Published in:
Journal of Vegetation Science, 31 (2020) 1, Seite 14-25
Language:
English
DOI:
10.1111/jvs.12822
ISSN:
1654-1103;
1100-9233
Origination:
Footnote:
Description:
AbstractQuestionsLong‐term programs monitoring the impact of climate change on alpine vegetation necessarily involve changing observers. We aim at quantifying observer errors and ask if the signal of alpine vegetation transformation due to climate change exceeds pseudo‐changes caused by observer errors.LocationTwo mountain regions in the Alps, Schrankogel and Hochschwab (both Austria), and one in the High Tatra Mountains (Slovakia).MethodsVascular plant species presence and cover were recorded on 10–12 1‐m2 plots by 13–14 observers per site. Observer errors were calculated as species turnover, and deviations of species cover and the plot thermic vegetation indicator (which is correlated with temperature) from the mean over all observers. Observer errors in estimating species cover were split into a random and systematic part. The influence of plot and species characteristics on observer errors was investigated using (generalized) linear mixed‐effect models. Changes over time from three surveys in species turnover, cover and the thermic vegetation indicator were related to the amount of observer error using a bootstrap approach.ResultsSpecies cover was the most influential factor affecting observer errors in recording species lists and in species cover estimation. Plot attributes and observer identity had a weak but significant influence on errors in the thermic vegetation indicator. Systematic errors in estimating species cover were ≤5%. Changes over time in estimating species cover, as well as in species turnover and the thermic vegetation indicator exceeded observer errors in all cases where the observation period was ≥10 years.ConclusionsThe thermic vegetation indicator, which combines species composition and cover with species’ elevational distributions, provides a reliable estimate of warming‐related vegetation changes. Our results underline the importance of long‐term monitoring and long observation periods, which enable us to account for short‐term fluctuations and observer errors alike.