Community structure determines the predictability of population collapse

Abstract

Early warning signals (EWS) are phenomenological tools that have been proposed as predictors of the collapse of biological systems. Whilst a growing body of work has shown the utility of EWS based on either statistic derived from abundance data or shifts in phenotypic traits such as body size, so far this work has largely focused on single species populations.However, in order to predict reliably the future state of ecological systems which inherently could consist of multiple species, understanding how reliable such signals are in a community context is critical.Here, reconciling quantitative genetics and Lotka-Volterra equations which allow us to track both abundance and mean traits, we simulate the collapse of populations embedded in mutualistic and multi-trophic predator-prey communities. Using these simulations and warning signals derived from both population- and community-level data, we show that the utility of abundance-based EWS as well as metrics derived from stability-landscape theory (e.g., width and depth of the basin of attraction) are fundamentally linked, and thus the depth and width of the stability landscape could be used to identify which species should exhibit the strongest EWS of collapse.The probability a species displays both trait and abundance based EWS is dependent on its position in a community, with some species able to act as indicators species. In addition, our results also demonstrate that in general trait-based EWS appear less reliable in comparison to abundance-based EWS in forecasting species collapses in our simulated communities. Furthermore, community-level abundance-based EWS were fairly reliable in comparison to their species-level counterparts in forecasting species level collapses.Our study suggests a holistic framework that combines abundance-based EWS and metrics derived from stability-landscape theory that may help in forecasting species loss in a community context.

Ranking species based on sensitivity to perturbations under non-equilibrium community dynamics

“…For example, whereas the linearized dynamics can be used to compute a recovery rate under equilibrium (Arnoldi et al ., 2018, Medeiros et al ., 2021, Strogatz, 2018), we show that they can be used to derive the time-varying expected sensitivity of different species to perturbations under non-equilibrium dynamics. Similarly to previous studies, our results confirm the utility of tracking variances and covariances of perturbed abundances to anticipate changes in ecological communities (Baruah et al ., 2022, Chen et al ., 2019, Dakos, 2018). In addition, we use the leading eigenvector, which has been previously employed to decompose community responses into species responses to perturbations under equilibrium dynamics (Dakos, 2018, Ghadami et al ., 2020, Patterson et al ., 2021, Weinans et al ., 2019).…”

Abrupt transitions and its indicators in mutualistic meta-networks: effects of network topology, size of metacommunities and species dispersal

“…When an ecological network collapses both at the local as well as at the spatial scale, signals of global meta-network instability could also be manifested in species in an ecological community. It is relatively unknown, however, which species could exhibit signals of instability, but see Dakos (2017) , Baruah et al (2022) or Patterson et al (2021). Indeed, the species that exhibited strong increases in standard deviation as networks collapsed, were the ones which had on average a larger number of interactions (Fig.…”

“…Studies have indicated that in addition to monitoring dynamics of species abundance, phenotypic traits should be monitored as well. Including information from phenotypic dynamics such as body size could improve forecasts of biodiversity collapse (Clements and Ozgul 2016), although such an accurate forecast of biodiversity collapse is dependent on the type of environmental perturbation and the type of interactions prevalent in the community (Baruah et al 2019(Baruah et al , 2022. When changes in the environment impact communities both locally and spatially alike, ecological networks could abruptly collapse.…”

“…However, such signals and their efficacy have been questioned recently both in numerical simulations studies (Hastings and Wysham 2010; Clements et al 2015; Baruah et al 2020) and experimental studies (Wilkinson et al 2018; Baruah et al 2021). Their application on multi-species communities (Dakos and Bascompte 2014; Dakos 2017; Patterson et al 2021; Baruah et al 2022), particularly in the context of a spatial structure has been lacking. Here, I evaluated how the efficacy of commonly used statistical signals measured at the species level (autocorrelation at first-lag, standard deviation), community level (patch variability, spatial variability) and at the metacommunity level (temporal regional variability) in relation to network topology and spatial scale of interactions.…”

“…However, there are statistical tools that been developed to forecast such impending transitions which are commonly known as “early warning signals.” Commonly used signals are temporal autocorrelation and variance that could derived using a sliding window approach (see [Dakos et al (2012a) for details) from state based temporal data such as abundance or biomass. However, the utility of such signals are dependent on a host of factors that includes sampling requirements (Arkilanian et al 2020), data quality (Clements et al 2015), eco-evolutionary factors (Baruah et al 2020, 2021) and type of species interactions (Dakos 2017; Patterson et al 2021; Baruah et al 2022). One important challenge is to test the utility of such signals in a multispecies context embedded within a spatial scale of species interactions.…”