procedures for qa/qc on the air quality monitoring data

Libor Černikovský, Blanka Krejčí Czech Hydrometeorological Institute, CZ

http://www.chmi.cz [email protected], [email protected]

Procedures for QA/QCon the air quality monitoring data

2

Content

Procedures for AQ/QC on the air quality monitoring data

•QA/QC definitionobjectiveplan

•Legislation•Data quality objectives (DQO)•Collecting and reporting data•Examples

3

Definition


Quality assurance (QA): all planned and systematic actions necessary to provide adequate confidence that a product, process or service will satisfy given requirements for quality

Quality control (QC): operational techniques and activities that are used to fulfil given requirements for qualityISO Publications, ISO 8402, Quality Management and quality assurance - Vocabulary,

International Organization for Standardization,1994

In relation to air quality network operation:Quality assurance refers to the overall management

of the process involved in obtaining the data,i.e. relates to the measurement process

Quality control refers to the activities undertaken to check and optimise data accuracy and precision after collection,

i.e. concerned primarily with outputs

4

Objective = usable data


The measured data must allow an objective and quantitative assessment of AQ, whether it complies with relevant national and European standards, limit values, guidelines and other rules.

All tools used for AQ data acquisition, processing and evaluation has to generate data which is acceptable by the national authority, European Commission and European Environmental Agency in terms of quantity, quality and format in which is data stored, presented and transmitted.

5

QA/QC plan 1/3


A document that• shall specify all the QA/QC activities required

to achieve the data quality objectives (DQOs)• should describe how the data is assessed for

precision, accuracy, representativeness, completeness (combined data capture and time coverage) and comparability

• mechanisms used when corrective actions are necessary

The QA/QC plan should assure that• the quality of the data is known• the total measuring uncertainty can be

quantifiedand is available to users of the data

6

QA/QC plan 2/3


The QA/QC plan of the AQ monitoring network must explicitly define

• the unambiguous responsibility and authority for each of the activities contributing to the data quality

• co-ordination between them

Key elements are• organizational rules• operational rules• appropriate staff

7

QA/QC plan 3/3


Measured data QA/QC has to be guaranteed by • methods of measurements• standard operating procedures (SOPs)• hardware and software tools• maintenance, calibration and emergency plans• appropriate staff in terms of quality and quantity• personnel training and education

Network design, station siting and instrument selection are crucial before the monitoring launch!

Correct sampling is crucial, as well as storage and transport of samples...

8

Legislation


… QA/QC plan shall specify all the QA/QC activities required to achieve the data quality objectives (DQOs)

The system for acquisition, processing, evaluation and reporting AQ data has to be in accordance with the EU legislation on AQ as well as with EU standards, regulations and existing guidelines, i.e. primarily with• New Air quality directive 2008/50/EC• AQ Framework Directive (FWD)• Daughter Directives (DD 1-4)• Exchange of Information Decision 97/101/EC• Commission Decision 2004/461/EC (annual reporting

on ambient air quality assessment)

Directive On ambient AQ and cleaner air for Europe (AQD)

see http://ec.europa.eu/environment/air

9

New Air quality directive 2008/50/EC


… includes the following key elements:The merging of most of existing legislation into a single

directive (except for the fourth daughter directive) with no change to existing air quality objectives*

New air quality objectives for PM2.5 (fine particles) including the limit value and exposure related objectives – exposure concentration obligation and exposure reduction target

The possibility to discount natural sources of pollution when assessing compliance against limit values

The possibility for time extensions of three years (PM10) or up to five years (NO2, benzene) for complying with limit values, based on conditions and the assessment by the European Commission.

* Framework Directive 96/62/EC, 1-3 daughter Directives 1999/30/EC, 2000/69/EC, 2002/3/EC, and Decision on Exchange of Information 97/101/EC.

10

Data quality objectives (AQD, Annex 1) 1/4


Specify• uncertainty• minimum data

capture• minimum time

coverage for each pollutant

covered by the Directive

Apply to individual analysers and samplers at individual station

11



Notes:

1.Member states may apply random measurements instead of continuous measurements for benzene, lead and particulate matter if they can demonstrate to the Commission that the uncertainty, including the uncertainty due to random sampling, meets the quality objective of 25 % and the time coverage is still larger than the minimum time coverage for indicative measurements. Random sampling must be evenly distributed over the year in order to avoid skewing of results. The uncertainty due to random sampling may be determined by the procedure laid down in ISO 11222 (2002) "Air Quality – Determination of the Uncertainty of the Time Average of Air Quality Measurements". If random measurements are used to assess the requirements of the PM10 limit value, the 90.4 percentile (to be lower than or equal to 50 µg/m³) should be evaluated instead of the number of exceedances, which is highly influenced by data coverage.

2.Distributed over the year to be representative of various conditions for climate and traffic.

3.One day's measurement a week at random, evenly distributed over the year, or 8 weeks evenly distributed over the year.

4.One measurement a week at random, evenly distributed over the year, or 8 weeks evenly distributed over the year.

