Document XzvYDj5wNRZNXVqjk6LX9DoNB

TO: Doug Skokna ^ CP*' K FROM: DATE: Interoffice SUBj Communication T. G. Grumbles August 5, 1987 EMERGENCY RELEASE MODELING MEETING VIST/ As we have discussed, the subject meeting will be held on Tuesday, May 18, from 9:00 - 12:00 at the plant. Attached is a preliminary agenda for the meeting, a document describing the generalities of the project, and the actual modeling basis letter used at the LCCC. Please give me a call if you have specific questions on the agenda items. Please distribute this information to others you want involved, other than those I have copied. ----T. G. Grumbles ajo .201 cc R. W. Seymour D. L. Mahler L. A. Mauerman J. 0. Gibson P. F. Fetizanan M. J. Heller ) OOOOZ^52 AGENDA MODELING MEETING TUESDAY, AUGUST 18, 1987 9:00 - 12:00 I. Introduction Overview Tom Grumbles A. Purpose of Meeting B. Modeling Needs/Uses C. Regulatory Concerns (TSCA, SARA Title III) 15 II. Modeling Technology A. Capabilities B. Limitations Mary Jo Heller 15 III. Vista Capabilities A. Software Programs B. Schedule Mary Jo Heller 15 IV. What to Model A. Chemical Selection B. Data Needs Tom Grumbles/ Mary Jo Heller 30 V. The LCCC Experience A. Process Flow of Project B. Review of Results Mary Jo Heller 45 VI. Discussion of Action Steps/Schedule Group VII. Future Plans 00022933 TO: Distribution Interoffice Communication FROM: DATE: SUBJ: T. G. Grumbles November 12, 1986 DISPERSION MODELING PLANNING MEETING vise* The subject meeting will be held Friday, November 21 at the LCCP Safety Department conference room, from 9:00 a.m. - Noon. An agenda for the meeting, a general description of the modeling project, and a background letter is attached. Please note Item IV and be thinking about what would be useful for aiding the plant in the reporting of emergency releases to local and federal authorities and community right-to-know aspects of recent regulations. If you have questions before the meeting, please call me (X3445) or Mary Jo Heller (X6884) in Houston. ) Thomas G. Grumbles ajo/9 Attachments cc T. H. Huffman -ICC/? L. R. Bauer - L-A& R. A. Conrad -Jcycp') '* J. A. DeBernardi -- LCCC- DISTRIBUTION: -V S. Pitts - 4 CC? Cnffhj DiLtdiar * M. Hayes t t P. Carrico -LCCP PUnf &/>/r_ + D. Friesner -- n N Ptf /' v M. Brady /1 AUo/vol * " + S. Ashby LCrVCAA IDtrfsfsri' M. Ashby t Oft**" t G. Hopkins tt Chf*f Preesrd. P. Warner i A-e '< /. + K. Fogg // S'aJc.^, ptrcdfor V R. Johnson Y M. Heller PE.O M'auac , VP. Fetizanan PEZ) Ztj \ flriLec- h / AGENDA MODELING MEETING FRIDAY, NOV El ID Ell 2A~^ -9:-00-------1^60 I. Introduction Overview l<:> cr ou~ ^' O0 A. Purpose of Meeting B. Modeling Needs/Uses C. Regulatory Concerns (TSCA, Tom Grumbles II. Modeling Technology A. Capabilities B. Limitations III. Vista Capabilities A. Software Programs B. Schedule Iv'SvEnd Products N A. \Format ^lon^raphs, Numbers, Overlays?) B. Uses Mary Jo Heller Mary Jo Heller ^ Group V What To Model A. Chemical Selection B. Data Needs Tom Grumbles/ ,, Mary Jo Heller 1 IahJ. L C c c <t^n. yi. "? V3) r-- T) NeviuioJ V<- 2c-*b. wAtS. fl-V "O'Cr.i ,"i X \ Sok*sL)._A ---- -JV\ s VA 4s Wv 000022935 3* MODELING PROJECT GOAL To produce data that will allow Vista to assess future releases in terms of potential for off-site concentrations of hazardous chemicals. METHOD Model release scenarios to give quantitative ranges of contaminent levels under specified conditions. Variables will include contaminant, release rate and points, and atmospheric conditions. STEPS 1. Obtain capability, including evaluation of models, software, and training to model in-house. 2. Evaluate use of the modeling data for future use in release response and determine format of end-product as indicated. 3. Identify chemicals to be modeled at each plant. These should be prioritized based on hazard* volume, use, storage, and delivery (transfer) conditions. Hazard evaluation should include volatility, flammability and toxicity. 4. Based on the above choose release scenarios to evaluate off-site concentration potentials. vvv 000022936 5. Determine criteria levels for action (notification, evacuation, etc.) for those contaminants modeled. GENERAL CONSIDERATIONS Many of the steps outlined can be done concurrently. The tendency will be to use intuition to pick the contaminants of concern. A consistent technique of hazard prioritization should be used to pick the chemicals to be modeled. The end result may be the same but we may be surprised. A screening method example is attached. This method was contained in EPA's Chemical Emergency Preparedness Program. The action level will take a good deal of discussion on any specific contaminant. Basically the affect to be avoided (irritation, lung damage, etc.) and the level where those effects occur must be