Document pBm8ymV37gJREnXqYKMJxp6dD

VC DISPERSIO N MODELING ucc 023983 VC DISPERSION MODELING 4 Ctc<? . Draft Report i/c. &~r V<;u//a 5^-i vcu ~"* MAY 26 1978 *-N. WHEELER, JR DISPERSION MODELING ANALYSES OF VINYL CHLORIDE EMISSIONS FROM THE VINYL CHLORIDE MONOMER/POLYVINYL CHLORIDE PRODUCTION COMPLEX NEAR LONG BEACH, CALIFORNIA FOR THE BF GOODRICH COMPANY AND THE STAUFFER CHEMICAL COMPANY- ucc 023984 Draft Report 4/20/78 TABLE OF CONTENTS SUMMARY 1. INTRODUCTION 2. DETERMINATION OF VC EMISSIONS INVENTORY 2.1 General Approach 2.2 B. F. Goodrich PVC Plant Emissions 2.2.1 Point Sources 2.2.2 Area Source t 2.3 Stauffer Chemical VCM-PVC Plant Emissions 2.3.1 Point Sources 2.3.2 Area Sources 2.4 Relief Valve Discharge Events 2.5 Consolidation of Sources for the CRSTER Code 3. MODELING METHODOLOGY 3.1 General 3.2 Meteorological Data 3.3 Sensitive Receptor Locations 3.4 Annual Average Concentrations 3.5 24-Hour Concentrations 3.6 Sensitivity Tests 4. MODELING RESULTS 4.1 Annual Average Concentrations 4.2- 24-Hour Concentrations REFERENCES (Not included in this submittal.) FIGURES APPENDIX I - Description of Computer Programs A* APPENDIX II - Listings of Emissions Input Data APPENDIX III - Listings of Meteorological Input Data APPENDIX IV - Listings of VC Concentration Output Data. (Not included in this submittal) UCC 023985 Draft Report ; 4/20/78 SUMMARY Computerized dispersion modeling was performed of the vinyl chloride (VC) emissions emanating from the vinyl chloride monomer (VCM) and poly vinyl chloride (PVC) production complex near Long Beach, California. The spatial variation of annual average VC concentrations and the spatial and temporal variation of 24-hour VC concentrations were estimated, both in the general area surrounding the complex as well as at certain sensitive locations such as proximal residences, schools, and hospitals. Stateof-the-art methods were used to develop the VC emissions inventory and to perform the dispersion modeling, in an effort to simulate realistically the ambient VC levels which will obtain after compliance with the EPA emissions standards for VC is achieved at the facility. The comprehensive development of the emissions Inventory resulted in estimated routine emissions substantially lower than previous esti mates developed by the EPA for so-called typical VCM and PVC plants. Total routine (continuous) VC emissions including both point and fugi tive sources for this complex after compliance were estimated to be about 1.08 g/s (8.6 Ib/hr). These continuous emissions were evaluated as to their effects on ambient VC concentrations, and in addition both non-continuous routine releases and non-continuous non-routine releases were simulated. Modeling was performed by means of EPA computer codes recommended for applications of this type (which involved numerous point and area sources of an essentially inert contaminant), primarily the new RAM suite of codes and the AQDM. In addition, CRSTER and PTMAX were used for certain of the analyses. The primary meteorological data input to the dispersion modeling consisted both of hourly meteorological data and the stability/wind direction/wind speed frequency distribution from the Long Beach airport for the years 1960-1964. This station is some 8 km from the site on flat terrain, and the data therefrom are adequate to characterize the dispersion of VC both at the site and in the surrounding area. UCC 023986 Draft Report , S 4/20/78 S~' Sensitivity tests were performed to evaluate the effects on the results of the following: 1) of uncertainties in the specifications of fugitive VC emissions* 2) of the application of urban vs. rural versions of the RAM codes, and 3) of the inability of the modeling codes to treat explicitly the recirculation of VC under sea-breeze/drainage flow conditions. The first two of these tests established a range of apnual and 24-hour values within the bounds of which the actual concentrations probably will lie. The recirculation effects were shown to be insignif icant as regards annual average and 24-hour average concentrations. The modeling results indicate that the maximum annual average VC concentrations at the closest sensitive locations at 0.7 km to 1 km from the plant will be on the order of 2-3 ppb. At distances greater than 2 km from the plant, annual average concentrations will be less than 1 ppb, and at distances greater than 4 km, annual average concentrations will be less than 0.2 ppb. The modeling results indicate that the maximum expected 24-hour concentrations resulting from routine, continuous VC emissions at the C closest sensitive locations (0.7 km - 1 km) will be on the order of 20- 27 ppb, with the number of 24-hour values per annum greater than 10 ppb at these locations in the range of 18 to 32. If the routine, noncontinuous release at the VCM plant associated with the oxychlorination vent scrubbing once per annum for a full week were to coincide with the worst case 24-hour meteorological conditions for the routine releases, the maximum expected 24-hour concentration at a sensitive location would be increased to about 28 ppb, with the remainder of the 24-hour frequency distribution essentially unchanged. The maximum 24-hour concentration resulting from the worst of the five expected relief value discharges per annum from the complex could be on the order of 500 ppb at a specific point off-site. ucc 023987 Draft Report 4/20/78 1. INTRODUCTION The objective of this study was to provide estimates of ambient vinyl chloride (VC) concentrations in the area surrounding the vinyl chloride monomer (VCM) and polyvinyl chloride (PVC) production complex near Long Beach, California. The complex consists of a PVC plant and a VCM plant owned and operated by the Stauffer Chemical Company, and a PVC plant owned and operated by the BF Goodrich Company (see Figure 1). This complex has been in operation for nearly twenty years, and it has recently become subject to emissions standards for VC promulgated by the U. S. Environmental Protection Agency (EPA). These standards were promulgated on October 21, 1976, and they require compliance thereto by all existing and new plants which emit VC by October 21, 1978. The California Air Resources Board (ARB) is currently considering a statutory limitation on ambient concentrationsof VC. The modeling analyses presented herein were designed to provide realistic estimates of the ambient VC concentrations in the Long Beach area, so as to assist the ARB in the matters now under consideration in this regard. The computerized dispersion modeling was performed on the basis of highly specific source parameters prepared jointly by BF Goodrich, Stauffer and Dames & Moore. Because of the critical importance of an accurate emissions inventory to the realism of the dispersion modeling, an unusually refined quantification of emissions was performed, in accordance with the methods described in Section 2. The dispersion modeling methodology was arrived at after in-depth discussions with scientists from the Modeling Section of the ARB. The specifics of the approach are described in Section 3, and descriptions of each EPA dispersion modeling code used in the analysis are set forth in Appendix I. Although the general approach was to simulate the releases by means of various standard EPA modeling codes in accordance with Long Beach Airport meteorological data, where uncertainties were identified and/or professional judgement was applied, sensitivity analyses also were performed to clarify the range of possible outcomes subject to these uncertainties. UCC 023988 Draft Report 4/20/78 The modeling results are presented in Section 4, in terms of the annual average VC concentration field, and the frequency distributions of 24-hour VC concentrations at identified sensitive locations around the complex, such as the most proximal residences, schools, and hospitals. Tables are integrated within the textual material, and all figures are presented following the text. In addition to the Appendix describing the computer codes, subsequent appendices set forth listings of certain emissions input data, meteorological input data, and VC concentration output data. UCC 023989 Draft Report 4/20/78 s 2. EMISSIONS INVENTORY 2.1 GENERAL APPROACH At the start of this study it was tentatively presumed that the vinyl chloride (VC) emissions from the BF Goodrich-Stauffer plant complex would be roughly comparable to those estimated by the U. S. Environmental Protection Agency (ERA) for typical vinyl chloride monomer (VCM) and polyvinyl chloride (PVC) plants operating under the regulations prom ulgated in October, 1976. These sets of VC emissions are referred to subsequently in this report as those from "EPA-typical" plants. Disper sion modeling of the appropriately sized EPA-typical plants in the location and under the meteorological conditions of the Long Beach, California site was thus expected to provide a reference point to which the BF Goodrich-Stauffer plant complex could be compared. For the purposes of this analysis, emissions for the Long Beach plants were calculated for operations after October 21, 1978, after which compliance with the EPA emissions standards is required. The VCM emissions sources In such plants may be conveniently divided into two classes: (1) "point sources", whose locations are well-defined, and for which estimates of emissions can be made by measurements of VC concentrations in the product stream at the location, and (2) "fugitive sources", whose locations are poorly defined, and for which accurate estimates of the emission rates are more difficult to make. An example of a fugitive source would be the sum of the small leaks from a large number of valves and flanges in certain areas of a plant. Upon receipt by Dames & Moore of Initial estimates of VC emission rates from the process engineers of the BF