Premium
Extreme flood disturbance effects on multiple dimensions of river invertebrate community stability
Author(s) -
Eagle Lawrence J. B.,
Milner Alexander M.,
Klaar Megan J.,
Carrivick Jonathan L.,
Wilkes Martin,
Brown Lee E.
Publication year - 2021
Publication title -
journal of animal ecology
Language(s) - English
Resource type - Journals
SCImago Journal Rank - 2.134
H-Index - 157
eISSN - 1365-2656
pISSN - 0021-8790
DOI - 10.1111/1365-2656.13576
Subject(s) - disturbance (geology) , stability (learning theory) , pairwise comparison , environmental science , ecosystem , flood myth , ecological stability , ecology , event (particle physics) , mathematics , statistics , computer science , geography , geology , biology , physics , paleontology , archaeology , quantum mechanics , machine learning
Abstract Multidimensional analysis of community stability has recently emerged as an overarching approach to evaluating ecosystem response to disturbance. However, the approach has previously been applied only in experimental and modelling studies. We applied this concept to an 18‐year time series (2000–2017) of macroinvertebrate community dynamics from a southeast Alaskan river to further develop and test the approach in relation to the effects of two extreme flood events occurring in 2005 (event 1) and 2014 (event 2). Five components of stability were calculated for pairs of pre‐ or post‐event years. Individual components were tested for differences between pre‐ and post‐event time periods. Stability components’ pairwise correlations were assessed and ellipsoids of stability were developed for each time period and compared to a null model derived from the permuted dataset. Only one stability component demonstrated a significant difference between time periods. In contrast, 80% of moderate and significant correlations between stability components were degraded post‐disturbance and significant changes to the form of stability ellipsoids were observed. Ellipsoids of stability for all periods after the initial disturbance (2005) were not different to the null model. Our results illustrate that the dimensionality of stability approach can be applied to natural ecosystem time‐series data. The major increase in dimensionality of stability observed following disturbance potentially indicates significant shifts in the processes which drive stability following disturbance. This evidence improves our understanding of community response beyond what is possible through analysis of individual stability components.