reviewed. Acutely toxic materials (i.e., HF or HC1) are easier to deal with than VCM, or chronic hazards. VVV 000022937 APPENDIX C QUANTITY DETERMINATION METHOD This appendix describes a method that can be used to determine whether the quantity of acutely toxic chemical substance on a site, if accidentally released into the air, can produce potentially harmful concentrations in a nearby community. This method is not designed to provide an upper exposure limit for the release of a certain amount of an acutely toxic chemical substance, or to be used to establish a "maximum allowable" release amount for setting limits for storage or release under an ordinance or regulation. Its intended use is as a screening tool to allow community work group members to determine whether there is enough of a particular acutely toxic chemical substance on a site that, in the event of an accidental release, potentially harmful amounts might reach the community under the worst meteorological conditions that might occur. Tr.is procedure can help the work group rank various sites that handle acutely toxic chemical substances, so it can deal first with those sites and substances that pose the greatest potential for causing harm should an accidental leak occur. Communities are invited to contact their Regional Office of EPA for more detailed assistance on airborne contaminant modelling for purposes of developing specific levels, target populations, or evacuation areas as part of contingency planning. More sophisticated models that include the specific area meteorology, topography, and specific acutely toxic chemical substance physical/chemical data may be availble for these purposes. C.1 BACKGROUND EPA believes that communities may need assistance in determining whether enough of an acutely toxic chemical substance is present on a ssite that, if accidentally released, it could cause severe consequences in the community. Many models are available to predict the downwind concentration of a release into the air. However these models are very complex, not easily used, and not all are appropriate for this particular situation. In addition, each site is unique in topography, weather conditions, and distance to the community, and each chemical has its own unique characteristics and behavior. Since it is not reasonable to expect each community to have the expertise to develop a model simulation for every situation that might occur, EPA developed this method for use as a screening tool. EPA performed an analysis based on reasonable worst case conditions (like low wind speeds, ground level release, limited mixing and ambient temperature conditions) that could occur at any site using an EPA model that simulates dispersion from an instantaneous source. The model, known as INPUFF, is part of EPA's UNAMA? airborne contaminant modelling system. The model ran airborne releases associated with various quantities of chemicals over varying meteorological conditions, and the highest 30 minute average concentration at downwind distances ranging from 100 feet to 10 miles was calculated. The source of release was also simulated VVV 000022938 * * * September 23, 19S5 Draft -- Do Not Quote or Cite it tV C-2 at a low level (3 feet above the ground) and at ambient temperature (higher temperatures cause the vapor cloud to rise and result in lover concentrations in the air) for release tiroes ranging from 1 minute to 30 minutes. The 1 minute release condition was selected for use in the quantity determination because it generated the highest concentration in air per quantity of substance released. This condition simulates a rupture or large spill where a large cloud of material is generated quickly. The results of these model simulations were arranged in the form of a graph that can be easily used. The method for using the graph is described below. Note, however, that the method has significant limitations and remember that the quantity determination should be made in light of the kind of activities at each particular site associated with acutely toxic chemical substances. C .2 METHODOLOGY To use the graph to determine whether there is enough acutely toxic chemical substance on a site to generate airborne concentrations that could reach a level of concern beyond the site fence line, three pieces of information are needed: (1) the distance from the point where the acutely toxic chemical substance being evaluated is stored, processed or handled on site to the site fence line or the nearest population; (2) the molecular weight of the acutely toxic chemical substance; and (3) the level of concern for the substance. The distance can be determined from maps or from discussions with a technical contact at the site. The molecular weight of the chemical may be