc.) shall  be considered on
a case-by-case basis and discussed during the public review process.  Exclu-
sion of the flagged data would only be allowed if the responsible control
agency determines in conjunction with a public review that the flagged
data are inappropriate for use.
                                      11

-------
                                SECTION 4
                DEFINITION AND APPLICATION OF EXCEPTIONAL EVENTS

4.1  DEFINITIONS OF EXCEPTIONAL EVENTS
     This guideline provides general  definitions for these events and general'
criteria for their use in flagging air quality data.  The application of a
definition may vary from area to area because of differing air quality and
control situations.  For example, salting and sanding of streets for snow and
ice control may be an exceptional event in the southern sections of the country,
whereas they may be routine, controllable events in the northern sections.
Therefore, the following definitions are only a national guide and are not
meant to replace reasonable judgment on the part of the Regional, State, and
local air pollution control agency officials in defining and identifying
exceptional events for the purpose of flagging data.
      In situations where it can  be shown that the national criteria are
generally inappropriate, a State Agency may propose alternate criteria to their
EPA  Regional office.  These proposals would have to be  subjected to puolic
review within the  State.  The Regional office would be  responsible for review
and  preliminary  approval of the  alternative criteria.   The OAQPS will have
final  approval authority and  if  approved, will  include  the alternative criteria
as a supplement  to this  national guideline.
      With the above limitations  in mind, the following  general  definitions-have
been developed to  promote  consistency with  respect  to  flagging  data that  have
been collected during an exceptional  event.  Whenever  possible,  specific  cri-
teria or  terms  have been  used to define  the  event  to minimize inconsistent
interpretations.   Of  course,  no  term  can  be  absolutely  defined  so  that there is
                                        12

-------
no room for interpretation.  Thus, these definitions and associated criteria
provide for some flexibility in their application to an individual  event.
What may be unusual  or exceptional for one part of the country may be typical
for another and this variability requires flexibility in national  definitions
and criteria.
     The definitions that follow have in some cases specific numbers included
within the definition.  Although they are not fully supported by technical
studies, they are included as a practical alternative to deleting weakly sup-
ported values and, to the extent possible, reflect the comments received during
development of this guideline.  These numbers may require adjustments as
experience is obtained with the guideline.

4.1.1  General Exceptional Event Criteria
     Before one can define an individual exceptional event, one must have a
general definition of "exceptional."  In a sense, this definition also serves
as an overriding criterion with regard to the specific definitions presented
in 4.1.2..
     Webster defines "exceptional" as forming an exception, rare, uncommon,
extraordinary, deviating from the norm.  With respect to air quality consider-
ations in this guideline, an exceptional event is defined as an event that is
not expected to recur routinely at a given location, or that is possibly uncon-
trollable or unrealistic to control through the SIP process.  As noted previously,
what is exceptional in one area of the country may not be exceptional in another.
Therefore, some judgment is needed in identifying whether an event is exceptional
in the area of the country where  it has occurred.
                                       13

-------
4.1.2  Specific Definitions
     The following definitions are provided for the purpose of identifying
an exceptional event.  Some additional discussion is provided in each case
to aid 1n the application of these definitions.

HIGH HINDS (PH)
     Definition:
          An  hourly windspeed of greater than or equal to 30 mph or gusts
          equal to or greater than 40 mph, with no precipitation^ or only
          a trace of precipitation (observed as scattered drops that do not
          completely wet or cover an  exposed area up to a rate of 0.01
          inch/hrj).
     The high wind condition with no  precipitation or only light precipitation
and  dry soil  must be associated with  a significant contribution (estimated to '
be >_ 8S% by weight) of crustal material on the PM sampling medium.  High winds
without unusually high PM  levels due  to the suspension or resuspension of
crustal materi-al  should  not be considered  an exceptional event.  Additionally,
fugitive emissions or  dust from any part of an industrial source should not be
flagged.
STRATOSPHERIC OZONE  INTRUSION  (0-j)
      Definition:
           A stratospheric ozone intrusion  occurs  when  a parcel  of  air
           originating in the  stratosphere, average  height  20  km (12.4 mi),3
           is entrained directly to  the  surface  of the  earth.
      Although this event is exceptional, the circumstances or the  criteria
 under which it occurs are difficult to  measure  or document given current

