Regulatory Guide 1.145

Atmospheric Dispersion Models for Potential Accident Consequence Assessments at Nuclear Power Plants

ML12216A014
Person / Time
Issue date:	08/31/1979
From:	Office of Nuclear Regulatory Research, NRC/OSD
To:
References
RG-1.145
Download: ML12216A014 (14)

U.S. NUCLEAR REGULATORY COMMISSIONAugust 1979)REGULATORY GUIDECOFFICE OF STANDARDS DEVELOPMENTREGULATORY GUIDE 1.145ATMOSPHERIC DISPERSION MODELS FOR POTENTIAL ACCIDENTCONSEQUENCE ASSESSMENTS AT NUCLEAR POWER PLANTS

A. INTRODUCTION

Section 100.10 of 10 CFR Part 100, "ReactorSite Criteria," states that meteorological condi-tions at the site and surrounding area shouldbe considered in determining the acceptabilityof a site for a power reactor. Section 50.34 of10 CFR Part 50, "Domestic Licensing ofProduction and Utilization Facilities," requiresthat each applicant for a construction permit oroperating license provide an analysis andevaluation of the design and performance ofstructures, systems, and components of thefacility with the objective of assessing the riskto public health and safety resulting from theoperation of the facility. Section 50.34 of 10CFR Part 50 also states that special attentionshould be directed to the site evaluationRegulatory Guide 1.3, "Assumptions Used forEvaluating the Potential adiological Con-sequences of a Loss of 'tlant Accident forBoiling Water Reactors,"'. gulatory Guide1.4, "Assumptions Use fo aluating thePotential Radiological seque es of a Lossof Coolant Accident Pressurized WaterReactors." A nn ther regulatoryguides also inclu e endations for orreferences to r olo analyses of potentialaccidents. The lp of the specific cri-teria discusse inAo these other analyseswill be conide a case- by- case basis.Until suc pe generic guidelines aredeveloped h analyses, the methodologyprovid in .s ide is acceptable to the NRCstaff.factors identified in 10 CFR Part 100 in the -"

B. DISCUSSION

assessment of the site.spheric diffusion' models describedThe regulatory positions presented in this t__gde reflect review of recent experi-guide represent a substantial change from pro- ata on diffusion from releases atcedures previously used to determine relative n level without buildings present andconcentrations for assessing the poten ro releases at various locations on reactoroffsite radiological consequences for a range cility buildings during stable atmosphericpostulated k.accidental releases of radioacti ditions with low windspeeds (Refs. 1material to the atmosphere. These procedure rough 6). These tests verify the existence ofnow include consideration of plume me r, I ffluent plume "meander" under light wind-directional dependence of rs-ion speed conditions and neutral (D) and stableconditions, and wind frequencies for rious (E, F, and G) atmospheric stability conditionslocations around actual exclusion area o (as defined by the AT criteria in Regulatorypopulation zone (LPZ) boundaries. Guide 1.23, "Onsite Meteorological Programs").Effluent concentrations measured over a periodThe direction-dependent approach was devel- of 1 hour1.157407e-5 days <br />2.777778e-4 hours <br />1.653439e-6 weeks <br />3.805e-7 months <br /> under such conditions have beenoped to provide an improved basis for the Part shown to be substantially lower than would be100-related review of propose ctor and site predicted using the traditional curves (Ref. 7)considerations. Accordingl i de pro-vides an acceptable meetho Lo deter-mining site-specific relativ ~ncentrations(x/Q) and should be t de abk tining x/Qvalues for the eov atiohte discussed inUSR U 'ORY GUIDESRatb" Ginesa we rs ai d mako wsilota to thoe pcblimohotids foceeoivd withoI Na " mort1 part oe thecommisuion's higiuido, the ned the son in oark-of F"It ' l " or or so Provide guidanc inpinlm g ih diorn is Im rowuirodMLod& and soluions d"Word fromso out in to guids will b acceptbli N mw wdo a boo for O ---goConUT6iorcommet and sugndons for irpovrwo es on