Goodrich and Stauffer plants, it was noted that the emission rates for the point sources were well below the values specified for the EPA-typical plants. Both the Stauffer and the BF Goodrich PVC suspension plants are roughly the same size as the 68 million kg/year EPA-typical plant. The Stauffer plant point source emission rates were estimated as roughly one-half those given for the EPA-typical plant, while the BF Goodrich point source emission rates UCC 023990 Draft Report 4/20/78 f were estimated to be less than five percent of the EpA-typical values. (The difference between the two plants is largely determined by the different products manufactured. The polymer-water slurry for the BF Goodrich product is much more easily "stripped" of its VCM component than is the case for the slurry developed at the Stauffer plant.) For the Stauffer VCM plant, which has 23 percent of the EPA-typical plant production capacity, the only routine point source release occurs during scrubbing and venting of the oxychlorination vent, a condition that occurs routinely but infrequently (a maximum of one week per year). In contrast to the case for the point source inventory, both BF Goodrich and Stauffer Chemical presented estimates of fugitive emissions which agreed with the EPA-typical plant values. For BF Goodrich the EPA-typical values had simply been assumed, while for Stauffer the esti mate had been calculated using an actual valve count at the plant and a published value for an average leak rate for valves in petroleum refineries. Under this initial set of assumptions concerning the fugitive emission rates, more than 95 percent of the total VC emissions from the plant complex were assigned to the very inaccurately known fugitive emission sources. Thus it became apparent that any attempt to realis tically model the effects of the VC emissions from the plant complex would require a much more precise determination of the fugitive emission rates. Accordingly, a comprehensive effort was performed in an attempt to describe these fugitive emissions as accurately as possible. The details of the development of the emissions inventory as finally used In the dispersion modeling are given in the remaining sub-sections of Section 2. The full VC emissions inventory is tabulated in Table 2-1, and the relative locations of the point sources and fugitive area sources are indicated in Figure 2. ucc 023991 Draft Report 4/20/78 2.2 BF GOODRICH PVC PLANT EMISSIONS 2.2.1 Point Sources A description and characterization of the VC emissions from BF Goodrich point sources (source numbers 1 through 12 in Table 2-1) was provided by BF Goodrich. The emissions are determined, or estimated, to be those that will result from plant operation after implementation of all controls required by the federal NESHAPS standard for vinyl chloride. These controls are to become fully implemented by October 21, 1978. Lwith the_exception of the cooling tower source7^11 of the point sources indicated are due to regulated releases from the process stream of PVC slurry, storage of the product PVC, or scheduled testing of the quality of the PVC slurry. The values of the emission rates from these sources are thus determined with a considerable degree of confidence. The emission rate from the cooling tower is an estimate of possible contamination of plant cooling water by leaks from the product stream in heat exchangers--in this case, primarily the cooling jackets of the PVC reactor vessels. Measurements of vinyl chloride concentrations in the cooling water at the inlet to the cooling tower and in the cooling tower sump were supplied by BF Goodrich. The initial data indicated a concen tration of 0.07 ppm in water going to the tower, and 0.03 ppm in the tower sump. The difference of 0.04 ppm was presumed to be released to the atmosphere in the cooling tower, leading to the emission rate given for this source in Table 2-1. A similar loss rate could be inferred from data supplied by Stauffer Chemical, which indicated that 5 percent of samples of cooling water showed vinyl chloride concentrations of 1 2 ppm, with the remainder at zero measured concentration. This impli an average concentration of approximately 0.05 ppm. If this entire concentration is presumed lost from the cooling tower, then one is le to substantially the same emission rate as was calculated from the initial BF Goodrich data. BF Goodrich personnel later made additional measurements of coolmy water vinyl chloride concentrations. These measurements indicated a uniform concentration of 0.04 ppm VCM in water both approaching and UCC 023992 Draft Report 4/20/78 leaving the cooling tower; i.e., apparently zero loss of vinyl chloride from the cooling tower. Nonetheless, it was decided to retain the emission rate as originally calculated in the modeling of dispersion of VC gas from the plant complex. This decision was made in order to take a conservative approach to a source whose emission rate could not be accurately determined. 2.2.2 Area Sources Fugitive emissions from four classes of sources were considered in the development of an area VC emission rate for this plant. These source classes, and the contribution of each to the total area emission rate, are indicated in Table 2-2. The area being considered is shown as source A4 in Table 2-1 and Figure 2. The first three of the contributions in Table 2-2 were estimated by BF Goodrich. The materials handling and transfer source is due to transfer of VCM from tank cars into the plant process stream. This is a regulated source, and thus the emission rate is controlled and determined. The "miscellaneous emissions in the neighborhood of the poly building" figure is a best estimate by plant personnel of possible losses from routine plant operation in the area of the PVC reactors. The emission rate given for "miscellaneous equipment opening and repair" is based on a figure of an average loss of 5 lbs VC per maintenance opening. It was estimated that such openings would occur on an average of once a month. This frequency was arbitrarily increased to once a day in the calculation of the emission rate given in Table 2-2, simply to provide a con servative estimate. No emissions were assumed to exist from pump or compressor shaft seals, because only double-sealed shafts would be in operation in the plant after October, 1978. The major fugitive source is that due to leakage from valves and flanges in VCM or PVC slurry service. An estimate of this emission rate required a very detailed inventory of all valves and flanges in such service within the plant. The data and assumptions used in developing the emission rate used are described below. UCC 023993 ucc 023994 SOURCE NAME BFG Tall Stack BFG Centrifuge BFG Blend Tanks BFG Dryer BFG Dryer BFG Silos BFG Silos BFG Silos BFG Silos BFG Silos BFG Cooling Tower BFG Lab Vent SC Qxychlorination SC PVC No. 1 SC PVC No. 2 SC PVC No. 3 SC Blend Tanks SC Blend Tanks SC Dryer Sg SC Dryer TABLE 2-1 VC EMISSIONS INVENTORY FOR BF GOODRICH/STAUFFER CHEMICAL VCM-PVC PLANT COMPLEX * 3U * Z?/ SOURCE NUMBER 1 2 3 4 5 6 7. 8 9 10 11 12 13 14 15 16 17 18 19 20 UTM COORDINATES EAST NORTH 385.453 3743.000 385.469 3743.010 385.494 3742.995 385.508 3742.987 385.509 3742.960 385.513 3743.009 385.545 3743.007 385.559 3742.949 385.508 3742.980 385.495 3742.891 385.491 3742.960 385.460 3742.938 385.668 3743.082 385.661 3743.126 385.668 3743.103 385.714 3743.102 384.633 3743.105 385.747 3743.069 385.617 3743.103 385.617 3743.120 EMISSION RATE . (9/s) 0.0072 0.0053 0.0045 0.0063 0.0096 0.0006 0.0003 0.0006 0.0003 0.0006 0.0162 0.0027 Q.3780b 0.0009 0.0014 0.0040 0.1184 0.0202 0.0743 0.0378 STACK HEIGHT Cm) 24.4 6.0 18.3 15.2 9.1 15.2 15.2 18.3 12.2 24.4 * 5.5 5.5 33.0 11.0 11.0 11.0 9.0 9.0 . 3.0 7.0 EXIT TEMPERATURE (K) Ambient a 310.9 322.0 360.9 360.9 Ambient Ambient Ambient Ambient Ambient Ambient Ambient 310:9 327.6 327.6 327.6 355.4 355.4 338.7 . 338.7 . . Page 1 of 2 EXIT VELOCITY (m/s) 14.33 0.00 15.33 33.09 21.16 1.07 1.07 1.07 1.07 1.07 7.93 1.10 43.69 5.83 5.83 10.36 .8.74 8.74 6.78 11.44 STACK DIAMETER (m) 0.76 4.69 0.62 0.68 0.95 0.89 0.63 0.89 0.63 0.89 10.20 0.60 0.20 0.25 0.25 0.48 0.22 0.13 0.84 1.54 - 4/20/78 D raft Report TABLE 2-1. (Continued) SOURCE NAME SC Dryer SC Centrifuge SC Centrifuge SC Silos SC Silos SC Silos SC Cooling Tower 1 SC Cooling Tower 2 SC Cooling Tower 3 SC Lab Vent SC Area SC Area SC Area BFG Area SOURCE NUMBER 21 22 23 24 25 26 27 28 29 30 A1 A2 A3 A4 -UTM COORDINATES EAST NORTH 385.747 3743.063 385.641 3743.112 385.747 3743.063 385.593 3743.138 385.600 3743.086 385.724 3743.063 385.540 3743.072 385.704 3743.076 385.695 3743.050 385.681 3743.148 385.660 3743.099c 385.435 3743.024 385.585 3743.024 385.435 3742.949 EMISSION RATE (g/s) 0.0416 0.0592 o.oior 0.0140 0.0060 0.0070 0.0207 0.0080 0.0139 0.0053 0.2163 0.1869 0.0794 0-1003 STACK HEIGHT Cm) 15.0 10.0 10.0 22.0 22.0 22.0 14.9 13.7 15.2 5.5 6.0 6.0 6.0 * 6.0 EXIT TEMPERATURE (K) 338.7 338.7 338.7 Ambient Ambient Ambient Ambient Ambient Ambient Ambient Page 2 of 2 EXIT VELOCITY (m/s) STACK DIAMETER (m) 24.27 0.20 0.15 0.85 0.85 0.85 7.93 7.93 7.93 1.10 0.61 0.25 0.20 0.74 0.49 0.53 12.80 6.40 9.10 0.60 D ra ft Report a The dispersion model computer codes require these to be not less than the highest ambient temperature in the meteorological data. Plume temperatures of 310.0 K were assigned for runs using one year of meteoro logical data. b This emission rate occurs for a maximum of one week per year. An emission rate of 0.0001 g/s was assigned for routine cases, and the higher rate was used only in runs on days of interest defined by the highest pre dicted concentrations. c Location of southwest corner of each square area, 75 m on each side. oo 4/20/78 023995 Draft Report 4/20/78 Table 2-2. BF Goodrich Fugitive Source Contributions Type Of Source Contribution To Emission Rate (g/s) Materials Transfer and Handling 0.0002 Miscellaneous Emissions ' . Qn~the Neighborhoodzof the Poly Building 0.0026 Miscellaneous Equipment Opening and Repair 0.0263 Leakage From Valves and Flanges 0.0712 TOTAL 0.1003 O O ucc 023996 Draft Report 4/20/78 A total of 895 valves and 3085 flanges were included in the initial inventory supplied by BF Goodrich and used in this calculation. These valves and flanges were classified by height above ground (four height classes: ground level, 15 feet, 30 feet, and 45 feet), and by concentra tion of VCM in the product stream (three VCM concentration classes: 100 percent, 30-40 percent, and 5 percent or lower). The valves were further classified by type and size of the valve. An emission rate for each of the four levels could then be estimated by multiplying the valve and flange counts by the appropriate loss rates and VCM concentrations. Loss rates from valves in refineries in Los Angeles County were the object of a study by Palmer, et al (1957). This study indicated an average loss rate for valves in high-volatility service (product vapor pressure greater than that of kerosene) of 0.49 lb/day per valve. The distribution of loss rates in the population studied was very skewed, with 0.2 percent of the valves contributing 4(f percent of the total loss rate, and 1.5 percent of the valves contributing 84 percent of the total loss rate. Eighty-eight percent of the tested valves showed no detectable loss rate, and elimination of the worst 1.5 percent oi the valves reduced the average loss rate to 0.06 lb/day per valve. As was pointed out by the authors of this report, the loss rates appropriate to any particular plant are extremely sensitive to the quality of the valve maintenance program at the plant in question. Another factor which has an important effect is the type of valve being considered, valves with rising stems being more leak-prone than those of the plug or ball type construction. A senior professional from Dames & Moore who has extensive experience with process streams of the type used in these plants visited the plant complex in order to assess the condition of the valves and to arrive at realistic estimates of valve loss rates based upon professional knowledge and judgment. As.a result of this professional judgment, the following loss rate formula was used in calculating the VC leakage from valves in the BF Good rich PVC plant: For all rising-stem valves a loss rate of 0.02 lb/day * ucc 023997 Draft Report 4/20/78 per valve was assumed, and for all non-rising-stem valves a loss rate of 0.01 lb/day per valve was assumed. Rising-stem type valves constitute 12.5 percent of the valves in the BF Goodrich plant. Following Palmer, et al (1957), the loss rate for flanges was taken to be one-tenth the plant-wide average valve loss rate. This yielded a loss rate of 0.0011 lb/day per flange. The computer dispersion codes used in this study can only accept a single height for a given area source. For this reason the height resolution of the valve and flange emission rate data could not be completely utilized in the modeling procedure. The mean height of the 100 percent VCM equivalent valves for this plant was slightly more than 4 m. As this height is not greatly different from the 6 m area source height used routinely by the EPA, the effective stack height for this area source was assumed to be 6.0 m. UCC 023998 Draft Report 4/20/78 2.3 STAUFFER CHEMICAL VCM-PVC PLANT EMISSIONS 2.3.1 Point Sources A description of the VC emission characteristics from Stauffer Chemical point sources (source numbers 13 through 30 in Table 2-1) was provided by Stauffer Chemical. These emission rates are determined, or estimated, to be those in effect after October, 1978. With the exception of the cooling tower sources, all of the point sources indicated are due to regulated releases from the process stream of VCM or PVC slurry, product storage, or scheduled testing of VCM and PVC slurry quality. The values of the emission rates from these sources are thus determined with a considerable degree of confidence. Source number 13, the oxychlorination vent, is normally vented to the incinerator, and thus has an emission rate of zero. However, for regularly scheduled, but infrequent, periods 'during which this vent is being scrubbed (occurring a maximum of one week per year), this source emits VC gas at the rate indicated in Table 2-1. This source was handled by assigning it a pro-rated emission rate (0.0001 g/s) for routine calculations of concentrations in the neighborhood of the plant. On the days which were subsequently identified as having the highest concentra tions, the calculations were reperformed using the larger value of the emission rate for this source. The emission rates from the cooling towers (source numbers 27, 28, and 29) are an estimate of possible contamination of the plant cooling water by leaks from the product stream in heat exchangers--in this case, primarily the cooling jackets of the PVC reactor vessels. Data supplied by Stauffer Chemical from routine measurements of the cooling water VCM content Indicated that one to two ppm concentrations were present in the water in approximately 5 percent of the measurements, with undetectably small concentrations in the remaining 95 percent of the measurements. From this it was assumed that an average concentration of 0.05 ppm was lost from water passing through the cooling towers, for the purposes of the emissions inventory. This estimate is consistent with the value DOC 023999 Draft Report 4/20/78 determined independently for initial data on cooling water VCM concen trations in the BF Goodrich plant, as described in Section 2.2.1. As indicated in that section, the value is quite likely overestimated, but is a reasonably conservative approach to a source for which the emission rate cannot be accurately determined. 2.3.2Area Sources Three areas, identified as Al, A2, and A3 in Table 2-1 and Fig ure 2-2, were needed to model the fugitive contributions to the Stauffer Chemical plant's VC emissions. Area Al includes the PVC reactors. Area A2 includes the VCM units 1 and 2, and Area A3 includes VCM unit 3. Five categories of VC sources were considered for each of these areas, as listed in Table 2-3. The emission rates given for the first four categories of sources in Table 2-3, i.e., pump seal losses, sampling-losses, losses from mis cellaneous equipment openings, and losses from sewer vents, were explic itly supplied by Stauffer Chemical. The major fugitive source is that due to leakage from valves and flanges in VCM or PVC slurry service. An estimate of this emission rate required a very detailed inventory of all valves and flanges in such service within the plant. The data and assumptions used in developing the emissions rates used are described below. A total of 2482 valves and 984 flanges (not associated with valves) were included in the inventory supplied by Stauffer Chemical and used in this calculation. This inventory was categorized by valve type (risingstem or non-rising-stem), by height class (6 classes ranging from zero to 60 feet), by concentration of VCM (9 classes ranging from 60 ppm to 100 percent), and by area within the plant (17 localized areas). This very detailed inventory provides a basis for a fugitives emissions inventory of high resolution, if taken in conjunction with the appro priate loss rates for the various valve types and conditions. Loss rates from valves in refineries in Los Angeles County were the - object of a study by Palmer, et al (1957). This study indicated an average loss rate for valves in high-volatility service (product vapor UCC 024000 Draft Report 4/20/78 Table 2-3. Stauffer Chemical Fugitive Source Contributions (g/s) Type of Source Pump Seal Losses Sampling Miscellaneous Equipment Opening Sewer Vents Valve & Flange Losses Area A1 0.0004. -0- 0.0001 -0- 0.2158 AREA TOTALS 0.2163 Area A2 0.0002 0.0005 0.0001 ' 0.0227 0.1634 0*1869 Area A3 0.0001 0.0005 0.0001 0.0240 0.0547 0.0794 (?. \ \ ucc 024001 Draft Report 4/20/78 pressure greater than that of kerosene) of 0.49 Ib/day per valve. The distribution of loss rates in the population studied was very skewed, with 0.2 percent of the valves contributing 40 percent of the total loss rate, and 1.5 percent of the valves contributing 84 percent of the total loss rate. Eighty-eight percent of the tested valves showed no detectable loss rate; and elimination of the worst 1.5 percent of the valves .reduced the average loss rate to 0.06 Ib/day per valve. As was pointed out by the authors of this report, the loss rates appropriate to any particular plant are extremely sensitive to the quality of the valve maintenance program at the plant in question. Another factor which has an important effect is the type of valve being considered, valves with rising stems being more leak-prone than those of the plug or ball type construction. A senior professional from Dames & Moore who has extensive experience with process streams of the type used in these plants visited the plant complex in order to assess the condition of the valves and to arrive at realistic estimates of valve loss rates based upon professional knowledge and judgment. As a result of this professional judgment, the following loss rate formula was used in calculating the VC leakage from valves in the Stauffer Chemical VCM and PVC plant: 1. For all non-rising stem valves a loss rate of 0.01 lb/day per valve was assumed. 2. For rising-stem valves in plant local areas 11 and 12 (in cluded in area source A2), 50 percent were assigned a loss rate of 0.49 lb/day per valve, and the remaining 50 percent were assigned a loss rate of 0.02 lb/day per valve. 