                                        14

-------
measured meteorological  parameters.  Stratospheric ozone intrusions are
infrequent and very localized events of short duration,  which  makes it  dif-
ficult to use currently available airport data to determine whether a strato-
spheric ozone intrusion has occurred.  Stratospheric ozone intrusions are
typically associated with strong frontal  passages or severe thunderstorms and,
thus, may occur primarily during the spring of the year.  The  above definition
is provided only as a general guide for differentiating  between stratospheric
ozone intrusion, which is an exceptional  event for the purpose of flagging
data, and other nonexceptional meteorological events.  Although data have been
identified in the past as being the result of stratospheric ozone intrusion,  no
standard definition or criteria have been established for concrete identifica-
tion.  Therefore, determining whether a stratospheric ozone intrusion has
occurred should be a case-by-case decision based on reasonable judgment regarding
the Reason of the year, time of day, and accompanying meteorological conditions
associated with the ozone measurement in question.  The  EPA concurrence or
non-concurrence authority for flagging of ozone data for stratospheric intrusion
is the responsibility of OAQPS upon review of the documentation submitted by
the State or local agency flagging the ozone data,

VOLCANIC ERUPTIONS (CO, SO?, NO?, PM)
      Definition:
          The emission or ejection of volcanic materials at the Earth's
          surface from a crater or fissure.9
      Emissions from volcanic eruptions have a large-scale areawide impact on
air quality.  Excessive air quality concentrations resulting from volcanic
eruptions should be flagged regardless of the frequency of eruptions.  In most
                                       15

-------
cases, however, the eruptions and resulting impact on air quality would be



very infrequent.





SANDBLASTING (PM)



     Definition:



          Sandblasting or gritblasting refers to the temporary use of



          abrasive blasting with pressurized air for surface preparation



          purposes at a given location.



     Methods currently available are used to control these operations to



minimize their  impact on air quality at fixed point sources where routine



applications are part of the facilities' operations.  Ambient air quality



levels influenced by these activities should not be flagged.  However,



completely effective control techniques are not necessarily available or



possible for portable sandblasting operations.  Therefore, data collected at



a site within a micro or middle scale distance from a temporary (generally 3



weeks or less)  sandblasting operation may be flagged if all reasonable control



measures have been employed to minimize adverse impacts on air quality.





FORESTFIRES (CO, PH)



      Definition:



          An uncontrolled  fire  in vegetation or associated flammable material



          that  requires suppressive  action to protect natural resources or



          "values  associated with natural  resources  or that is destructive to



          natural  resources.10
                                        16

-------
 Class
   A
   B
   C
Size, acres
  £0.25
   0.26 - 9
     10 - 99
Cl ass
  D
  E
  F
  G
Size, acres
  100 - 299
  300 - 999
 1000 - 4999
>5000
     Some forest fires are unpreventable and because uncontrollable emissions
from forest fire can adversely affect air quality concentrations over a large
area, forest fires should be considered an exceptional  event for the purpose of
flagging air quality data.  In. general, Class A and B fires tend to have a more
localized impact; therefore, only data collected at downwind monitors within 3
miles of these type fires should be flagged.  For Classes C through 6, the
impact of the fire is more widespread and the location  with respect to the
monitor should be supported by receptorll«12,13,14 or dispersion modeling.15,16

STRUCTURAL FIRES (CO. PM)
   -  Definition:
          Any accidental fire involving- some kind of structure.   In
          general, a structural fire involves a building having  at
          least 500 square faet.
     The structural fire should be within a micro- or middle-scale distance (up
to 500 meters as defined by 40 CFR 58, Appendix D) of a monitor  in most instances
for  the data from that monitor to be flagged.  However, a much greater distance
may be appropriate for large fires such as a refinery,  industrial, or commercial
business area fire provided the causal relation is supported by  receptor or
dispersion modeling.
                                      17

-------
HIGH POLLEN COUNT (PH)
     Definition:
          A pollen count index above 25 grains/cm2 or 1000 grains per
          cubic meter.17
     The pollen count index is usually obtained by use of a coated slide
mounted on a circular plate that is generally mounted on the top of a seven
or eight-story building with, an unobstructed air flow.  The index is in units
of grains of pollen per cm,2  Another method of measuring the concentration
of pollen is volumetric, i.e., counting the number of grains per cubic meter.
Other recognized methods for measuring pollen levels-may be used.  For a
high-pollen count to  be flagged as an exceptional event, the pollen count
index should be greater than 25, or 1000 grains per cubic meter and the high-
volume filters  on which the samples were collected should be analyzed micro-
scopically to ensure  that  significant amounts of pollen  (i.e., 50% or greater
than the  normal pollen count for a typical  sample) were  collected on the day
or days being considered for flagging.  Where high pollen is considered to be
a  problem, State  or  local  agencies should develop procedures for ascribing
the effects of  high  pollen count on filters and should  submit them to the
respective Regional  Office for  approval.

