Comparing observer performance in vegetation records by efficiency graphs derived from rarefaction curves. (February 2020)

• Record Type: Journal Article
• Authors: Seidling, Walter; Hamberg, Leena; Máliš, František; Salemaa, Maija; Kutnar, Lado; Czerepko, Janusz; Kompa, Thomas; Buriánek, Václav; Dupouey, Jean-Luc; Vodálová, Anna; Canullo, Roberto
• Is Part Of: Ecological indicators. Volume 109(2020)
• Publication Date: 2020-02
• Subjects: Species richness -- Observer error -- Overlooking rate -- Misidentification rate -- Singleton species -- ICP Forests; Environmental monitoring -- Periodicals; Environmental management -- Periodicals; Environmental impact analysis -- Periodicals; Environmental risk assessment -- Periodicals; Sustainable development -- Periodicals; 333.71405
• DOI: 10.1016/j.ecolind.2019.105790
• Languages: English
• ISSNs: 1470-160X

Abstract

Highlights: Observer error is a serious problem in vegetation monitoring and surveys. Rarefaction curves provide a sound bases to describe observer performance in detail. Singletons can differentiate overlooking and misidentification in observer error. Observer error is for the most part explained by location and scale. Abstract: Species richness is a key variable in measuring diversity of ecological communities. It is crucial to get reliable estimates for the number of plant species in space (mapping) and – even more important in the context of monitoring – over time. Therefore, knowledge on error rates related to recordings of species numbers should be considered in such inventories. The performance of observers in four field tests to capture species numbers carried out in forest ecosystems in central and southern Europe were compared. Observer-related species accumulation (rarefaction) curves and derived efficiency curves were analysed, resulting in mean error rates of 29.7% and 39.4% over series of plots sized 4 m 2 and 100 m 2 respectively. As a new approach individual rarefaction and efficiency curves reveal site-specific and spatially differentiated capabilities of observers to register plant species. Since expertise and individual searching strategies are difficult to parametrise, reasons for variation in error rates remain largely unknown. However, statistical modelling with site- and scale-specific mean error rates gave an overview on important influential factors like location, scale, spatial integration, and their interactions. Our results underline the importance to incorporate specific training and inter-comparison measures in monitoring programs and critical perception of results on temporal changes of species richness.