Document 3Q3OQ1kvrXqg23yvEL7oe06kO

AR226-2361 September 2002 Throdgb August 2003 A ir Dispersion Modeling Analysis o f APFO Emissions (Revised) May 24,2004 Page 1 of 9 SEPTEMBER 2002 THROUGH AUGUST 2003 AIR DISPERSION MODELING ANALYSIS OF APFO EMISSIONS -------------------------------------- (Revised)-------------------------------- -- -- DuPont Washington Works Facility Parkersburg, West Virginia Prepared for: West Virginia Department of Environmental Protection Division o f Air Quality 7012 MacCorkle Ave, SE Charleston, WV 25304-2943 Prepared by: DuPont Engineering Technology (DuET) Environmental Section Wilmington, DE 19898 C:\c8\c8_0803_report_revised May 24,2004 I E ID 775177 EID 775177 September 2002 Through August 2003 Air Dispersion Modeling Analysis o f APFO Emissions (Revised) May 24,2004 Page 2 o f9 1. Introduction DuPont conducted a it dispersion m odeling o f APFO* em issions from its W ashington W orks facility 'located finff PaHrefsSurg] W V T tfciffiB ^V M itiitE ic63to predict long-term am bient air concentrations o f APFO resulting from actual plant emissions that occurred during the period o f September, 2002 through August, 2003. This report describes the APFO em issions inventory used in the modeling analysis, the meteorological data, the dispersion model and m odeling procedures, prediction locations (receptor grid), and the results o f the modeling analysis. Compared to the modeling report subm itted on October 17,2003, this revised report incorporates several revisions to stack param eters that were the result o f stack testing, and the refinem ent o f UTM coordinates. 2. Em issions Inventory The following emission inventory information has been assembled in order to conduct the air quality modeling: 1. Stack locations 2. Stack heights 3. Stack diameters 4. Stack gas exit temperatures 5. Stack gas flow rate or exit velocities 6. Detailed plant layout, including all building dimensions 7. Sept 1,2002 Aug. 31,2003 estimated actual APFO emissions All o f the stack parameters are presented in Table 1, which shows the source representation for modeling purposes. The estimated actual emission rates o f APFO, per source, are also presented in Table 1. Figure 1 presents the general locations o f the APFO sources. 3. M eteorological D ata One year o f on-site meteorological data for the calendar year 1996 was used in this study. Concurrent twice-daily upper air data from the upper air observation station located in W ilmington, OH was used along with on-site surface temperatures to obtain hourly mixing depths. M issing data and measured wind speeds of less than 1.0 m /s were treated consistent with the recommendations made in the F.PA's "Meteorological M onitoring Guidance for Regulatory Modeling A pplications(I). An anemometer height o f 10 meters was used for the modeling analysis E1D775178 *"APFO" means ammonium perfluorooctanoate, and for the purposes o f this report includes the anion o f the acid periluorooctanoic acid (PFOA). C:\c8\e8_08O3jreport_revisetl 2 EID 775178 , 4. M odel Selection Septem ber 2002 Through August 2003 A ir Dispersion M odeling Analysis of APFQ Emissions (Revised) May 24,2004 Page 3 of9 The area surrounding W ashington W orks is primarily non-urban. The U. S. EPA procedures classify land use within 3 kilometers o f the site by the Auer method. Previous review o f U. S. Geological Survey (USGS) maps, aerial photographs, and site visits clearly indicated that the area is well over 50% n o n -u rb an -T l significant terrain features on both sides o f this river valley. As a result, terrain elevations were considered in the modeling analysis. The Industrial Source Complex Short Term Model (ISCST3) was used as the primary model to estimate long-term pollutant concentrations. ISCST3 is a steady-state Gaussian model recommended by the U.S. EPA. It is included in the "Guideline on A ir Quality Models'1, which is codified as Appendix W to 40 CFR Part 51. It is appropriate for modeling o f pollutant emissions from multiple, industrial-type sources subject to significant building downwash. The downwash algorithms in the ISCST3 model provide a representation o f the aerodynamic downwash o f a stack plume caused by complex building configurations typical o f industrial facilities. Refined ISCST3 modeling was conducted using one year (1996) of sequential hourly meteorology from the on-site observation facility, as described above. 