CHEMICAL  SPILLS AND  INDUSTRIAL  ACCIDENTS  (CO, SO?. NOg,  PH)
     Definition:
           Emissions  that  result from  accidents  such  as  fire,
           explosions, power outages,  train  derailment,  vehicular
           accidents, or  combinations  of these'.18
     The  spill  or accident must, of course, not  be a routine occurrence.   Any
 of these  situations  that do occur  routinely should be critically  evaluated and

-------
stopped.  Data from sites that can be unequivocally related to a non-routine
accident or spill  riiay be flagged,

UNUSUAL TRAFFIC CONGESTION (CO)
     Definition:
          A condition resulting from a major accident (rather than
          frequent minor accidents,) or short-duration obstruction,
          such as demolition or construction.  During these conditions
          the level of traffic may increase until it exceeds the
          maximum capacity of a given street or highway.  Speeds are
          reduced substantially and stoppages may occur for short or
          long periods of time because of downstream congestion.  In
          extreme cases, both speed and volume can drop to zero.19
     As a general rule of thumb, congestion must occur within five hundred
meters of a monitoring site (i.e., micro- to middle-scale) and not be a regular
occurrence for the data from that site to be flagged.

CONSTRUCTION/DEMOLITION (PM)
     Definition:
          The building/destroying/renovation of any residential,
          institutional, commercial, or industrial building (including
          apartment buildings with more than four dwelling units),
          structure,  facility, or installation that-lasts for only a
          short period of time and  is  reasonably controlled.
     The construction or demolition activity must take place within a reasonable
distance of the monitoring site  and all reasonable control measures must be
                                        19

-------
in use before the data from the site can be flagged.  Flagged data should be
limited to sites classified as micro- or middle-scale (up to 500m) and downwind
with respect to the construction activity.

AGRICULTURAL TILLING (PM)
     Definition:
          The act of preparing dry soil for cultivation or for
          controlling the growth of weeds by the use of mechanical
          "devices during periods with an hourly average windspeed
          of greater than 20 mph.
     Generally, agricultural tilling operations must occur within a reasonable
distance  (500 meters) of the monitoring site and tilling must have occurred at
the same  relative location while the hourly average windspeed is  greater than
20 rnph for the monitoring data to be flagged.  Flagged data must  have been
collected during or  immediately after the day tilling occurred and should be
limited to sites which would be classified as micro- or middle-scale with
respect to the  tilling operations,

HIGHHAY CONSTRUCTION (PM)
      Definition:
          The  act of building  a new, or  repairing  an existing,
          highway,  road  or  street.
      Particulate matter  resulting  from  reasonably  controlled  highway
construction  for short time  periods  may  be  flagged provided  that  a microscopic
analysis  of  the filter  indicates  that  85  percent of material  on the  filter  is
related to  construction  activities  and  all  reasonable  control measures  have
been  utilized. ,
                                       20

-------
 REROUTING OF TRAFFIC (CO)
      Definition:
           A temporary deviation or detour of vehicular traffic
           because of an accident, construction, or demolition.  The
           detour must be for no more than.l week.
      The rerouting of traffic should be within a reasonable distance of a
 monitoring site and last for no more than 1 week for the monitoring data to
 be flagged.  Flagged data should be limited to sites classified as micro- or
* middle-scale (within 500 meters) with respect to the detour.

 SALTING/SANDING OF STREETS (PH)
      Definition:
           The application of salt and/or sand to the road surface to
           increase traction and/or prevent the surface water from
           refreezing after it has melted.
      The salting and sanding must occur within a reasonable distance (up to
 500 meters) of the monitoring site if the data from the site are to be flagged,
 Flagged data should be limited to sites classified as micro- or middle-scale
 with respect to the salting/sanding operations and microscopic examination
 shows that 85 percent of the material on the filter is salt and/or sand.  Data
 collection is not limited to the date of salting or sanding of the street.
 All  reasonable control measures must be taken to minimize the adverse air
 quality impact of the salting or sanding operations in order for the data to
 be  flagged.  In general  this type of data  should not be flagged in areas
 which experience a  significant  use of salt and/or sand.
                                        21

