{"Bibliographic":{"Title":"Urban Air Pollution Modelling","Authors":"","Publication date":"1970","Publisher":""},"Administrative":{"Date created":"08-17-2023","Language":"English","Rights":"CC 0","Size":"0000031182"},"Pages":["A\nQC\n880\nA4\nno 1.37\nNOAA Research Laboratories\nAir Resources\nAtmospheric Turbulence and Diffusion Laboratory\nOak Ridge, Tennessee\nDecember 1970\nURBAN AIR POLLUTION MODELLING\nF. A. Gifford, Jr.\nSteven R. Hanna\nU.S. DEPARTMENT OF COMMERCE\nNATIONAL OCEANIC AND ATMOSPHERIC ADMINISTRATION","A\nQC\n880\nA4\nno.37\nURBAN AIR POLLUTION MODELLING\nF. A. Gifford, Jr. and Steven R. Hanna\nAir Resources\nAtmospheric Turbulence and Diffusion Laboratory\nNational Oceanic and Atmospheric Administration\nOak Ridge, Tennessee\nATMOSPHERIC SCIENCES\nLIBRARY\nJUL 19 1972\nN.O.A.A.\nU. S. Dept. of Commerce\n(Presented at 1970 International Air Pollution Conference of the Inter-\nnational Union of Air Pollution Prevention Associations.) .\nATDL Contribution No. .\n37\n.\n'72\n3746","URBAN AIR POLLUTION MODELLING\nF. A. Gifford, Jr. and Steven R. Hanna\nAir Resources\nAtmospheric Turbulence and Diffusion Laboratory\nNational Oceanic and Atmospheric Administration\nOak Ridge, Tennessee\nABSTRACT\nA simple but physically realistic model of the ground level concentration\ndistribution resulting from area sources of pollution is presented. It is shown\nthat the results are not greatly dependent upon the form of the vertical concen-\ntration distribution. This area-source model, which satisfies the two-dimensional\nequation of diffusion with height-variable wind and diffusivity, is well adapted\nto simple, hand computation. The ground level concentration in any grid square\nof a two-dimensional grid covering the area sources is given by the sum of the\nsource strengths of all the grid squares, each multiplied by a simple weighting\nfactor. This factor depends mainly on distance from the receptor square, the\nfrequency with which the wind blows from that square to the receptor square,\nwind speed, and atmospheric stability. The matrix of weighting factors is inde-\npendent of the location of the receptor square, Comparisons with urban air\npollution computations based on more complex urban diffusion models (e.g. Lamb's\nmodel applied to Los Angeles, Fortak's model applied to Breman, and Martin and\nTikvart's model applied to Atlanta) illustrate that this simple model adequately\nrepresents urban diffusion for most air pollution problems.\nINTRODUCTION\nof the three great natural sinks for pollution, the air, the water, and\nthe land, only the air cannot be purified or controlled by man, once it is\ncontaminated. Material emitted into the air is removed or diluted only by\nnaturally occurring processes. Whereas river water can, and often does, undergo\nmany stages of pollution alternating with purification as it travels downstream,\nthere is no such possibility for the air. Also pollutants from all kinds of\nsources, once they are emitted to the atmosphere, mix together and can't be\ndistinguished. This means that the problem of calculating air pollution becomes\na particularly crucial one.","- 2 -\nSpecific requirements for calculating levels of urban air\npollution range from the need to design \"air quality control regions, 11\nthrough the establishment of air pellution control regulations, with\ntheir related enforcement activities, to operational (i.e. real-time)\nair resource management. All can involve quantitative estimation\nof air pollution levels, and it is the business of the air pollution\nmeteorologist to make such estimates.\nFollowing Lucas'' early study a series of simple urban air pollu-\ntion models were described in papers by Leavitt2, Pooler 3, Clarke4,\nand Miller and Holzworth5.\nThese studies have much in common.