Literature Cited (1) Daly, N. Enuiron. Sci. Technol. 13, 1979, 1373-76. (2) Post, K.; Bilger, R. W. Atmos. Enuiron. 1978, 12, 1857-65. (3) Post, K. Atmos. Enuiron 1979,13, 783-90. (4) Post, K.; Bilger, R. W. Sydney Oxidant Study, Annual Report for the University of Sydney for 1977-78. Charles Kolling Research Laboratory Technical Note ER-29, Sept 1978. (5) Post, K. “Ozone Concentration Distributions for Sydney”,Proceedings, “Air Pollution into the Eighties-The National Quest for Emission Controls”; clean Air Society of Australia and New Zealand: Canberra, Oct 1979. (6) Bilger, R. W. Enuiron. Sci. Technol. 1978,12, 937-40. (7) Bilger, R. W. “The Hydrocarbon Route to Control of Photo-
SIR: The paper entitled “Use of Frequency Distributions of Potential Ozone in Evaluating Oxidant Controls” describes a method for combining photochemical models and real airshed data to provide frequency distributions of potential ozone. The method is exemplified by analysis of the published data for the Sydney airshed ( I ) . Post accepts the technique but rejects the results of the analysis (2).The rejection is based on the assertions that the technique does not use either a representative precursor distribution or a proven model (2). These assertions are examined below. The data analyzed consisted of -400 sets of 0.5-h averaged concentrations of NMHC and NO, concentrations originally described by Post as “typical of the morning precursor measurements made in Sydney on both trajectory and nontrajectory days” ( 3 ) .Subsequently, Post has published a more complete set of concentrations measured by the SPCC of New South Wales and Sydney University which contains 836 0.5-h averaged concentrations of NMHC and NO, ( 4 ) , and he points out that the values originally analyzed exclude most of the low-concentration data (2).The data originally excluded by Post were those which did not satisfy one of the following conditions: (1)The maximum recorded oxone concentration was greater than 5.0 pphm. (2) The maximum recorded NO, concentration was greater than 5.0 pphm. (3) The maximum recorded NMHC concentration was greater than 0.5 ppm. The bias built into the original set by the method of selection chosen by Post is a bias toward the precursor sets more likely to lead to oxidant problem days. Accordingly the distributions of potential ozone relate to such days and are those formable “for days of high oxidant or precursor concentrations” ( I ) . The identification of selection criteria for concentrations enables the threshold for these oxidant problems to be estimated. On the basis of the modified Dodge model, the limit of 0.05 ppm NO, and the limit of 0.5 ppm NMHC are, within the precision of the models, close to a limit of 0.12 ppm ozone. The discarding of concentrations below either of these precursor limits approximates the discarding of precursor sets with an ozone potential of less than -0.12 ppm of ozone. Post asserts (2) that the potential ozone distributions are not based on a proven model. They are as discussed fully ( I ) , based on the Dodge model both in its original form and in the 2/3 modified form used by Post and Post and Bilger (2,3). The factor */3 used by Post and Bilger ( 3 )is an empirical correction used in the attempt to improve the fit of ozone values calculated from Dodge to 16 sets of NMHC, NO,, and ozone data measured during the Sydney Oxidants Study. The sets of precursor concentrations have NMHC/NO, ratios in the range 5-20 and are presented ( 3 ) as validation of the Dodge isopleths providing that the ozone concentrations calculated from Dodge are multiplied by 0.67. No test of the goodness of fit is reported, and no verification of the model outside the NMHC/NO, range of 5-20 is established. Adoption of these 0013-936X/80/0914-1533$01.00/0
chemical Smog”;Symposium Proceedings of the Proposed Controls on the Evaporation of Solvents and the Storage and Transfer of Volatile Organic Liquids; Clean Air Society of Australia and New Zealand: Sydney, March 1978. (8) Daly, N. S.; Fuller, G. J. “Evaluating Options for Controls”, Proceedings of the Conference on Air Pollution into the Eighties-The National Quest for Emission Controls; Clean Air Society of Australia and New Zealand: Canberra, Oct 1979. Keith Post Department of Mechanical Engineering University of Sydney New South Wales 2006, Australia
: 0
5 4
5 5
5 IO 50 2 00 DAYS PER YEAR FOR WHICH C O N C E N T R A T I O N EXCEEDED STATED VALUE 1
Figure 1. Predicted and observed ozone distributions in the Sydney Basin: ( - -)values measured across the SPCC network, 1975-77; (0) distribution predicted by Post; ( 0 )distribution predictable from potential
ozone (see text).