found on the Profile sheet included with the E?A list of acutely toxic chemical substances or, if the substance is not on the E?A list, from the site technical contact. The level of concern should be taken as the IDLH (Immediately Dangerous to Life and Health) level. This level (established by the National Institute of Occupational Safety and Health) represents the maximum level to which a person can be exposed to for 30 minutes and escape without suffering irreversible health effects. The EPA Profile sheets include IDLH values, when available. There are problems inherent in using the IDLH as a measure of the level of concern: The IDLH is based upon the response of healthy, male worker population and does not take into account exposure of the elderly, children, or people with other health problems. The IDLH is based upon a 30 minute exposure time frame which may not be realistic for accidental airborne releases. IDLH values do not exist for all acutely toxic chemical substances. VVV 000022939 * * fr September 23, 1985 Draft -- Do Not Quote or Cite * * * C-3 Thus, LDLH values used in this way do not necessarily indicate "safe levels," and a working group using them for screening purposes may wish to make appropriate allowances, e.g., for sensitive populations (such as the elderly) that may live close to the site boundary. If the acutely toxic chemical substance you are evaluating does not have an IDLH in the Profile or there is no profile available on the substance, use the LC(low) value which may be obtained from the site technical contact or from published toxicology sources. The LC(low) is the lowest lethal concentration observed in tests on laboratory animals. If an LC(low) is not available, use the LC50 value for the substance. Tne LC50 is the level for which 50 percent of the test animals died when exposed for a specified time period. This value may also be obtained from the Profiles, the site contact or toxicological literature. Once the level of concern is known, it can be used on the graph to determine quantity. Toxicological levels may be given, in units of parts per million (ppm), milligrams per cubic meter (mg/m3), milligrams per liter (cng/1), or grams per liter (g/1). Levels given in parts per million can be converted to grams per cubic meter (g/m3) as follows: Concentration, g/c3 = (level in tom) x MV 1000 24.5 where MW is the substance molecular weight. For example, chlorine has a level of concern (IDLH) of 25 ppm and a molecular weight of about 71. Thus the level of concern in grams per cubic meter is: Concentration, g/m3 = 25 x 71 = 0.07 g/m3 1000 24.5 If the level of concern is given in milligrams per cubic meter, divide it by 1000 to obtain grams per cub meter. If the level is given in milligrams per liter or grams per liter, convert liters to cubic meters by using 1000 liters equivalent to one cubic meter. See Table C-l for a summary of conversion factors. TABLE C-l CONVERSION FACTORS mg/m3 x 10 5 = mg/liter (mg/liter) x (24,5/molecular weight) = ppm 1000 tag 1 gram VVv 00940 * * * September 23, 1985 Draft -- Do Not Quote or Cite * * * C-4 Once the level' of concern in grams per cubic meter and the distance in feet are known, use the graph to determine the quantity. Using the left-hand scale on the graph, mark the proper value for the level of concern for the chemical substance you are considering. For example, chlorine should have a mark at 0.07. Next, move to the center scale and mark the distance from the source of release to the site fence line or to a particular target of concern such as a hospital or nursing home. Using a ruler or other straight-edge, draw a straight line from the mark on the left scale through the mark on the center scale to the scale on the right. Where the line hits the scale on the right is the quantity of substance that would result in the concentration on the left scale at the distance on the center scale if it were all released. For example, using chlorine, if a distance of 200 feet were used, the quantity given on the right scale would be about 10 pounds. In other words, if 10 pounds of chlorine were released under poor meteorological and wind conditions, harmful levels of chlorine would occur 200 feet away. The procedures and example given above are appropriate for vapors released from gases or volatile liquids. If the acutely toxic chemical substance being evaluated is a dust or powder, the same nomogram procedure can be used assuming that fine dusts or powders behave like vapors in air. The level of concern may need adjustment, however. Solids are often tested for lethality on laboratory animals by administering the substance orally. The results are given as the lethal dose for 50 percent of the test animals or LD50 in units of milligrams per