Finding appropriate reference sites in large-scale aquatic field experiments Susanne I. Schmidt Marie Ko ¨nig-Rinke Katja Kornek Carola Winkelmann Markus A. Wetzel Jochen H. E. Koop Ju ¨rgen Benndorf Received: 17 November 2006 / Accepted: 29 November 2007 / Published online: 21 December 2007 Ó Springer Science+Business Media B.V. 2007 Abstract Defining the reference condition is one of the most critical aspects of ecosystem investigations since it describes the baseline against which the experimental sites will be evaluated and compared. In large-scale ecosystem experiments, this reference is ideally another ecosystem which is similar to the experimental system. We investigated two streams for their potential as experimental sites for a full-size pairwise ecosystem experiment. Temporal (2 years) and spatial (pool, riffle) variabilities of abiotic factors and as biotic element the structure of the macroinver- tebrate communities were investigated. Criteria of similarity that we applied at the two streams were: (1) high similarity in abiotic factors, (2) only small differences in the faunal assemblages (abundance structures, composition, feeding types), and (3) that the differences between the two systems should not exceed the temporal and spatial differences within each system. Among the abiotic factors investigated, only the inorganic nutrients (nitrate and soluble reactive phosphorus), major ions (magnesium, cal- cium), electric conductivity, and pH showed significant differences between the two streams. Discharge rate, current velocity, temperature, and oxygen concentra- tions did not significantly differ between the streams. Also, the community structure did not differ in species richness, abundance, and biomass; and only small differences in dominance structure and feeding-type composition were observed. The differences between habitats within each stream were always higher than those between the streams. Thus, both the streams are characterized by a similar structure of the macroinver- tebrate community, a main component of the stream food-web, which make them suitable for a full size pairwise ecosystem experiment. The present case study can form a basis for other full-size field experiments. Keywords Experimental design  Field study  Macrozoobenthos  Fish  Similarity  Stream ecology Introduction Food-web interactions in ecosystems can be gener- ally investigated using two different experimental approaches. One option is to establish the typical food-web components in artificial and simplified model systems (field or laboratory mesocosms). The second approach is an ecosystem experiment where a large part of a certain natural ecosystem can be manipulated. The advantage of a simplified, small S. I. Schmidt  M. A. Wetzel  J. H. E. Koop (&) Department of Animal Ecology, Federal Institute of Hydrology, Koblenz 56068, Germany e-mail: koop@bafg.de M. Ko ¨nig-Rinke  K. Kornek  C. Winkelmann  J. Benndorf Institute of Hydrobiology, Dresden University of Technology, Dresden 01062, Germany 123 Aquat Ecol (2009) 43:169–179 DOI 10.1007/s10452-007-9155-6