-------
-IB-
for their completeness .
If the data submitted to the Regional Office fron.
the States arr.% in the -row:! or punched cards, the !V: <;; on a 1
Office can visually inspect the Ijo.tch to i;i«Ice .sure tiia-i;
pertinent col 1:1.1.15 are punched anc! aligned correctly. The
Regional Office may finr' it o'osirpb'le to actually ^rii-.t
out or list the data from selected agencies be for: for-
warding the cards to MPAD. If the: data are sent on mag-
netic tape, thnre is litt'lt: the Regional Office can <'o,
at present, but forward it on.
B. Interrogate Data Ban!', Data Requests and Manual
Examination
The M'DAD has some existing SAROA.D outputs which-".
the Regional Office tnay find helpful in evaluating their
air quality data. The Regional Office can request out-
put from the data bank and get quarterly and yearly fre-
quency distribution.. 1 i sts for each srunpling station. The
output includes the site description at the.top of each
page and a frequency distribution fo;; ( ach pollutant, year
or quarter-year. The number,.of"obsei vi. fcio.ns, minimum,
maximum, and the percenti 1e, valuer. a;\; ;•!.-;ted for each
pollutant-quar.ter-y-e.ar. The ar'iti;n,:Gi. ^ ?..»tin,., geometric
mean, and geometric stan^^rc! djfiVvi a;.;; ..Ji .^.yi/en only for
those pollutant-quarter-years wh(ic,!i ,i,.?ct National Aerometric
Data Bank criteria. . ,:, .
The freq-u-:!!::.> :.•;•:';••';, . ... ... • o. aii/ibie on a na-
tional, EPA regional and stale basis. Other options in-
-------
-in-
clude tin? ability Lo request the d Istribu'tion for 1 iini l:;:-o'
numbers of pollutants, years or quarters.
These c;.nd other oiu.'Hits arid remote b^r.t.h urn'
interactive access mi-thoris arc1 more fully defined and dis-
•"•' ?
cussed in the SARO'',D Terminal User;> ,'••;;/n UK T; oiu! the Poc; ? •:;•:;• T
Office f!i:nS/SARO/'i) contact should he conl«cU^ •'•':•.•>- addi-
tional information.
The Regional Office will* in the future bo .:;!:.u;
to make comparisons bot;/«tn measured air quality data srio
• "'f
that which they, arid/or t/ie 5rate and local ?
station and po11u t a n t.
C. Check Anomalous I)at-a
Anomalous or questionable data values nay arise
from the data flow system as a result of the following pro-
cedures: edit checks, validation screen and the application
of the flagging technique. The Regional Office has the 're-
sponsibility of either accepting, roj;:cting or modifying
the data value'or average in -question. In this regard, the
Regional Office has the option of req^Lsting that the origi-
nating agency determine the'validity c the data or provicie
"''..' i' *
certain information and d&cun;e'ntfii:ic;". J.at they may rcnke
i • •' . ~'
the final determination.
; ,'' > I
The procedure usad to chacl; out a.-iy specific data
value prior to the init-'at ioii of B.-, r.noi.ialy request to NADS'
-------
could d c | • '•> n cl on:, the Re 'j i o ; i a 1, 0 f f i c o ' ;; a s s e <•: s 1.1 o n t o f v: ,: . '••
TiiO c'-rrt:,1 :>hou'lci i,-e tr&cori viii c-'.'fi !•:. •.•.;:
trdl prorj; t>!D, aru! previous performance . I'lPAD su.orest?
Ih tit int. fol "I fiwi tii) j; e q u & : >. •:. e of : steps te fol lev.'?c! in or-J;;/
to Cliei. .': OM'1: •'. , ';» ••;';•.•.• 'I OUS d ••'(.:'. V /• 1 V; •;: •-, 0!" • ;• OMipOf. i it: :.\ 0!V,.> . •; ,
In all •:; '•< i: C; ^ , it shrriilJ (••:: rceegn1;-0(! that ''any Ayi-ncy
which alters, msnipu'latos or tr,an scribes a o'ati; volut: in
any v/ey is potenti ol ly capsa'l,^; of introducing an error.