kilogram (rg/kg) of body weight. If the test results already give an- 1DLK or include an LC value, use the procedures above. If only an LD value is given for an oral test, then it must be converted to an inhalation value for use on the quantity determination graph. The National Research Committee has developed a simple method for determining the inhalation level associated with an LD oral dose. Convert the LD oral dose in milligrams per kilogram (mg/kg) to inhalation in grams per cubic meter (g/a*) as follows: LD oral, mg/kg x 213 = Inhalation, g/mJ Compare the quantity determined from the graph to the quantity stored in any one container or the quantity that could possibly be released from a process on the site. If the quantity is much higher than the quantity determined from the graph, the work group should continue gathering information for contingency planning for this substance and site. If the quantity on the site is less than the quantity determined from the graph, the work group may want to defer further data gathering efforts for this substar.ee at the site, depending on the kinds of operations at the site involving the acutely toxic chemical substance. VVV 000022941 * * ' * September 23, 1985 Draft -- Do Not Quote or Cite * * * LbveL . cP C-5iJC^LCV (g/mJ) 10 1.0 dr 0.5 0.2 ) 0.05- 0.02 0.01 0.005' ).002--1-- 0.001 Figure C-l GRAPH FOR QUANTITY DETERMINATION Distance (ft) 10 20 50 100 200 H-500 1,000 '2.000 '5,000 10.000 20.000 50,000 100,000 Quant, (lbs) i.o 10 20 50 100 200 500 1.000 2,000 5,000 10,000 20,000 WV 000022942 50,000 100,000 To: DISTRIBUTION Interoffice Communication From: Date: Subject: M. J. HELLER, Houston, TX. February 23, 1987 CAER MODELING - BASIS VISP CAER scenarios were received from each Lake Charles plant over the past two weeks. Before the actual modeling begins, a basis for modeling must be set. I recommend that the evaluation of each scenario involve: A. stability classes D-day and F, B. wind speeds of 5, 15, and 25 MPH, FtB c . C. an ambient temperature of 70F, D. a mixing depth of 5000 meters. E. rural dispersion coefficients, and F. criteria levels as outlined on page three. -. Co;-, .. ...... . .... - -....... - .... -- Based on the above, only six runs will be needed for each scenario. STABILITY CLASS Atmospheric stability is a measure of turbulence and is classified as A through F with F being very stable. Figure I is a graph of stability class versus time of day for a typical day in January in Lake Charles. This figure shows that the atmosphere is very stable during most of the night and becomes less stable during the day. In fact, categories A through D-day rarely occur during the night and categories D-night through F rarely occur during the day. Figure II shows the frequency of each stability class in Lake Charles for four three-month periods during 1980 and 1981 and also for the entire year. The two most frequent classes are D-day and F. Of the daytime categories (A through D-day), D-day gives the most conservative results and is the most frequent. Of the night time categories (D-night through F), F is most conservative and most frequent. Because of these facts and because it is easier to determine whether it is day or night than to determine a stability class, D-day and F were chosen for modeling. (1) Daytime is defined as one hour after sunrise to one hour before sunset. 009/PED16 vvv 0000Z29*3 CAER MODELING - BASIS February 23, 1987 Page Two WIND SPEED Several runs were made on Che ISCST program to determine the effect of wind speed on dispersion. Wind speed affects how quickly a plume will reach a neighborhood more than how high the steady state concentrations will be in the neighborhood. However, there is enough effect on the steady state concentrations, that wind speed should be considered as a variable. Figure III shows the frequency of several wind speed categories in Lake Charles for four three-month periods and also for the entire year. Category 5 is the most frequent category. To cover the range without modeling every category, I suggest using, five, fifteen, and twenty-five miles per hour for each scenario. TEMPERATURE Two runs were made using the ISCST model to determine the effect of ambient temperature on downwind concentrations. One run was at 40F and the other at 80F. There was no noticable difference in dispersion between the two. One temperature, 70F (average for Lake Charles), should be sufficient for modeling purposes. MIXING DEPTH Mixing depth is the height of the atmosphere available to dilute the pollutant. Pollutants will not normally travel above this height. Mixing depth data cannot be taken in the plant but must be obtained through the nearest national weather service upper air station. The nearest station to Lake Charles is