12



The uncertainty (expressed at a 95 % confidence level) of the assessment methods will be evaluated in accordance with the principles of the CEN Guide to the Expression of Uncertainty in Measurement (ENV 13005-1999), the methodology of ISO 5725:1994 and the guidance provided in the CEN report "Air Quality – Approach to Uncertainty Estimation for Ambient Air Reference Measurement Methods" (CR 14377:2002E)...

The uncertainty for modelling ...The uncertainty for objective estimation is defined as the

maximum deviation of the measured and calculated concentration levels, over the period considered, by the limit value (or target value in the case of ozone), without taking into account the timing of the events.

The requirements for minimum data capture and time coverage do not include losses of data due to the regular calibration or the normal maintenance of the instrumentation.

13



To ensure accuracy of measurements and compliance with the DQO the appropriate competent authorities and bodies ... shall ensure the following:

• that all measurements ... are traceable;• that institutions operating networks and individual stations

have an established QA/QC system which provides for regular maintenance to assure the accuracy of measuring devices;

• that a QA/QC process is established for the process of data collection and reporting and that institutions appointed for this task actively participate in the related Community-wide quality assurance programmes;

• that the national laboratories ... are taking part in Community-wide intercomparisons... are accredited according to EN/ISO 17025 ... or are in the process of accreditation.

All reported data shall be deemed to be valid except data flagged as provisional, i.e. all faulty data, zero/span checks, calibrations etc. must be removed from the reported dataset.

14

AQD, Annex VIIOzone target values and long-term objectives


The following criteria shall be used for checking validity when aggregating data and calculating statistical parameters:

15

AQD, Annex XILimit values for the protection of human health


Without prejudice to Annex I, the following criteria shall be used for checking validity when aggregating data and calculating statistical parameters:

(1) The requirement for the calculation of annual mean do not include losses of data due to the regular calibration or the normal maintenance of the instrumentation.

16

DQO - remarks 1/2


In order for the measurements to constitute a compliant overall assessment of AQ in the Member State also need to be met

• requirements for the appropriate numbers of monitoring points in Zones and Agglomerations (Annex V and IX)

• locations and macro and micro siting of monitoring points (Annex III and VIII)

• reference methods for assessment (Annex VI)• calculation of uncertainty: the CEN standards provide a

specific methodology for calculation of the uncertainty of measurement for direct comparison with the Directive DQO. The CEN standard requires that this is calculated annually and DEG is likely to require that this is calculated individually from the QA/QC test data for each analyser and reported annually to the Commission along with the corresponding annual data set.

17

DQO - remarks 2/2


Standardisation (Annex VI, C.):all the results for gaseous pollutants

have to be expressed at the following conditions of temperature and pressure:293 K and 101.3 kPa.

for particle bound components, data shall be reported at ambient conditions.

Correct metadata (geographical co-ordinates, altitude, station’s classification,...) are important, too.

18

EU Data Exchange Group (DEG)


Implementing Provisions (IP) for reporting under AQD are currently being developed by the DEG (current status: draft, under preparation).

Subject: • to provide the (technical) requirements for the AQ information

flow under AQD• to serve as the main basis for the Commission's preparation of

the IP for reporting under AQD and any related guidance

Main IP elements:• specification of the reported information• information flow requirements (deadlines/periodicity,

reporting scheme etc.)• common data format and metadata description (all

data-flows)• description of tools for checking the format, data and

metadata consistency and integrity• description of tools for merging, aggregation and

rendering of the data

19

EMEP QA/QC


http://www.nilu.no/projects/ccc/qa

• EMEP Manual for sampling and chemical analysis

• Data Quality Objectives• Flagging data • Detection limits and precision• Results from laboratory intercomparison• Results from field intercomparison• Data quality reports

20

EMEP data quality objectives


for the acidifying and eutrophying compounds10 % accuracy or better for oxidized sulphur and oxidized nitrogen in single analysis in the laboratory

15 % accuracy or better for other components in the laboratory

0.1 units for pH 15-25 % uncertainty for the combined sampling and

chemical analysis (components to be specified later) 90 % data completeness of the daily values the targets, with respect to precision and detection

limit follow the DQO of the WMO/GAW precipitation programme (WMO, 2004)

the targets for the wet analysis of components extracted from air filters are the same as for precipitation. For SO2 the limit above for sulphate is valid for the medium volume method with impregnated filter. For NO2 determined as NO2- in solution the accuracy for the lowest concentrations is 0.01 mg N/l.

for heavy metals90% completeness 30% accuracy in annual average accuracy in laboratory...

21

Collecting and reporting data


Theory: all of the QA activities are undertaken correctly, in compliance with the relevant CEN standards and SOPs the measurements will fulfil the requirements of the EU Directives without further checking

Practice: there is a need to QC the data by careful data management and checking, analyser / sampler faults must be identified and addressed quickly in order to fulfil the DQO for data capture

Prior to submission of data to the data user, any suspect data must be identified and investigated;in addition there is the need to ensure that the data are reported correctly

22

QC scheme - data flow & feedback


1. Monitoring station2. Measurement data

centrea) Laboratory (manual meas.)b) Automatic measurements

The roles and responsibilities must be unambiguous as well as the feedback between persons / institutions.