-------
INFREQUENT LARGE GATHERINGS (CO, _PH)
     Definition:
          A gathering of more than 10,000 people (5000 cars)  at any
          one time and at a -single location.   Unusual  traffic congestion
          must be associated with the event.
     A large gathering in and of itself without the associated traffic
congestion would not qualify as an unusual event for the purpose of flagging.
(See discussion of unusual traffic congestion for more details on the cri-
teria for flagging data because of unusual traffic congestion).  The event
would also be expected to occur less  than once per year, and  the event should
be at a location not regularly used for such purposes.  Flagged data should be
limited to sHes classified as micro- or middle-scale with respect to the
gathering.

ROOFING OPERATIONS (PH. SO?)
     Definition:
          The process of building, repairing, or recoating the external
          upper covering of .a house or building that involves the
          application of a  petroleum-based material (usually haavy
          residuals from a  refining operation) to a roof.  The material
          is heated and then sprayed or rolled onto the surface.
          Generally, a roof so covered would be resurfaced or treated
          no more than once every 3 to 5 years.  Unless major damage
          has occurred, some surfaces would require treatment only every
          5  to  10 years.
                                       22

-------
     Roofing operations must occur within a micro-scale distance and upwind
of the monitoring site, and all  reasonable control  measures must be applied
for the data to be flagged.

PRESCRIBED BURNING (PH, CO)
     Definition:
          A controlled fire of vegetative material  that is used
          to improve range lands, agricultural  or forestry resources,
          or associated values,
     Prescribed burning is generally a controlled activity that is limited .to
those days or periods when the meteorological conditions' are conducive to good
dispersion.  Thus, the emissions which affect air quality are dispersed to the
extent possible.  However, the resulting emissions occasionally adversely affect
air quality concentrations over- a large area.  In areas of the country where
prescribed burning is used regularly and extensively for agricultural and/or
forestry land management, prescribed burning-may not be considered an excep-
tional event for the purposes of flagging air quality data.  Prescribed
burning in these areas is usually subject to rules and regulations, including
smoke management plans, under which a regulatory agency permits burning after
deciding where, and to what extent, the smoke will be allowed to impact air  -
quality.
     However, in many areas of the country,  prescribed burning for agricultural,
forestry land management, or other purposes  is an infrequent but necessary
activity and may be impractical  to fully control.  These practices may.be con-
sidered an exceptional event for the purposes of flagging air quality data.
                                       23

-------
  However, it must be demonstrated through receptor or dispersion modeling,
  that the burning operations have a substantial  impact on  the monitored air
  quality data.

  CLEAN UP ACTIVITIES AFTER A MAJOR DISASTER (PH. CO, SO?)
       Definition;
            For the purposes of flagging, major disasters are serious
            public misfortunes for which State or Federal  relief has
            been granted.
       PM, CO, S02 or other pollutant data affected by and  collected during, or
  for a reasonable period after, the clean up activities following a major
  disaster may be flagged.

  4.2  DEMONSTRATION OF CAUSAL RELATION
       Excluding the use of valid air quality data from regulatory purposes
  is  a serious action.  Accordingly, a clear demonstration of the relationship-
  between the  exceptional event and the measured air quality must be provided.
.  As  a minimum, this demonstration should:  (a)  include all relevant raw data
  (e.g.,  air quality data, meteorological data, traffic counts, etc.);  (b) show
  that the monitor did not record high concentrations before and after the
  period  of the exceptional  event;  (c) show that the local  wind direction was
  such that the monitored pollutant was transported  from the exceptional event
  source  to the monitor during the  period in question;  (d) include  as appro-
  priate  receptarll>12,13,14  or dispersion modeling!5,16 connecting the monitored
  concentrations  with the exceptional source emissions;  (e)  include, as appro-
  priate, microscopic  filter  analyses (for  particulate  emissions);  and  (f)
  include documentation supporting  the existence of  the exceptional event.
                                        24