3. For rising-stem valves in the remaining areas of the plant, 35 percent were assigned a loss rate of 0.49 lb/day per valve, and the remaining 65 percent were assigned a loss rate of 0.02 lb/day per valve. Following palmer, et al (1957), the loss rate for flanges was taken to be one-tenth the plant-wide average loss rate. This yielded a loss rate of 0.0048 lb/day per flange. ucc 024002 Draft Report 4/20/78 The RAM suite of dispersion model computer codes require that area sources be handled in a very uniform manner, in terms of area source sizes and placement. For this reason the original areas indicated by the Stauffer Chemical valve and flange inventory, which were of varying sizes, were consolidated into the uniformly-sized area sources Al, A2, and A3. Also, the computer codes allow only a single height to be stipulated for a given area source. In the Stauffer Chemical plant, the mean height of 100 percent VCM-equivalent valves was found to be 4.1 m. This is close to the EPA Area Source Standard height of 6 m. An effective stack height of 6.0 m was therefore assigned to each of the area sources. ucc 024003 Draft Report 4/20/78 2.4 RELIEF VALVE DISCHARGE EVENTS Non-routine releases of VC may occur at the plant complex due to process upsets leading to the opening of safety relief valves. These occurrences are infrequent and unpredictable. Information concerning the characteristics of such releases were obtained from BF Goodrich and Stauffer Chemical, and the types of events with an estimated frequency of occurrence of once per year or greater were included in the disper sion modeling. These events are listed in Table 2-4. These events were modeled in two different ways for their contri butions to annual average and 24-hour concentrations in the neighborhood of the plant complex. For the annual average calculation using code AQDM, all three types of events were included, the total emissions from these emissions then averaged over the year's timq. In order to estimate a short-term contribution from these sources, the largest emission amount or the three events was used, with an assumed duration of 5 minutes used to calculate an emission rate, and this source modeled by means of the PTMAX code. UCC 024004 TABLE 2-4 ESTIMATED CHARACTERISTICS OF RELIEF VALVE DISCHARGE EVENTS SOURCE BF Goodrich VCM RELEASE (kg) 46 FREQUENCY OF OCCURRENCE 3 per year DURATION (min) 1 to 5 HEIGHT (m) 12,2 TEMPERATURE (k) 333 VERTICAL VELOCITY (m/s) 26 DIAMETER (m) 0.07 Stauffer, Reactor #1 120 1 per year 1 to 5 12.2 333 26 0.07 Stauffer, Reactor #2 283 1 per year 1 to 5 12.2 333 26. 0.07 D raft Report 4/20/7B 024005 Draft Report 4/20/78 2.5 CONSOLIDATION OF SOURCES FOR THE CRSTER CODE The dispersion model code CRSTER was used to compare the concentra tions produced under 1964 meteorological conditions with those for the years 1960 through 1963. The CRSTER code admits only co-located point sources to a maximum number of nineteen. Thus some consolidation of the thirty point sources and four area.sources of this plant complex was necessary before input to the CRSTER code. Sources with similar exit characteristics (height, temperature, exit velocity and diameter) were collected into consolidated point sources. In this manner an emissions inventory consisting of fourteen point sources was produced for use only in the CRSTER code. V =1 ucc 024006 Draft Report 4/20/78 3. MODELING METHODOLOGY 3.1 GENERAL In order to predict realistically the ambient VC levels in the area surrounding the two plants, two basic data sets are required as input for the most appropriate dispersion modeling codes. The two required input data sets are the emissions inventory and the meteorological' data, with each resolved to the requirements of the concentration averaging period of interest. The development of the detailed VC emissions in ventory is described in the preceding Section 2. Published studies have indicated that the suspected effects on health of VC are associated with long term exposures thereto; there fore, the expected annual average concentration patterns were regarded to of interest in this study. Inasmuch as the ARB is considering a limitation on 24-hour average concentrations, the patterns of variations of 24-hour average values also were addressed herein. The following sub-sections describe the meteorological data used as input to the modeling codes, the selection of the receptor grids, the selection and the manner of application of appropriate modeling codes, and the sensitivity tests which were performed. UCC 024007 Draft Report 4/20/78 3.2 METEOROLOGICAL DATA The Long Beach airport is located some 8 km due east of the plant site, as is shown in Figure 1. Both long term summarized data and hourly data are available for this station in the format required by standard Gaussian dispersion modeling codes. The plant site is extremely flat, as is the area lying between the plant site and the airport: Therefore, meteorological measurements taken at the airport are suffi ciently representative of conditions at the plant site, even though there is a hill small in areal extent (Signal Hill) lying some 2 km to the south-southwest of the airport. The maximum relief of this hill is about 300 feet, which could cause some channeling of flow along a westnorthwest/east-southeast axis. The wind rose for this location for the period 1960-1964 is displayed as Figure 4. Although this wind rose indicates that the most prevalent direction is in fact west-northwest, this direction (as,well as west) is associated primarily with sea-breeze flow conditions. In order to perform calculations of annual average concentrations, the stability/wind speed/wind direction frequency distribution for the period was obtained for input. In order to calculate the frequency distribution of 24-hour concentrations at sensitive receptor locations, hourly data for the same period of 1960-1964 were also obtained. The processing of this latter data set is described in sub-section 3.5 below. ucc 024008 ucc 024009 Table 3-1 (Continued) __________ RECEPTOR Catskill Ave. School Broad Ave. Elem. Banning H.S. Delores St. School St. Philomena School Carson St. School School, Unknown Los Cerritos School SOURCE NUMBER 21 22 23 24 25 26 27 28 TYPE 11 II n ii n ij ii ' ti Page 2 of 2 UTM COORDINATES EAST NORTH 382.40 383.10 383.10 382.26 381.82 381.92 384.23 388.91 3741.72 3740.47 3739.60 3742.88 3743.72 3744.06 3746.06 3743.55 DISTANCE FROM CENTER OF COMPLEX (km) 3.44 3.54 4.19 3.35 3.86 3.85 3.39 3.35 ` ' 4/20/78 D raft Report Draft Report 4/20/78 3.3 SENSITIVE RECEPTOR LOCATIONS Inasmuch as the plant complex is located within an area of highly diverse land use, it was of interest not only to calculate the general patterns of ambient VC concentrations but also concentrations at speci fic locations in the immediate area. Three categories of "sensitive" locations were identified, on the basis either of potential long term exposure or of possible vulnerability. In the former class were placed residences, and in the latter schools and hospitals. A detailed area map was surveyed to identify in. each sector from the plant the locations of the closest residences, schools and hospitals. On this basis, some.28 sensitive locations were specified, within about A .km of the plant. Figure 5 displays the locations of these points relative to the plant site, on a standard UTM coordinate system on which one unit represents 1 km. Table 3-1 provides a brief description of each numbered location, of which the first eight are the closest resi dences, the following two are the closest hospitals, and the last eighteen are the closest schools. Subsequent dispersion modeling runs used these locations in addition to regular cartesian grids. ucc 0240^0 Draft Report 4/20/78 3.4 ANNUAL AVERAGE CONCENTRATIONS The ERA computer code used to obtain the annual average VC con centration fields was the Air Quality Display Model (AQDM). c0(je has been probably the most frequently applied modeling program in air quality evaluations since its development in 1969, and it is specifically recommended for use by the ERA in the "Interim Guideline on Air Quality Models.(EPA, 1977). tt is specifically designed to simulate the disper sion of a large number of point and area sources with respect to annual average concentrations over a regular 15 x 15 square receptor grid, with ad hoc receptors specifiable at off-grid locations. Appendix I presents a further description of AQDM. ; The meteorological input to AQDM was the 1960-1964 STAR Program listing for the Long Beach airport. Two regular 15 x 15 grids were applied, one being a 1 km grid and the second,a 2 km grid displaced so as to not duplicate receptor locations of the 1 km grid run. Additionally, the identified sensitive receptor points were used as ad hoc receptor locations. The sensitivity of the annual average concentrations to uncertainties in the VC fugitive emission inventory was tested by performing runs on both grids on the assumption that the actual mass rate of fugitive emissions could be 50 percent higher or lower than the final calculated values. Therefore, the sensitivity runs used as input total fugitive emission rates increased by 50 percent. UCC 024011 TABLE 3-1 SENSITIVE RECEPTOR LOCATIONS IN NEIGHBORHOOD OF VCM/PVC COMPLEX NEAR LONG BEACH (Plant Center at E:385.61, N:3742.96) Page 1 of 2 D raft Report RECEPTOR Wilmington Ave. & S.D. Fwy. Arlington St. 223rd St. Alameda St. 223rd & Lucerne Spring St. Wilmington & Sepulveda Avalon Village Pacific Hospitl Seaside Man. Hosp. Del Amo Elem. Daniel Webster Elem. Stephens Jr. H.S. Andrew Carnegie Jr. H.S. Bonita St. Dominguel Elem. Muir School Eliz. Hudson Elem. St. Lucy School Wilmington Park Elem. SOURCE NUMBER 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 TYPE Residence II It It II n M n Hospital It t School II 11 11 11 tl II II II II UTM COORDINATES EAST NORTH 385.03 3743.39 386.62 3742.88 384.52 3743.19 386.43 3743.72 384.16 3743.19 386.91 3741.77 383.75 3741.36 383.22 3742.21 389.53 3741.24 