Technical ReportPDF Available

EU MARS Project D4.1 A synthesis of river basin case studies - Final Report

Authors:

Abstract

The modelling undertaken across the study basins allow us to have an understanding of the trends displayed by the data gathered at the various case-studies, encompassing a global view of European riverscapes. Overall, the studies highlight a general difficulty to identify significant pair-wise multi-stressor interactions, even if considering that suitable data is not always available. The significance and strength of the interactions revealed by empirical data treatments also depend on the length of the environmental and disturbance gradients displayed by the data, in fact, enough gradient is needed to obtain good responses, as was shown in the Elbe basin case, or a proper scale of disturbance to fit the time scale of the biotic response, as was shown in the Thames basin study. Therefore, some guidance and basic requirements are needed to detect multi-stressor trends and treat the data, for each basin, in order to account for the multiple stressor interactions. In other words, it is not easy to post hoc detect interactions based on prior monitoring schemes. The multi-stressor interactions were addressed using a common methodology of a set of empirical models including correlation analysis, general linear models, random forrest and boosted regression trees allowing for interpretable common outputs. Abiotic states showed a moderate to high capacity to explain changes in biotic indicators and similarly predict changes in ecosystem services indicators. In Otra case study, SO4 concentration (the main driver of pH) was found to be the single most important predictor of salmon abundance. At Vansjø case study, on the other hand, temperature and total phosphorus were found to have a synergistic influence on phytoplankton biomass. Nevertheless, few significant multiple-stressor interactions were found, in spite of all biological elements and metrics being investigated (26 significant interactions in all basins: 11 antagonistic positive or negative, 14 synergistic and 1 additive). Indices/combined and trait based metrics were generally more responsive to multi-stressor combinations than sensitivity or tolerance metrics. In some case studies it could be clearly demonstrated multi-stressor interactions, e.g temperature and nutrient stress for fish abundance in Regge and Dinkel; however, in other cases, pair-wise interactions could be identified as significant but the interaction type could not be defined, e.g. water residence time ratio x total phosphorus for Lower Danube. In many cases, however, no interactions could be identified, e.g. in Beysehir and Pinios case studies. It is clear that interaction signaling (type and direction) vary a lot across basins, even for similar biological indicators or stressor combinations. In Welsh uplands case study, acidification has not generally interacting with nutrient enrichment or climate change. Rather, these three stressors may display isolated or additive effects depending on the spatial and temporal scale studied. Interactions were searched for, in each case study, using unique combinations of indicators, sole process-base or empirical based modeling, or both, in a search for best predictive results. The Welsh uplands case study used combined long-term and spatially extended data, which resulted in a powerful strategy to analyse multiple stressors. This seems to indicate that for each basin, a combination of ecological expertise and modeling skills are needed, in order to obtain the trends of ecological status’ response that have to be delivered to the water administration in order to guide the PoM. Multi-pressure interactions also seem to be indicator-specific and therefore, the parameters that better illustrate or predict multi-stressor processes should be prioritized in the monitoring. An increase in the number of significant interactions seem to depend also on the existence of large empirical biological data sets, e.g. 7 significant interactions in the Ruhr basin. Although responses and interactions were found for all WFD biological elements, many come from macroinvertebrate traits; nonetheless, this trend may simply reflect the available data for empirical models’ use. A large number of these significant interactions detected by the empirical data treatments were strongly affected by natural variables such as basin size, fish zonation, or the natural vegetation cover of the basin. For example, in Welsh uplands, catchments exposed to acidification are usually nutrient poor, probably as a result of being less exposed to agriculture or pastures, however, land-use intensification to secure food provision may result in acid sensitive sites exposed to nutrient enrichment in a near future as revealed by the most intensive scenarios (Techno world and Fragmented world). In Sorraia, the strongest relationship with biological indicators was natural environmental variability, followed by land use variables, and then hydrological and nutrient variables. In fact, frequently in river case studies, land use variables/stressors (likely as proxies of multiple stressors in their own), have stronger relationships with biological indicators than single hydrological and nutrient stressors. No multi-stressor interactions were found for indicators of ecosystem services, in the case studies where these were studied. Full list of Authors and co-authors: Authors and co-authors: Yiannis Panagopoulos, Kostas Stefanidis, Maria Mimikou (NTUA. Greece), Jenică Hanganu, Adrian Constantinescu (DDNI, Romania), Meryem Beklioğlu, Tuba Bucak, Şeyda Erdoğan, Ayşe İdil Çakıroğlu, Emel Çakmak, Jan Coppens (METU, Turkey), Carina Almeida, Paulo Branco, Ramiro Neves, Pedro Segurado (University of Lisbon, Portugal), Eugenio Molina-Navarro, Shenglan Lu, Dennis Trolle, and Hans Estrup Andersen (Aarhus University, Denmark), Lilith Kramer, Marijn Kuijper, Perry de Louw, Vince Kaandorp, Erwin Meijers, Ellis Penning and Dimmie Hendriks (Deltares, Netherlands), Ute Mischke, Judith Mahnkopf, Andreas Gericke and Markus Venohr (Leibniz Institute of Freshwater Ecology and Inland Fisheries, Germany), Christel Prudhomme, Mike Hutchins, John Bloomfield, Alex Elliott, Corinna Abesser, Olivia Hitt, Majdi Mansour and Mike Bowes (Natural Environment Research Council, U.K.), Alexander Gieswein (University of Duisberg-Essen, Germany), and Rafaela Schinegger, Christiane Aschauer and Stefan Schmutz (University of Natural Resources and Life Sciences, Austria), Raoul-Marie Couture, Richard F. Wright, Yan Lin, Øyvind Kaste, José-Luis Guerrero, Anne Christiansen, Jonas Persson, Petra Mutinova (Norwegian Institute for Water Research NIVA), Katri Rankinen, Ninni Liukko, Seppo Hellsten (Finnish Environment Institute SYKE), Fabien Cremona (Estonian University of Life Sciences), Paul G. Whitehead, Gianbattista Bussi (Water Ressource Associates WRA), Cayetano Gutiérrez-Cánovas, Steve Ormerod (Cardiff University).
A preview of the PDF is not available
ResearchGate has not been able to resolve any citations for this publication.
ResearchGate has not been able to resolve any references for this publication.