-------
4,3  APPLICATION OF DEFINITIONS AND/OR CRITERIA
     As noted earlier, the purpose of this guideline is to provide guidance
and direction for flagging air quality data associated with exceptional
events and thus promote national  consistency in the flagging of such data.
      Having established general  definitions for the exceptional events  and
the criteria for applying flags to the data, the State or local agency should
consider flagging of any data meeting the criteria of this guideline and
prepare adequate evidence to clearly demonstrate the causal relationship
between the exceptional event and the flagged data.  The flagged data with
adeq-uate documentation of the causal relationship should be submitted to the
appropriate Regional Office following the procedures outlined in Sections 2.1
and 2,2.  The Regional Office, after review, should notify the State or  local
agency of any disagreements, their reasons for disagreeing, and seek to  resolve
the conflict.  Following their determination (concurrence or non-concurrence),
the Regional Offices should notify AIRS and the State or local agency of their
action so that the. appropriate notation may be made to the AIRS data base
and/or the Annual SLAMS Report.
     The only exception to this policy is the flagging of stratospheric  ozone
intrusion.  Because of the technical complexity of determining stratospheric
ozone intrusion, OAQPS will, upon request by the State or local agency through
the Regional Office, either concur or non-concur based on documentation
submitted by the requestor.
     The criteria (definitions) presented in this guideline will serve as the
basis for reviewing determinations associated with exceptional events.  If  a
State or local agency chooses to apply criteria that are less stringent  than
those contained in the guideline, such recommendation should be subjected

                                      25

-------
to public review.  The recommendation, along with the summary of comments,
should then be submitted to the respective Regional Office for review and
approval.  OAQPS will have final approval authority and, if approval is granted,
will include the alternative criteria in a supplement to this national  guideline.

4.4  DOCUMENTATION
     All decisions regarding the flagging of data because of exceptional
events must be documented so that the EPA and the public can review these
decisions.  The documentation may take several forms.   It may include reports
from the National Weather Service; a copy of a newspaper clipping or news
report  indicating that a chemical spill or industrial accident has occurred; a
report  from the local health department on the pollen count for a given day,
supported by filter  analyses; special reports prepared  by the State or a con-
sultant; or special  statistical- analyses of the  air quality data and other
key parameters associated with  the event.  The actual form of the documentation
depends on the event and the extent of publicly  available reports or documents
that would support the determination  of  its occurrence.  As appropriate, the
State or local agency is encouraged to rely on available reports and documen-
tation.  When  such documentation  is not  available, the  State or local agency
must  prepare the  necessary  material sufficient (see  Section 4.2) to support
its decision  regarding  the  flagged data.
      The State or local  agency  should retain  copies ,of  the  necessary supporting
material or documentation  in  its  offices  for  review  upon request.   Copies  of
this  material  will  have  to  be  available  in  the record during the public
review  process and must  be  provided to  EPA  for concurrence.   It is  not  intended
that  a  public  review process  (comment period  or  public  hearing) be  conducted
                                       26

-------
solely for the purpose of determining whether data should or should not be
flagged.  Instead, the public review process referred to  in this guideline
refers to the public review process which is otherwise required for regulatory
actions taken in accordance with the Clean Air Act.   Although the main  purpose
of the documentation is to support the decision to flag a given piece of data,
it also provides air quality analysts with background and supporting information
regarding the events associated with the data on which the analyst can  rely if
and when the data are needed to make a particular air quality decision.   The
air quality analyst will also be awara of any limitations that should be con-
sidered in the use of such data.
                                      27

-------
                                    APPENDIX
                 DEVELOPMENT Of CRITERIA FOR EXCEPTIONAL EVENTS

     The purpose of this appendix is to summarize the development history of
this guideline and to assist users in understanding the information contained
herein,
     The Standing Air Monitoring Work Group (SAMWG) first addressed the issue
of exceptional values and data flagging in December 1981.  Subsequent to that
time, SAMWG solicited input from EPA Regional Offices and State and local
agencies to identify those exceptional events that could have any adverse
effect on air quality values measured during.their occurrence.  The SAMWG also
solicited input on how air quality measured during the exceptional event should
be flagged.  An initial  list of possible exceptional events was compiled based
on comments by selected  Regional Offices and State and local  agencies for
further consideration.

