Beyond mean fitness: Demographic stochasticity and resilience matter at tree species climatic edges

Abstract

Aim Linking local population dynamics and species distributions is crucial to predicting the impacts of climate change. Although many studies focus on the mean fitness of populations, theory shows that species distributions can be shaped by demographic stochasticity or population resilience. Here, we examine how mean fitness (measured by invasion rate), demographic stochasticity and resilience (measured by the ability to recover from disturbance) constrain populations at the edges compared with the climatic centre. Location Europe: Spain, France, Germany, Finland and Sweden. Period Forest inventory data used for fitting the models cover the period from 1985 to 2013. Major taxa Dominant European tree species; angiosperms and gymnosperms. Methods We developed dynamic population models covering the entire life cycle of 25 European tree species with climatically dependent recruitment models fitted to forest inventory data. We then ran simulations using integral projection and individual‐based models to test how invasion rates, risk of stochastic extinction and ability to recover from stochastic disturbances differ between the centre and edges of the climatic niches of species. Results Results varied among species, but in general, demographic constraints were stronger at warm edges and for species in harsher climates. Conversely, recovery was more limiting at cold edges. In addition, we found that for several species, constraints at the edges were attributable to demographic stochasticity and capacity for recovery rather than mean fitness. Main conclusions Our results highlight that mean fitness is not the only mechanism at play at the edges; demographic stochasticity and population capacity to recover also matter for European tree species. To understand how climate change will drive species range shifts, future studies will need to analyse the interplay between population mean growth rate and stochastic demographic processes in addition to disturbances.

Forest storm resilience depends on the interplay between functional composition and climate—Insights from European‐scale simulations

“…One key originality of our study was to analyse resistance and recovery together. Indeed, due to the difficulty of capturing species‐specific sensitivity to disturbance, studies simulating forest resilience tend to focus on recovery (Guyennon et al., 2023; Schmitt et al., 2020) without considering tree species resistance, despite the existence of possible trade‐offs between resistance and recovery (Nimmo et al., 2015) and the fact that both metrics contribute to explaining overall forest resilience (Holling, 1973; Lloret et al., 2011). Our SEM analysis tends to support the existence of such a trade‐off by showing that resilience and recovery are associated with different edges of the same functional axis: communities dominated by conservative species (slow radial growth, high wood density) were more resistant, whereas communities dominated by fast‐growing species (fast radial growth and high height to diameter ratio) had higher recovery.…”

“…(2020) and in Guyennon et al. (2023), without repeating in detail the calibration process. We present in detail all the new additions to the model.…”

“…We used an IPM, that had already been calibrated with National Forest Inventory (NFI) data. Here we present the main structure of the model, which is presented in Kunstler et al (2020) and in Guyennon et al (2023), without repeating in detail the calibration process. We present in detail all the new additions to the model.…”

Spatial replication can best advance our understanding of population responses to climate

“…We suggest new studies focusing on sites with similar climate located at the leading or rear edge of species ranges, because these locations are ecologically more important, and because they should provide higher statistical power. Leading or rear range edges are the locations where populations are most sensitive to climate (Morley et al 2017, Amburgey et al 2018, Kleinhesselink and Adler 2018); accordingly, range edges also show the highest variability in population growth rates (Sexton et al 2009, Csergő et al 2017, Guyennon et al 2023). From an ecological perspective, such high climate sensitivity provides an opportunity to understand range limit formation and to forecast range shifts (Parmesan 2006, Ehrlén and Morris 2015).…”

“…2c) is constant across replicates also reflects current knowledge. Specifically, review (Sexton et al 2009) and synthesis studies (Csergő et al 2017, Guyennon et al 2023) indicate that the variability of population growth rates should become larger only at range edges. Third, so long as the spatial replicates are sufficiently spaced (e.g.…”