\nThey each approach the urban area-source concentration problem\nby way of the usual Gaussian point source diffusion model. Differ-\nences occur only in the details of how the area source summation is\ncarried out and in how various meteorological parameters are included.\nMore recently urban diffusion models of far greater intricacy\nhave been described, for instance by Turner6, Martin and Tikvart7,\nShieh, Davidson, and Friend8, Roberts and Fortak10. These models\nuse much the same basic point source scheme as the earlier group,\nbut differ in the detail with which meteorological factors are\nincluded. The execution of these later models requires appreciable\namounts of high speed digital computer time. Even apart from the\nexpense involved, this is a practical difficulty. If urban dif-\nfusion models are to be used in real-time pollution control, the\ntime to run the model must not be so great that it conflicts with\nthe requirement for prompt control action. Furthermore the meteor-\nological calculation is only one element in the control or analysis\nsystem. It should not consume a disproportionate share of the available","- 3 -\ncomputer time. So the question naturally arises, is this elaboration\nnecessary? Are the numbers produced by the current crop of models\nbettor than thoco produced by the earlier ones? Should we anticipate,\n11\nas Stern suggests, yet a third round of urban diffusion models,\npresumably of still greater complexity and which will consume even\nmore computer time? At the moment there is no easy answer. Probably\ndetailed models will always be studied in research applications, where\nthe need. is to understand detailed complexities of urban pollution;\nand perhaps this is justifiable. However for many operational uses,\nincluding legislative and enforcement activities as well as engineering\ndesign, it seems crucial to retain essential simplicity in urban dif-\nfusion models so that the meteorological factor can be introduced\ninto the environmental pollution problem in as uncomplicated a way\nas possible. There remains of course the possibility that a quite\nsimple model may give all or nearly all of the precision that is\navailable.\nIn the following paragraphs a simple model of urban air pollution\nis presented, based on reasonable first principles and incorporating\nsome attempt at rigor. The result, a simple, easily applied area.-\nsource concentration formula, is compared with several of the pre-\nvious formulas and the conclusion is that it performs well.\nTHE ATDL AREA-SOURCE MODEL\nWe assume, in the first place that the problem of urban air\npollution can be simplified by considering separately the isolated\npoint sources such as tall stacks and lumping the contribution of\nthe multitude of lesser sources of all types into a spatially variable","- 4 -\narca-source concentration XA This is assumed to obey the steady-\nstate diffusion equation in two independent variables,\n(=)\n( 2 )\nwhere the mean wind, u(z), blows in the x-direction. Neglect of the\ny-component of the diffusion is equivalent to the observation that\npoint-source plumes in the atmosphere tend to be long and narrow,\nand so the concentration at a point can be influenced only by sources\nin a narrow, plume-shaped upwind sector [see Gifford12.. If u and\nthe eddy-diffusivity, K, are assumed to obey the usual power laws,\n(2)\n(3)\nthe \"partial\" solution to equation (1) for the ground level concen-\ntration distribution, XAO' due to an area source has been shown by\n&ifford\nbe\n1-s\n(21-1)\ni=l\nwhere S = (m+1)/(2+m-n) and the distance, X, from the receptor point\nto the upwind edge of the city is given by X = N is the number\nof upwind grid squares in the source inventory, and Ax is the grid\nsize, as given by the usual \"checkerboard\" source=inventory pattern.\nSource strength, Q,is constant for each square.","- 5- -\nThe receptor point is assumed to be located at the center of a\nsource square. There is no particular difficulty involved in\nadapting the model to other geometries, including irrogular area-\ncource patterns.