!.'.h«n a data vali/a- is ide:i ,;i f 1 cd . <.s by ing qufe s ti :..:.• a :-1 e , i'!;0
responsible Aysncy :nnst dotcrininc whethci- o-r not the chtr.
value indintai nod its integrity .thrcuaiiout the Aijcncy's
d?ta r-cqui si ti on /nv.! proccss.irifj . syste;.:.
!!!•:. S AS •.
systci:', th:: Rf.-gior.al Offices., St«vto Aoi.. f:cy ari:i/or lc--cr'i
Ac;oiK.y tu its original recording, v/het!urr it bo a value-
from a computer readout, paper tape printatC','1 strip chart,;,
or a report from the chemist in, ;the :l-;a;.»orfif6ry. The type.'?
of errors usually found in the/^tnterrval cireck are: typino,
key punching, tabulating and transposition, mathematical
(such as addition, muHipl icet.ion and transcribing) .
Further discussion of these errors and methods to reduce
their frequency may be found in .already published guide-
line documents. 3»^»5 , v; = , ,s
If no errors have been identified in the internal
check, at all agency levels, the -verification and evaluation
process s h c u 1 d •; o n t i > i u e •-, o w n t w o s f m f 1 a r b u t s - : p « ;•• o. t ft p a t h :' .
-------
- ?.. I •
'..'In'eh p<';t!i is c ho cirri dc-ji^ods on w^o ih^r the; ;i,-;La in
questior; is a single vn):.... or a•/.:o:.; posit S ayc>r<-f;e.
i. V e ri f'y ing and Ffva 1 ua!.:i n;j Speci f i c Ai r
Quality Date flu lues
Instrument C; 1 U-r;; ti on 4 Speci f i c <: t i o :i s S;K
Operations
The oporation and calibration or continuoi.:f>
instrunients is of the utmost ii!ipo)-t.3Mco 'in
the production of valid air quality data. The
instrument calibration should be reviewed fur
the time period in question, both before <<\\<\
after the suspect data point. It should b-:- de-
termined if the instrument We,s operating within
pre-deternritvd performance speci fie,:; ti oi-'is ^-.(cl'i
as drift, operating temperature Fluctuations,
unattended operational periods. etc. Those
performance specifications for automatic monitor
are defined andvpu'bl ished in the Federal
Register and summarized in various guideline
*3 A
documents. .'These specifications are likely,
however, to be superceded by those published in
the October 12, 1973 issue pf the Federal
Register on proposed Equivalency Regulations.
Guidelines on air quality co;ntrol practices and
p-
error trcd'":; techniques are als-: Av:; 11'.-.bl ..^
-------
Before and After i-.cadi figs
If the in'* trident generating the data was
found to he "in control," the values ;M--
modlately before and after should be os-
tc nn i nod . Cc;mpc r i sons bo twefcn tho >.; t- rc t;n I
and/or gross deviations could be inn do.
Ideally, this difference in concentration
should be determined through a statistical
analysis of historical data. For example,
it may be determined that a difference of
0.05 ppnMn SOo concentration for successive
hourly averages occurs very rarely (less than
one percent of the time). The criteria for.
what constitutes an excessive change may. also
be linked to the tirna of day. For example,
an hourly change of CO-of .10 ppp between' 6 £.!•;
a.nd 7 AM may be common but would be suspect
• *• C
if it occurred between 2 AM and 3 Af!.0>t)
Other Instruments at the Same Location
Observing the behavior of other instruments
at the same location would give the evaluator
a qualitative insight into the possible reasons
for the anomalous reading. If all of the in-
struments showed a 'general increase, meteorolog-
ical factors-might" be considered while n dra^s".
-------
-23-
• devMtion Over-the same short, period of,t'hi;e
may indicate an electrical problem or an eir1'
condftioninr! mel function. Oh the other, hancf.s
if the other instruments behaved normally, a
temporary •infTu'fjM'ce of a single pollutant or:
single polluta/nt source may be suspected.