• gulds we -at am ,M and gulden will be revisd, as approrio, t ooo Corn-inee and tor now vrow siden or S '0 Howeve. omfwf anMb guide, N MOO" abo dut too 01 ft IMIM Y. 130Psitiolef unlul In owb~n ft need for an a* rolsotIn discussions throughout this regulatory guide, atmos-pheric dispersion w/il be considered as consisting of two compo-nents: atmospheric transport due to organized or mean airflowwithin the atmosphere and atmospheric diffusion due todisorganized or random air motions.Comments should be sent to the Secreta" of tohe Commilsio U.S. NudesRegulatory Commission. Washington. D.C. 2M Attention: Docketing andService Branch.The guides am issued m the following ton broed divisions:1. Power Reactors 6. Products2. RPsemch end Teat Reactors 7. Transporttion3. =e mid Materias Faclties .&HOccupetiol'Health4. .end Sti 9 Antitrust and Financial Review5. Materii nd Pn Prootection 10. GeneralRequests for singto copies of issued guides 1Iwiuich mey be rrocdior forPlanant soan en autoutfeic distribution list for @in& Uopese Of future gudeaIn l: e 11ic divsionasahould be indef In vuviting to U.S. NWoolt RegulatoryCommi ,ion. WNington, D.C. 2056, Attenaion: Director, Division ofTesdts" kormetlon nd Document Control.

of lateral and vertical plume spread, which arefunctions of atmospheric stability and down-wind distance.The procedures in this guide also recognizethat atmospheric dispersion conditions andwind frequencies are usually directionallydependent; that is, certain airflow directionscan exhibit substantially more or less favorablediffusion conditions than others, and the windcan transport effluents in certain directionsmore frequently than in others. The pro-cedures also allow evaluations of atmosphericdispersion for directionally variable distancessuch as a noncircular exclusion area boundary.

