Sampling protocol

  The sampling protocol used to develop the predictive models inside MEDPACS, is based on the Rapid Bioassessment Protocol of the IBMWP/IASPT index (formerly named BMWP’, Alba-Tercedor & Sánchez-Ortega, 1988), widely used in Spain (e.g. Zamora-Muñoz et al. 1995; Alba-Tercedor 1996; Zamora-Muñoz & Alba-Tercedor 1996; Alba-Tercedor & Pujante 2000; Alba-Tercedor et al. 2005; Alba-Tercedor, 2007) and recently adopted by most of the Spanish water agencies as the protocol to evaluate wadeable rivers and streams.
     To evaluate any test site by using the predictive models develop in this web application, it is essential to follow this sampling protocol, that ensures a standardized dataset as well as ensures to register the maximum information about the composition of the aquatic macroinvertebrate communities.
     Previously the sampling, it is important to survey the study reach (approximately 100m, or a distance equals 20 times the width of the reach) identifying all microhabitats (riffles, pools, macrophytes, roots, gravel, mud, etc.). At the same time, it is important to register all those shy aquatic macroinvertebrates that live on the water surface (e.g. Gyrinidae, Gerridae, Hydrometridae, etc.), due to once into the water they used to hide. Afterwards, a multi-habitat sample should be taken with a kick net (250-400 μm mesh). In riffles and runs, we situated the net downstream of stones, which were agitated and cleaned to collect dislodged organisms. In lentic habitats, we swept marginal vegetation, gravel, and mud. We emptied collected material into trays, and identified organisms to family level in situ (except for Hydracarina, Oligochaeta, and Ostracoda). A maximum of 3 individuals from each family should be stored in a vial with alcohol (70%) as field records. Organisms not collected but seen in the field should be also registered. The sampling finished when no new taxa are recorded. The remaining material should be stored in 4% formalin (or alcohol), and once in the laboratory it should be sorted with the aim of complete the final taxa list with those organisms that could be unseen during the field sorting.
    The WFD demands semi-quantitative samplings. The current predictive models do not need this kind of data, but they are completely compatible with the IBMWP semi-quantitative (Jáimez-Cuéllar et al., 2006) to perform the ecological assessments.