“modified” isopleths is in fact adoption of the Dodge model, since the isopleths are assumed to be parallel to the Dodge isopleths a t all times, especially outside the NMHC/NO, range of 5-20. The isopleths of the Dodge model were regarded originally as applicable to the evaluation of relative ozone rather than absolute concentrations ( 5 ) . Post applies the modified model to calculating absolute concentrations of ozone (2) by using conditions considerably different from those in the original model. The assumptions in the treatment are set down elsewhere (6). The technique of evaluating distributions of potential ozone by using airshed data treats the problem as one of controlling the ambient concentrations of NMHC and NO, in such a way as to minimize the frequency with which the oxidant has the potential for exceeding the stated goal ( I ) . Thus the ozone potential of each set of precursor concentrations is calculated as if existing airshed conditions were akin to those of the smog chamber. In practice the meteorological conditions favoring oxidant formation occurs only at random intervals so that ont would expect that the frequency with which potential ozone exceeds a stated value to be greater than the frequency with which ozone actually exceeded that value. This basic postulate is not appreciated by Post (21,who maintains that “the predicted ozone distribution is clearly not compatible with the observed ozone distribution and reflects the bias of the incomplete precursor distribution toward higher precursor concentrations.” In fact, it is more likely to reflect that meteorological conditions do not favor oxidant formation on every occasion. When the distribution of potential ozone is subjected to the same operations as those assumed by Post, it leads to a distribution which agrees with the distribution of ozone reported by Ferrari et al. (7). The previous analysis
@ 1980 American Chemical Society
Volume
14, Number 12, December 1980
1533
reported that the sets of precursors for the high-oxidant days have the potential for exceeding 0.20,0.14, and 0.10 ppm wzone for 13.4,42, and 64% of reported sets. If, following Post ( 2 , 4 ) , it is accepted that oxidant formation occurs only during the oxidant season, taken as 0.5 yr, then these calculated percentages of values can be halved. Conversion of these frequencies to days per year, following Post, generates the ozone distribution shown in Figure 1, which shows a distribution of potential ozone comparable with the values observed by Ferrari e t al. (7) and the distribution of Post. In fact, it is possible to make the potential distribution fit the observed by use of a suitable scaling of the frequency scale. Post’s scaling factor of 2 achieves reasonable fit for his distribution and also achieves reasonable fit to convert the potential distribution to a prediction of airshed values. The fit might perhaps be improved by choosing an equally likely factor of, e.g., 2.2. The empiricism of the choice of the numerical factor can be appreciated since it is definable as the ratio (days per year)/(days in oxidant season), which of course may vary from season to season. Such a scaling is not necessary to the technique of comparing controls using distributions of potential ozone but does highlight the difference between an airshed distribution and a potential distribution. However, since the corresponding operations used by Post generate the corresponding fit to the airshed distribution, the criticism that the potential distribution is not compatible with observed data is not valid. Post ( 2 )reports that analysis of the full set of data leads to results which show that 33% reduction in NMHC is as effective as 33% reduction in both NMHC and NO,, which contradicts the result obtained for the analysis reported earlier for the higher concentration of precursors ( I ) . This result not only contradicts the previous analysis but also is in conflict with the fact reported by Post ( 4 ) and confirmed (6) that the observed Sydney ozone values across the whole range of observed concentrations are described by the empirical relationship
( O s ) = k[(NMHC)(N0,)]0.36 which shows ozone reponds equally to both NMHC and NO, concentrations. The implications for further studies are described elsewhere (6).
An analysis of the 836 sets of precursor concentrations has been carried out previously (6) and makes an appreciation of the limits to the present description of the Sydney airshed. The analysis shows that, for all the values of the NMHC/NO, ratio observed in the airshed, combined controls of NMHC and NO, are more effective than claimed by Post (2) and justifies the consideration of NO, controls along with NMHC controls. I t has been stressed ( I ) that the results obtained for the analyses are valid “only within the limits with which the models describe the airshed.” For both the modified Dodge model and the empirical model also proposed by Post ( 4 ) the verifications rest on the ability to reproduce airshed data. The minimum requirement is that the observed ozone distributions for Sydney be described. Spatial and Temporal correlations of episodes should also be established although this has yet to be achieved. No verification of the models exists outside the range of NMHC/NO, ratios of 5-20. Within the range for which verifications exist there is no clear case that control of NMHC only represents the optimum controls for Sydney, although there is a case (6) for additional studies of the system.
Literature Cited (1) Daly, N. J. Enuiron. Sci. Technol. 1979,13, 1373. (2) Post, K. Enuiron. Sci. Technol., preceding paper in this issue. (3) Post, K.; Bilger, R. W. University of Sydney, Sydney, New South
Wales, 1977, Charles Kolling Research Laboratories Technical Note ER-20. (4) Post, K. Atmos. Enuiron. 1979,13, 1979. (5) Dodge, M. C. Research Triangle Park, NC, 1977, US Environmental Protection Agency Report EPA 600-3-77948. (6) Daly, N. J. “A Guide to the Control of Photochemical Smog”; Australian Government Publishing Service: Canberra, ACT; in press. (7) Ferrari, L. N.; Hayes, R. A.; Johnson, D.; Michalk, G. Clean Air 1979, 13, 1.
Neil Daly Department of Chemistry Australian National University P.O. Box 4
Canberra, ACT 2000, Australia
Correction 1979, Volume 13 J o h n M. Ondov,* Richard C. Ragaini, and A r t h u r H. Biermann: Elemental Emissions from a Coal-Fired Power Plant. Comparison of a Venturi Wet Scrubber System with a Cold-Side Electrostatic Precipitator. Page 600. The numbers on the scale of the ordinate in Figto ure 1B should read from lo-’ to lo-’ instead of In Figure 1C the curve for manganese from the ESP unit should read Mn instead of Mn X lo-’. In Figure 1D the curve for thorium should read T h X 5 instead of T h X 0.5. J o h n M. Ondov,* Richard C. Ragaini, and Arthur H. Biermann: Emissions and Particle-Size Distributions of Minor and Trace Elements at Two Western Coal-Fired Power Plants Equipped with Cold-Side Electrostatic Precipitators. Page 949. In Figure l a the curve for samarium from plant
A should read Sm X 10:’ instead of Sm X lo4. In Figure l b the numbers on the scale of the ordinate should range from lo-’ to lo-* instead of to lo-:]. In Figure l e the curve for manganese from plant B should read Mn instead of Mn X 10-1.
1534
Environmental Science & Technology
