SEARCH

SEARCH BY CITATION

Keywords:

  • adaptive management;
  • butterfly ecology;
  • conservation;
  • distribution modelling;
  • global warming;
  • metapopulation dynamics;
  • patch area;
  • realised niche

Summary

  1. Top of page
  2. Summary
  3. Introduction
  4. Materials and methods
  5. Results
  6. Discussion
  7. Conclusion
  8. Acknowledgements
  9. References

1. There is a pressing need to understand how to facilitate species’ range shifts under climate change. However, few empirical studies are available to inform decision-making, particularly at fine spatial and temporal resolutions.

2. We present a case study of a thermally constrained habitat specialist, the silver-spotted skipper butterfly Hesperia comma, at its expanding range margin in south-eastern Britain. Using data from 724 habitat patches over a 9-year interval (2000–2009), we examined local and landscape determinants of colonization, survival and population density. We then predicted probabilities of colonization and survival for habitat patches under the observed 1982 and 2009 distributions to investigate how the factors limiting range expansions change through space and time.

3. Between 2000 and 2009, Hesperia comma continued to expand its range in Britain, but the 67 recorded colonization events were offset by 48 local extinctions. Extinctions were strongly linked to climate, occurring predominantly in cooler regions and on shaded north-facing slopes.

4. Population density and probability of survival were closely related to conditions within a site, whereas probability of colonization was largely determined by functional connectivity. Survival probability was also influenced by connectivity, suggesting that immigration helped to support extinction-prone populations (a ‘rescue effect’).

5. Patch occupancy beyond the range margin was primarily constrained by colonization, but close to the expanding front, population survival became the key limiting factor. This pattern was conserved during range expansion, altering management priorities at individual sites.

6.Synthesis and applications. Previous studies on facilitating range shifts have stressed the need to increase landscape-scale connectivity to remove constraints on colonization, and our data substantiate this advice. However, we show that enhancing population survival can also help to facilitate range expansions, because populations at leading range edges face high extinction risk. Population survival can be improved directly through local management actions, such as enlarging patch size and increasing habitat quality, or indirectly by improving connectivity. Thus, local management can secure vulnerable populations at the range edge and provide larger and more stable migrant sources for future expansion and deserves consideration when facilitating range shifts under climate change.


Introduction

  1. Top of page
  2. Summary
  3. Introduction
  4. Materials and methods
  5. Results
  6. Discussion
  7. Conclusion
  8. Acknowledgements
  9. References

Climate is one of the primary determinants of species’ distributions, both directly through impacts on physiology and indirectly through its influence on interacting organisms (Parmesan & Yohe 2003; Hickling et al. 2006; Thomas 2010). As climates advance polewards and uphill under global warming (Loarie et al. 2009), the suitability of species’ current habitats will change (Thomas et al. 2004; Araujo, Thuiller & Pearson 2006). This presents potential risks for species at their upper thermal limits, and potential benefits for species at their lower thermal limits, generating needs and opportunities for conservation (Hulme 2005; Thomas et al. 2010). Consequently, adapting management to climate change has become a priority (Hulme 2005; Mitchell et al. 2007).

Broadly, there are two potential courses of action for conserving climate-sensitive species: improve species’ abilities to cope with climate change in habitats within the existing range that are becoming less climatically favourable, and shepherd species towards uncolonized habitats that are becoming climatically suitable (Heller & Zavaleta 2009; Lawler 2009; Thomas et al. 2010). As global temperatures continue to rise (Brohan et al. 2006), there will be a shift in emphasis from preserving species’ current distributions towards the need to facilitate range shifts (Galatowitsch, Frelich & Phillips-Mao 2009). However, as other factors aside from climate constrain species occurrence, the fragmented distribution of suitable habitats can impair range shifts, particularly for species with highly specialized habitat requirements (Warren et al. 2001).

In fragmented habitats, the dual issues are how species survive as existing local populations within fragments (or ‘habitat patches’), and how they spread (colonize) from patch to patch across the landscape, as they expand their distributions (Hanski 1999). Current thinking on managing range expansions centres on increasing the number of colonization events by boosting dispersal among habitat fragments (i.e. increasing connectivity or, more controversially, assisted colonization; Heller & Zavaleta 2009; McLachlan, Hellmann & Schwartz 2007; Hoegh-Guldberg et al. 2008; Richardson et al. 2009). Commonly, coordinated landscape-scale schemes are proposed in which per capita dispersal success is enhanced through increased habitat aggregation, corridor creation or improved matrix permeability (Lovejoy & Hannah 2005; Manning, Gibbons & Lindenmayer 2009; Krosby et al. 2010). However, connectivity may also be improved by augmenting the number of dispersing individuals through measures that increase the number or size of extant populations (Hodgson et al. 2009).