\nThe parameter B depends on the vertical concentration distri-\nbution. The total concentration distribution XA is related to the\nground-level value, XAO' by\n( 5)\nHere the reference height, z, essentially represents the \"top\" of\nthe polluted air, and increases with distance from the upwind edge\naccording to the formula\n( 6 )\nThis assures that a \"variables separable\" solution of the type of equation\n(5) will satisfy equation (1). Then, with 5 - 2/2, the value of B follows\nfrom the continuity condition\nfuce) (x,z)dz =\n( 7 )\nFrom this and equation (5),\nB = precises\n(8)\nThe constant C1 is determined from the relation C1 - 3a 1/ (m+1)","- 6 1\nwhere a is defined by the usual power-law formula for the standard deviation\nof the vertical concentration distribution\nb\n02 = a x\n( 9 )\n3\nand Z 12 302. Values of a and b based on extensive observational data have been\n14\n15\nsummarized by for instance Slade\nand Smith\nTable I is based on the values\n,\ngiven by Smith and includes our estimate of the value corresponding to Pasquill's\ntype-D (slightly stable) condition. We believe that this, rather than Smith's\n\"stable\" value is more appropriate to urban conditions; unfortunately few data\nare as yet available on diffusion over cities,\nTable I\nMeteorological Conditions:\nis\n(1-b)\na (1-b)\na\nVery unstable\n0.40\n0.91\n0.09\n.036\n:\nUnstable\n0.33\n0.86\n0.14\n.046\n:\nNeutral\n0.22\n0.80\n0.20\n.044\nEstimated Pasquill \"D\"\n0.15\n0.75\n0.25\n.037\n:\nStable\n0.06\n0.71\n0.29\n.017\n:\nFor Gaussian vertical distribution, with the above assumptions,\na\nFor a linear decrease in concentration, B : 0.4. In general it is unlikely that\nB will differ much from these values.\nEquation (4) can be generalized in the usual ways, by introducing wind\ndirection and speed frequency class intervals, and varying the meteorological\nparameter, S. The only real problem that arises in the extension of equation (4)\nto the case of annual average concentrations is that of adapting the basic recti- -\nlinear source configuration to radial wind directions other than the cardinal\nones. Our arbitrary but simple scheme for doing this is illustrated in Figure\n1.","- 7 -\nCOMPARISON WITH OTHER AREA SOURCE MODELS\nThere are at the moment no urban air pollution observations that are\naccepted as definitive for the purpose of testing diffusion models,\ngo the comparisons to follow are with the results of calculations using\n16\nseveral other urban diffusion models, those of Fortak, Lamb , and\nMartin and Tikvart. Our intention is to show that the comparatively\nsimple, easily and quickly performed calculations required by equation\n(4) give area-source concentration values comparable with these more\ncomplex models.\nFortak's model: In common with several area source diffusion\nmodels, Fortek's is based on the integration over a plane area of\nthe Gaussian plume formula. Other models employing the same idea\nare those by Turner, and Martin and Tikvart. Models such as those by\nRoberts, et al. and Shich, ot al., which employ an instantaneous Gaussian\npuff as the basic diffusion element, do not seem to us to be essentially\ndifferent from these. Fortak assumes a constant mean wind speed and\nSO the comparable special case of equation (4) is given by the values\n1/2\nThe working equation is\nm = O, of = S, and B\n1-b\n1-b\n1-b\n7r(10)\n(21-1)\n-\n101\nwhere u is the (constant) mean wind speed, and fik 1s the frequency of\nwind of this speed from direction K.","- 8 -\nEquation (10) has been compared with a representative area\nsource calculation for Bremen, Fortak's Figure 22. Using the values\nu = 3 m sec-1 , a = 0.15, and b = 0.75, corresponding to neutral\nconditions, and for a south wind, concentrationvalues predicted\nby equation (10) for each source square are given in Figure 2\nand the source data, the sum of the source strengths given in\nFortak's Figures 13 and 14, appear in Figure 3. Using these\nsource data the concentrations can be reproduced from equation\n(10) in a few minutes. The concentration pattern as well as the\nmaximum values are seen to correspond well with Fortak's isopleths,\nwhich are shown in Figure 4.