Model For Relating PM 2.5 Federal Reference Method & Nephelometer Measurements Presented by Jenny Parikh
12/6/ EPA AQI Conf. 2/3-5/02 Agenda What was done Value to Public How the model is developed What’s Next
12/6/ EPA AQI Conf. 2/3-5/02 GOAL Develop Statistical Linear regression model relating PM2.5 FRM and Nephelometers (continuous monitors)
12/6/ EPA AQI Conf. 2/3-5/02 WHY n FRM is used for NAAQS compliance (FILTER based) 1999-to date PM 2.5 measurement are below NAAQS Continuous Monitor data is available n Temporally resolved data : Diurnal cycles n Need to estimate PM 2.5 in real time (AQI) n Push from EPA to complete Real time network n QA TOOL
12/6/ EPA AQI Conf. 2/3-5/02 VALUE TO PUBLIC n Ability to represent unhealthy air conditions while it is occurring n Part of 40CFR requires all MSAs with population of 350,000 or greater to report daily AQI
12/6/ EPA AQI Conf. 2/3-5/02 METHODOLOGY n Determine required environmental measurements, suitable data and data reduction n Perform exploratory data analyses Histogram Scatter plots and outliers Time series analyses n Determine the “best” model n Evaluate the ability of the model to estimate PM2.5
12/6/ EPA AQI Conf. 2/3-5/02 Guidance Document
12/6/ EPA AQI Conf. 2/3-5/02 Guidance document n /aqidqorept.pdf /aqidqorept.pdf
12/6/ EPA AQI Conf. 2/3-5/02 Nephelometer Monitoring Sites Nephelometer Sites 2005
12/6/ EPA AQI Conf. 2/3-5/02 MODEL DEVELOPMENT n Eastern Sites Chewelah, Coleville, Pullman, Ritzville, Walla Walla, Kennewick, Yakima, Twisp, Moses Lake…… n Western Sites Longview, Lacey, North Bend, Yew Street, Beacon Hill, Redmond, Bellevue Goldfarb, Vancouver…
12/6/ EPA AQI Conf. 2/3-5/02 RESULTS: Chewelah
12/6/ EPA AQI Conf. 2/3-5/02 Next Step n Strong FRM- CM relationship indicated it was worth considering a comparison of CM and FRM that are not collocated
12/6/ EPA AQI Conf. 2/3-5/02 How the E. WA model was developed n FRM and Neph Data for the sites started from 04/01 till todate, about 75 days of data 44 days where all five sites had data
12/6/ EPA AQI Conf. 2/3-5/02 Comparison of Nephelometers
12/6/ EPA AQI Conf. 2/3-5/02 Comparison of FRMs from all E. WA sites
12/6/ EPA AQI Conf. 2/3-5/02 Analysis of Data n Strong CM-FRM relationship was observed n Compared CM and FRM data that are not co-located n Strong relationship indicated a possibility of developing a model using the average of the set of FRM measurements with CM measurements n This also gives measurements that are better representative of the overall region
12/6/ EPA AQI Conf. 2/3-5/02 Colville and E. Washington Model
12/6/ EPA AQI Conf. 2/3-5/02 Pullman and E. WA Model
12/6/ EPA AQI Conf. 2/3-5/02 Eastern Washington Model
12/6/ EPA AQI Conf. 2/3-5/02 Ritzville and E. Washington Model
12/6/ EPA AQI Conf. 2/3-5/02 WallaWalla and E.WA Model
12/6/ EPA AQI Conf. 2/3-5/02 Models n All Eastern Sites PM2.5 correlated = 10 ( (logBscat) ) This model is applied to Lacrosse, Rosalia, Starbucks, Mesa) n Final NorthBend PM2.5 correlated = 10 ( (log Bscat) ) n Previous North BendPM2.5 correlated = 10 ( (log Bscat) ) n Previous Beacon Hill PM2.5 correlated = 10 ( (log Bscat) ) ( R 2 = 0.84) n New Beacon Hill PM2.5 correlated = 10 ( (log Bscat) ) ( R 2 = 0.90) n NWAPA Model PM2.5 correlated = 10 ( (log Bscat) ) This is applied to Mt. Vernon
12/6/ EPA AQI Conf. 2/3-5/02 IMPLICATIONS n Valuable tool for Eastern WA – covers large geographical area with Nephelometer n Divest from PM2.5 FRM network- optimization and modification n Resource allocations Workforce demands Monitor costs
12/6/ EPA AQI Conf. 2/3-5/02 WHAT’S NEXT n Refine and update with additional data for Western Washington Model Quarterly / Seasonal basis n Additional sites e.g; Stevenson (Bonneville, Mt. Zion, Wishram), Leavenworth
12/6/ EPA AQI Conf. 2/3-5/02 Nephelometer Monitoring Sites Nephelometer Sites w/correlated PM2.5 Nephelometer Sites 2005
12/6/ EPA AQI Conf. 2/3-5/02 Nephelometer/Correlated PM 2.5 n Most of our nephelometer site data has been analyzed to create a correlation to PM 2.5. n Those sites with a PM 2.5 correlation are shown on the site operators’ page AND the public page as NPM