Document JJ6pZwYwkr3MNQ19wqdm1KkvO

AR226-2346 'WSi9fvasay Air Dispersion Modeling Verification Study Oct 22,2003 Objective Compare results from ambient air sampli air dispersion modeling to understand ho modeling predictions relate to measured air concentrations Verification Study Scope Air Sampling Program Analytical Methods Modeling Analysis of Results Air Sampling Locations Simultaneous coverage of all 4 major directions (N, S, E, W) Capture point of predicted maximum concentration from modeling Include co-located samplers at predic maximum Total of six locations ASH026725 EXPQ15794 ASH026726 EXP015795 Air Sampling Strategy Measure ambient air levels i 24-hour sampling duration Seven (7) sampling events scheduled ove weeks Vary day of week to capture variations in operations (none expected) and variation meteorology Record site-specific meteorological and operational data ^ Monitoring Strategy: Samples taken every 6 days f MONTH 1 QZ 345 6 8 9 10 11 12(^14 15 16 17 18^9)2021 22 23 2^5)26 27 28 29 3( MONTH 2 1234 5^)7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 Monitoring Strategy: Samples taken every 6 days f On site training For internal use: potential schedule Oct/Nov S M T W Th F S -28 2931 1 678 2 3 4^^ 9 10^^12 13 14 If 16 <Ql8 19 20 21^ ^5)24 25 26 if^n^ 30 December 1 2 3 4Q)6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 Total Samples (1) field blank per event (5) primary samples per event (1) duplicate sample per event* (7) samples x (7) sampling events = 49 tota * Sample location 3 is co-located with location 2 and se as a duplicate sample Equipment "Sorbent" sampling method to captur PFOA on ion-exchange media OSHA Versatile Sampling tube (OVS Captures particles and vapors Remote power source will be provide Weatherproof stations for sampling equipment OVS Tube Air flow (1 L/min) Analytical Method Validated method being used Acetone extraction followed by liquid chromatography/mass spectrometry (LC/MS) Detection limit of 0.1 ug/m3 Modeling Program Using U.S. EPA ISCST3 Model Model will simulate actual field conditions - Site-specific surface meteorological data (win direction, temp, stability) - Off-site upper air meteorological data - Emissions data Site specific operational data will be used to relate data to emissions levels using "capacity factor" Capacity factor based on prior stack testing and m - 24-hour averaging time - Predictions at actual height and location of ai intake Comparison of Results Comparing 35 pairs of measured and mod - (5) primary sampling points x (7) sampling e Will use statistical tools to evaluate correlation/differences between measured modeled data ^i^is^ ^ Internal? Questions we are trying to an Are modeled and measured data wellcorrelated? (+/- X%) What is the probability or confidence that the model predictions will be equ or greater than the measured value? Does model "accuracy" change at diff concentration levels? Statistical Analyses Plot of measured vs. modeled data to test lin correlation Statistical goodness of fit criteria (R-square Simple sign test for residuals - Calculate (measured - modeled) - Trend in residuals: positive, negative, neutral Plot of residuals vs. modeled data to test go fit at different concentrations Test specific hypotheses; examples - Model predicts within 20% of measured value. - Model over predicts 50% of the time. Project Timing Mid-October Analytical Method Development Site Mobilization Weeks 1 - 6 Weeks 2 - 9 Monitoring/Data Collection I. Laboratory Analysis i Met Data Processing .WsakaJ-dLl-.JData-Quality Review ISC3 Modeling L Comparison of modeling and monitoring results Week 12 Presentation of Results and Recommendation to Business I Weeks 1 Documentation remaining internal question Are there site communications we nee think about prior to sampling? EXTRA SLIDE Feasibility of Deposition Mod Wet and dry deposition rat^s are small - ma rate 0.18 gram/m^yr for year 2000 Direct field measurement on 24-hr basis immeasurable, so direct comparison with sh term modeling not feasible Use model verification to add degree of confidence EXTRA SLIDE Deposition Modeling Both wet and dry deposition algorithms depend on the ability of the model to calc ambient air concentrations - Dry deposition flux = Ambient hourly concentration x deposi velocity i - Wet deposition flux = ^scavenging coefficient x ambient con integrated in the vertical) Verification of ambient air model will he I confirm basis for deposition calculations M . CTl ^1 ^