\nMartin and Tikvart's model: This straightforward model is\nclosely related to Turner's area source model; both are based on\nthe Gaussian plume. The many \"Reports for Consultation\" issued\nby DHEW to establish air quality control regions in the United\nStates used a version of Martin and Tikvart's model. These con-\nsultation reports differ considerably among themselves in the\namount of air pollution source data included. The report for the\n17\nAtlanta, Georgia, region is representative of this large body\nof literature and includes a reasonable amount of source strength\ndata. We made a calculation of annual average particulate concen-\ntrations for Atlanta, based on the source strengths given in the\nAtlanta report. The annual average wind speed was used, but\nequation (10) was modified by multiplying by the wind frequency\nin each direction and summing. Neutral conditions ( a = .15, To = .75)\nwere again assumed.","- 9 -\nThe results of our Atlanta calculation are shown in Figure 5.\nWe include the concentration isopleths appearing in the DHEW report\nfor comparison. The Atlanta source data included emissions from\nnumber of tall stacks. Concentration patterns for these were\na\ncalculated separately by the usual plume rise and dispersion\n18\nformulas [see Gifford , and Briggs 19 I\nand added to the area\nsource values of equation (10) to give the totals in Figure 5,\nso that these would be comparable with the isopleths of the DHEW\nreport. Equation (10) reproduces the absolute values and general\npattern of the DHEW isopleths reasonably well. The DHEW isopleths\nseem, however, to have been smoothed. The pattern of our values\nindicates a somewhat greater elongation toward the south and the\nnorthwest, and our maximum value in Atlanta is considerably higher.\nIt should be remembered that the purpose of the DHEW studies was\nto delineate regional areas, and not necessarily to determine maxima.\nLamb's model: Having demonstrated, at least to our own satisfaction,\nby the above and several similar calculations that equation (10) re-\nproduces the results of area source calculations based on the Gaussian\nassumption, we wished to compare it with a model based on solution\nof the diffusion equation. An ambitious attempt along\nthis line is the interesting study by Lamb at UCLA,","- 10 -\nof diffusion in the Los Angeles basin. Lamb's model, although\nit employs constant eddy-diffusivities, and wind not varying with\nheight, is in other respects the most complete and flexible model\nwe have examined. For instance it includes time-variable sources,\nspace-variable winds, ground absorption, and even simple chemical\nreactions.\nWe compared Lamb's model with equation (10) using source data\non natural gas emissions in the Los Angeles Basin, for a particular\n16-hour period. Calculations of ground-level concentrations using\nthese data were kindly provided to us by Mr. Lamb. This is a very\nsevere test of equation (10). For annual or seasonal concentrations,\nsuch as in the previous comparison, it is not too surprising that\nequation (10) performs well. Over any long period the average ground\nconcentration from an area source is obviously strongly weighted\nby the local source-strength. But for a short period such as 16\nhours all possible complexities come into play.\nThe results of these calculations are illustrated in Figure 6 .\nOur model seems to be giving area-source concentration values of the\nsame order as the UCLA model, perhaps a factor of two higher.