Similar•Instruments at Adjacent Locations
Comparing ths behavior of other instruments
In the vicinity which monitors the same
pollutant could'further elucidate the ;
situation. For example, if the adjacent in-
struments (upwind and downwind) exhibited
the same general trend, an area problem in
which the maximum effect was over tha station
of interest, would be indicated. However, if
;-i
the adjacent stations seemed to peak either •>
before or after the time the Suspect value
' " ' ' ' f 6 ,.
was recorded, the station may have been,under
the Influence of plume fumigation which wand&reo
• • '!"' is;' '
according to wind direction influences. Micror
meteorological, influences should not be over-
.. , ; -. '.,' ;.;{,] ,'
looked either,^ The station may be under the
-•;.-•'' ''''••,/ f C .' •
Influence of subsidence effects from the urban
.. . '?(,•
'':' 7 ,°
heat island or upslope-downslope Influences.-'"
Meteonilo'j'/o i .'pnditions
No attempt ..», explain an anqmalous air quality
-------
-24-
data point 'would be complete, without :-•. con-
. sideration ,of the meteorological conditions
present a'|,;'.the tiir.e of the reading. '..''A'.pelsing
front/and strong inversion, extended' ealia.s or
strong winds are conditions which ha.ve a ore:- -'•'•' .
may change end adversely affect a .ci.vc-n air .
' v "*;•'*
monitoring -station's representativeness. Ex-
! a ...-.• i -,;. v'5 '
amples of this might be an adjacent apartment
-------
-25-
house or supermarket changirtg tfroin garbage
haul-away to an Incinerator, t'lrhan renewal
may also render the location temporarily •
un'representative. It may be benofi.cial for
each Agency or .Regional Office to maintcin-
a map and photograph of each site showing in
fluencing site characteristics. These could
be updated on a periodic basis* The site
location, sampling probe material and con-
figuration should also be within the bounds
o
of those,specified in pub!ished guidelines.
* ' " ^ '*'
Figure 4 presents a stepv/ise. rei!vie>/Jand guide
to ,the verification of specific data values.
It should provide the Regional Offices with
an overall picture of the suggested processing;
of State and local air quality data.
ii. Verifying and Evaluating Annual Air Quality Averages
Summary Statistics
If no calculation or recording errors' have been
'..}'.
found, those, summary statistics which describe
the average should be checked. These may in-
clude both geometric apd arlthme/tic; me#ns,
statid'ard 3€v?atipns, and the frequency dis-
tributibn in porcentiles. Both th* standard
deviations -P:. tf.a magnitude of the .difference :
between th,a .;•.;• OVK.trie ,and' th& ar^thirietfic' means
-------
ifled•
f i • ' t *
. Aero- ;•'';
v <•
, Reject
Error
Found
Error
found
-Inteirna^.
office',
.-!•• n -.i>
Error . •: • , •>'
CorreeiKJd> :,
r?'E:iirpr>. rip't
'•• '' '"
Regional _ ^
, nfei^* - Error
: ^Olf ^q^ *: ; -.- ,. •,„...
Intor.iai Corrafcted
'
r not
.; fpiaind .; • r
i;'.' ,'• 'f{. • . ',
and/or Local
' '
Reject
Error
Error
- ..............
Foand
' Statp and
— ---.-,,-i Calihrat
.4. , DATA VERIFICATION
DATA VALUES
; CHAR^P
Error
-- • ...-,
°rrected
SPEC1F.IC
-------
V
NO
j.'sr.cM
. ';/
*
J.;:,
LOCATIOI-:
S1113 S';'.-!. I • 'j.1 J. A T .C *. G
> \' ' V * " V*l T <^ "• f"' '•"•*>
• J. ,\.>.\ **.: . -Z^O
-------
-"28.
are no re sensitive to a few oxf.remely hicih
vfiluf-s thai't to many moderately high level?..
Ins pc c t i on o i t !i e v a 1 uc s c o r r c s po n d i n g t o t i'ie
h ~\ o! i c r- p e i' c. e < i' 1:11 o::- w o u 'i d a 1 s r i s h o >.•/ t h a i n -
f'lhence of abnormally irigh values. On the
average, standard devistions do not cjenorally
change much from yer.r-to-year.
List I nci i v i c! u •?, 1 Values
If the summary statistics indicate that the
mean was heavily influenced by a few hi oh
values, or in the absence of sun.mary statistics,
the individual data values which comprised the
averacie shouU1 be listed. From inspection of
this list, it can be determined if the a'verege
was influenced by a relatively few large values
or whether the bulk of the data appears to he
consistently high. If the former appears to be
the situation, each individual data value should
be treated according to the guidelines for
specific air quality data points presented above
In the latter case, proceed to the next step in
the verification of annual averages.
