Balancing effort and benefit - How taxonomic and quantitative resolution influence the pesticide indicator system SPEAR.

Biological indices aim to reflect the ecological quality of streams based on the community's species or trait composition. Accordingly, the capability to predict the ecological quality depends on (i) the knowledge on the association of taxa or traits with stressors and (ii) the taxonomic and quantitative resolution of taxa. Generally speaking, a higher resolution is associated with a better linkage between environmental condition and biological response but also with higher efforts and costs. So far it is unknown how the taxonomic and quantitative resolution affect the ecological quality assessment of streams related to pesticide effects when applying the invertebrate-based indicator SPEAR. We investigated the ecological quality of 101 streams considering four taxonomic levels (species, genus, family, order) and three quantitative resolutions (abundance, three abundance classes, and presence-absence). In a multiple linear regression analysis between 13 investigated stressors and SPEAR, the full models' explained variance remained fairly constant with decreasing taxonomic and quantitative resolution. As expected, the highest association between pesticide pressure and SPEAR was reached at a species/abundance resolution yielding an R of 0.43. In contrast, the lowest quantitative resolution of order level combined with presence-absence information revealed an explained variance of 0.28 R. We suggest the family/abundance class resolution (R = 0.38) as the best trade-off between effort and accuracy for large-scale monitoring. Due to a comparable linear regression at family/abundance class resolution, the assigned ecological quality classes were largely congruent (69 %) to species/abundance resolution. We conclude that the ecological quality assessment with SPEAR at family/abundance class resolution can be used to link pesticide contamination and invertebrate community structure with less taxonomic expertise and less quantification effort.