Introduction
The factors affecting community stability are important to understand as community stability ultimately underpins the stability of ecosystem function (Millennium Ecosystem Assessment 2005; Cardinale et al.2012). A central problem is how species diversity (richness) contributes to stability (Elton 1958; May 1972; Tilman & Downing 1994; Tilman 1999). Community stability is influenced by two properties of the community: average population stability and asynchrony (Doak et al. 1998; Tilman et al. 1998; Yachi & Loreau 1999; Thibaut & Connolly 2013). Briefly, average population stability describes the variability of each population over time and can be influenced by factors such as mean-variance scaling, where larger populations are relatively less variable (Taylor 1961; Kilpatrick & Ives 2003) as well as density-dependent (e.g. pathogen-induced mortality) and density-independent processes (e.g. mortality induced by extreme weather events). Population asynchrony constitutes negatively correlated population dynamics, which can be driven by species interactions (e.g. competition) or through varying responses to environmental conditions. Average asynchrony between the populations in a given community is expected to increase with species richness (Thibaut & Connolly 2013), but associations between richness and average population stability are less obvious (Jiang & Pu 2009).
Though the theoretical background of community stability is well developed, the mechanisms have not been tested in natural animal communities and much experimentation has been conducted with plant or aquatic communities (Dı́az & Cabido 2001; Craven et al. 2018; van der Plas 2019). For example, asynchrony increases with species richness in grassland communities (Roscher et al. 2011; Isbell et al. 2019), but less is known about how richness impacts asynchrony or its importance in animal communities. Previous work suggests that the relative importance of the mechanisms can vary between taxa, with some studies finding a prominent role for asynchrony in driving community stability (Jucker et al. 2014; Ma et al. 2017). For example, asynchrony is observed as key to the stability of arthropod and plant communities between different habitat types (Blüthgen et al. 2016), while work on intertidal and algal communities suggest average population stability can explain differences in community stability (Pennekamp et al. 2018; White et al. 2020).
To understand why the impact of these mechanisms varies between communities, it is important to quantify both how asynchrony or population stability contribute to community stability and also the factors leading to differences in synchrony and stability. Three main forces influence correlations among the dynamic of species populations: intra- and inter-specific interactions, differing responses to environmental variations, and demographic stochasticity (Roscheret al. 2011). A key factor for insect communities is the response to environmental fluctuations (Ives et al. 1999). The distributions of butterflies are strongly influenced by climate (Setteleet al. 2008) and population dynamics by weather (WallisDeVrieset al. 2011; Palmer et al. 2017). Consequently, responses to environmental variation could be important in understanding differences in stability. However, a challenge when linking weather to synchrony in natural communities is that the importance of environmental variables will vary among species (Roy et al. 2001; WallisDeVrieset al. 2011; McDermott Long et al. 2017).
Additionally, at local scales species richness and abundance may vary due to the suitability of sites, or at large scales, due factors such as latitudinal gradients in richness (Hillebrand 2004). For example, the abundant centre hypothesis posits that species will be most abundant at locations near the centre of their range (Andrewartha & Birch 1954; Brown 1984; Lawton 1993), and so sites located near many species’ range centres should have higher population stability due to mean-variance effects (i.e. larger populations being relatively less variable over time) (Oliver et al. 2012, 2014). The importance of asynchrony or population stability may then vary due to community structure, e.g. differing levels of interspecific competition (Lehman & Tilman 2000). Finally, varying sensitivities to environmental variation (Ives et al. 1999) may then drive differences in population variability and contribute to the degree of asynchrony in a community.
To tackle this complexity, we apply the ecological niche concept, i.e. a species tolerance to different environmental variables, can be represented as an n-dimensional space (Hutchinson 1957). Fundamental to niche modelling, at both the species and community scales (Hirzel & Le Lay 2008; Poggiato et al. 2021), is that species occurrence under local conditions is dictated predominantly by the niche space. Consequently, notwithstanding nuances of extinction debt and colonisation credit (Tilman et al. 1994; Kuussaari et al.2009), species richness should be predictable from the location of a site relative to the niches of the species in the regional species pool. Similarly, mean abundance should be informed by site position relative to the niches of the species at the site (Osorio‐Olvera et al.2020). Species at the edge of their niche may also have lower average population stability (Oliver et al. 2012; Mills et al.2017) and species with larger niche breadths may be more robust to local environmental variation. Finally, when a site is in a different niche position for two species, they may have differing responses to environmental variation at the site – generating asynchrony.
We apply the niche-based approach to study 140 butterfly communities spanning a European scale from Finland to Northeast Spain. We utilise a computationally fast approach to generate bioclimatic niche hyper-volumes (Blonder et al. 2014, 2018) for each species. We then derive niche-based metrics to test potential factors affecting first species richness, mean abundance, and then the mechanisms of community stability. Specifically, we test five expected responses associated with key mechanisms: 1) species richness will be higher if the site is located at shorter distances to the niche centres of the total species pool; 2) mean abundance will be greater when the site is nearer to the niche centres of the species at the site; 3) average asynchrony between pairs of species in a community will be greater if the niches of species are more dissimilar at a site, and will also increase with species richness; 4) average population stability will be highest when the site is nearer the centre of the species niches, when the average niche breadth of species is larger (i.e. more resilient to local weather anomalies), and when the average abundance of the species is greater; 5) asynchrony and average population stability will explain differences in community stability.