389.17 3738.74 385.22 3744.47 387:34 3742.37 387.22 3741.43 383.53 3743.75 383.32 3743.65 387.49 3744.47 387.92 3741.87 387.00 3740.40 387.36 3740.40 384.55 3738.90 DISTANCE FROM CENTER OF COMPLEX (kml 0.72 1.01 1.11 1.12 1.47 1.76 2.45 2.51 4.28 5.53 1.56 1.83 2.22 2.22 2.40 2.41 2.55 2.92 3.11 4.19 4/20/78 ucc 024012 Draft Report 4/20/78 3.5 24-HOUR CONCENTRATIONS The primary code selected to calculate the 24-hour average VC concentrations was the Real-Time Air Quality Simulation Model (RAM). This suite of codes was recently developed by the EPA, and it is speci fically designed to address the effects on short-term concentrations of multiple point and area sources. RAM actually includes four basicmodeling codes, i.e., RAM, RAMR, RAMF and RAMFR. RAM and RAMR use as input 24 hours or less of hourly meteorological data, whereas RAMF and RAMFR are programed to run on a full year of hourly data. RAM and RAMF simulate dispersion in accordance with the urban dispersion coefficients of McElroy and Pooler (1968), and RAMR and RAMFR use the rural dispersion coefficients referred to as the PasquillGifford curves (Turner, 1969). Appendix I presents additional informa tion on this set of codes, which is also one of those specifically recommended by the EPA in the guideline document (EPA, 1977). The versions of the codes used herein were the "official" versions recently released by the EPA. Although no User's Manual is available yet. Dames & Moore is very familiar with these codes, having previously assisted the EPA in their debugging stages. Inasmuch as an objective was to calculate a representative frequency distribution of 24-hour VC concentrations, the use either of RAMF or RAMFR was dictated, i.e., the codes which run on long period hourly data. Initially it was presumed that RAMF would be more appropriate to this application than RAMFR, because the McElroy-Pooler urban curves were developed on the basis of low level tracer releases in St. Louis. Particularly owing to the importance of the fugitive VC emissions from this complex, which are non-buoyant and are released at low levels, it was anticipated that the urban version (RAMF) would be more appro priate. Both versions were tested to ascertain the sensitivity of results to this selection, and RAMFR was ultimately used for reasons discussed in sub-section 3.6 below. ucc 024013 Draft Report 4/20/78 Five years of hourly meteorological data for 1960-1964 were available from the Long Beach airport. However, the resolution of the 1964 data was better than that for 1960-1963, due to the universal change in the observing practice of the Weather Bureau. Prior to 1964, wind direction was reported by meteorological sector, i.e., resolved to the 22.5 of the sixteen such sectors. Beginning in 1964, wind direction was reported to the nearest 10, sector, i.e., resolved within the thirty-six sodefined sectors. Furthermore, for the years after 1964, the hourly data resolved to 10 sectors are not routinely available in machine-readable form, because only the three-hourly data are systematically placed on tape. Therefore, in general, the 1964 data are the only data generally available in the required format and degree of resolution suitable for the RAM codes, as was the case of the Long Beach data. This necessitated reliance upon 1964 as the hourly data set used as input to the RAM codes in this study. However, sensitivity tests were performed to compare the representativeness and degree of conservatism of the 1964 data, by comparison with the 1960-63 data, as discussed in sub-section 3.6 below. An analysis of the annual wind rose (Figure 4) indicates that calm winds are reported 15.5 percent of the time at Long Beach. This is a relatively high percentage, particularly because such a hourly observa tion lacking a specified wind speed and wind direction is not directly usable for performing dispersion calculations. The EPA preprocessing program treats the reported calms as in the following way. The wind speed is arbitrarily assigned as 1 m/s, and the assigned wind direction is that of the last reported wind direction, (which then is randomized within the assigned 10 sector). With a 15.5 percent general incidence of reported calms, it can be expected that on some days, many consecutive reported calms would occur. Professional judgment suggests that appli cation of the same wind direction (whthin 5) for long consecutive strings of calms is both unrealistic and excessively conservative. Therefore, for the purposes of this evaluation, only the first calm in a reported consecutive string was assigned to the same sector as the last UCC 024014 Draft Report 4/20/78 \ reported direction, and then randomized within 10. Subsequent calms in a string were randomized within a 40 sector, i,e. 20, to reflect more realistically yet still conservatively the highly variable actual wind direction under these light wind conditions. In most other respects, the data preprocessor was applied without modification to the hourly data, with the exception of the mixing-height treatment. For the low level releases treated in this analysis, mixing k height variation has a negligible effect on ground level concentrations at the short travel distances of interest. Therefore, mean seasonal mixing heights for the area were used as input. -K 1. ucc 024015 Draft Report 4/20/78 3.6^_SENSITIVITY TESTS_ There were several aspects of this study in which data uncertain ties were identifiable and where there were known limitations in the available analytical methods, thereby necessitating the exercise of professional judgment. In the most important of these, sensitivity testing was applied to assess the effects on the modeling results'of such imponderables. As indicated previously in sub-section 3.4 above, the uncertainty associated with the quantification of the fugitive VC emissions was thought to be at most +50 percent. Therefore, AQDM runs were performed with the total, fugitive emission rates increased by 50 percent! In * general, this resulted in increases of the calculated annual average concentrations of 25 percent to 30 percent, not an unexpected result inasmuch as the fugitive VC sources represent about half of the total inventory. If the fugitives were decreased by 50 percent, then the effect on calculated annual average concentrations would be a decrease of 25 percent to 30 percent. As discussed in sub-section 3.5 above, only the 1964 data were available for Long Beach resolved to 10 sectors. This degree of re finement is desirable for input to the RAM codes, and the 1960-1963 hourly data are resolved only to the 22.5 sectors. Two tests were per formed upon the 1964 data, to ascertain whether they were either re presentative of longer term dispersion conditions or a conservative specification thereof. Because the low level fugitive VC sources dominate in relation to maximum short term concentrations at close locations, the worst disper sion cases in general will be under low wind speeds. Figure 4 indi cates that the frequency of occurrence of calm winds for the full period of 1960-1964 was 15.5 percent, and an analysis of the frequency of calms in 1964 showed the value of 18.3 percent pertaining thereto. Therefore, at least from this point of view, 1964 affords a conservative characteri zation of longer term dispersion conditions in this area as regards low level sources. ucc 024016 Draft Report 4/20/78 A second test of the sensitivity of using only 1964 hourly data was made by means of CRSTER, the EPA single source model. This code also runs on hourly data for a year at a time, but all input sources must.be co-located, and area sources cannot be treated explicitly. Appendix I contains a more detailed description of CRSTER. All point sources from the complex were artificially co-located, and area sources were simu lated as point sources at the same location. The 1964 hourly wind directions were randomized within 22.5, so that they would be resolved to the same extent as the 1960-1963 hourly data. Each year in turn was used as input to CRSTER, and the calculated annual average concentra tions, highest 24-hour concentrations, and second highest 24-hour con centrations at several locations of interest were compared. Generally the concentration values for the year 1964 were the median value (third highest) of the five years, with exceptions usually being the fourth highest value. Therefore, on the basis of this and the previous com parison, it was concluded that the year 1964 is sufficiently repre sentative of a longer period of record in this area. All of the dispersion modeling codes used in these analyses employ Gaussian, straight-line trajectory techniques, i.e., RAM, AQDM, CRSTER and PTMAX. If a flow reversal of 180 occurs, such as under daytime sea-breeze and nighttime drainage flow conditions, VC could recirculate of a location over which it had previously passed, adding to the concen tration previously registered. None of the computer codes applied in this study are capable of treating the recirculated aspect of the plume(s). Therefore, the possible importance of this deficiency was investigated, and it was found to be insignificant as regards 24-hour average and annual average VC concentrations from an areally small plant in this locale, for two reasons. The first of these reasons was that in general the sea-breeze in this area has a westerly component, rather than a southerly component, due primarily to the local topography. The main drainage flow at night from the north (San Gabriel Mountains) is essentially normal to this sea-breeze flow, and thus the flow reversal simply does not take place, thereby obviating recirculation. UCC 024017 Draft Report 4/20/78 However, it was assumed that a flow reversal could conceivably occur, either E-W or N-S. Dispersion calculations were performed under different postulated