How do low-abundance taxa affect river biomonitoring? Exploring the response of different macroinvertebrate-based indices

The contribution of rare taxa to aquatic bioassessments remains a subject of debate, and generates contrasting positions among researchers. Very little is known about the effect of low-abundance taxa (LAT) for calculating both single and multimetric macroinvertebrate-based indices, as well as the ecological status classification. In this study, we aimed to: i) identify the aquatic macroinvertebrates that need special attention during index applications given their low abundance; ii) analyse the effect of excluding LAT on single (IBMWP and IASPT) and multimetric (STAR_ICMi) biological indices; and iii) investigate the influence of LAT on river ecological status assessments. To this end, two different river basins in SE Spain and N Italy with contrasting climatic conditions and river types were selected. Our results showed that almost all the taxa at the family level can act as low-abundance taxa. In particular, the LAT belonged mainly to orders Diptera, Trichoptera, Coleoptera, Gastropoda and Hemiptera. The IndVal analysis stressed Tabanidae, Cordulegasteridae and Hydroptilidae as the most characteristic low-abundance families in the Spanish data set, while Dryopidae and Athericidae were identified mostly in N Italy. Excluding LAT affected the studied index values and the resulting bioassessment classification, except for the IASPT index. Loss of the entire LAT pool reduced the ecological status for 78% of the samples for the IBMWP index. Changing took place in 41% of the samples when considering the STAR_ICM index. Relevant changes were detected even when considering loss of 50% of the LAT, especially with the IBMWP index. Similar values and patterns were obtained in each considered quality class and river type. Our results provide useful information about controversial taxa and stress the significance of LAT in river biomonitoring. Excluding LAT is discouraged, although different responses according to the considered index were detected. The IBMWP index always obtained lower values (coupled mainly with an underestimation of ecological classes), while STAR_ICMi and, especially the IASPT index, were less affected by excluding LAT. Paying special attention to all the protocol application stages is recommended, with emphasis placed when using the IBMWP index.