I.   IDENTIFICATION OF EVENTS
      In general, the events given further consideration were  grouped  into three
major - categories:  (1) natural events  (meteorological and other natural events),
(2)  unintentional  anthropogenic events, and  (3)  intentional  anthropogenic events.
The  events  within  each major category  were  as follows;
      1,   Natural  Events
          (a)  Meteorological Events
              o   Sustained  high windspeeds  (PM)
              o   Stagnations/inversions  (all pollutants)
              o   Unusual lack  of  precipitation  (PM)
              o   Stratospheric ozone intrusion  (03)
                                        28

-------
         (b)  Other  Natural  Events
             o  Volcanic  eruption  (CO,  S02,  PM)
             o  Forest  fires  (CO,  PM)
             o  High  pollen count  (PM)
     2.   Unintentional Anthropogenic  Events
         o   Large  accidental structural  fires  (CO,  PM)
         o   Major  traffic  congestion  due to accident  or  nonrecurring
            obstruction  (CO)
         o   Chemical spills  (SC>2,  NOZ,  PM, CO)
         o   Industrial accidents  ($02,  N02, PM,  CO)
     3.   Intentional Anthropogenic  Events
         o   Short-term construction/demolition  (PM)
         o   Sandblasting (PM)
         o   High-sulfur  oil  refining  (SOg)
         o   Roofing  operations (PM, $03)
         o   Salting  or sanding of  streets (PM)
         o   Infrequent large gatherings (PM,  CO)
         o   Soot blowing from  ships (PM)
         o   Agricultural tilling  (PM)
         o   Prescribed burning (CO, PM}
         o   floncompl iance—point  source (CO,  $02»  NOg ,  PM)
     Those  whose comments and  suggestions were  solicited generally agreed that
natural  events,  other  than meteorological events,  and unintentional  anthropo-
genic events are,  by their very nature, exceptional  events.  The control  of
emissions from such  events is  usually not Included in the general control
strategy for a given pollutant, although it  is sometimes handled as part  of an
                                       29

-------
emergency response action immediately after the event has occurred to minimize
possible adverse health impacts on local residents.  Reasonable precautions
would not guarantee against recurrence or that the air quality would not be
affected by these events in the future.
     Some disagreement was expressed with regard to intentional anthropogenic
events.  Some argued that these events occur routinely and the data therefore
should not be flagged.  Others argued that, although intentional anthropogenic
events occur routinely, they are exceptional with respect to normal activities
around a particular monitoring site.  Still others argued that  intentional
anthropogenic events can and should be controlled; they further argued that,
unless these events or activities occur within the immediate vicinity of a
monitoring site, they should not be considered as exceptional and the data
collected during their occurrence should not be flagged.
     Finally, some strong disagreement was expressed with respect to two
meteorological  events—stagnations  and  inversions.  Many argue  that these  two
events routinely occur.  They  further  argue, that because these  events are
climatological  factors rather  than  exceptional events, pollutant levels
measured during their occurrence should not be flagged.
     The list of  suggested  exceptional  events was  reviewed  and  evaluated to
determine whether each of the  events  listed should be considered exceptional  and
whether  other events  should be added.   The  review  indicated that  this list was
comprehensive and that no  additional  events should be considered at this time.
     After  considerable  discussion  over an  extended  period  of  tim'e, a general
consensus was reached  that  15  of  the  events initially considered for designation
did generally satisfy the  criteria  for defining  an exceptional  event.   Six of
the events  initially  considered (implementing transportation controls,

                                        30

-------
stagnation/inversions, high-sulfur oil refining, sootblowing from ships,
noncompliance—local sources, and unusual lack of precipitation) however, were
finally  rejected.   The rationale for  rejecting these events is summarized
below.

II.   EVENTS  NOT CONSIDERED EXCEPTIONAL FOR DATA FLAGGING PURPOSES
      1.   Implementing Transportation  Controls
      Transportation control measures  are not considered exceptional, and data
collected  during the implementation of transportation controls should not be
flagged.   If,  however, traffic must be temporarily rerouted during the imple-
mentation  of the transportation control measures or some congestion occurs
.due  to  initial startup of the transportation plan, the data collected at
monitors  near  the  rerouted traffic or congestion may be flagged.
      2.   Stagnati ons/1nversi ons
      Stagnations and inversions are frequent climatological occurrences that
must be  considered in evaluating whether a control program is adequate to
attain  and maintain the  NAAQS.  An inversion is said to occur at a point, or
through  a layer, where temperature increases with increasing height,20»21
Surface-based  inversions are those that  extend vertically from the surface to
some altitude  aloft.  One  study found that surface-based inversions generally
occur about  32 percent of  the time.22 They are usually short-lived and disperse
shortly after  sunrise.   Because inversions are expected to occur frequently and
are  part of  weather patterns, they are not considered exceptional events for
the  purpose  of flagging  data.
      Stagnation  episodes are periods  of  4 or more days with surface wind speeds
of generally 4 ra/sec or  less and no precipitation or frontal passage.   In some
                                        31

