Definition of sampling units begets conclusions in ecology: The case of habitats for plant communities

• Published in: PeerJ
• Main Authors: Mörsdorf, Martin Alfons, Ravolainen, Virve, Støvern, Einar, Yoccoz, Nigel Gilles, Jonsdottir, Ingibjørg, Bråthen, Kari Anne
• Format: Article in Journal/Newspaper
• Language: English
• Published: PeerJ 2015
• Subjects: Sampling design; Expert knowledge; Formal rules; Sampling frame; Snowbed habitat; Mesic habitat; VDP::Mathematics and natural science: 400::Zoology and botany: 480::Ecology: 488; VDP::Matematikk og Naturvitenskap: 400::Zoologiske og botaniske fag: 480::Økologi: 488; Tundra
• Online Access: https://hdl.handle.net/10037/8657; https://doi.org/10.7717/peerj.815

In ecology, expert knowledge on habitat characteristics is often used to define sampling units such as study sites. Ecologists are especially prone to such approaches when prior sampling frames are not accessible. Here we ask to what extent can different approaches to the definition of sampling units influence the conclusions that are drawn from an ecological study? We do this by comparing a formal versus a subjective definition of sampling units within a study design which is based on well-articulated objectives and proper methodology. Both approaches are applied to tundra plant communities in mesic and snowbed habitats. For the formal approach, sampling units were first defined for each habitat in concave terrain of suitable slope using GIS. In the field, these units were only accepted as the targeted habitats if additional criteria for vegetation cover were fulfilled. For the subjective approach, sampling units were defined visually in the field, based on typical plant communities of mesic and snowbed habitats. For each approach, we collected information about plant community characteristics within a total of 11 mesic and seven snowbed units distributed between two herding districts of contrasting reindeer density. Results from the two approaches differed significantly in several plant community characteristics in both mesic and snowbed habitats. Furthermore, differences between the two approaches were not consistent because their magnitude and direction differed both between the two habitats and the two reindeer herding districts. Consequently, we could draw different conclusions on how plant diversity and relative abundance of functional groups are differentiated between the two habitats depending on the approach used. We therefore challenge ecologists to formalize the expert knowledge applied to define sampling units through a set of well-articulated rules, rather than applying it subjectively. We see this as instrumental for progress in ecology as only rules based on expert knowledge are transparent and lead to results reproducible by other ecologists.