AS 3580.6.1:2016 Methods for sampling and analysis of ambient air
10.4 Data acquisition and transfer
The integrity of the data being recorded by the data capture system shall be checked annually. The integrity of the data transfer process, whether via manual data entry or by automated telemetry, shall also be checked annually. If an external data storage medium is used in conjunction with the ozone analyser, e.g. an external data logger, and this involves the conversion of electrical signals, e.g. voltage to digital signals, then this shall be calibrated annually. The measurements stored on the external system shall be within ±0.2% of the operational full scale range of the analyser.
10.5 Data validation
Data collected in accordance with this Standard’s method shall be validated prior to its use. The data validation process shall consist of a review of the data by trained personnel using data screening criteria to identify possible incorrect values (e.g. calibration data mixed with sample data). This process should be conducted on a regular basis to avoid the invalidation of large quantities of data. The screening process is usually based on historical data or realistically expected values. Other values may be appropriate for a given location, therefore site-specific screening values shall then be developed. If the data does not fall within the screening criteria, the validity of the data shall be investigated. Relationships between different variables are considered when evaluating data because other parameters may either confirm or deny recorded events (e.g. wind direction changes, low visibility, etc.). The data review is an ongoing process that includes—
(a) reviewing calibration information, the recorded data, and any status flags that could affect data; and
(b) reviewing all calibrations, operational precision checks, station notes and maintenance sheets.
When validating the data, the underlying assumption shall be that all data is considered valid unless evidence or sound scientific principles can be given to support its invalidation. Copies of original (unvalidated) data sets shall be kept for audit purposes should any aspect of the validation process be brought into question.