-------
parts of the United States stagnations usually persist for an extended period
of time, and they can affect an entire air basin; therefore, they are not
generally considered exceptional' for the purpose of flagging-data.
     3,  High-Sulfur Oil Refining
     High-sulfur oil refining refers to the process of refining crude oil
with a sulfur content that is 20 percent or greater than the design capacity
of the refining operation.  Because this is a common practice at many
refineries, and procedures and control methods are used to minimize 502
emissions, it is not considered to be an exceptional event for the purpose
of flagging data.
     4.  Sootblowing from Ships
     Sootblowing from ships is a method in which air is used to remove
deposits that may build up on the walls of the vessel's boiler tubes.  This is
a common practice that  is either controlled or limited (in many areas of the
country) by establishing opacity limits.  Because these activities are common
and  steps  can be taken  to minimize  associated emissions, sootblowing from
ships  (like general  sootblowing from utility and industrial  boilers) is not
considered-an exceptional event for the purpose  of  flagging  data.
     5.  Noncompli ance—Local Sources
     Limited noncompliance  of  local sources can  be  expected  from  time to time
as  a result of  process  upsets or malfunctioning  control equipment.   These
events are usually  classified  as "upsets" or  "malfunctions"  as defined by
the  applicable  State or local  agency  regulations, or they may be  considered
a violation of  applicable enission  or  opacity limits.   If these events are
caused by  upsets or malfunctions, they  should be so noted and reported to the
 appropriate control agency.   If  they  constitute  a  violation, the  appropriate

                                        32

-------
legal remedies will be taken.  If legal action is taken, the air quality data
collected 1n the vicinity of the source will  in all likelihood be used in the
legal proceedings, and any appropriate limitations associated with the data
would be reviewed and evaluated as part of the legal process.  Because, data
collected during noncompliance conditions have special uses and the source
is required to notify the .State of the upset  or malfunction, noncompliance of
local sources is not considered an exceptional event for the purpose of
flagging data.
     5.  Unusual Lack of Precipitation
     Lack of precipitation in and of itself would not be considered an
exceptional event because it has very little  impact on PM air quality levels.
Lack of precipitation or drought conditions combined with high winds, however,
would be considered an exceptional event.  Therefore, unusual lack of precipi-
tation is not considered an exceptional event for the purpose of flagging
data.
                                       33

-------
                               REFERENCES


 1.   "Guideline for  Interpretation of Air Quality Standards," U. S. Environmental
     Protection Agency, Office of Air Quality Planning and Standards, Research
     Triangle Park,  N.C. OAQPS No. 1.2-008 (revised February 1977).

 2.   40 CFR  51.12d,  Code of Federal Regulations, July 1, 1984.

 3.   Federal Register,  Vol. 49, March 20, 1984 pp. 10433-10435.

 4.   Clean Air Act as  amended August 1977.

 5.   ByerSj  H. R.  General Meteorology.  McGraw Hill, New York, New York, 1974,

 6.   Anthes, R. A.,  et al.  The Atmosphere.  Charles E. Merril Publishing
     Co., Columbus,  Ohio,  1975.

 7.   Federal Meteorological Handbook No. 1, Surface Observations.  Washington,
     O.C.  1976.

 8.   Dorn, W. I.  Meteorology.  Third edition. McGraw Hill, New York,
     New  York, 1965.

 9.   Bates,  R. L., and J.  A. Jackson.  Glossary of Geology.  Second edition.
     American Geological  Institute,  Falls Church, Virginia, 1980.

10.   Personal Communication with  Linda Farnsworth of the USDA  Forest Service,
     Aviation and Fire Management Division, Fire  Department,  November 1983.

11.   Receptor Model  Technical  Series, Volume I, Overview of Receptor Model
     Application  to  Particulate Source Apportionment, U.S. Environmental
     Protection Agency, Office of Air Quality  Planning  and Standards, Research
     Triangle Park,  N  C.  EPA Publication No. EPA-450/4-81-Q16a. July 1981.

12.   Receptor Model  Technical  Series, Volume II,  Chemical Mass  Balance, U.S.
     Environmental  Protection  Agency, Office of Air Quality Planning and Stand-
     ards,  Research  Triangle Park, N C.  EPA Publication No. EPA-45Q/4-81-Q16b.
     July 1981.