C. REGULATORY POSITION

This section identifies acceptable methods for(1) calculating atmospheric relative concentra-tion (x/Q) values, (2) determining x/Q valueson a directional basis, (3) determining x/Qvalues on an overall site basis, and (4)choosing X/Q values to be used in evaluationsof the types of events described in RegulatoryGuides 1.3 and 1.4.Selection of conservative, less detailed siteparameters for the evaluation may be sufficientto establish compliance with , regulatoryguidelines.I. CALCULATION OF ATMOSPHERIC RELATIVECONCENTRATION (x/Q) VALUESEquations and parameters presented in thissection should be used unless unusual siting,meteorological, or terrain conditions dictate theuse of other models or considerations. High-quality site-specific atmospheric diffusion testsmay be used as a basis for modifying the equa-tions and parameters.1. 1 Meteorological Data InputThe meteorological data needed for x/Q cal-culations include windspeed, wind direction,and atmospheric stability. These data shouldrepresent hourly averages as defined in regu-latory position 6. a of Regulatory Guide 1. 23.Wind direction should be classed into 16 com-pass directions (22.5-degree sectors, centeredon true north, north-northeast, etc. ).Atmospheric stability should be determinedby vertical temperature difference (AT)between the release height and the 10-meterlevel or by other well-documented parametersthat have been substantiated by %diffusion data.Acceptable stability classes are given in Table2 of Regulatory Guide 1.23.Calms should be defined as hourly averagewindspeeds below the vane or anemometerstarting speed, whichever is higher (to reflectlimitations in instrumentation). If the instru-mentation program conforms to the regulatoryposition in Regulatory Guide 1.23, calms shouldbe assigned a windspeed equal to the vane oranemometer starting speed, whichever is 0higher. Otherwise, consideration of a con-servative evaluation of calms, as indicated bythe system, will be necessary. Wind directionsduring calm conditions should be assigned inproportion to the directional distribution ofnoncalm winds with speeds less than 1.5 metersper second. 21.2 Determination of Distances for x/Q CalculationsFor each wind direction sector, x/Q valuesfor each significant release point should becalculated at an appropriate exclusion areaboundary distance and outer low populationzone (LPZ) boundary distance. The followingprocedure should be used to determine thesedistances. The procedure takes into considera-tion the possibility of curved airflow tra-Jectories, plume segmentation (particularly inlight wind, stable conditions), and the poten-tial for windspeed and direction frequencyshifts from year to year.For each of the 16 sectors, the distance forexclusion area boundary or outer LPZ bound-ary x/Q calculation should be the minimumdistance from the stack or, in the case ofreleases through vents or building penetra-tions, the nearest point on the building to theexclusion area boundary or outer LPZboundary within a 45-degree sector centeredon the compass direction of interest.For stack releases,, the maximum ground-level concentration in a sector may occurbeyond the exclusion area boundary distanceor outer LPZ boundary distance. Therefore,for stack releases, x/Q calculations should bemade in each sector at each boundary distanceand at various distances beyond the exclusionarea boundary distance to determine themaximum relative concentration for considera-tion in subsequent calculations.1.3 Calculation of X/Q Values at Exclusion Area BoundaryDistancesRelative concentrations that can beassumed to apply at the exclusion areaboundary for 2 hours2.314815e-5 days <br />5.555556e-4 hours <br />3.306878e-6 weeks <br />7.61e-7 months <br /> immediately following anaccident shouid be determined.3 Calculationsbased on meteorological data averaged over a1-hour period should be assumed to apply forthe entire 2-hour period. This assumption isreasonably conservative considering the smallvariation of x/Q values- with averaging time(Ref. 8). If releases associated with a postu-lated event are estimated to occur in a period2Staff experience has shown that noncalm windspeeds below1.5 meters per second provide a reasonable range for definingthe distribution of wind direction during light winds.3See 100.II of 10 CIR Part 100.1.145-2 of less than 20 minutes, the applicability of themodels should be evaluated on a case-by-casebasis.Procedures for calculating "2- hour" x/Qvalues depend on the mode of release. Theprocedures are described below.1.3.1 Releases Through Venzts fn Othee Ruilding P-enetrationsIhis class of release modes includes allrelease points or areas that are effectivelylower than two and one-half times the height ofadjacent solid structures (Ref. 9). Within thisclass, two sets of meteorological conditions aretreated differently, as follows:a. During neutral (D) or stable (E, F,or G) atmospheric stability conditions when thewindspeed at the 10-mete.r level is less than 6meters per second, horizontal plume meandercan be considered. X/Q values may be deter-mined through selective use of the following setof equations for ground-level relative concen-trations at the plume centerline:A is the smallest vertical-plane cross-sectional area of the reactor build-ing, in m2.