sea-breeze cum drainage flow cases, e.g., 16 hours (or 12 hours) of D with wind speed of 3 m/s, followed by a reversal with 8 hours (or 12 hours) of F with 3 m/s. Both the point and fugitive sources from the plant were so simulated, with the effects on receptors at 0.5 km and at 3 km calculated. Concentrations for 24-hour averages were obtained at each such receptor, both excluding and including the contribution from recirculation. The relative differences were shown to be primarily a function of receptor distance and wind speed, but in any case insignificant (<1 percent) at the distances of primary interest (0.5 km - 8 km) for the averaging periods of interest (24-hours and annual). Therefore, the neglect of recirculation by the dispersion codes used in this analysis was shown to be unimportant, on two counts. * Perhaps the major area in which professional judgment was applied in this study was in the selection of the version or option of the RAM codes applicable to this study, i.e. either RAMF (urban dispersion coefficients) or RAMFR (rural dispersion coefficients). Intuition suggested that RAMF was more appropriate, but analyses indicated that RAMFR would provide more realistic concentration estimates, and it was used for the final applications runs. To begin with, RAMF and the urban coefficients have undergone little operational validation. An explanatory note is in order at this point. The urban vertical dispersion coefficients are far greater than the rural vertical dispersion coefficients, for the corresponding sta bility class and distance. Therefore, RAMF generally predicts higher maximum concentrations for elevated sources and for lower concentrations for low level sources than does RAMFR, given the same emissions and meteorological input. The writer has recently performed a major study of sulfur dioxide (SOg) dispersion in three urbanized areas in Florida (Dames & Moore, 1978). One conclusion of the study was that RAMF overestimated maximum observed 24-hour $0^ concentrations by factors of 3 or 4, whereas RAMFR UCC 024018 Draft Report 4/20/78 yielded more far realistic concentration estimates. Of course, these Florida analyses primarily involved elevated sources, in contradistinc tion to this study. However, these Florida results suggest that for low level sources, if anything RAMF would be expected to underestimate concentrations. To analyze this possibility further, RAMF and RAMFR were run-with the VC sources from the plant on 1964 data, and a receptor grid consisting of sensitive receptors 1 and 2 (see Figure 5). Annual average VC con centrations were calculated, as well as the frequency distribution of 24-hour VC concentrations. An AQDM run also was performed, with these two ad hoc receptor locations, to generate two additional annual average values for comparison at the same locations; AQDM has been used and evaluated so often that the results therefrom can be reasonably thought of as providing a "ball park" reference point. The results of the annual average comparisons showed that the AQDM values were intermediate between those of RAMF and RAMFR at each of the two locations. At location 1, the values were as follows: RAMF, 0.71 ppb; AQDM, 1.79 ppb; RAMFR, 2.34 ppb. At location 2, the values were as follows: RAMF, 0.59 ppb; AQDM, 1.36 ppb; RAMFR, 2.37 ppb. When it is considered 1) that the meteorological data input to AQDM was the STAR listing for 1960-1964, and the meteorological input to the RAM codes was 1964 hourly data, and 2) that the dispersion algorithms differ, it can be concluded that the RAMFR and AQDM results are reasonably consistent with each other. The RAMF annual concentrations are not consistent with the other two sets, and are lower by a factor of 3 or 4. This ratio essentially held for the maximum 24-hour concentrations from RAMFR and RAMF, i.e. RAMF predicted lower maxima by a factor of about 4. There fore, because 1) available model validations indicate that the urban dispersion coefficient used in RAMF unrealistic, 2) results from RAMFR and AQDM were reasonably consistent, and 3) RAMFR provides far more conservative concentration estimates for the source types under con sideration herein, RAMFR was selected for the applications runs. UCC 024Q19 Draft Report 4/20/78 4. MODELING RESULTS 4.1 ANNUAL AVERAGE VINYL CHLORIDE CONCENTRATIONS Figures 6 and 7 display the results of the AQDM runs on the 19601964 Long Beach data. All continuous routine vinyl chloride releases were simulated, and non-routine releases were also included in a con servatively pro-rated manner. To prepare Figures 6 and 7, the calculated VC concentrations in 3 ug/m were converted to ppb and plotted at each indicated grid point. The indicated isopleths were drawn by hand. Figure 6 displays the results of the 1 km grid run, and Figure 7 those of the 2 km grid run. These results indicate that the maximum annual average concentrations greater than 1 ppb are restricted to locations within about 1.5 km of the plant. At distances greater than 2 km, concentrations are less than 0.5 ppb. These AQDM results are consistent in pattern with if somewhat lower in magnitude than the annual average concentrations predicted by RAMFR. RAMFR indicated maximum annual average concentrations at several of the sensitive receptor locations within 0.7 km - 1 km of the plant of 2-3.4 ppb. Figures 6 and 7 indicated that annual average VC conditions rapidly drop off with distance from the plant site. A synthesis of the AQDM and RAMFR results is that the maximum annual average VC concentrations at the closest sensitive receptor locations at 0.7 km - 1 km can be expected to be on the order of 2-3 ppb, that at distances greater than 2 km, annual values will be less than 0.5 ppb, and at distances greater than 4 km, annual concentrations will be less than 0.2 ppb. Appendix II sets forth a listing of the emissions input to the AQOM runs. Appendix III lists the meteorological data input thereto. Appendix IV sets forth a listing of the VC concentration output listings from AQDM. ucc 024020 Draft Report 4/20/78 4.2 24-HOUR VINYL CHLORIDE CONCENTRATIONS Figure 8 displays the results of the RAMFR runs on the 1964 hourly Long Beach data. All continuous routine vinyl chloride releases were simulated, and non-routine releases were also included in a conserva tively pro-rated manner. To prepare Figure 8, the calculated maximum 24-hour VC concentra tions at each sensitive receptor were converted to ppb and plotted at the appropriate location. The indicated isopleths were drawn by hand. Figure 8 is a composite worst case 24-hour concentration plot, because the indicated concentrations do not occur during simultaneous 24-hour periods. This plot indicates that maximum 24-hour VC concentrations ' greater than 20 ppb can be expected out to distances less than 2 km from the plant complex, and that 24-hour maxima greater than 10 ppb can be expected out to distances less than 4 km from the plant. In addition to calculating the maximum expected 24-hour concen trations at sensitive receptor locations, full annual frequency distribu tions of the 366 consecutive 24-hour concentrations at these locations were calculated, in order to define the expected variability in 24-hour concentrations. Maximum 24-hour concentrations at receptors 3, 1, and 5 were greater than 20 ppb, and these three locations are situated within the 20 ppb isopleth at distances less than 1.5 km from the plant. The frequency distributions at these locations are summarized in Table 4-1. At receptor 3, where the overall 24-hour maximum value of 27.1 ppb, 32 predicted values exceeded 10 ppb. At receptor 1, at which the maximum was 20.9 ppb, 18 values exceeded 10 ppb; and at receptor 5, the max imum was 20.3 ppb and 18 values exceeded 10 ppb. In order to define the expected temporal variation of 24-hour VC concentrations, the sequence of the 366 calculated values were listed for the most heavily affected location, i. e., receptor 3. The se quential plot of these values is presented in Figure 9. An Inspection of the listing of these 366 values indicated that during this data year, there were three occurrences when two or more consecutive 24-hour cal culated values exceeded 10 ppb, during one of which three consecutive UCC 024021 Draft Report 4/20/78 TABLE 4-1 FREQUENCY DISTRIBUTIONS OF 24-HOUR VINYL CHLORIDE CONCENTRATIONS IN ppb AT THREE HIGHEST RECEPTOR LOCATIONS Maximum Value (ppb) Range (ppb) 25-30 20-25 15-20 10-15 5-10 0-5 Receptor 3 Receptor 1 27.1 20.9 10 41 93 18 * 14 66 48 268 300 Receptor 5 20.3 0 1 .1 16 48 300 ucc 024022 Draft Report 4/20/78 values greater than 10 ppb occurred, on days 34, 35, and 36. There were no occurrences when consecutive days exceeded 15 ppb. In order to examine more explicitly the effects on possible 24-hour concentrations of the oxychlorination vent scrubbing, the associated VC emissions were assumed to occur on the worst identified day at the worst receptor, even though this release occurs only for about one week per annum. The identified day was that which produced the value of 27.1 ppb at receptor 3, and the resimulation by RAMR of this day with the additional vent scrubbing emissions increased the calculated value to about 28 ppb. In order to evaluate the effects of the five postulated relief value discharges per annum, the associated emissions thereof were used as input to PTMAX (see Appendix I). These releases are generally of the order of 5 minutes in duration, and therefore any high concentrations would be of very short duration. PTMAX uses a full spectrum of hypothetical mete orological conditions as input, and therefore a wide range of calculated concentration values for these rare events is possible. Selection of roughly the 90th percentile worst of these conditions in conjunction with the worst of the five postulated relief valve discharges, and averaging the results to 24-hour concentrations yielded an estimated value on the order of 500 ppb at very close-in locations (0.2 km) offsite. Appendix II sets forth a listing of the emissions in put to the RAMFR runs. Appendix III lists the hourly meteorological data of that day which produced the highest 24-hour VC concentration of 27.1 ppb from the routine releases. Appendix IV sets forth listings of the 24-hour frequency distributions at the 10 worst sensitive receptor lo cations, and the sequential listing of values at receptor 3. ucc 024023 REFERENCES Draft Report 4/20/78 Not yet available. O UCC 024024 Draft Report 4/20/78 9 ( 3 3 4 B 6 T 8fcn>. FIGURE 1. GENERAL LOCATION OF THE PLANT SITE. UCC 024025 D ra ft Report 4/20/78 ucc 024026 D ra ft Report 4/20/78 Q tt 024027 ,.Dr'M>-RepQir-t- 4/20/78 WNW W WSW SPEED (KNOTS) FIGURE 4. ANNUAL WIND ROSE FOR LONG BEACH AIRPORT. 