Assessing the quality of biomonitoring via signal-to-noise ratio analysis

The analysis of the signal-to-noise ratio of a survey is introduced as a means to assess the quality of a survey. Here, the survey signal is defined as the survey variance, and the survey noise is determined by measurement of the local variance (variance per site). The signal-to-noise ratio, and thus the quality of the survey, can be improved by factor analysis aided processing of the data. This is illustrated by processing data from various biomonitoring programs on trace element air pollution. Clean-up of the data-set, in which a bias-introducing contribution is removed (for example, the soil factor) or source profile isolation, in which attention is focussed on a single source, yield strikingly different indications of the quality (or suitability) of the biomonitoring species when compared with the indications obtained from the original dataset. The approach presented here stresses that the quality of a survey largely depends on the decisions taken with respect to the selection of the biomonitor material; these decisions should be based on optimization of the signal-to-noise ratio rather than on minimization of the noise only.