The effect of taxonomic resolution on the assessment of ecological water quality classes

Within the ecological assessment of running waters based on benthic macroinvertebrates different levels of taxonomic resolution (species, genus, family and higher) are in use. Although assessment systems are often developed with detailed data on species level, water managers and other end-users could like to use data on higher taxonomic levels to assess the ecological quality of a water body because of limited human or money resources. The question that arises is, if an assessment system built with species level data is also applicable using data with a higher taxonomic resolution. Within the AQEM project a multimetric assessment system was developed to evaluate the ecological quality classes (from bad (1) to high (5) ecological quality) of different stream types throughout Europe. The present study focuses on the question whether the resulting water quality class changes using the AQEM Assessment Software (AAS) with different taxonomic resolutions and if yes, how large the deviations of ecological quality classes from the original classes are and if the deviations are unidirectional. For analyses data from four Austrian and two Dutch stream types were used. It is demonstrated that the assignment of a site to an ecological quality class may change if different taxonomic levels are used. Deviations in both directions (higher/lower ecological quality class) were observed. In most cases the divergence was only one ecological quality class, but also larger deviations occasionally occurred. The causes of changes in the assessment were investigated by separately looking into the underlying metrics of the multimetric system. Some of the evaluated metrics rely on autecological information on species level and are simply not applicable on higher taxonomic levels. Other metrics worked on higher taxonomic levels as well and showed more or less good distinctions between ecological quality classes. It is concluded that the AQEM Assessment Software is not applicable if data on higher taxonomic levels are used. As the deviations were not unidirectional and ranged from one to three ecological quality classes, it is not possible to include a correction factor for using the software with higher taxonomic resolution data.