1960-1964 PERIOD OF RECORD. UCC 024028 HOSPITAL W SCHOOL Draft Report 4/20/78 HI ''N -W in FIGURE 5. LOCATIONS OF THE 28 SENSITIVE RECEPTORS, INCLUDING 8 RESIDENCES, 18 SCHOOLS AND 2 HOSPITALS WITHIN ABOUT 4 KM OF THE SITE. UCC 024029 Draft Report 4/20/78 | i FIGURE 6. ISOPLETHS OF ANNUAL AVERAGE VINYL CHLORIDE CONCENTRATIONS ,IN PPB.FROM AQDM WITH 1 KM GRID. 1960-1964 LONG BEACH DATA. EMISSIONS FROM BF GOODRICH/STAUFFER PLANT AFTER 10/21/78. UCC 024030 Draft Report 4/20/78 O UTM - NORTH {krtO FIGURE 7. ISOPLETHS OF ANNUAL AVERAGE VINYL CHLORIDE CONCENTRATIONS IN PPB FROM AQDM WITH 2 KM GRID. 1960-1964 LONG BEACH DATA. EMISSIONS FROM BF GOODRICH/STAUFFER PLANT AFTER 10/21/78. UCC 024031 Draft Report 4/20/78 ( (unt) H> M O N -Win 024032 Q 2 4 -HOUR VC CONCENTRATIONS O -- 4^ O CoCOO o FIGURE 9. SEQUENCE OF DAILY CALCULATED 24-HOUR VC CONCENTRATIONS BY RAMF ON HOURLY 1964 LONG BEACH DATA, EMISSIONS AFTER 10/21/78. RECEPT0R_3 Draft Report 4/20/78 APPENDIX I 1.1 INTRODUCTION COMPUTER CODES Brief descriptions are provided of the EPA computer codes used for the quantitative dispersion modeling performed in this study. The codes are AQDM, the RAM series, CRSTER and PTMAX. The format for the discussion of each code is as follows: 1. Reference 2. Abstract 3. Input 4. Formulation 5. Output UCC 024034 Draft Report 4/20/78 1.2 AIR QUALITY DISPLAY MODEL (AQDM) Reference: U. S. Department of Health, Education and Welfare, Ai.r Quality Display Model, User's Manual prepared for HEW, Washington, D.C., NTIS PB 189 194, November 1969. Abstract: Program AQDM is a multiple point and area source disper sion code which calculates annual average concentrations over uniform 15 x 15 square grid with user-specified grid spacing. Modeling is by means of standard Gaussian techniques. The version used in this study incor porated the Briggs plume rise equations. Input: Up to 1000 point and area sources with specified locations and emissions parameters. Joint frequency distribution of stability/ wind speed/wind direction. 15 x 15 regularly spaced grid with specified spacing and up to 12 ad hoc receptor points. Formulation: Coning dispersion with Pasquill-Gifford (rural) coefficients for classes 1-5. Horizontal sector-ayeragcd concentrations. Briggs plume rise. Output: Annual average concentrations at each grid point. Partial concentrations by source at 5 highest receptor points. UCC 024035 Draft Report 4/20/78 1.3 REAL-TIME AIR QUALITY SIMULATION MODEL (RAM) Reference: Turner, D. B., and Novak, J. H., "User's Guide for RAM." Environmental Protection Agency, Research Triangle Park, North Carolina 27711, 1977 (Not yet officially available). Abstract: RAM is a set of codes incorporating a steady state Gaussian plume model for estimating concentrations of relatively stable pollutants i for averaging times from an hour to a day from point and area sources. Level or gently rolling terrain is assumed. Calculations are performed for each hour. Both rural and urban versions are available. RAM (urban and RAMR (rural) ace used to run on 24 hours or less of meteorological data. RAMF (urban) and RAMFR (rural) run on a year of data. Input: Up to 26 point sources and up to 26 area sources with specified source characteristics. Up to specified 60 receptor points. Formulation: Coning dispersion in accordance with Turner Workbook (rural) coefficients and Briggs equation. Constant horizontal wind field.. Perfect reflection from surface and mixing layer, until a = 1.6 times mixing height, then uniform distribution in vertical. Power law vertical wind profile as per DeMarrais, with reference height of 10 m. Chemical reactivity and physical removal treated by separate user-specified half- life values. Output: Hourly and average (up to 24 hours) concentration at each receptor. Limited individual source contribution list. Cumulative frequency dis tribution based on 24-hour averages and up to 1 year of data at a limited \ number of receptors. .. : : .. UCC 024036 Draft Report 4/20/78 I*.4 CRSTER (EPA SINGLE SOURCE MODEL) Reference: Environmental Protection Agency. "User's Manual for Single Source (CRSTER) Model." Publicaton No. EPA-450/2-77-013 (NTIS PB 271360). Office of Air Quality Planning and Standards, Research Triangle Park, North Carolina 27711, July 1977. Abstract: CRSTER is a steady state Gaussian plume technique thought to be applicable to both rural and urban areas in uneven terrain.. The purpose of the technique is: (1) to determine the maximum concen trations, for certain averaging times between 1-hour and 24-hours, over a one year period due to a single point source of up to 19 stacks, (2) to determine the meteorological condition's which cause the maximum concentrations, and (3) to store concentration information useful in calculating frequency distributions for various averaging times. The concentration for each hour of the year is calculated and midnight-tomidnight averages are determined for each 24-hour period. Input: Source.parameters for up to 19 co-located point sources. Monthly variations in emissions possible. Receptor grid is 180 points specified by 5 input distances and the 36 10 sectors. Meteorological input is one year of hourly values of wind speed and direction, Pasquill stability class, mixing height and temperature. Formulation: Coning dispersion in accordance with Turner Workbook (rural) coefficients and Briggs equations. Constant horizontal wind field. Perfect reflection from surface and mixing layer, until = 1.6 times mixing height, then uniform distribution in vertical. Power law vertical wind profile in accordance with exponents of DeMarrais, with reference UCC 024037 Draft Report 4/20/78 height of 10 m. Terrain treated by reduction of calculated effective stack height by elevation differential between plant and receptor point. Output: Highest and second highest concentrations for the year at each receptor for averaging times of 1-, 3-, and 24-hours, plus a user-selected averaging time which may be 2-, 4-, 6-, 8-, or 12-hours. Annual arithmetic average at each receptor. For each day, the highest 1-hour and 24-hour concentrations over the receptor field. Hourly concentrations for each receptor on magnetic tape. UCC 024038 Draft Report 4/20/78 1.5 PTMAX Reference: Turner, D. B., User's Guide to UNAMAP, Environmental Ap plications Branch, Meteorology Laboratory, Mail Drop 80, EPA, Research Triangle Park, NC 27711, 1975. Abstract: Program PTMAX calculates maximum short-term (approximately 10-minute), ground-level concentrations resulting from the emissions of a single-point source and the distances from the source at which these maxima occur. Rather than using actual meteorological conditions as input, a spectrum of 49 simultaneous combinations of wind speed and stability conditions are tested, so that the resulting calculations of concentration and distance reflect a range of possible atmospheric conditions. PTMAX is based on standard Gaussian dispersion concepts and incorporates dif fusion coefficients found in Turner, 1970. Plume rise calculations are made using a form of the Briggs method (Briggs, 1971). Input: Emissions input consists of stack parameters of a single source. Meteorological input is contained internally within the code, i.e., 49 combinations of wind speed and Pasquill stability class. Formulation: Coning dispersion in accordance with Turner Workbook coefficients and Briggs equations. Output: Magnitude and location of maximum 10-minute concentrations and associated effective stack height for each meteorological combination. UCC 024039 Draft Report 4/20/78 APPENDIX II LISTINGS OF EMISSIONS INPUT DATA This appendix presents the emissions inventory as used as input data to the computer programs AQDM and RAMFR. Table II-l lists the input emissions data to AQDM, and Table II-2 lists the input emissions data to RAMFR. UCC 024040 Draft Report 4/20/78 TABLE 11-1 EMISSIONS INPUT DATA TO PROGRAM AQDM Page 1 of 2 (XYC: FUGITIVES INCKLA^FD 6Y FACTOR 1.6). r-F bbOUWICH CHEMICAL COMPANY (WITH STAUF'FtK CHEMICAL COMPANY) VCM-hv'C CHEMICAL PLANT COMPLEX* LUNG 6F.ACH CA (MtT-64) VCH XVC * I Nu A T A '.'S09CE = 37, Si-rtLF (1,1) =385.453 ,3743.000 . 0.0 U 0 0 ,0.00066 .0 .00068 P'+,4 385.469 ,3743.010 .0.0000 ,0.00050 ,0 .00060 ,06.0 385.494 ,37 4?,995 ,0.0000 ,0.00043 , 0 .00043 , lb. 3 385.508 3742.987 0.0000 , 0.0 0 0 6 0 .0 .00060 * ) 5.2 385.509 ,3742.960 .0.0000 .0.00091 . 0 .0009) ,09.1 3H5.513 ,3743.009 ,0.0000 ,0.00006 .0 .00006 .15.2 385.545 .3743.007 , 0.0 0 0 0 ,0.00003 *.0 .00003 .15.2 385.559 .3742.949 .0.0000 0.0U0O6 0 .00006 ) 6 3 385.508 ,3742.980 . 0.0 0 0 0 ,0.00003 .0 .U0U03 12.2 . 385.495 .3742.b91 .0.0000 ,0.00006 .0 .00006 ,24.4 385.491 .3742.960 .0.0000 ,0.00154 .0 .00154 , ns.5 385.480 .3742.936 .0.0000 ,0.00026 .0 .00026 , U5.5 385.668 .3743.082 . 0.0 0 0 0 ,0C 0 Ob9 ,0 .00069 , 33.0 385.661 .3743.126 0.000 0 .0.00009 .0 .