Precursor Gas Quality Assurance Implementation Dennis K Mikel

Precursor Gas Quality Assurance Implementation Dennis K. Mikel EPA Office of Air Quality Planning and Standards EPA Precursor Gas Training Workshop

Outline • QA System Development • QA Requirements for the NCore Program – Changes to the 40 CFR Appendix A – Overview of the New Calculations • Data Assessment Statistical Calculator (DASC) • Comparing the DQO/MQOs to the Burden’s Creek data • National Performance Audit Program 2 EPA Precursor Gas Training Workshop

EPA Your Program What is a Quality System ? A structured and documented management system that describes how and by whom an organization assures quality in its work. Quality System Goals – – Make correct decisions Optimize resource use The Project! 3 EPA Precursor Gas Training Workshop

Where are we? Reporting 4 EPA Precursor Gas Training Workshop

Changes in the QA Regulations (40 CFR Part 58 Appendix A) • Ensured regs reflected current EPA QA policies and requirements • Combined Appendix A and B (PSD) • Institute the Primary Quality Assurance Organization (PQAO) to help lower the QA burden on SLT agencies and provide adequate level of independence • Measurement quality objectives (MQOs) changing for gaseous pollutants 5 EPA Precursor Gas Training Workshop

Gas Instrument QA Requirements* * Table A-2 is from 40 CFR 58, Appendix A EPA Precursor Gas Training Workshop 6

MQOs for Gas Pollutants Has Changed • Ozone: – Coefficient of Variation (CV): 7% – Bias: +/- 7% • SO 2, NO 2, CO: – CV: 10% – Bias: +/-10% Nationally, 97% of all ozone monitors report Coefficient Variance of less than 7%. • For PM: – No Changes EPA Precursor Gas Training Workshop 7

“Reporting Organization” vs. “Primary Quality Assurance Organization” • What’s the Difference? – RO: “an entity, such as State, local or Tribal monitoring organization that collects and reports air quality data to EPA. ” – PQAO: Monitoring Organization or “other” organization that is responsible for a set of stations that monitor the same pollutant and which data quality assessments can be pooled. ” • PQAOs are responsible for implementing QA EPA Precursor Gas Training Workshop 8

“Reporting Organization” vs. “Primary Quality Assurance Organization” (cont. ) • PQAO: – Measurement uncertainty “reasonably homogeneous” – Common team of field operators – Use of a common QAPP and SOPs – Oversight of QA functions – In some cases, RO will equal PQAO • Must demonstrate both Adequacy and Independence 9 EPA Precursor Gas Training Workshop

Adequacy • Performing audits at 20% of monitoring sites/instruments • Audit data submission to AQS • TTP Delivery system • Follow NPAP field/lab SOP critical performance criteria • Audit gases that are NIST certified and validated at least once a year • Validation/certification with the EPA NPAP program • Incorporated in QAPP 10 EPA Precursor Gas Training Workshop

Independence • No part of the organization is directly performing and accountable for the work being assessed (i. e. , PQAO) • Management structure that allow for the separation of its routine sampling personnel from its auditing personnel by two levels of management • Submission of a plan demonstrating independence to the EPA Regional Office. 11 EPA Precursor Gas Training Workshop

New Statistics in 40 CFR 58 Appendix A • Focus Workgroup of the QA Strategy team formed to review and revise the old statistics in Appendix A • Proposed that the stats be based on confidence intervals to meet the MQOs • This approach allows flexibility for analyst and agencies to tighten or loosen their test if they meet criteria • Allows analysts to look at “pooled” results rather than individual or small sets of data for which the old stats were designed • For gases, both the precision and bias calculations use the bi-weekly QC 1 -point check (i. e. , Precision checks) 12 EPA Precursor Gas Training Workshop

Precision and Bias Estimates start the same way & both use the 1 -point QC check : Monitor value : Target concentration (standard) : Percent Difference (individual bias) 13 EPA Precursor Gas Training Workshop