Physical SiU Inspection
The physical site location should be evaluated
in terms ci •;•„•!. i ipresentativenass of the pol-
lutant of hvi ••.•;-e!;t, the averaging time of in-
-------
-29-
terest, the averaging time of interest and the
pollutant receptor. The operation of i:ne site
should bo evaluated in terns of samp'! ~< ;u.!
methodology, maintenance procedures, cr,Ti~
bratioii procedL:ras arid qua'] i;ty contro"I practi : :•;i
The actual scinpling probe am.; manifold nu>.t^r ' - 1
configuration and placement should also be
evaluated. Guidelines describing in detail
these aspects of air quality inoni torinfj have
been published. J/!'''
Plot Data
Comparing i\ visual plot of t:he current data to
that of prior ears on a typical annual pattern
could further pinpoint reasons to accept or re-
ject the annual average in question. Note thct;
however, seme year-to-year variation is expecte'.
Figure 5 presents a typical 502 annua^ pattern
based on expected mcntii'iy averages (exaggerated
for purposes of i 11 us ;..r«v; :i on ). Figure 6 also
shows this sane pattern v/ith a constantly in-
creasing baseline drift. A pattern of this
type suggests a contirnir ionn-terrn failure
(change) in a cor.ipo.is.n of the instrument,,
deterioration in the supplies being used or a
subtle civ. (•:••• ,v' *'K: " -v? v'oryr?!-;t. Figure 7
i
presents I..:, vy: ,-•'• ; ,, -torn with an abrupt dis-
location of i hr- ')r>3? line. This nay be "inch"cat.i
-------
-30-
of a clu'ncic in instruments, method o:f «u,: r a t o r;, n .••• y b e
following tire procedure correctly for the first
time. Figure 8 presents a seasonal a l:> n o rnu-1 i vy
in the expected pattern. It should be kypt i r:
mind that a deviation from the expected pattern
can be negative as well as positive. Figiny
9 demonstrates how the expected pattern can };•::
smoothed (masked) by a nearby source v/h-ose
emissions are fairly constant throughout the ye
"normal" and part of the year "masked" if there
are pronounced seasonal wind direction chafes.
For those pollutants su>-:h as oxidants whose pc-••;:'
values occur during a ?i;t:le season a plot of
weekly or bi-weekly averr?-es through the per loo1
of interest '-/ould provid:- IT re information on
the cyclical pacterns in i i;unthly avertiges.
Check Prior- DC': te f o - "ivoiiL,
Plotting at "locst four ;>rt:vious annual averages
a 1 o n g v i t! -, t '.\;-. c i! r re •;;; y:.: = r and visually i n -
S|J(JCt'i ''i;.! I •«'. •: ' •"•: r : : •; 31 y? the GValuntOl"
-------
m
'' " ! ".' '.' J "'.'. I: '.'.".'.:'. I. i.'...".'.'.".'.: '•>%•<•/ *•/••'< V-' V rri'ftn •>••>' ••«« :'''.•»!'• i:» li ' 'r.;', r.-tri:;--v i I';" ';.i 'I;
j i..' •' i- •"-•! — •- •.-—.—. _.-. ,~_..i.i.. .-,•• i _—. »-..,J*^.>J... .Ji..k»i.rvU—.uL.^.. ..>- .—.-.._
i n_ i .mi, ir» i 11 j mi in inn Wi» M i i r n "til - n —ni i
.- ....;...... . . .— _...'_....-_ ... j . . ------ - -• • -- — - • --...(.—...-.-._.--.,..-»_...-. ..._t ..*-........ .. .... .«.«... ,., -*-.. -,-._ | --,....-...', ..-,.. x ; _t.— .. — _j
• ,
.f .
r:r:
_ :•. , ., _.;.. , <<..„• • T.^ -^-.,1 -1....—._.:_ i;:. r;~.~..
-i--": -•••• :*-K;— • •--- • -~ - -f—•--—••-••••• -:•-:--•" :•-•— •:" : >•-. •--—- > • -•--..:.- '••• ^.1^ -it ••• - — < , •'
^JK;h!=l£H3"^^li^^^^^^^
" i.*" r r TTT^I™ ? * ^r'.
f........,,,.-. -^
.i:r=:jj..:::".:i:--t-.-.::::.T.i::7.