(Other structures and/or : directional consideration maybe justified when appropriate. )x/Q values should be calculated usingEquations 1, 2, and 3. The values from Equa-tions I and 2 should be compared and thehigher value selected. This value should becompared with the value from Equation 3, andthe lower value of these two should be selectedas the appropriate xiQ value. Examples and adetailed explanation of the rationale for deter-mining the controlling conditions are given inAppendix A to. this guide.b. During all other meteorological condi-tions [unstable (A, B, or C) atmosphericstability and/or 10-meter level windspeeds of 6meters per second or more], plume meandershould not be considered. The appropriate x/Qvalue is the higher value calculated fromEquation 1 or 2.1.3.2 Stack Releasesx/Q =1UIo(1OyOz + A/2)(1) This class of release modes includes allrelease points at levels that are two and one-half times the height of adjacent solid struc-tures or higher (Ref. 9). Nonfumigation and(2) fumigation conditions are treated separately.X/Q -1Uio(3u y a Z)X/Q -IUloltly aza. For nonfumigation conditions, theequation for ground-level relative concentrationat the plume centerline for stack releases is:(3)wherex/Q is relative concentration, in sec/ms,n is 3.14159,U10 is windspeed at 10 meters aboveplant grade,4 in m/sec,a is lateral plume spread, in m, aY function of atmospheric stabilityand distance (see Fig. 1),o is vertical plume spread, in m, az function of atmospheric stabilityand distance (see Fig. 2),Y is lateral plume spreaswith meanderY and building wake effects, in m, afunction of atmospheric stability,windspeed U10, and distance [fordistances of 800 meters or less,I = Mo , where M is determinedfrvom Fil. 3; for distances greaterthan 800 meters, y = (M -1)ay800m + y]I, and4the 10-meter level is representatve of the depth throughwhich the plume is mixed with building wake effects.x/Q 1 r-h 1nyzwhere(4)Uh is windspeed representing conditionsat the release height, in m/sec,he is~effective stack height, in m:h = ht,heh is the initial height of the plume(usually the stack height) aboveplant grade, in m, andht is the maximum terrain height aboveplant grade between the releasepoint and the point for which thecalculation is made, in m; ht cannotexceed hs.b. For fumigation conditions, a "fumiga-tion x/Q" should be calculated for each sectoras follows. The equation for ground-level rela-tive concentration at the plume centerline forstack releases during fumigation conditions is:1.145-3 x/Q = 1 , h > 0 (5)(2701/2 Uh ayheeywhereEh is windspeed representative of thee layer of depth he , in m/sec; in lieuof information to the contrary, theNRC staff considers a value of 2meters per second as a reasonablyconservative assumption for h ofabout 100 meters, and eo is the lateral plume spread, in m,y that is representative of the layer ata given distance; a moderately stable(F) atmospheric stability condition isusually assumed.Equation 5 cannot be applied indiscrimi-nately because the x/Q values calculated, usingthis equation, become unrealistically large ash becomes small (on the order of 10 meters).Tie x/Q values calculated using Equation 5must therefore be limited by certain physicalrestrictions. The highest ground-level x/Qvalues from elevated releases are expected tooccur during stable conditions with low wind-speeds when the effluent plume impacts on aterrain obstruction (i.e., h = 0). However,elevated plumes diffuse upv$ard through thestable layer aloft as well as downward throughthe fumigation layer. Thus ground-levelrelative concentrations for elevated releasesunder fumigation conditions cannot be higherthan those produced by nonfumigation, stableatmospheric conditions with h = 0.. For thefumigation case that assumes F stability and awindspeed of 2 meters per second, Equation 4should be used instead of Equation 5 atdistances greater than the distance at whichthe x/Q values, determined using Equation 4with he = 0, and Equation 5 are equal.1.4 Calculation of x/Q Values at Outer LPZ BoundaryDistancesTwo- hour x/Q values should also be cal-culated at outer LPZ boundary distances. Theprocedures described above for exclusion areaboundary distances (see regulatory posi-tion 1.3) should be used.An annual average (8760-hour) x/Q shouldbe calculated for each sector at the outer LPZboundary distance for that sector, using themethod described in regulatory position 1.c ofRegulatory Guide 1.111, "Methods for Estimat-ing Atmospheric Transport and Dispersion ofGaseous Effluents in Routine Releases fromLight-Water-Cooled Reactors." (For stack re-leases, h should be determined as describedin regulaeory position 1.3.2.)These calculated 2-hour and annual averagevalues are used in regulatory position 2.2 todetermine sector X/Q values at outer LPZboundary distances for various longer timeperiods. 