\nFigure 7 is a scatter-diagram presentation of the same information.\nThe open points come from the top three grid-rows of Figure 6.\nWe suspect that the UCLA model is computing higher values there\nbecause it takes into account flow convergence caused by the ring\nof mountains. Lamb uses a streamline and isotach analysis of","- 11 -\nhourly records from 32 wind observation stations in the LA basin,\nvarying the wind field each hour. We simply used the annual average\nLoo Angeles wind direction frequencies and a aingla moan wind spood,\nchoosen 30 as to agree with the average of Lamb's wind spood data.\nDoubtleoo agreement between the two models could be improved by\nrecomputing ours for each hour, using the actual wind data. 01\ncourse we don't know which model is giving the best values as there.\nare as yet no entirely satisfactory verification data.\nDISCUSSION AND CONCLUSIONS\nThe above comparisons lead us to conclude that our area source\nmodel performs well, producing ground level concentration values\ncomparable with those from other, more complex models. We also\nbelieve that the success of these comparisons amply justifies our\nbasic physical assumption, namely neglect of lateral dispersion\n`Calder20 refers to this as the\"narrow plume hypothesis.\" ]\n[\nIn several respects our model is more general than other steady-\nstate area-source models. It permits both u and K TO vary with Z\nand makes no a priori assumption about the form of the concentration\ndistribution. Since our area source model is quite simple to apply,\nrequiring little computational effort (a few minutes on a desk cal-\nculator or several seconds of high-speed digital computer time) it\nshould be of considerable use in air pollution applications.\nOne problem with area source diffusion models that depend on numeri-\ncal integration of a point source diffusion equation (all the models in the\nreferences are of this type) is that it is not obvious how any single","- 12 -\nvariable in the basic formula influences the final result. This has led\nto several analyses of \"sensitivity,\" \" in which parameters are varied in an\nattempt to establish their influence on the ground level concentration.\nSuch studies have been carried out by Hilst21 9 and Milford, et al. 22, and\n23\nanother, by Thayer , is in progress.\nIt is a virtue of the present, explicit solution of the problem that the\nparametric behavior of the result is obtained essentially by inspection. Our\nbasic physical assumption, which appears to be quite reasonable, is that\nground-level area source concentration is essentially independent of the\nlateral dispersion. The behavior of the ground concentration, X AO' with\nrespect to the remaining parameters of equation (4) is summarized in Table II,\nin which the fractional change of X is given by\nAO\n( 11 )\nSXAO / XOO are (Coefficient) 5P/P\nwhere P stands for any parameter on the right hand side of equation (4) .","- 13 -\nTable II\nSXAO/XAO\nCoefficient of SP/P\nRange\nParameter\n1-30m sec-1\n-1\n1 Su/u\nu\nSAX\n(1-b)\n5-50 km\nAx\n(1-b)x1-b ln x\ns\n(1-b)\nX\n(1-b)\n.09 - .29\n1-b\n(for af ed\nS\n(N+1)\n(1-b)\n(1-b)\n5 - 10\nN\n(2N+1)\nCo\n12 0.6 for type D\no\nMany orders\nand 04=00\n(Central\nsource box)\nQj\n+\n.10 to .15 for type D and\nMany orders\n=\n(i'th\n01=00 =\nsource box)\ni=1\n/\n-1\n- SB/B\n0.4 - 1\nB\n1\n8£1/f1\n.01 - 1\nfi\nIn the above, F - (2i + 1)1-b-(21-1)1-b - The result for (1 - b) was\nobtained, assuming for simplicity that Of Co = constant, from the continuous\nform of equation (4) 9\n% =\n( 12)\nIn view of the small variability of a(1-b) over the expected range, from un-\nstable to type-D conditions, this product was assumed constant in evaluating\nthe effect of (1-b) 5","- 14 -\nTable II displays the behavior of this area source model rather completely.