Precision “A measure of mutual agreement among individual measurements of the same property, usually under prescribed similar conditions” * Where is the 10 th percentile of a Chi-Squared Distribution This represents a 90% upper confidence limit on the CV estimate EPA Precursor Gas Training Workshop 14

Bias “A systematic or persistent distortion of a measurement process which causes errors in one direction” Absolute Bias Point Estimate: Absolute Bias Upper Bound: Where t 0. 95, (n-1) is the 95 th quantile of a Student’s t distribution with n-1 df and sd_abs is the standard deviation of the absolute value of the relative percent differences This represents a 95% upper confidence limit on the Bias estimate 15 EPA Precursor Gas Training Workshop

P&B Guidance and Data Assessment Statistical Calculator (DASC) Software 16 EPA Precursor Gas Training Workshop

AMP 255 Report – Monitoring Orgs. Can run this as often as they wish – OAQPS will run annually – Box–and-whisker plots included in annual summary http: //www. epa. gov/ttn/amtic/parslist. html EPA Precursor Gas Training Workshop 17

Data Quality Objectives for the Precursor Gases • DQOs are used to create quality parameters so that the data collected is of sufficient quality for its use • OAQPS staff began the DQO Process 1. 5 years ago • OAQPS staff employed Battelle Inc. to run simulation models to examine how the different QA indicators would affect data quality • From these the Measurement Quality Objectives are defined 18 EPA Precursor Gas Training Workshop

DQO Table for Precursor Gas Parameters Carbon Monoxide Urban Site Power to Detect Annual Trend at a Single Site Estimated Fraction of Data Falling in Prediction Interval Power to Detect Annual Trend Across 50 Sites Annual Completeness Hourly Measurement Precision (CV) Site. Specific Bias 5% Reduction 10% Reduction Multiplicative 1/2 Width for 95% Prediction Interval at an Individual Site 90% 7%-20% 5%-20% 75% 100% 9. 6 27% 70% 75% 7%-20% 5%-20% 71% 100% 16. 1 40% 75% Lower Bound Upper Bound 19 EPA Precursor Gas Training Workshop

Burden’s Creek Station • AAMG staff have been operating PG/trace gas instruments at BC for 2 years • We have been operating like an NCore Station • Since we perform 1 -point QC tests, as described in 40 CFR 58, we can assess the data • How do our results compare to the Battelle model? 20 EPA Precursor Gas Training Workshop

Burden’s Creek MQOs (May 06 – April 07) Coefficient of Variance Bias Data Completenes s O 3 API 2. 1 +2. 1 91. 1% CO–API 6. 9 +9. 4 76. 7% CO-TEI 5. 8 +/-5. 4 90. 1% NO-API 2. 8 -4. 7 84. 1% NOy-API 3. 3 -5. 7 84. 1% SO 2 -TEI 3. 6 +/-3. 0 92. 5% 21 EPA Precursor Gas Training Workshop

NPAP for Precursor Gas Sites • Precursor monitoring will be part of NPAP • Monitoring organization are responsible for implementing adequate and independent audits – Allows for continued Federal implementation with STAG funds • NPAP program testing TTP capabilities for precursor gasses at Burdens Creek – Some equipment still being purchased – Expected testing in 1 st quarter 2007 22 EPA Precursor Gas Training Workshop

Expansion of Audit Levels • Low levels added to accommodate Trace Level Instruments • At Least three consecutive audit levels • The audit levels selected should represent or bracket 80 percent of ambient concentrations 23 EPA Precursor Gas Training Workshop

Summary • 40 CFR 58 Appendix A regulations have changed • PQAOs vs. ROs • There are new statistical calculations – New Tools to assist EPA and stakeholders • EPA has performed a DQO modeling exercise • The BC data illustrates that we can collect data with the MQOs • NPAP TTP audit vehicles are now gearing up 24 EPA Precursor Gas Training Workshop