52. DETERMINATION OF MAXIMUM SECTOR x/QVALUESThe x/Q values calculated in regulatory posi-tion 1 are used to determine "sector x/Qvalues" and "maximum sector x/Q values" forthe exclusion area boundary and the outer LPZboundary.2.1 Exclusion Area Boundary2.1.1 General MethodUsing the x/Q values calculated for eachhour of data according to regulatory posi-tion 1.3, a cumulative probability distributionof x/Q values should be constructed for each ofthe 16 sectors. Each distribution should bedescribed in terms of probabilities of given x/Qvalues being exceeded in that sector duringthe total time. A plot of x/Q versus probabilityof being exceeded should be made for eachsector, and a curve should be drawn to forman upper bound of the data points. From eachof the 16 curves, the x/Q value that isexceeded 0.5% of the total time should beselected (Ref. 10).. These are the sector x/Qvalues. The highest of the 16 sector values isdefined as the maximum sector x/Q value.2.1.2 Fumigation Conditions for Stack ReleasesRegulatory position 1.3.2 gave proce-dures for calculating a fumigation x/Q for eachsector. These sector fumigation values, alongwith the general (nonfumigation) sector valuesobtained in regulatory position 2.1.1, are usedto determine appropriate sector x/Qs for fumi-.gation conditions, based on conservativeassumptions concerning the duration of fumiga-tion. These assumptions differ for inland andcoastal sites, and certain modifications may beappropriate for specific sites.a. Inland Sites: For stack releases atsites located 3200 meters or more from largebodies of water (e.g., oceans or Great Lakes),a fumigation condition should be assumed toexist at the time of the accident and continuefor 1/2 hour (Ref. 11). For each sector, if thesector fumigation x/Q exceeds the sector non-fumigation x/Q, use the fumigation value forthe 0 to 1/2-hour time period and the nonfumi-gation value for the 1/2-hour to 2-hour timeperiod. Otherwise, use the nonfumigationsector value for the entire 0 to 2-hour timeperiod. The 16 (sets of) values thus deter-mined will be used in dose assessments requir-ing time-integrated concentration considera-tions.'58M 5100.11 of 10 CFR Part 100.01.145-4 b. Coastal Sites: For stack releases atsites located less than 3200 meters from largebodies of water, a fumigation condition shouldbe. assumed to exist at the exclusion areaboundary at the time of the accident andcontinue for the entire 2-hour period. For eachsector, if the sector fumigation x/Q exceedsthe sector nonfumigation x/Q, use the fumiga-tion value for the 2-hour period. Otherwise,use the nonfumigation value for the 2-hourperiod. Of the 16 sector values thus deter-mined, the highest is the maximum sector x/Qvalue.c. Modifications: These conservative as-sumptions do not consider frequency and dura-tion of fumigation conditions as a function ofairflow direction. If information can be pre-sented to substantiate the likely directionaloccurrence and duration of fumigation condi-tions at a site, the assumptions of fumigation inall appropriate directions and of duration of1/2 hour and 2 hours2.314815e-5 days <br />5.555556e-4 hours <br />3.306878e-6 weeks <br />7.61e-7 months <br /> for the exclusion areaboundary may be modified. Then fumigationneed only be considered for airflow directionsin which fumigation has been determined tooccur and of a duration determined from thestudy of site conditions. 