located in Shreveport. For the worst case, the mixing depth should be set equal to the release height. The same quantitative results may be obtained within the model by using a very high mixing depth such as 5000 meters and doubling the resulting concentrations. Since immediate mixing depth information is not available in a control room, and since this information is difficult to predict, I suggest we make the worst case assumptions. DISPERSION COEFFICIENTS The equation for basic dispersion contains dispersion coefficients to describe the extent of dispersion. Separate sets of these coefficients are available for rural and urban areas. Dispersion in urban areas is more rapid than in rural areas. (1) Wind speeds for modeling purposes should be measured at 10 meters above ground level. 009/PEDI6 VVV 000022944 CA.ER MODELING - BASIS February 23, 1987 Page Three According to two EPA tests, the area around the Lake Charles complex is rural. One test is based on population density and the other on land use. Neither depends on zoning. CRITERIA LEVELS Based on the levels in the literature, and making conservative assumptions, Tom Grumbles has proposed the following levels for the initial modelling contours. Benzene Chlorine EtO HC1 HF VCM 25 ppm and 1,000 ppm 1 ppm, 3 ppm, and 25 ppm 1 ppm, 5 ppm, and 20 ppm 5 ppm and 100 ppm 3 ppm and 20 ppm 1 ppm, 5 ppm and 20,000 ppm (1/2 LEL) The levels above will give us a ceiling value contour, a conserva tive IDLH contour, and for carcinogens a range of levels relative to occupational limits. Where a NRC/EEG level exists it will be modeled. Actual levels for notification criteria to the local emergency response commission warrant further discussion. For your information. Table I lists criteria levels from the literature for the chemicals listed above. OTHER CONSIDERATIONS For scenarios which involve the release of more than one chemical simultaneously, multiple overlays will be generated. Each overlay will represent a single chemical with its own criteria levels shown. The effect of being exposed to more than one chemical at a time will be determined by Tom Grumbles as the occasion arises. Based on the scenarios received so far, the model we have now, ISCST, will need to supplemented with a puff model. This puff model will handle very short releases better than the ISCST pro gram. Exactly which puff model to buy will be decided and ordered in the immediate future. After receiving the Puff model, a sepa rate basis letter may be needed for very short releases. 009/PED16 Vvv oooozzo*5 CAER MODELING - BASIS February 23, 1987 Page Four Please let me know if you have comments or questions about this basis by 3/11/87 so that we can proceed with the modeling* Mary Jo Heller /mb File: P-18 Distribution: Sid Pitts Steve Ashby Keith Fogg Paul Warner Gary Foshee George Hopkins Mike Hayes Ed Taylor Brian Davis Tom Grumbles Pat Fetizanan Charlie Dutra Tom Heller 009/PED16 A FIGURE la) STABILITY CLASS VS TIME OF DAY STABILITY CLASS TIME OF DAY NOTE: (1) Figure Denotes a typical January ; ,, t .,i. /"M. -i I-- X >< o QZ a z < QQ O Q Q U L UJ 0 Li i F IG U R E II STABILITY CLASS FREQUENCY TIME OF YEAR i----------------- 1----------------- i------- ----------1----------------- 1----------------- 1----------------- 1----------------- 1 'ot inoon tnaa<n not- tno (%) A0N3fl03yj VVV 000022948 FIGURE III WIND SPEED FREQ UENCY 18.9 2 3 .9 IX X a. X X X a. X Q. Q_ x X riX 2 o 03 n cm X X N t- | | aj Q Z 2 -I < 1 X-- 1 o o x-- Iq CM 0) CM N r- V" A U o Li (%) A0N3nD3HJ VVV 000022949 TABLE X BENZENE CHLORINE ETO HC1 HF VCM CRITERIA LEVELS FROM THE LITERATURE FOR SELECTED CHEMICALS PEL/TLV STEL IDLH^ 10ppm(30mg/m^ ) PEL lOppm TLV Ippm NIOSH 3 lppra(3mg/m ) (ceiling) Ippm (ceiling) PEL TLV 3 lppm(2mg/m ) Ippm PEL TLV 3 5ppm(7mg/m ) (ceiling) 5ppm (ceiling) PEL TLV 3 3ppm(2.5mg/m ) PEL 3ppm TLV (ceiling) 50ppm 25ppm None 3 3ppm(9mg/m ) None None 5ppm(NIOSH) None None 3 6ppm(5mg/m ) None 2 jOOOppm 25ppm 800ppm lQOppm 20ppm 1.0ppm(3mg/m^)PEL 5. Oppra (15mg/nt ) TLV 5ppm None None Found nrc/eeg^2^ 1,000ppm 3 ppm 20ppm - NOTES: (1) IDLH is defined by NIOSH as the 30 minute maximum concentration from which one could escape without escape impairing symptoms or any irreversible health effects. (2) NRC/EEG - These are the levels proposed, or recommended by the National Research Council, committee on Emergency Exposure Guidelines. These are considered as a ceiling limit for a one time exposure, not to exceed 60 minutes. Comments: In general ceiling limits are "not to be exceeded" limits for industrial workers. They would be conservative estimates for the normal population. HF, HC1, and Chlorine are all detectable and potentially irritating at levels of 1/2 to 2 times the PEL/TLV. 009/PED16 vw 000022950