T-G-:.^c:-d7 ~
_rj:.... n-^r — :
._......Jpi o^»- *
~7T:;t.'r; M w&~*
~f::-:i:-"---
:r::::™£n£==:f
:.-::: .-TIT trrHrrr
r:u!::';: Mr::ri:'
:.;.:,•{ :"4-" S2.f;«(jji
::-:;r::-!'ABtf03'
H-^^-j^, — • ;J-
: .-.;.;: r ~^i ~—~.~.~ ':: iv.::: :~zr}r:.:r r\" \ ::.r;-:..:;r: : ™i::.-? : ;•• r::":r .' • • :r: •• ~" ::~r" •
" T ' ^TT^T' ---"'-- r ^.i .:';ar^". / f .^'^.r ri- ^rmj mn.mc i m-T-irT^ , ;JTI rm r~-_ .*r^u 3
. .._._.-. _......._- ... .»-._. _.___-^.— . — . — _r*,... A.*.. ~i.i* — ... — „_..- ......
. ., .—_ — , — ... T— .--,' *-( »••
.^..'.^-..^j^j'i:.';:^..:1-. — -..,..,— -^"rrr. :..... — ^.•.^•^.j.'^y.ti:^: ":"....: :;;.::".:.:., .
::t;!^pi7^:;^^\~^:;H^HHi^^
..-.—.« «^ ^4 . . ... ,..._.. ^»^ / y y y / / t : s -»....-..- ... » - - . .. ^ . . . —-,......- . . >• ». ^ . ,._ ...,«.»...^j
i£~?:?^:^H;^f:^^
nT,JTY;. :;::::"::::: :::r^:!:"-:::::r' . ..:r.:vi :.-;:;: :r: ;-.:-:• "Tj1:-::::: •-:?!;.:•.-.:"-]
^_-_-...i-. •;,•!_— _^_.:j-j ;...::-.
.T:Tr.T'i.--i(i-wjav*iiriirt.w*Af-«'w-»*»» ca
- - f~: -~ ; ' - : -. :-:-J . -. - Jj d ;::
:;^::::.:::::ir- :;::;; ::;;?!::::
. ir -.mir »-~ r r**~ t* * r ^ m i^T-.'
-::::^:::j:"*:.:;7;;:::::K:;::
•K**ra.i«MM»oNW»iirrA'iMr-« i-*^***
iV~~juL~T
-------
-32-
e q LI ft 1 i t a I i v e i n s i c h t 1 n t o w h e 1 h e r th e. c u r r «L> r.;.
annual average; is a significant deviation frcr
or an exUns-ion of the projected trend.
Compare ;'!ith Surrounding Stations
If there arc cnowoh surrounding sites to do-
ve 'I op air quality isop'ieths of the area, the
evaluator could see how the annual average in
question fits in with the overall picture;, i'cr
instance, if the point in question wes i.iic.ay
between the isopleth lines ropresentino 80 ynri
60, but the recorded value was 50% greater '
than expected, i.e., 105, an abnormality PS ay
be expected.
This comparative technique may also bs used 'in
areas where there are not enough sites to dire "I
plot air quality isopleths but where .a pre- \
dictive air quality model has been de-velopad-'
and verified with a limited number'of "actual
data values. In these cases, for exawpley da-
viations of _+ 100% could be suspect. ;
Meteoroloay
The annual average should be interpreted1 in c-p-n-
;,
junction with meteorological conditions for the
year in question. For example, if.,the w-int.er
of the.yer.-" : ovestion were the coldest in
-------
-33-;
above the 20-year horw, ah' ihtireasejjSQ2: aversgs
would be; expectod., Suspended partituVate.
values can bG greatly affected1';by wind direc-
tion and a disproportionate wind rose {atypical
•'-"' ' • , .--',v "'"••' ' ' 'lf.~ ''•''' 'i - - ' IN
for the; area) cdyld help explain unusualti^fiiifes,
. ,. •..,,-•. ^•••.- • ' •;.• ., . .,;'::•. ••..•.•• .••••: V& -••• ''Sv V
Corapar i rig -^h&^j^fi^ -'f^^ -'"
meters suchr^ai^xi'ltTfa^TV-^nd ^jj,e.ed,' nuinbd'r;'"|" •"i;:;
and length of fnv^rsfon, ^JiSnipefature artd'i
' • .-,. 'I ., .;•• ' •":;
days to their •lonig-"te¥m'<'a:yf'r:ag't^^--*1 :'^e'..,''!'