62.2 Outer LPZ Boundary2.2.1 General MethodSector x/Q values for the outer LPZboundary should be determined for varioustime periods throughout the course of thepostulated accident. " The time periods shouldrepresent appropriate meteorological regimes,e.g., 8 and 16 hours1.851852e-4 days <br />0.00444 hours <br />2.645503e-5 weeks <br />6.088e-6 months <br /> and 3 and 26 days aspresented in Section 2.3.4 of RegulatoryGuide 1.70, "Standard Format and Content ofSafety Analysis Reports for Nuclear PowerPlants--LWR Edition," or other time periodsappropriate to. release durations.For a given sector, the average x/Qvalues for the various time periods should beapproximated by a logarithmic interpolationbetween the 2-hours sector x/Q and the annualaverage (8760-hour) x/Q for the same sector.The 2-hour sector x/Q for the outer LPZboundary is determined using the generalmethod given for the exclusion area boundaryin regulatory position 2.1. The annual average6For example, examination of site-specific information at a lo-cation in a pronounced river valley may indicate that fumigationconditions occur only during the downvalley "drainage flow"regime and persist for durations of about 1/2 hour. Therefore,in this case airflow directions other than the downvalley direc-tions can be excluded from consideration of fumigation condi-tions. and the duration of fumigation would still be consideredas 1/2 hour. On the other hand, data from sites in open terrain(noncoastal) may indicate no directional preference for fumiga-tion conditions but may indicate durations much less than 1/2hour. In this case, fumigation should be considered for alldirections, but with durations of less than 1/2 hour.?See §100.11 of 10 CFR Part 100.*The X/Qs are based on 1-hour averaged data but are as-sumed to apply for 2 hours2.314815e-5 days <br />5.555556e-4 hours <br />3.306878e-6 weeks <br />7.61e-7 months <br />.x/Q for a given sector is determined asdescribed in regulatory position 1.4.The logarithmic interpolation procedureproduces results that are consistent withstudies of variations of average concentrationswith time periods up to 100 hours0.00116 days <br />0.0278 hours <br />1.653439e-4 weeks <br />3.805e-5 months <br /> (Ref. 8).Alternative methods should also be consistentwith these studies.For each time period, the highest of the16 sector x/Q values should be identified. Inmost cases, these highest values will occur inthe same sector for all time periods. These arethen the maximum sector x/Q values. However,if the highest sector x/Qs do not all occur inthe same sector, the 16 (sets of) values will beused in dose assessments requiring time-integrated concentration considerations. Thex/Q values for the various time periods withinthat sector should be considered the maximumsector x/Q values.2.2.2 Fumigation Conditions for Stack ReleasesDetermination of sector x/Q values forfumigation conditions at the outer LPZboundary involves the following assumptionsconcerning the duration of fumigation for in-land and coastal sites:a. Inland Sites: For stack releases atsites located 3200 meters or more from largebodies of water, a fumigation condition shouldbe assumed to exist at the outer LPZ boundaryat the time of the accident and continue for 1/2hour. Sector x/Q values for fumigation shouldbe determined as for the exclusion area bound-ary in regulatory position 2.1.2.b. Coastal Sites: For stack releases atsites located less than 3200 meters from largebodies of water, a fumigation condition shouldbe assumed to exist at the outer LPZ boundaryfollowing the arrival of the plume and continuefor a 4-hour period. Sector X/Q values forfumigation should be determined as for theexclusion area boundary in regulatory posi-tion 2.1.2.c. The modifications discussed in regula-tory position 2.1.2 may also be considered forthe outer LPZ boundary.3. DETERMINATION OF 5% OVERALL SITE x/QVALUEThe x/Q values that are exceeded no morethan 5%. of the total time around the exclusionarea boundary and around the outer LPZboundary should be determined as follows(Ref. 10):Using the x/Q values calculated accordingto regulatory position 1, an overall cumulativeprobability distribution for all directions com-bined should be constructed. A plot of x/Qversus probability of being exceeded should be1. 145-5 made, and an upper bound curve should bedrawn. The 2-hour x/Q value that is exceeded5% of the time should be selected from thiscurve as the dispersion condition indicative ofthe type of release being considered. Inaddition, for the outer LPZ boundary themaximum of the 16 annual average x/Q valuesshould be used along with the 5% 2-hour x/Qvalue to determine -X/Q values for theappropriate time periods by logarithmicinterpolation.4. SELECTION OF x/Q VALUES TO BE USED INEVALUATIONSThe x/Q value for exclusion area boundaryor outer LPZ boundary evaluations should bethe maximum sector x/Q (regulatory position 2)or the 5% overall site x/Q (regulatory posi-tion 3), whichever is higher. All direction-dependent sector values should be presentedfor consideration of the appropriateness of theexclusion area and outer LPZ boundaries andthe efficacy of evacuation routes and emer-gency plans. Where the basic meteorologicaldata necessary for the analyses describedherein substantially deviate from the regula-tory position stated in Regulatory Guide 1.23,consideration should be given to the resultinguncertainties in dispersion estimates.