\nAs to sensitivity to small changes it. reveals nothing very spectacular, except\nfor the fairly large variation of the coefficient of SP/P which, over the\nrange of the stability parameter (1-b), changes by a factor of about 20 under\nstable conditions. This means that, with this exception, small changes in any\nof the parameters produce only small changes in ground-level concentrations,\nX\nConsequently extremes in X AO must be sought in connection with extreme\n*AO\nvalues of the quantities P. For example, high values of X will be associ-\nAO\nated with large Oo' the central area source strength, or with high values of\nthe source strength Oj for nearby source areas. High X AO also is associated\nwith low u and, where a long-term average is involved, with high values of f\ni\nComparatively large changes in XAO will accompany changes in stability, as\nmeasured by (1-b), particularly during stable conditions and for large values\nof \"fetch\" over the city, X.\nAs to future research on area source models, we believe that improve-\nments are required on three aspects. Models should be extended to account\nfor irregular terrain, chemical reactions and removal effects, and unsteady\nconditions. Some work has been done on each of these but more is required.\nThe success of such a simple approach as we have outlined leads us to hope\nthat adequate means can be developed to introduce these additional features\nwithout sacrificing simplicity. In this connection, it is a definite implica-\ntion of Table II that area-source pollution is not very sensitive to the form\nof the vertical concentration distribution In the extreme and unlikely case\nof a uniform vertical concentration distribution, B - 1, i.e. not very different\nfrom the values we have used. Thus we are not inclined to regard uncertainty\nabout the precise form of this quantity as being much of a problem.","- 15 -\nThe usefulness of any new area source diffusion model depends on its\nperformance compared with other area source models. In the past, as new\nmodels appeared in the literature, there was no comparison with other\nmodels. Clearly, if the concentrations predicted by a complicated\nmodel are not significantly better than the predictions of, for example,\nthe simple \"box\" model of diffusion, then there is no practical justi-\nfication for the new model. In fact ours is the first model we\nknow of that has been compared with other models.\nFinally, we wish to record a plea. Very few published area-\nsource models have included data on source strengths and predicted\nconcentrations in a form that makes it easy or even possible to\nreproduce and compare results. It would be very helpful if authors\nwould; 1) include the area-source strength data that they use; 2)\nprovide calculated arca-source concentration values in the same\ngrid system as for the source data.\nACKOWLEDGMENT\nThis research was performed under an agreement between the\nU.S. . è Atomic Energy Commission and Environmental Science Services\nAdministration.","- 16 -\nREFERENCES\n1. D. H. Lucas, \"The Atmospheric Pollution of Cities,\" \" Int. J. of Air\nPoll., 1, 71-86 (1958).\n2. J. M. Leavitt, \"Meteorological Considerations in Air Quality\nPlanning, J. Air Poll. Control Assoc., 10, 246-250 (1960).\n3. F. Pooler, \"A prediction Model of Mean Urban Pollution for Use\nwith Standard Wind Roses, \" Int. J. of Air and Water Poll., 4,\n199-211 (1961).\n4. J. F. Clarke, \"A Simple Diffusion Model for Calculating Point\nConcentrations from Multiple Sourcesy,\" J. Air Poll. Control\nAssoc., 14, 347-352 (1964).\n5. M. Miller and G. Holzworth, \"An Atmospheric Diffusion Model\nfor Metropolitan Areas, \" J. Air Poll. Control Assoc., 17, 46-50,\n(1967).\n6. D. B. Turner, \"Urban Atmospheric Dispersion Models -- Past, Present,\nand Future, \" USPHS (1969).