Evaluation of data quality in lichen biomonitoring studies: the Italian experience

A total of 65 operators involved in lichen mapping studies in central and northwestern Italy underwent quality control tests during five lichen biomonitoring workshops organized between 1999and 2000. The results showed that 75% quantitative accuracy and90% quantitative precision can be regarded as satisfactory levels for lichen biodiversity data; 65% proved to be sufficientfor accuracy of taxonomic identification in the field. Average correct assignment of the interpretative naturality/alteration class was only 48.7%. The results indicated the need for taxonomic training.     

Visual Assessment of Foliar Injury Induced by Ozone on Indicator Tobacco Plants: A Data Quality Evaluation

A group of 50 inexperienced scorers were asked to estimate –using standard reference photos – foliar injury induced by the gas pollutant ozone on the supersensitive indicator planttobacco (Nicotiana tabacum L.) cv. Bel-W3. Theaverage accuracy level was 56.1% (95.9% when theclasses nearest to the reality were also consideredcorrect) and the average repeatability was 65.4%. Theextreme classes were easily scored. Central classesproved to be more difficult to be evaluated: this maydepend on the fact that two leaves may have similaractual total injured area, but substantially differentnumbers and spatial distribution of the lesions. Insome cases we observed a prevalence of overestimationerrors in the high classes and underestimation in thelow classes: this is in contradiction with theWeber-Fechner law. It is noteworthy the very shorttime required by operators to score, regardless of the results.