Development of a multimetric index based on benthic macroinvertebrates for the assessment of natural wetlands in Southwest Ethiopia

Biotic indices are widely applied for conservation and management of aquatic resources since they allow water resources monitoring agencies to get insight in complex biological data and yield policy relevant information. Despite the worldwide popularity of biotic indices, little information on their use and applicability in Eastern Africa is available. Here, we develop a multimetric index based on macroinvertebrates to assess the ecological condition of natural wetlands in Southwest Ethiopia. Index development was based on a dataset of 222 samples collected during two consecutive years from 63 sites located at eight different wetlands. We used physico-chemical and hydro-morphological variables (land use pattern, habitat alteration, hydrological modification and chemical water quality) to classify sites as reference or degraded. We tested a total of 58 potential metrics representing various aspects of macroinvertebrate assemblages including family richness, composition, tolerance measures and presence and abundance of functional feeding groups. Metrics were selected for the development of a final index based on their sensitivity in discriminating reference from impaired sites, strength of correlation with the anthropogenic disturbance gradient, chemical measurements, and the degree of redundancy. Metrics retained for the final index were:overall family richness, family richness of Ephemeroptera, Odonata and Trichoptera (EDT), and percentage of filterer-collectors. The final index, derived from the sum of three metric scores, was divided into five water quality classes (very bad, bad, moderate, good and very good). Our final multimetric macroinvertebrate index (MMI) distinguished well between reference and impaired wetland sites and showed a significant negative response to a gradient of disturbances (R-2 = 0.86, p < 0.05). Moreover, it classified a validation dataset accurately with a correctly classified instances of 80% and a Cohen's Kappa value of 0.6. This MMI can be considered as a robust and sensitive tool that can be applied to evaluate the ecological condition of natural wetlands in Ethiopia, where wetland resources are under high pressure as a result of agricultural activities such as grazing and urbanization. (C) 2013 Elsevier Ltd. All rights reserved.