A comparison of different biotic indices based on benthic macro-invertebrates in italian lakes

Benthic macroinvertebrates samples were taken from Italian lakes with different geological, morphological and chemical characteristics. Thirty-two lowland small and large lakes sampled using a grab in soft substrate were selected to develop biotic indices. Diversity indices based on species numbers - abundances and indices using species sensitivity values were compared. The lakes selected were all situated in the Alpine Ecoregion below 800 m a.s.l. and had similar chemical composition but different levels of anthropogenic pressure. Lakes with data available in different years were included as separate lakes in the analysis; littoralsublittoral samples of large lakes were also separated from profundal samples yielding a total of 41 sites for analysis. Seven different biotic indices were compared: (1) Shannon diversity index (H), (2) weighted Shannon diversity index (Hw) including in the calculation a sensitivity value assigned to each species, (3) a benthic quality index based on means of three different environmental variables, measuring trophic status, weighted by species abundances (BQITS), (4) an index based on weighted means using a larger set of environmental variables (BQIENV), (5) a modified BQITS, which included both species numbers and total abundance of individuals (BQIES), (6) an index calculated according to a rarefaction method (ES), (7) an index considering indicator species based on experts judgment (BQIEJ). The indices were compared with a trophic status index (TSI) constructed by joining three environmental variables: O2% saturation in the hypolimnion during summer stratification, total phosphorous and transparency during full circulation. Comparisons were also made with another environmental stress index (ENI) constructed on a larger number of variables. All the biotic indices had significant correlations with both TSI and ENI. BQIES, WFD compliant and well correlated with TSI and ENI, was selected to tentatively assign the investigated lakes into 5 quality classes: high (H), good (G), moderate (M), poor (P) and bad (B). The statistical power of the classification was estimated. Assuming tentatively equal intervals for each of the five quality classes, 2 lakes were classified at high status, 7 lakes were classified as good, 13 were classified as moderate, 13 were classified as poor and 4 were classified as bad. Fifteen lakes were classified with a power less than 80%. Some of the lakes resampled in different periods displayed a shift of class in the different years. Future work should focus on extending the database to test the indices in other lake types subjected to different pressures.