Abstract

A method is presented to assess measurement errors associated with sensor accuracy and field practises in the environmental monitoring industry. These measures of uncertainty are combined to present a total value estimate for measurement uncertainty. This measurement uncertainty is then used to define an objective range of quality codes that can be applied to data.

Quality codes are an assessment of the standard to which a measurement has been made and are usually interpreted in terms of useability of the value. Increasingly, they are being used to qualify the value of the collection process.

In general, the result of a measurement is only an approximation or estimate of the value of the measurement and thus is complete only when accompanied by a statement of the uncertainty of that estimate. In practice, the specification or definition of the measurement is dictated by variability introduced during the measurement process. Any measurement should be qualified with sufficient completeness, with respect to the measurement accuracy, to define the value in terms of the errors surrounding the measurement process. This is as true of environmental measurement as it is of any other measurement.

The people that collect the data are not necessarily those that use the data and thus there must be standards for collecting and validating the data before it is available for use by others.

Clients need to understand whether results are measured or estimated, actual or adjusted. Resource managers need to know how close the recorded value is to a fixed reference. Modellers need to know how to quantify model uncertainty.

The crux of identifying and explaining measurement accuracy, then, is in the definition of identifiable errors inherent in the process. The application of that error is defined by the quality coding that we place on the data once the errors have been defined and quantified. These quality codes should be quite clear to the end user and not suffer from the imposition of philosophical judgements.