250 likes | 283 Views
Learn about EPA's new gas quality assurance system, calculation changes, and data assessment tools to optimize resource use and ensure accurate air quality data.
E N D
Precursor Gas Quality Assurance Implementation Dennis K. Mikel EPA Office of Air Quality Planning and Standards
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
What is a Quality System ? EPA A structured and documented management system that describes how and by whom an organization assures quality in its work. Your Program Quality System Goals • Make correct decisions • Optimize resource use The Project!
Reporting Where are we?
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
Gas Instrument QA Requirements* * Table A-2 is from 40 CFR 58, Appendix A
Nationally, 97% of all ozone monitors report Coefficient Variance of less than 7%. MQOs for Gas Pollutants Has Changed • Ozone: • Coefficient of Variation (CV): 7% • Bias: +/- 7% • SO2, NO2, CO: • CV: 10% • Bias: +/-10% • For PM: • No Changes
“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
“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
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 Adequacy
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.
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)
Precision and Bias : Monitor value : Target concentration (standard) : Percent Difference (individual bias) Estimates start the same way & both use the 1-point QC check
* Where is the 10th percentile of a Chi-Squared Distribution Precision “A measure of mutual agreement among individual measurements of the same property, usually under prescribed similar conditions” This represents a 90% upper confidence limit on the CV estimate
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 95th 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
P&B Guidance and Data Assessment Statistical Calculator (DASC) Software
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
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
DQO Table for Precursor Gas ParametersCarbon Monoxide Urban Site
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?
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 1st quarter 2007
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
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