\n7. D. O. Martin and J. A. Tikvart, \"A General Atmospheric Diffusion\nModel for Estimating the Effects of One or More Sources on Air\nQuality,\" US DHEW, NAPCA mss (1968).\n8. L. J. Shieh, B. Davidson and J. P. Friend, \"A Model of Diffusion\nin Urban Atmospheres; SO, in Greater New York, \" preprint, Symposium\n2\non Multiple Source Urban Diffusion Models, 54 pp (1969).\n9. J. J. Roberts, E. J. Croke and A. S. Kennedy, \"An Urban Atmospheric\nDispersion Model,\" Presented at the Symposium on Multiple Source\nUrban Diffusion Models (Oct. 1969).\n10. H. G. Fortak, \"Munerical Simulation of the Temporal and Spatial\nDistributions of Urban Air Pollution Concentrations, preprint,\nSymposium on Multiple Source Urban Diffusion Models, 16 pp (1969).\n11. A. C. Stern, \"Summary of Symposium; Symposium on Multiple Source\nUrban Diffusion Models, proceedings to be published (1969).\n12. F. Gifford, \"Computation of Pollution from Several Sources,\"\nInt. J. of Air Poll., 2, 109 (1959).\n13. F. Gifford, \"Atmospheric Diffusion in an Urban Area,\" presented at\n2nd IRPA Conference in Brighton, England, May 1970, 5 PP, 2 figs.","- 17 -\n14. R. Lamb, \"An Air Pollution Model of Los Angeles,\" Master's\nthesis, UCLA, viii and 104 pp mimeo. (1968).\n15. D. Slade (Editor), \"Meteorology and Atomic Energy, 1968,\"\nUSAEC, Rep. No. TID-24190, x and 445 PP (2968).\n16. M. E. Smith (Editor), \"Recommended Guide for the Prediction of\nthe Dispersion of Airborne Effluents, \" ASME, ix and 85 pp (1968).\n17. U. S. Department of Health, Education, and Welfare, \"Report for\nConsultation on the Metropolitan Atlanta Intrastate Air Quality\nControl Region (Georgia), ix and 57 pp mimeo (1970).\n18. F. Gifford, \"An Outline of Theories of Diffusion in the Lower\nLayers of the Atmosphere,\" [Ch. 4 in reference 15 7 (1968).\n19.\nG. Briggs, 1969, \"Plume Rise,\" USAEC rep. no. TID-25075,\n81 pp, (1969).\n20. C. Calder, 1969, \"A Narrow Plume Simplification for Multiple\nUrban Source Models,\" unpublished mss, 11 pp mimeo (1969).\n21. G. Hilst, 1969, \"The Sensitivities of Air Quality Predictions to\nInput Errors and Uncertainties,\" paper presented at the NAPCA\nSymposium, Multiple Source Urban Diffusion Models, Oct. 27-30, 1969,\nChapel Hill, N.C.\n22. S. N. Milford, G. C. McCoyd, L. Aronowitz, J. H. Scanlon, and Conrad\nSimon, 1970, \"Air pollution models of the New York, New Jersey,\nConnecticut Air Quality Region,\" paper presented at the 63rd Annual\nMeeting of the APCA, June 1970.\n23. S. D. Thayer, 1970, \"Sensitivity analysis and evaluation of multiple\nsource urban air pollution diffusion models,\" \" Geomet Report No. 7517-1\nprepared under Contract No. CPA 70-94, NAPCA.","CAPTIONS\nScheme for combining rectilinear source-grid squares with\nFigure\n1.\nradial wind directions.\nFigure 2. Calculated wintertime ground-level SO2 concentrations,\nin mg SO2/m3, based on equation (10), for Breman: wind\ndirection, south; wind speed 3 m sec-1; atmospheric stability,\nneutral.\nFigure 3.\nSource strength data for calculation shown in Figure\nmean emission rates in kg SO2/km2 hr for heating period\n(space-heating plus small industries).\nFigure 4. Isopleths of wintertime ground-level SO2 concentration,\nmg SO2/m3, calculated by Fortak.\nFigure 5. Calculated wintertime ground level particulate concentrations,\nmg/m3, based on equation (10), for Atlanta (numbers in squares).\nIsopleths are the calculated values, mg/m3, presented in\nreference 17.\nFigure 6.\nGround level concentrations of natural gas in Los Angeles,\ncc/m3, as calculated using Lamb's model (numbers above the\nline) and equation (10) (numbers below the line). Source\ndata, 103 ft3/day mi2, are indicated at the bottom of each\nsquare.\nFigure 7. Scatter diagram of the information presented in Figure 6.