D. IMPLEMENTATION

This proposed guide has been released toencourage public participation in its develop-ment and is not intended to foreclose other op-tions in safety evaluations. Except in thosecases in which an applicant proposes anacceptable alternative method for complyingwith specified portions of the Commission'sregulations, the method to be described in theactive guide reflecting public comments will beused in the evaluation of applications tenderedon or after the implementation date to bespecified in the active guide (in no case willthis date be earlier than November 1, 1979) asfollows:1. For early site review applications.2. For construction permit applications (in-cluding those incorporating or refer-encing a duplicate plant design and thosesubmitted under the replicate plantoption of the Commission's standardiza-tion program).For the following cases, either the proposedguide or the procedures described in StandardReview Plan Section 2.3.4 (1975) may be used:1. Construction permit applications tend-ered before the implementation date.2. Operating license applications whose con-struction permits precede the implemen-tation date.3. Operating reactors.This proposed guide does not apply to thefollowing options specified in the Commission'sstandardization policy under the referencesystem concept:1. Preliminary design approval applications.2. Final design approval, Type 1, appli-cations.3. Final design approval, Type 2, appli-cations.4. Manufacturing license applications.1.145-6

1035210z005--. -. --~-----~ -.-4--~-I IjIj Ii-T- IiilliI/ic,-4----- --4- -4--4-4-4-4-4-4---~--, 4 -4~'4- 4 f 41,D~LE' IA- EXTREMELY UNSTABLEMODERATELY UNSTABLE.C -SLIGHTLY UNSTABLET -NEUTRALE- SLIGHTLY STABLEF MODERATELY STABLE-2 lI01=_, 0 [4.10 102 25 103 2 5 104 2DISTANCE FROM SOURCE (W5 105Figure 1. Lateral diffusion without meander and building wake effects, oa, vs. down-wind distance from source for Pasquill's turbulence types (atmosphericstability) (Ref. 7).For purposes of estimating u during extremely stable (G) atmosphericstability conditions, without pl~ne meander or other lateral enhancement,the following approximation is appropriate:Oy(G) = 3-y(F)1.145-7

3. 032-z0,.210S0Ib"2010lo2 5 103 2 5 101 2 5DISTANCE FROM SOURCE (m)105Figure 2. Vertical diffusion without meander and building wake effects,z, vs. downwind distance from source for Pasquill's turbulencetypes (atmospheric stability) (Ref. 7).For purposes of estimating oz during extremely stable (G) atmosphericstability conditions, the following approximation is appropriate:az(G) = Vz(F)1.145-8 Stabi I ityClass6Ga-0E3-1 2 3 4 5 6 10WINDSPEED (m/sec)Figure 3. Corect!on factors for Pasquill-Gifford a values by atmospheric stability class(see Appendix A to this guide)1.145-9 APPENDIX AATMOSPHERIC DIFFUSION MODEL FOR RELEASES THROUGH VENTSAND BUILDING PENETRATIONSRationaleThe effects of building wake mixing and am-bient plume meander on atmospheric dispersionare expressed in this guide in terms of condi-tional use of Equations 1, 2, and 3.Equations 1 and 2 are formulations that havebeen acceptable for evaluating nuclear powerplant sites over a period of many years (Ref. 7and Regulatory Guides 1.3 and 1.4) but haverecently been found to provide estimates ofground-level concentrations that are consist-ently too high during light wind and stable orneutral atmospheric conditions for 1-hour re-lease durations (Refs. 1 through 6).Equation 3 is an empirical formulation basedon NRC staff analysis of atmospheric diffusionexperiment results (Ref. 2). The NRC staffexamined values of lateral plume spread withmeander and building wake effects (I ) byatmospheric stability class (based on ATY, cal-culated from measured ground-level concentra-tions from the experimental results. Plots ofthe computed Y values by atmospheric stabil-ity class and downwind distance were analyzedconservatively but within the scatter of thedata points by virtually enveloping most testdata. The resultant analysis is the basis forthe correction factors applied to the Pasquill-Gifford a values (see Fig. 3 of this guide).Thus, Eq~aation 3 identifies conservatively thecombined effects of increased plume meanderand building wake on diffusion in thehorizontal crosswind direction under light windand stable or neutral atmospheric conditions,as quantified in Figure 3. These experimentsalso indicate that vertical building wake mixingis not as complete during light wind, stableconditions as during moderate wind, unstableconditions although the results could not bequantified in a generic manner.The conditional use of Equations 1, 2, and 3is considered appropriate because (1) horizon-tal plume meander tends to dominate dispersionduring light wind and stable or neutral condi-tions and (2) building wake mixing becomesmore effective in dispersing effluents thanmeander effects as the windspeed increases andthe atmosphere becomes less stable.Examples of Conditional Use of Diffusion EquationsFigures A-l, A-2, and A-3 show plots ofxUo/Q (x/Q multiplied by the windspeed Ulo)versus downwind distance based on the condi-tional use (as described in regulatory posi-tion 1.3.1) of Equations 1, 2, and 3 duringatmospheric stability class G. The variable Mfor Equation 3 equals 6, 3, and 2 respectivelyin Figures A-l, A-2, and A-3 (M is as definedin regulatory position 1.3.1). The windspeedconditions are those appropriate for G stabilityand M =6, 3, and 2.In Figure A-l, the XU1o/Q from Equation 3(M = 6) is less than the higher value fromEquation I or 2 at all distances. Therefore, forM = 6, Equation 3 is used for all distances.In Figure A-2, the xUo/Q from Equation 3(M = 3) is less than the higher value fromEquation 1 or 2 beyond 0.8 kln. Therefore, forM = 3, Equation 3 is used beyond 0.8 km. Fordistances less than 0.8 kin, the value fromEquation 3 equals that from Equation 2.Equation 2 is therefore used for distances lessthan 0.8 km.In Figure A-3, the x-uo/Q from Equation 3(M = 2) is never less than the higher valuefrom Equation 1 or 2. Therefore, for M = 2,Equation 3 is not used at all. Instead, Equa-tion 2 is used up to 0.8 km, and Equation 1 isused beyond 0.8 km.1.145-10

