How it is studied
Quality assessment uses calibration records, temporal continuity, identification and checks. An isolated value needs context: unit, interval, instrument or model, and validation status. Also retain corrections made to a series.
An example to interpret it
Example: 10, missing and 20 have a mean of 15 across the two available observations. Replacing the gap with zero would produce 10 and invent a measurement. Always report how many values contributed to a summary.
What to check
A plausible value can still be wrong. Automated checks detect some anomalies; changes in instrument or location also require a review of the series' comparability.