(0009 ,11.0 385.668 .3743.103 ,0.0000 ,0.00013 ,0 .00013 ,11.0 365,714 .3743.102 , 0.0 0 0 o ,0.00038 .0 .00036 ,11.0 365.633 .3743.106 .0.0000 ,0.0112b , 0 .01125 ,09.0 385.747 .3743.069 ,0.0000 ,0.0019? i 0 .0019? ,09.0 385.617 .3743.103 ,0.0000 ,0.00706 ,0 .00706 ,03.0 385.817 .3743.120 0.0 0 0 0 0.00359 , 0 ,00.359 .07.0 385.747 374 i . 0 6 3 ,0.0000 ,0.0039b .0 .00395 ,15.0 365.641 . 3743.112 ,0.0000 ,00 Qb63 , u .00663 .10.0 385.747 .3743.063 ,0.0000 ,0.00096 . 0 .0009o ,10.0 385.593 3743.136 , 0.0 0 0 0 .0.00133 ,0 .00133 ,22.0 385.600 ,3743.086 .0.0000 .0.00057 .0 .00057 ,22.0 365.724 3743.063 0.0000 0.00067 .0 .00ub7 22.0 385.540 3743.072 ,0.0000 0.00197 .0 .00197 ,14.9 3fa5.704 ,3743.076 ,0.0000 .0.00076 ,0 .00076 .13.7 385.695 .3743.050 .0.0000 ,0.00132 ,0 .00132 . 15.2 385.661 .3743.146 .0.0000 .0.00050 ,0 .00050 ,06.6 385.661 .3743.126 . 0.0 0.0 .0007,0. 0007,12. 2, 385.668 ,3742.103 0.0 0.0 .0017,0. 0017.12. 2, 385.453 .3743.000 .0.00.0 .0006,0. 0008,12. 2, 385.696 ,3743.137 .0.0056 ,0.020b. 0. 0309,06.0, 385.473 ,3743.062 .0.0056 0.0176, 0. 0267,06.0, 385.623 ,3743.06? ,0.008b ,0.0075. 0. 0113,06.0, 385.47 1 ,3742.987 0.0066 ,0.0095, 0. 0l4j,(>b.0, ucc 024041 Draft Report TABLE 11-1 (Continued) Pl.Uwt < 11) =14,33 *0.76 .244.0, 00.UO *4.69 ,810.4, 15.33 *0,62 322.0* 33.09 ,0.68 3 6 0.9 * 21.16 .0.45 *360.9* 01.07 *0.89 *294.0* 01.07 *0.63 ?9<*.n* 01.07 * 0. H9 294.0* 01.07 ,0.63 *294.0* 01.07 *0.89 *294.0, 7.93* 10.20 *296.7. 01.10 *0.60 294.-0* 43.64 *0.20 310.9. 05.63 *0.25 *327.6* 05.H3 0.25 ,327.6* 10.36 *0.48 *327.h. OH.74 ,0.22 *355.4, 08. 74 *0. 13 355.4. 06.76 *0.84 * 3.J6.7. 1 1.44 *1.54 *338.7. 24.27 *0.61 *338.7. 00.20 *0.26 836.7* 00.16 *0.20 * 3 3 r , 7 , 0 0.85 .0.74 *294.u no. hs *0.49 * 294.0, 00. b5 .0.63 *294.0, 7.93, 12.80 *300.0, 0 7.93 .6.40 *300.0. 07.93 .6.10 300.0. 01.10 ,0.60 *?94.n* . * . * . *26 0 0 07 333 2 , * . * . *26 0 0 07 333 2 . * . * . .26 0 0 07 333 2 4/20/78 Page 2 of 2 ucc 0240-42 m D raft Report ucc 024043 TABLE II-2 EMISSIONS INPUT DATA TO PROGRAM RAMFR BF iGOODRICH t STAUFFER VCM-PVC PLANT COMPLEXt LONG BEACH CA 30 4 21 2 0,075 1.0 BFG STK 01 385.4533743,000 0,0072 9.9 24.4 310.0 BFG CENT 02 BFG BLTK 03 BFG DRY 04 385.4693743,010 385.4943742.995 385.5083742.987 0,0053 0.0045 0.0063 9,9 9,9 9,9 06.0 18.3 15,2 310.9 322.0 360,9 BFG DRY 05 BFG SILO 06 BFG SILO 07 385.5093742.960 385.5133743.009 385.5453743.007 0,0096 0,0006 0,0003 9,9 9.9 9.9 09.1 15,2 15.2 360.9 310.0 310.0 BFG SILO 08 385.5593742.949 0.0006 BFG SILO 09 385.5083742.980 0,0003 9,9 9,9 18.3 12.2 310.0 310.0 BFG SILO 10 385.4953742,891 0,0006 9,9 24.4 310,0 BFG CT 11 385.4913742.960 0,0162 BFG LAB 12 385.4603742.938 0.0027 9,9 9.9 05.5 05.5 310.0 310.0 SVC OXY 13 385.6683743,082 0.0001 9.9 33.0 310,9 SPV N01 14 385.6613743.126 0.0009 SF'V N02 15 385.6683743.103 0.0014 9,9 9,9 11.0 11.0 327,6 327.6 SPV N03 16 335.7143743.102 0.0040 9,9 11.0 327,6 SPV BLTK .17 SPV BLTK 18 SPV DRY2 19 385.6333743.105 385.7473743.069 385.6173743.103 0.1184 0.0202 0.0743 9,9 9,9 9.9 09.0 09.0 03.0 355.4 355,4 338.7 SPV DRY3 20 SPV DRY4 21 SPV CENT 22 385.6173743.120 385.7473743,063 385.6413743.112 0.0378 0.0416 0.0592 9,9 9.9 9.9 07.0 15.0 10.0 338.7 338,7 338.7 SPV CENT 23 385,7473743,063 0.0101 SPV SILO 24 385.5933743.138 0,0140 9.9 9,9 10.0 22.0 338.7 310.0 SPV SILO 25 385.6003743.086 0,0060 9,9 22.0 310.0 SPV SILO 26 385.7243743.063 0.0070 SPV CT1 27 385.5403743.072 0.0207 9.9 9,9 22.0 14.9 310.0 310,0 SPV CT2 28 385.7043743.076 0.0080 9.9 13,7 310.0 SPV CT3 29 SPV LAB 30 END POINTS 385.6953743,050 385.6813743.148 0.0139 0,0053 9,9 9,9 15,2 05.5 310,0 310.0 385.660 385.435 385.585 3743.099 3743,024 3743,024 0.075 0.075 0,075 0,2163 0.1869 0.0794 9,9 9.9 9.9 6,0 6,0 6.0 385.43;s 3742.949 0.075 0,1003 9,9 6.0 0,76 4,69 0.62 0.68 0.95 0.89 0.63 0.89 0,63 0.89 10,20 0,60 0,20 0.25 0.25 0,48 0,22 0.13 0,84 1.54 0,61 0,25 0,20 0.74 0,49 0.53 12,80 6,40 9.10 0,60 14.33 00,00 15.33 33,09 21.16 01.07 01.07 01.07 01.07 01.07 07.93 01.10 43.69 05.83 05.83 10.36 08,74 08.74 06,78 11,44 24,27 00,20 00,15 00.85 00,85 00.85 07.93 07.93 07.93 01.10 Draft Report 4/20/78 APPENDIX III LISTINGS OF METEOROLOGICAL INPUT DATA This appendix presents meteorological data sets which, were used as input to the computer programs AQDM and RAMR, Table III-l lists the data used for AQDM. Table III-2 lists the twenty-four hours of meteoro logical data for the day (day 274 of 1964) on which the highest 24-hour concentration was predicted for the site referred to in the main body of this report as Receptor 3. UCC 024044 Draft Report 4/20/78 TABLE III-l METEOROLOGICAL INPUT DATA TO AQDM Page 1 of 2 U * 1 ) =U,0001 0.0 0 0 0 . 0.0 0 0 0 .0,0000 0.0000 *0.0000 O.OOOI 0.0000 .0.0000 .0.0000 *0.0000 .0.0000 0.0002 0.0000 .0.0000 *0.0000 *0.0000 *0.0000 0.0001 0.0001 . 0.0 U 0 0 .0.0000 *0.0000 0.0000 0.0001 0.0000 .0.0000 *0.0000 *0.0000 .0.0000 0.0001 0.0001 .0.0000 .0.0000 0.0000 .0.0000 o.ooos 0,0003 .0.0000 .0.0000 0.0000 ,0.0000 0.0007 0.0005 .0.0000 0.0000 *0.0000 *0.0000 0.0017 0.0013 .0.0000 *0.0000 .0.0000 .0.0000 0.0009 0.0007 0.UOQO *0.0000 *0.0000 .u.0000 0.0010 U.OOOtt .0.0000 . 0.0000 .0.0000 .0.0000 0.0009 0.0005 0.0000 .0.0000 *0.0000 .0,0000 0.0009 0.0005 .0.0000 .0.0000 *0.0000 .0.0000 0.0005 0.0004 *0.0000 *0.0000 *0.0000 .0.0000 0.0001 0.0000 .0.0000 *0.0000 *0.0000 .0.0000 0*0001 O.0000 *0.0000 .0.0000 ,0.0000 0.0000 0.0023 U.0013 .0.0001 *0.0000 ,0.0000 .0.0000 0.0007 0.0004 *0.0000 *0.0000 ,0.0000 0.0000 0.0011 0.0006 .0.0000 .0.0000 *0.0000 *0.0000 0.0016 0.0009 .0.0001 .0.0000 *0.0000 *0.0000 0.0022 o.oooa *0.0001 *0.0000 *0.0000 .0.0000 0.0017 0.0017 .0.0001 *0.0000 *0.0000 *0.0000 0.0015 0.0021 .0.0009 .0.0000 *0.0000 *0.0000 U.0014 0.0024 *0.0029 *0.0000 *0.0000 *0.0000 0.0027 0.0056 *0.0120 *0.0000 0.0 (' 0 0 .0.0000 O.OOlfa 0.0033 .0,0067 *0.0000 *0.0000 . 0.0 0011 0.0012 0.0017 0.001H 0.0024 0.0019 0.0014 0.0025 .0,0037 .0.0000 0.0000 .0.0000 0.0028 .0.0017 .0,0000 0.0000 . 0.0 0 0 i,l 0.0033 .0.0021 *0.0000 *0.0000 .0.0000 0.0035 * 0.003*. * 0.0 0 0 U *0.0000 .0.0000 0.0016 tU.0004 *0.0000 0.0000 * 0 . UO'JII 0.00C9 0.0002 .0.0000 * 0.1) 0 0 0 .0.0000 0.0009 0.0028 *0.0010 *0.0000 *0.0000 .0.0000 0.0004 0.0011 .0.000 .1 * 0.0 0 0 o 0.0000 *0.0000 0.0005 0.0008 .0.0002 *0.0000 *0.0000 *0.0000 0.0008 0.0021 .0.0005 . 0.0000 *0.0000 0.0 0 G 0 0.0013 0.0030 .0.0004 *0.0000 *0.0000 *0.0000 0.0009 0.0025 *0.0008 .0.0000 o.onoo *0.0000 0.0007 0.0018 *0.0011 *0.0001 0.0000 *0.0000 0.0003 0.0015 *0.0039 *0.0004 *0.0000 .0.0000 0.0010 0.0035 *0.0216 *0.005<* *0.0000 *0.0000 0.0004 0.0012 *0.0100 0.0024 *0.0000 .0.0000 0.0005 0.0 015 0.0O*rl 0.0006 O.i.OOO o.onoo 0.0004 0.0014 .0.0027 *0.0005 0.0000 ,0.0000 0.OOOH 0.0024 0.0109 *0.0025 *0.0001 .0.0000 C . 0 01 4 0.0 U M .0.0153 *0.0018 . 0 . (1(100 .0.0000 V . 0 0 0 9 0.0028 .0.0029 .0.0004 .0.0000 .0.0000 UCC 024045 Draft Report 4/20/78 TABLE 1II-1 (Continued) 0.00 09,0.0023,0.00 OH 0.0039,0.007=1<0.0010 0,0011?. 0.0024,1). 0003 0.0029,0.0G46,0.0013 0.00b7,0.010n,0.0033 0.0074,0.0162*0.0047 Oji.O.O 71_,J). 0JL 7 3.0.0 0 ft 9 0.0053,0.0120,0.00b3 0.0037,0.0090,0.0U67 0.0054.0.0143,0.0176 0.0019,0.0054,0.0066 0.001b.0.003b,0.0041 0.0016,0.0029,0.0047 0.0023,0.0053,0.0163 0.0056,0.0131*0.0135 0.0035,0.0062,0.0036 0.0019,0.0044,0.0012 0.008b.0.0067,0.0009 0.0026,0.0022,0.0003 0.0049,0.0041,0.0007 0.0101.0.0115,0.0012 0.0149,0.014?,0*0010 0.0121,0.0140,0.0009 0.0075*0.0064.0.0006 0.0032.y.0042,0.0010 0.0046,0.0077,0.0025 0.0016,0.0024,0.0011 0.0014,0.0024,0.0009 0.00 36.0.0 046,0.00-, 1 0.0090,0.0133,0.0144 0.0301*0.0453,0.0190 0.0 l,(). 0246,0.0066 0.01 OH,0.0130,0.0017 0 .0000 0 . 0004 0 . 0002 0 . 0006 0 . U024 0 0 0 16 0 ,0014 0 . 00?u 0 .0021 0 ,00ft? 0 . 0 0 42 (! .0019 0 .0060 0 . 0240 0 .0102 0 .0033 0 . 0 006 0 ,0000 0 .0000 f; . 0 0 0 0 0 .0000 0 ,0000 0 . 0000 0 . 0000 0 .0000 0 . 0000 0 .0000 0 .0000 0 . 0000 0 , 0000 0 . OOOU 0 . 000(1 0 . 0000 0.0000 0.0001 0.0000 0.0001 0.0 0 06 O.OOQft 0.0001 0.0002 0.0 0 Cl 1 0.0002 0.0001 0.0000 0.0005 0.0027 0.0006 0.0005 0.0002 0.0000 0 ^0 0 0 0 0.0000 0.0000 0.0000 0.0000 0.0000 0.0000 0.0000 0.0000 0.0000 0,0000 0.0000 0.0000 0.0000 0.0000 0.0000, 0.0 0 0 0 , 0,0000* 0.0000, 0.0002, 0.0002, 0.0000, o.oooo* 0 . U0Q0 o.oooo. 0.0000* 0.0000* 0.0001, 0.0006, 0.0001 , 0.0003, 0.0001, 0.0000, 0.0000, 0,0000, o.oooo. 0.0000, o.oooo. o.oooo. o.oooo. 0.0000, o.oooo. 0.0000, O.OOOO, O.OOOo, 0.0000 0.0000, 0.0000, < '2T6(t#, = 600.0 , 921 6r!/* = "30 0 0 , T o =29 3.15,^45=1000.0, lvJKlGGS=35 Page 2 of 2 ucc 024046 Draft Report 4/20/78 TABLE IU-2 TWENTY-FOUR HOUR METEOROLOGICAL INPUT DATA TO RAMR FOR DAY OF HIGHEST PREDICTED CONCENTRATION Iti- / Ot. JL/'L Ci'ii-. 'I* vl -rv(. tOr-'^LrA -on 474 1 \'H of ' 'c T i S A T A r - r '| (1 \ 1 >) 69/ 274 s r* F f; u 1 1 '.-i- (-1/5) Ht.I:'`-l ( ' ) 1 (- i-'H ( jtG-K ) r.'ILI 1 r C| OSS i 1.64 6IM (j 0 c 49,0U 3 7 7.00 I . 00 1.00 6 0 0.0 0 80 ' . DO 4 94 . iji) 1.00 6 in . I) 5 97.0 0 n 99. uu 1 . (HI 1 . <.! 0 * 0 o . I) U 16.95 7 96.0 0 r, 17.00 9 3i4.00 1 0 C64 , Oil i 1 229. >m. J if 2*^2 0 ij 1.0 II 2 ->7 2,57 c.ft 7 4,63 4 . `j 6 114,64 212.72 319.64 4 i' . 6i'6, jo 6 0`> .44 l o 3 0 6 . (> 0 14 343. t. o 1 / t' , 1. |> 4,14 b. c.* 6.1 7(--. 1 4 - 0 C . v MM;, (JO lu 3 0 7 . 1 : 1 7 -)01 . ill! 1 , > 7 . . t 1- 4 6 'I . U Ij 40 ? *5 4,0 0 ' J 2 *<*, c i> *- ' rlf . 0 0 6 . ri 4 <`.7-3 4.14 4.-7 1.1'U 1. 0 4.6 7 6 till . I) .* >(.<. 1' f , 1 .1 6 <' i' : 6 . ij ii 1<m . Ii| pi. .. . * O 141.00 , /T 1 l.i(i. oO 3 . "9 >. 'i9 6( , '.Ml o 71 >%! . e | i' 141 ' ! I JO - 29*5.9,1 466.93 465.37 f. .->4 /?7',, 4 c. -j** m hr 2'* 6.9 * fn'J.6 >i 2n9.1V 244.4 1 444.4# 4 4 5.9 . 49 7.1-4 .4M 29- . 4 4-)**. h- ; 4- 9 5.1' t` '111.9 / 4:9 . h f 2o9.He 4 . 2>' 4i>rt 7 1 4 CM ] ' 4 6.7 J f. s 6 (- 6 5 4 u 3 J 3 <4 S< r - 'U <i lI(*.ECT Ij:-j = 30'.. 0 2 kINIj 5td '.= e .7 1 n *-r h S I STFwft " 6E.9UL.1 i.vT i 1 ->> SHh.tO" 4VtM*.oE T. <?0 0s HOHAL >'( 1 L J 1 1* * h ucc 024047 Draft Report APPENDIX IV Not yet available. 4/20/78 'JCC 024048