-••• . " •• ' V.'. '• ' :/-:••-">• • : '.' :^
or 50-year norrns;^ -before tttem|>t1ng :to chf
• . • .• ... . ..: '• . • . • ••• • "" •:", • , ; -• ". .- ;;-,.n •(•••. . •' t
Implementation p1;4f>:» Is
.' • • ' ' '•.• , "' •• ' ' •'<••
0, Data Bank Updata/Repl-ac^yDfeleie
As Regional Office; tetrabtioil• •wltjl ttie SAROAl-'ria'ta
bank increases, there will
prof1c1en$ wlth"the
new data,-correct ej< 1 $ 11|ngf:••.•4^^%n'^.d;;a|»|fe'i
incorpprated. in the datalifrail^^^^a^^lfell1^
error. There are the,n |htree''||^']|e^|.;s '"* *^:^^-***^»^*^'*^v'-'^r
can be processed
. (^
and delete. In
raltted on a ,se^
t , > -r i •
fled .bath on thj
^,'
?i?«»l
NinjrA,
-------
-34-
Docuwentation of each'of the transaction typ^c,
describing the processing which the data goes, through and
inch ;:d.ing the''I imitation:; of each type of transaction iia$
.:-.?•;:, p'cvidod to the: Regional Office by MDAD (SI aym^hftv ' s
:r. •-.,.- :-f June 6, 1973).
The Regional Office should use the previously
discussed procedures to determine if identified suspsnt
data should be updated, corrected or deleted by means
of these transactions.
3.2 Future Areas or Responsibility
Future areas of Regional Office responsibility with .
respect to air quality data include:
A. Quality Control "••''
Quality control practices in the operation of air
monitoring instruments., laboratory analysis and data handl'in
procedures is of the utmost i.impbrtajice in producing valid
air quality data. The Regional Offices shouTci therefore
encourage quality control programs at the State and local
- ', ;.' i ; '.'
level. To afd the Regional Offices in this effort, the
Quality Assurance and Environmerital Monitoring Laboratory,'
\ '.*'•• * ; > f
NERC/RTP, has and is developing'various manuals describing
i
In detail, procedures to be folVowed during the course of,
sampling analysis and data h'andting for various pol-r
lutants?a'b'c'd
-------
-35
The Control Programs Development Division has
developed a general guideline for State arid local quality
control programs entitled "Quality Control Practices iu
Processing Air Pollution Samples."5 This guideline!
should help the Regional Office establish ja general quant,
.1
control program at the St?te and local level.
B. Edit arid Val idati on Checks
When MDAD develops the data validation program-,
and turns both the editor and data validation programs
over to the Regional Offices, it is expected that t!:e
Regional Offices will assume the lead in initiating
edit and validation checks on the incoming data. Hig.'i
quality data should then bs transmitted to the Nationa'!.
Aerometric Data Bank via upgraded remote access computer
terminals.
-------
REFERENCES
1. .S/)i'-:0/iO Users Ij^nua.'! , Office .of Air Programs P
No. APTD 0663, EPA, Research Triangle Park,. N.C.;-' July 1971.
?.. S/\\l0,'\n lerni nn 1 Usar '-s l-'ian u a 1., 01'f i c
-------
-37"
oriji.ce Method for the: £<.»,;••!: iiiuous Korn tors of Crirto;.
Monoxide in the Atii:-..-'sphere,
b. Refr:rer,co MetliOil Tor th? Df.icrnjina !.i on of 5uSpaded
Particulates in i;];? Atmosplicre (Mirjh Value lif.'tlsod).
c. Reference Method for Mcasureucnt of Pliotociieiiiical Oxide:!;
d. Reference Method for the Doierrofnatu.n of Sulfur Dioxide
in the A i.IT.'05,phere.
-------