QA/QC of Scottish Air Quality Data

1
QA/QC of Scottish Air Quality Data

• Ken Stevenson

2
Project Overview

• Aim
• To have all of the air quality data collected in
  Scotland in one accessible website
• Duration
• 3-year contract
• Key elements
• Data from National Network and LA sites in
  Scotland
• All quality assured to same standard
• Data available via open access website
• Long-term trend analysis
• Mapping of pollutants throughout Scotland
• Newsletter and Annual Report
• Annual Seminar

3
QA/QC of Scottish sites

• Current Situation
• Total 60 Sites on Scottish website (up from 44 in
  2007)
• 32 sites operated under this project
• 15 AURN
• 13 funded by LAs or Scottish Grant Scheme

4
Additional sites to be added

• 5 sites temporily closed for relocation or
  roadworks
• Fife Cupar
• Edinburgh Haymarket
• Aberdeen Market St
• N. Lanarkshire, Harthill
• N Lanarkshire, Ellis Road
• 4 sites still to join the database
• East Lothian, Mussleburgh North High St In
  progress
• East Renfrewshire, Clarkston In progress
• Shetland Stanley Hill telemetry issues
• North Ayrshire, Irvine New site recently
  installed

This will bring the total to 69 sites on the
Scottish Database.
5
QA/QC

• Daily collection of data from each site
• Scaling of data from last calibration results
• Daily checking of data
• On-going data update as new information becomes
  available
• Site audits
• Collection and storage of all calibrations and
  service records etc
• 6-monthly data ratification
• Quality Circle data review
• Update of ratified data to the website
• Assistance to LAs where-ever possible

6
Site Audits 2008/9
7
6-monthly site intercalibration and audit

• AEA visits every site every 6-months to undertake
  the intercalibration and site audit
• Check Analysers
• Check Calibration Gases
• Site Infrastructure
• Local Operator

0401
8
Problems found at 2008/9 audits

• Cylinders- 4 NO cylinders
• 1 SO2 Cylinder
• PM Analysers- 2 Ko out by more than 2.5 (2.8
  4.8)
• 5 Main flows out (12 20)
• 6 Leak test failed
• 1 Aux flow out (26)
• Gas Analysers 1 CO analyser fail leak test

9
Feedback from site visits
10
Data Ratification
11
Example of Data Summary Sheet
12
Measuring PM

• PM10
• Equivalent Analysers
• Partisol (with emfab filters)
• Beta (Met One) unheated, with correction factor
• Beta (Opsis SM200) with correction factor
• TEOM FDMS (not with C drier)
• (TEOM with Volatile Correction Model)

13
Partisol

• Automated sampling device - size-selected
  particulate matter on filters for later analysis.
  
• Unattended operation for up to two weeks between
  site visits.
• Filters are equilibrated, at 20 C and 50 R.H.,
  for 48 hours - before and again after sampling.
• Filters are carefully weighed using precision
  microbalances.
• Concentration is a function of air volume sampled
  and weight PM of material on filter.

14
Scottish Government Partisol Monitoring Programme
2007

• PM10 and PM2.5 at 5 sites for 1 year
• Data problems common to all UK Partisol
  measurements
• High field blank due to moisture
• Worst with Quartz filters even though these are
  specified in EU PM10 standard
• Data now corrected and reported
• Data used for PM Mapping in Scotland
• Changed to Emfab filters in AURN

15
BAM (Beta Attenuation Monitor)

• Size-selected particulate matter on filter tape
• Analysed by attenuation of Beta rays from
  Carbon-14 source
• Zero from clean filter compared to soiled filter
• Algorithm to calculate mass from attenuation
• Correction factor may be required for equivalence

16
FDMS
17
FDMS Operation Issues

• Filter Changing
• Possible long-term problems with FDMS C system
• Use B or CB type drier

18
PM10 Monitoring in the Scottish Database

• 45 monitors in total
• FDMS 12
• BAM 4
• TEOM 29

19
Conclusions

• QA/QC work is key to everything
• Audits and Ratification are a major part of the
  project work
• PM - Partisol issues resolved
• FDMS and BAM Equivalent PM analysers being
  deployed
• We strive to support and work closely with LAs
• respond to requests and enquiries
• offer advice
• feedback if we see problems with the data
• chase for calibrations
• The database is building rapidly
• 2006 20 sites
• 2007 44 sites
• 2008 60 sites