5. R eceptor Selection A Cartesian grid of receptors was utilized in this modeling analysis. This grid consisted o f the following: Fenceline receptors with a 100 m spacing between receptors Receptors beyond the fenceline w ith 100 m spacing on a 5 km by 7 km grid All receptors are located along or outside the plant fenceline. A Cartesian receptor grid o f this type is considerably more dense than recommended by the U.S. EPA in the Guidelines on Air Quality Models (U.S. EPA, 1998) for modeling a facility o f this type. Terrain elevations for each o f the receptors were imported from electronic files obtained from the U .S . Geological Survey (USGS) using the "highest" method to assign an elevation to each receptor. The receptor grid used in the modeling analysis is shown graphically in Figure 2. 6. M odeling Procedures The most recent version o f ISCST3 (version 02035) was used in the air quality dispersion modeling o f all receptors. All model options were set to the U.S. EPA regulatory default version o f ISCST3. The model was run in the rural mode since the land area in the immediate vicinity o f Washington Works is more than 50% rural. Any effects o f aerodynamic downwash caused by structures adjacent to each modeled stack were included in the ISCST3 modeling analysis along with a summary o f the building downwash input files (BPIP). A ir quality dispersion modeling vras conducted on an hour-by-hour basis using the one year o f meteorological data described above. The APFO modeling results were summarized for the annual averaging time period. C:\c8\c8_0803_repoit_revised 3 E ID 775179 EID 775179 ** 7. Results Septem ber 2002 Through August 2003 Air Dispersion Modeling Analysis of At*FO Emissions (Revised) May 24,2004 Tage 4 o f9 The results o f the modeling analysis indicate a maximum predicted annual average APFO concentration o f 0.70 ug/in3. This maximum is located along the northern properly feneeline, dong the Ohio River, at UTM 442043 fi, 4346883 N . The maximum predicted APFO concentration in an area where people may fesideis 0.1? tlg/m3. Thisprediction is located a t UTM: 442600 , 4347600 N , on th e Ohio side ofthe river. The results are presented graphically in Figure 3. C :\c8\c8_0803_report_revised 4 E ID 775180 EID 775180 Septem ber 2002 Through August 2003 Air Dispersion Modeling Analysis of APPO Emissions (Revised) t May 24,2004 Page 5 of 9 Vent ID 699 697 694 658 652 231 232 242 274 268 276 T able 1 *** PO lRT SOURCE DATA ** * I Stack Diameter UTM- 442098 442128 442101 441928 441926 441953 441952 441945 441787 441774 441842 UTM-N 4346843 4346829 4346815 4346757 4346758 4346766 4346776 4346746 4346744 4346753 4346772 ft 4 2.25 1.67 1.5 0.88 0.67 0.67 0.5 0.65 0.27 1.5 Stack Stack Stack Height Flow Velocity ft 170 45 45 63 64 92 99 114.5 110 72.5 75 ACFM 12,000 2,000 344 6,478 4,031 510 710 1,048 718 100 5,000 ft/sec 15.9 8.4 2.6 61.1 111.7 24.4 33.9 89.0 36.6 28.7 47.2 Stack Temp F 124 176 112 142 139 148 128 117 163 110 amb Actual C8 Emissions ib/yr 1,463 ...." 0.7 1.0 55 30 1,950 1,975 537 860 35 0.16 ActT C8 Emissions ib/hr 0.1670 0.0001 0.0001 0.0063 0.0034 0.2227 0.2254 0.0613 0.0982 0.0040 0.000018 R022EEF6 R022EEF86 R022EEF87 R022EEF89 442086 442069 442058 442063 4346624 4346627 4346634 4346635 2.5 2 2 2 47 8836 49 7540 49 1885 49 3770 80 12 80 0.3 80 3 80 0.6 0.00045 0.00045 0.00045 0.00045 C:\c8\c8_0803_report_revised 5 E1D775181 EID 775181 September 2002 Through August 2003 Air Dispersion Modeling Analysis of APFO Emissions (Revised) May 24,2004 Page 6 of 9 Figure 1 Sottrceatid BuildingLocations. Btitf C:\c8\c8_0803_report_revised EID775182 EID 775182 4249000 September 2002 Through August 2003 Air Dispersion Modeling Analysis of A PFO Emissions (Revised) May 24,2004 Page 7 o f9 Figure 2 Receptor Grid Used in the M odeling Analysis m eters 434800*$P +-K-4V'2 43470004346000 434SOOO- C:\c8\c 8_0803_rcport_reviscd EID 775183 E ID 775X 83 ' ' '' . ( September 2002 Through August 2003 A ir Dispersion Modeling Analysis of APFO Emissions (Revised) May 24,2004 Page 8 of 9 Figure 3 A nnual A verage C oncentrations (ug/m 3) C ontour In terv al 0.1 ug/m 3 E ID 775184 C:\c8\c8_QB03_report_reviscd EID 775184 Septem ber2002 Through August 2003 Air Dispersion Modeling Analysis ofA PFO Emissions (Revised) May 24,2004 Page 9 of 9 References 454/R-99-005, Office o f A ir Quality Planning and Standards, February, 2000. ---------- (2) A uer, A. H., "Correlation o f Land U se Cover w ith M eteorological Anomalies", Journal o f Applied M eteorology, Vol. 17, pp. 636-643,1978. (3) U . S. p p a niiiHplinp on Air Q uality Models (Revised), EPA-450/2-78-027R-C, 2001. C:\e8\c8_0803__report_rcvised EID 775185 EID 775185