\nOpen points refer to the top three rows of Figure 6.","ORNL-DWG 70-9553\nSOURCE\nGRID\nX\nNW\nNNW\nN\nNNE\nNE\nO\no\n0\no\nWNW\nENE\nc\n0\nRECEPTOR\nPOINT\nWe\nE\n0\no\n0\nWSW\nESE\n0\n0\n0\n0\nSW\nSSW\nS\nSSE\nSE\nFigure 1. is\nScheme for combining rectilinear source-grid\nsquares with radial wind directions.","based on equation (10), 3 for Bremen: wind direction, south; wind speed 3 in sec-1;\nORNL-OWG 70-9554 -\n8483\n166\n4\n2524\n78342\n83324\n6544\nFigure 2. Calculated wintertime ground-level SO 2 concentrations, X AO in mg SO2/m m 3 ,\n866681\n1842\n098622\n11\n75443\n116\n10\n25\n15\n12\n11\n7\n20\n24\n17\n15\n16\n28\n27\n26\n28\n15\n6333\n633\n1514\n241\n46\n10\n24\n37\n40\n24\n12\n14\n468887734\n70\n10\n20\n16\n12\n14\n31\n51\n8\n29\n52\n22\n72\n33\n40\n18\n10\n14\n18\n155465446\n9\n23\n34\n10\n63\n49\n44\n46\n73\n217\n13\nBREMEN 1962\n34\n10\n56\n55\n30\n33\n13\n21\n17\n39\n26\n10\n24\n19\n293\nit\n5\n323\n27\n25\n14\n15\n13\n12\n17\n112\n673225224\n2355\n15\n15\n411\n2\n2\n1\n3\n132\nI\n3\nI\n9\n8\nstability, neutral,\n69\n4\n7\n5\n,\n16\n13\n12\n17\n18\n8\nGRID=1X1.km\n19\n5\n6\n3\n22\n42314\n12\n410\n?\n86\n16\n2\n2\n3","ORNL-DWG 70-9555\nFigure 3. Source strength data for calculation shown in Figure 2, mean emission rates in\n3\n2524\n2\n62212\n4 61\n6964\n7144\n72512\nkg SO2/km2 hr for heating period (space-heating plus small industries).\n137\n619642\n22\n1\n20\n10\n24\n12\n8\n27\n16\n16\n16\n3\n6223\n21\n241\n30\n23\n18\n32\n27\n64\n13\n4\n84\n8\n23\n56\n17\n14\n31\n15\n983\n37\n27\n23\n68\n17\n21\n1\n34\n46\n19\n32\n2\n5\nBREMEN 1962.\n22\n28\n20\n2\n3\n4\n8 2\n3\n20\n10\n19\n2214\n1172\n3\n2\n11\n2\n10\n15\n411\nI\n1\n143\n33\n3\n11\n673223\n1\n5\n5 5 8 8\n10 2\n2\n10\n9\n18\n11\n17\n5\nGRID=1X1 km\n62\n84417 2\n16691\n4110\n31\n1\n1","smo-thing","ORNL-DWG 70-1357\nGWINNETT\nCOUNTY\nMARIETTA\nNORCROSS\nCOBB\nCOUNTY\n30\n30\n26\n21\n15\n9\n20\nC\n40\n40\n56\n36\n2\n38\n40\n60\nDECATUR\n80\nATLANTA\nDOUGLAS\n36\n127\n200\n50\n26\n18\nCOUNTY\nDE KALB\nCOUNTY\n17\n48\n36\n14\n11\n22\n971\n28\n14\n18\n7\n14\n18\n39\n16\n13\n7\nFULTON\nCOUNTY\nHENRY\nCLAYTON\nCOUNTY\nCOUNTY\nFigure 5. Calculated wintertime ground level particulate concentrations\nmg/m3, based on equation 10, for Atlanta (numbers in squares) \"\nIsopleths are the calculated values, mg/m3, presented in\nreference 17.","ORNL-DWO 70 - 9558\n250\n200\n10\n4.0\n5.0\n4.0\n4.0\n3.0\n4.1\n4.2\n2.4\n2.1\n1.5\n300\n300\n100\n100\n50\n5.2\n7.1\n6.2\n5.4\n5.7\n5.2\n4.8\n3.7\n5.4\n5.7\n5.4\n5.7\n5.7\n3.7\n2.8\n1.9\n400\n400\n350\n390\n400\n200\n150\n100\n6.0\n7.0\n8.0\n7.0\n7.0\n7.0\n6.0\n3.0\n2.5\n2.0\n4.5\n4.5\n6.5\n9,8\n7.7\n7.9\n7.7\n6.5\n4.3\n3.8\n4.3\n2.5\n300\n450\n800\n500\n550\n550\n450\n250\n250\n350\n150\n4.0\n7.0\n8.0\n7.0\n6.8\n7.0\n4.5\n3.3\n2.5\n2.0\n1.0\n6.2\n10.9\n14.4\n9,7\n8.2\n8.9\n6.6\n6.0\n5.6\n3.0\n2.8\n500\n900\n1200\n600\n500\n625\n400\n400\n400\n150\n200\n5.0\n5.0\n6.2\n6.2\n6.9\n5.7\n3.5\n3.0\n2.5\n0.5\n1.3\n12.5\n12.0\n40.4\n10.0\n10.0\n9,0\n6.9\n5.7\n6.5\n3.6\n2.8\n1170\n1000\n900\n700\n650\n600\n425\n350\n500\n200\n175\n0.5\n2.7\n4.6\n5.0\n4.5\n4.3\n3.3\n2.7\n2.0\n0,6\n2.7\n6.1\n9.8\n9,4\n9,9\n8.6\n6.9\n5.5\n5.4\n2.1\n150\n450\n800\n650\n650\n575\n425\n350\n400\n50\n2.7\n3.2\n3.0\n3.0\n3.0\n2,6\n1.0\n5.4\n8,4\n7.2\n7.4\n7.8\n6.3\n2.5\n450\n700\n500\n500\n550\n425\n50\n4.8\n1.7\n1.1\n0.9\n2.5\n1.5\n6.3\n5.5\n3.2\n1.9\n5.9\n3.6\nGRID - 4X4 miles\n600\n400\n150\n500\n200\n0.7\n0.9\n0.9\n0.5\nO.C\n0.5\n3.1\n3.3\n3,6\n4.3\n0,9\n07\n300\n250\n275\nLOS ANGELES RESIDENTIAL NATURAL GAS,\nGround level concentrations of natural gas in Los Angeles, cc/m3,\nFigure 6 -\nas calculated using Lamb's model (numbers above the line) and\nequation (10) (numbers below the line) e Source data, 103\nft3/day mi2 are indicated at the bottom of each square.","ORNL-OWG 70-9559\n16\n14\n12\n10\n8\n6\n4\nA TOP 3 ROWS\n2\nO\n0\n2\n4\n6\n8\n10\n12\n14\nCONCENTRATION cm3/m3 - UCLA CALC.\nScatter diagram of the information presented in Figure 6.\nFigure 7. .\nOpen points refer to the top three rows of Figure 6."]}