Skip to main content
SLU publication database (SLUpub)

Research article2004Peer reviewed

A multimetric macroinvertebrate index for detecting organic pollution of streams in southern Sweden

Dahl J, Johnson RK

Abstract

Bioassessment programs in Sweden at present commonly use relatively simple approaches for detecting ecological change (e. g. the presence/absence of single indicator taxa). This study introduces and tests a more contemporary approach for assessing the ecological quality of Swedish streams. Macroinvertebrate data from 281 stream sites in southern Sweden, sampled in the autumn of 1995, were used for calibration of a multimetric index for assessing organic pollution of streams. For validation a data set consisting of 15 sites, sampled twice (spring and autumn) in 2000 was used. Eighty-four candidate macroinvertebrate metrics, within six attribute groups (richness, composition, tolerance, feeding, habit and habitat measures) were tested for inclusion in the final multimetric. Sites were grouped into four regional categories: (1) all sites combined (n = 281), and stream sites situated in the (2) southern boreal (n = 63), (3) boreonemoral (n = 187), and (4) nemoral (n = 31) ecoregions. Within each category, tests were done for the association between macroinvertebrate attributes and organic pollution. Five metrics were selected for inclusion in the final multimetric index (the DJ index): (i) number of EPT (Ephemeroptera, Plecoptera, and Trichoptera) taxa, representing richness measures; (ii) percent Crustacea and (iii) percent EPT taxa, representing composition measures; and (iv) ASPT (Average Score Per Taxon) and (v) Saprobic index, representing tolerance measures. The DJ index was significantly correlated with environmental variables indicative of organic pollution stress, both at habitat- (silt substratum) and catchment-level (e. g. conductivity and percentage of the catchment classified as urban and arable). Using an external data set (the AQEM data set) for validation showed that the DJ index was significantly related to total phosphorous concentration both in spring (r(2) = 0.50, 1 SE = 0.30) and autumn (r(2) = 0.57, 1 SE = 0.43). Between-class discrimination was poor at moderate or low total phosphorous concentrations. However, the low number of sites (n = 15) used in this external validation may have resulted in low statistical power. Nonetheless, both the calibration and the validation showed that the final multimetric index developed here is a promising tool for detecting the effects of organic pollution on stream ecosystems

Published in

Archiv fur Hydrobiologie
2004, Volume: 160, number: 4, pages: 487-513
Publisher: E SCHWEIZERBARTSCHE VERLAGS