Epsilon Archive for Student Projects

Abstract

This study will deal mainly with the problem of assessing precision within the framework of the Swedish Integrated Monitoring Programme (IM). Data for the study were collected partly during the early part of the programme that started in 1982, partly during the field season in July 1999. The aspect of variation in species detection on a given sample plot was not considered in this study. Data are judged with reference to a general proposal that Swedish environmental monitoring should, with 95% significance, be able to demonstrate a change in any variable of minimum 20% during 5 years. The mean difference, in absolute values, between any two cover estimates was less than 5% in 95 per cent of the occasions. In relative values the mean difference was larger, i. e. 14% with species covering twenty per cent or more and 34% with species that cover less than one per cent. It is recommended not to consider relative changes when the cover is small, e.g. 10 per cent and less, because of the extreme values that may ensue from such calculations. The variation between estimates produced only a small effect on community sensitivity, “Ellenberg”, indices. Of sixteen species specially studied, the intrapersonal difference was significantly lower than the interpersonal one only for Pleurozium schreberi. Of three high-cover species only Vaccinium myrtillus showed a significant difference between series of paired observations. There was a comparatively high correlation between cover estimates in the field and cover measured on photographs.