3. Monitoring data centre

4. National data centre

5. European data centre

23

QC - level 1: monitoring station


The hardware and software tools on monitoring station must guarantee a correct data storage (results of measurements and also supported parameters such as temperature, pressure, sampled air volume etc.) and a transmission to the data centre

The first (automatic) check of completeness of data has to be implemented on automatic station

Software should use automatic data flags (valid data, faulty data due to incompleteness, zero / span checks, calibrations, maintenance,...)

provisional data

24

QC - level 2a: laboratory


The hardware and software tools in laboratory as well as organisational rules must guarantee correct samples storage and analysis, results of analysis evaluation, inspection and storage.

Evaluation of conditions and rules during: • sampling (temperature, air volume and flow

continuity,...)• storage and transportation of samples (temperature,...)• laboratory analysisshould be done, too

provisional data

25

QC - level 2b: automatic meas. data centre


The hardware and software tools as well as organisational rules must guarantee:

• assurance of the integrity of transmission of data from the station to data centre

• correct data storage• storage and use all supported information from station (e.g. zero / span

check and calibration records, information about (ir)regular analyzers inspection,...)

• review of data completeness and refill non-complete data series• review of data correctness from technical perspective:

identify and inspect out-of-range, negative, constant, extreme,above thresholds, rapidly changed and other suspicious data

reject invalid datacorrect data (e.g. if shift of measured level is known) flag the data

provisional data

All procedures must ensure that data capture is maximised, i.e. the data must be analysed frequently (ideally daily) so that measurements affected by instrument fault and other faults could be recognised quickly.

Some of reviews have to work automatically to allow real-time data release.

26

QC - level 3: monitoring data centre


The hardware and software tools as well as organisational rules must guarantee advanced review of data correctness, reliability and consistency:

• identify and inspect negative, constant, extreme, above thresholds, rapidly changed and other suspicious data

• consider data integrity• compare data series from different stations • compare data series from one station to inspect

relationship between different pollutants

provisional data – daily valid data - monthly, quarterly, yearly

All procedures must ensure that data capture is maximised, i.e. the basic review must be done frequently (ideally daily) so that measurements affected by instrument fault and other faults would be recognised quickly.

27

QC - level 4 & 5: national and European data centre


The hardware and software tools as well as organisational rules must guarantee advanced review of data correctness, reliability and consistency on national and European level:

• identify and inspect negative, constant, extreme, above thresholds, rapidly changed and other suspicious data

• consider data integrity• compare data series from different stations and

regions• compare data series from one station to inspect

relationship between different pollutants

Frequency: • national level: monthly, quarterly, yearly• European level: yearly

28

Examples: suspicious data invalid data


Zero / span check

29

Examples: suspicious data valid data


CO peak on traffic station

30



Emergency SO2 outflow from chemical factory (30min values)

31



O3 peak, 7th March 2005

32



24.3.2007 Dust above central Europe from eastern Ukraine

33

Examples: suspicious data checks


e.g. alien values

34



Simple outliers or outliers based on statistical methods

35



Cumulative concentrations

36

Examples: relationships


• some pollutant levels will be expected to rise and fall together or against

• diurnal peaks of NO are usually associated with traffic during rush hours

• diurnal behaviour of O3• ...

0

20

40

60

80

100

120

00 01 02 03 04 05 06 07 08 09 10 11 12 13 14 15 16 17 18 19 20 21 22 23

Kon

cent

race

v µ

g.m-3

T_TKAR T_TOFF T_TOPR T_TSTD

0

20

40

60

80

100

120

140

160

180

00 01 02 03 04 05 06 07 08 09 10 11 12 13 14 15 16 17 18 19 20 21 22 23

Kon

cent

race

v µ

g.m-3

CH_TBOM CH_TFMI CH_THAR CH_TKAR CH_TOFF

CH_TOPB CH_TOPR CH_TORA CH_TORV CH_TOZR

CH_TSTD CH_TVER CH_TOCB CH_TOMH CH_TOBA

37



NO2 [ppb] + NO [ppb] = NOX [ppb]

38



NO vs. O3

39



Different ratio PM10 vs.

PM2.5

40

Examples: especial events


• Special weather conditions (usefull meteorological data support)

• Service intervention• Unusual day run…

41

Thank you for attention...


http://ec.europa.eu/environment/air

• The CAFE Programme/ implementation of the Thematic Strategy on Air Pollution

• Ambient Air Quality • New Air Quality Proposal • Existing Air Quality Legislation • Implementation of existing AQ legislation• Meetings & Workshops - CIRCA website• EU Focus on Clean Air • Useful links • Feedback

procedures for qa/qc on the air quality monitoring data

Documents

terms of quality

quality management

thatthe quality

data accuracy

air quality network

quality assurance vocabulary

aq data acquisition

data qualitycoordination