QA/QC in monitoring - how to ensure data reliability
The most beautiful chart is worthless if the data behind it cannot be trusted. That is why professional monitoring rests on QA/QC - quality assurance and control - a set of habits that protect the reliability of results. They are worth knowing even when running monitoring on your own.
QA and QC - two sides of the same coin
- QA (quality assurance) - what we do in advance to make the data good: the choice of methods, procedures, training, the plan.
- QC (quality control) - what we use to check whether it actually is: calibrations, control samples, verification.
Practices that make a difference
- Equipment calibration - regular, with standards; an uncalibrated meter is a good-looking error.
- Blanks - they detect contamination during sampling/analysis.
- Duplicates - repeated measurements show repeatability.
- Chain of custody for the sample - from sampling to the laboratory: labelling, preservation, holding time (see sampling and preservation).
- Constancy of method and site - without it the data is not comparable.
Concepts worth understanding
- Measurement uncertainty - every result has a “margin”; reporting it is a sign of reliability.
- Limit of quantification - the lowest value a method can reliably measure; below it the result is reported by convention (and not as “0”).
- Accreditation (ISO/IEC 17025) - a formal confirmation of a laboratory’s competence; important when the data has official force.
Data validation
Before a result enters the summaries, it is worth reviewing it: unrealistic values, typos, unit mistakes. This is a simple step that catches quite a few errors.
In practice
Part of the QC is made easier by the tool itself. In LimnoLog the data has a uniform structure (stations, sessions, units), and the session approval mode lets you review and “close” the data before including it in a report. On how to present verified data, we write in the article A monitoring report.
← Knowledge base