13.   Receptor Model  Technical  Series, Volume III, User's Manual for Chemical
     Mass Balance Model, U.S.  Environmental Protection  Agency,  Office of Air
    'Quality Planning and Standards, Research  Triangle  Park,  N  C.  EPA Publica-
     tion No.  EPA-450/4-83-014. July 1983.

14.   Receptor  Model  Technical  Series,  Volume  IV,  Summary of  Particle Identifi-
     cation Techniques, U.S.  Environmental  Protection Agency,  Office of Air
     Quality Planning and Standards, Research  Triangle  Park,  N  C.  EPA Publica-
     tion No.  EPA-450/4-83-018. June 1983.
                                       34

-------
15,   Guideline on Air Quality Models,   OAQPS,  U.S.  Environmental  Protection
     Agency, Research Triangle Park,  N  C.  EPA  Publication  No.  EPA-45Q/2-78-
     027 (NTIS PB 288-783).   April  1978,  or  subsequent  revisions.

16.   Regional Workshops on Air Quality  Modeling:  A  Summary Report,  U.S.
    - Environmental Protection Agency, Office of Air Quality  Planning  and
     Standards, Research Triangle Park, N  C. EPA  Publication No.  EPA-450/4-
     82-015, (NTIS PB 83-150573), April 1981.   (NTIS documents  include  1982
     and 1983 addendums)

17.   Air Pollution Aspects of Aeroallergens  (Pollens).  National  Air  Pollution
     Control Administration  Consumer  Protection and Environmental Health
     Services, Department of Health,  Education, and Welfare, September  1969.

18.   Sonti, R. S.  Practical Design and Operation of Vapor-Depressuring
     Systems.  Chemical Engineering,  91(2);66,  January  23, 1984.

19.   An Introduction to Highway Transportation  Engineering.  Institute  of
     Traffic Engineering, 1978,

20.   .Mclntosh, 0. H.  Meteorological  Glossary.   Chemical Publishing,  New York,
     1972.

21.   Huschke, R. E.  Glossary of Meteorology. - American Meteorological
     Society, 1959.

22.   Seiger, R.  The Climate Near the Ground.   Harvard  University Press.
     1966.
                                      35

-------
                                    TECHNICAL REPORT DATA
                             (Please read Instructions on the reverse before completing)
 . REPORT NO.
 EPA 450/4-86-007
                                                             3. RECIPIENT'S ACCESSION NO.
d. TITLE AND SUBTITLE
 Guideline  on the Identification and Use  of Air Quality
 Data Affected by Exceptional  Events
5. REPORT DATE
 July 1986
6. PERFORMING ORGANIZATION CODE
 . AUTHOH(S)
                                                             B. PERFORMING ORGANIZATION REPORT NO.
                                                              EPA 450/4-86-007
 . PERFORMING ORGANIZATION NAME AND ADDRESS
 U.S. Environmental Protection Agency
 Monitoring and Data Analysis Division
 Monitoring and Reports  Branch
 Monitoring Section, RTP,  NC 27711
                                                             10. PROGRAM ELEMENT NO.
11. CONTRACT/GRANT NO.
12. SPONSORING AGENCY NAME AND ADDRESS

 Same as  9.
13. TYPE OF REPORT AND PERIOD COVERED
 Technical  guideline
14, SPONSORING AGENCY CODE
15. SUPPLEMENTARY NOTES
16. ABSTRACT
       This guideline  is  intended to  provide information to State  and local air
 pollution control agencies concerning the flagging of air quality data affected
 by exceptional events.   The guidance defines exceptional events,  establishes
 uniform criteria and procedures for flagging air  quality data, and provides a
 mechanism for allowing  consideration of excluding flagged data.
                                 KEY WORDS AND DOCUMENT ANALYSIS
                   DESCRIPTORS
                                                b..lDENTlF!ERS/OPEN ENDED TERMS
                                                                              COSATI Field/Group
  Exceptional events
  Data flag
  Air quality data
  Excluding flagged data
 13. DISTRIBUTION STATEMENT
  Unlimited
                                                19. SECURITY CLASS /Tin's Report)
                                                  Unclassified
                                                                            21. NO. OF PAG£S
                                                20. SECURITY CLASS (Tills page)
                                                  Unclassified
                                                                            22. PRICE
 EPA Farm 2220-1 (Rev. 4-77)   PREVIOUS EDITION is OBSOLETE

-------