Assessing the ecological status in species-poor systems: A fish-based index for Mediterranean Rivers (Guadiana River, SW Spain)

Abstract

The assessment of the ecological status of freshwater ecosystems is a key issue for many international laws such as the Water Framework Directive (WFD) in light of the actual impoverished status of these ecosystems. Different multimetric approaches have been successfully developed in different freshwater environments. However, multimetric indices are difficult to apply to Mediterranean rivers basins, characterized by freshwater fish communities very low in species richness and support a high number of endemics with generalist and opportunistic life strategies. Here we follow a site-specific approach to develop an adaptation of the multimetric procedure usually used in Indices of Biotic Integrity (IBI), resulting in the Index of Community Integrity (ICI). We modeled the presence–absence of 10 native freshwater fish species from the Guadiana River basin using the Assessment of Nearest Neighbor Analyses (ANNA) model, to assess the deviation of the observed and expected community composition at reference condition. ANNA, which is a multi-species model, allowed incorporating in the index some rare species, though not all of them present in the basin. Deviations were transformed into probabilities of belonging to a reference site and species by species measures were then integrated in a final score. The use of presence–absence only data reduces potential errors associated with estimations of species’ abundance and its seasonal changes, which is especially important in Mediterranean environments. The ICI was sensitive to both habitat degradation and the degree of dominance of non-native species within the freshwater fish community (or biotic perturbation), which have been traditionally overlooked in IBIs, but was unresponsive to natural sources of variation. Given the site-specific nature and the simplicity of the index that we propose here, we expect it to be applicable to other Mediterranean basins for which predictive models can be produced.