CY0.1 1.0 10PLUME TRAVEL DISTANCE (km)Figure A-1. xU10/Q as a function of plume travel distance for G stability conditionusing Equations 1, 2. and 3 (M = 6).1.145-11 o0.1 1.0 10PLUME TRAVEL DISTANCE (km)Figure A-2. x910/0 as a function of plume trvel distance for G stability usingEquations 1, 2, and 3 (M -3).1.145-12 Ik Eq. 3 (M=2)I II10-2--H10-oEq. I__ _i I _____ __ __ -q. 3 j(M=2).q. Eq. 2___ ___ I '!ii-410-10-s0.11.0PLUME TRAVEL DISTANCE (km)10Figure A-3. xUj10/Q as a function of plume travel distance for G stability conditionusing Equations 1, 2, and 3 (M = 2).1.145-13 REFERENCES1. Van der Hoven, I., "A Survey of FieldMeasurements of. Atmospheric DiffusionUnder Low-Wind Speed Inversion Condi-tions," Nuclear Safety, Vol. 17, No. 4,March-April 1976.2. Start, G. E., et al., "Rancho Seco Build-ing Wake Effects On Atmospheric Diffu-sion," NOAA Technical Memorandum ERLARL-69, Air Resources Laboratory, IdahoFalls, Idaho, November 1977, availablefrom Publication Services, EnvironmentalResearch Laboratories, National Oceanicand Atmospheric Administration, Boulder,Colorado-80302.-3. Wilson, R. B., et al., "Diffusion UnderLow Windspeed Conditions Near Oak Ridge,Tennessee," NOAA Technical MemorandumERL ARL-61, Air Resources Laboratory,Idaho Falls, Idaho, 1976, available fromPublication Services, Environmental Re-search Laboratories, National Oceanic andAtmospheric Administration, Boulder,Colorado 80302.4. Sagendorf, J. F., and C. R. Dickson,"Diffusion Under Low Windspeed, InversionConditions," NOAA Technical MemorandumERL ARL-52, Air Resources Laboratory,Idaho Falls, Idaho, 1974, available fromPublication Services, Environmental Re-search Laboratories, National Oceanic andAtmospheric Administration, Boulder,Colorado 80302.5. Gulf States Utilities Company, "Dispersionof Tracer Gas at the Proposed River BendNuclear Power Station," Preliminary SafetyAnalysis Report, Amendment 24, DocketNumbers 50-458 and 50-459, 1974.6. Metropolitan Edison Company, "AtmosphericDiffusion Experiments with SF6 Tracer Gasat Three Mile Island Nuclear Station UnderLow Wind Speed Inversion Conditions,"Final Safety Analysis Report, Amend-ment 24,' Docket Number 50-289, 1972.7. Gifford, F. A., Jr., "An Outline of Theoriesof Diffusion in the Lower Layers of the At-mosphere," Chapter 3 in Meteorology andAtomic Energy--1968 (D. H. Slade, Ed.),available as TID-24190 from the NationalTechnical Information Service, Springfield,Virginia 22151.8. Gifford, F., "Atmospheric Dispersion Modelsfor Environmental Pollution Applications,"Lectures on Air Pollution and EnvironmentalImpact Analyses, American MeteorologicalSociety, pp. 35-38, 1975.9. Snyder, W. H., and R. E. Lawson, Jr.,"Determination of a Necessary Height for aStack Close to a Building -A Wind TunnelStudy," Atmospheric Environment, Vol. 10,pp. 683-691, Pergamon Press, 1976.10. Memorandum from D. R. Muller to H. R.Denton, dated July 25, 1978, Subject:"Meteorological Model for Part 100 Evalua-tions," and August 2, 1978 reply.11. Van der Hoven, I., "Atmospheric Transportand Diffusion at Coastal Sites," NuclearSafety, Vol. 8, pp. 490-499, 1967.1. 145-14