The paradox of expert judgment in rivers ecological monitoring.

A great investment has been done in the last decades in the development of numerical and qualitative assessment methods to classify the ecological quality of water bodies. Yet, in spite of all attempts to avoid subjectivity, expert judgment is still used at numerous steps of the ecological classification and is considered by some authors as indispensible for management purposes. Thus, the aim of this study is to test the hypothesis that expert judgment, when done by the adequate experts (limnologists/river ecologist) with experience in the study area (i.e., natural conditions and expected communities), could be as good as quantitative indices and measures (i.e., result in the same classification), but quicker and with lower cost. For that we compared the classifications (on 13 aspects of rivers ecosystems) attributed by two experts to 20 sites (10 each) located in their study areas, with the classifications of ecological quality based on biological indices (for invertebrates and diatoms), hydromorphology and water chemistry, calculated by an independent team. Our results show that assessments made by experts and those calculated through indices (biological quality and hydromorphology) are globally very similar (RELATE test; Rho = 0.442; p < 0.001, 999 permutations). Most differences were of one class and experts tended to attribute a better condition than indices to the best quality sites but a worse condition to the worse quality sites. A Principal Components Analysis revealed that sites to which experts attributed a moderate quality had higher nitrate concentration and pH but were well oxygenated. The sites classified as poor and bad where those with stronger modifications in their habitats (given by the higher values of HMS). The difference between experts and indices is small but still represents 15% of sites, and includes both situations: the experts or the indices lead to the need of measures (i.e., classifications below class Good). Experts' evaluations on hydromorphological conditions of the channel and margins are also significantly correlated with the quality assessments made by the field team that has no experience in the study area (Rh0 = 0.518; p = 0.001; 999 permutation), indicating geographic independence in the expert judgment. We concluded that expert judgment could be used in the determination of streams and rivers ecological quality, saving money and time and helping to redirect monitoring funds to actual implementation of restoration measures. Yet, classification' scoring methods may still be useful for a better targeting of restoration measures.