Despite the potential benefits of region-wide conservation planning, in reality much of landscape structure emerges from the semi-independent actions of many different practitioners and landowners in separate habitat patches (Heller & Zavaleta 2009). Such site-specific or local management is thought to increase population size by improving the areal extent or quality (maximum population density) of a given habitat patch (Hodgson et al. 2009). As larger populations are known to be more resistant to extinction (Hanski 1999), local management is often prescribed to build resistance to climate change at trailing range margins to preserve species’ current distributions (Pearce-Higgins et al. 2011). In contrast, local management is rarely considered in discussions on facilitating range shifts, perhaps because many species distribution models assume that only dispersal limitation prevents species from occupying habitats that are predicted to become climatically suitable (Huntley et al. 2010). However, on short time-scales, inter-annual fluctuations in climate dwarf the long-term trend (Opdam & Wascher 2004), such that newly founded populations at the range edge may require protection.

In practical terms, the key issue is which management strategies to employ under different circumstances (Hoegh-Guldberg et al. 2008). Unfortunately, the existing literature on facilitating range shifts remains largely theoretical, with few empirical studies to guide management, especially at a spatial resolution fine enough to examine local-level impacts (Heller & Zavaleta 2009). This has left current advice lacking specificity (Heller & Zavaleta 2009), making it difficult for practitioners to determine the most appropriate actions for particular species or landscapes. Studies on species with specialized habitat requirements are particularly important in this context, as specialist species are likely to have greater difficulty tracking climate change because of the comparative scarcity of habitable land area (Warren et al. 2001; Wilson, Davies & Thomas 2010).

Here, we present a case study of a temperature-sensitive species (the silver-spotted skipper butterfly, Hesperia comma) with specialized habitat requirements, at its leading range margin in south-eastern Britain. Using data collected across 724 patches at a 9-year interval, we identify local and landscape determinants of population establishment, density and survival. This empirical evidence is used to draw inferences about the following questions:

  • 1
     Does local management (increasing the areal extent or quality of occupied habitat) have a role to play at expanding margins?
  • 2
     Does landscape management (increasing the connectivity of habitat patches) help to facilitate range expansions?
  • 3
     How do the chances of colonization and survival, and hence local management priorities, change in space across an expanding range margin and in time as a species expands?

Our results confirm the importance of landscape-scale connectivity in managing range shifts, but additionally show that improving survival through local management represents a key strategy at expanding range margins.

Materials and methods

  1. Top of page
  2. Summary
  3. Introduction
  4. Materials and methods
  5. Results
  6. Discussion
  7. Conclusion
  8. Acknowledgements
  9. References

Study System

The silver-spotted skipper butterfly Hesperia comma is a habitat specialist; in Britain, it is confined to calcareous grassland, laying eggs on a single host plant species, sheep’s fescue grass Festuca ovina (Thomas et al. 1986). Only short tufts (<10 cm) of F. ovina are selected for oviposition, restricting the butterfly to sites with intermediate to high levels of grazing (Thomas et al. 1986). In H. comma, both fecundity and egg-laying microhabitat availability are temperature-dependent: females show increased egg-laying rates at higher temperatures, and select warm microclimates, next to patches of bare ground, for oviposition (Davies et al. 2006).

Hesperia comma reaches its northern range limit in Britain (Fig. 1), where historically, populations have been confined to hotter south-facing slopes (Thomas et al. 1986). However, over the past 30 years, warming summer temperatures have broadened the range of microhabitats suitable for egg-laying (Davies et al. 2006) and permitted the colonization of cooler north-facing habitats (Thomas et al. 2001). This, together with more widespread grazing from rabbits and livestock (Thomas & Jones 1993), has increased the availability of suitable breeding habitat, catalysing a range expansion from fewer than 70 populations in 1982 to over 250 by 2000 (Davies et al. 2005; Wilson, Davies & Thomas 2010). However, the fragmented distribution of these habitat patches and the species’ limited dispersal abilities (Hill, Thomas & Lewis 1996) have constrained the rate of this expansion (Wilson, Davies & Thomas 2009, 2010).

image

Figure 1.  Maps of recorded changes in Hesperia comma site occupancy between 2000 and 2009 across south-eastern Britain, with chalk geology shown in white.

Download figure to PowerPoint

Survey Methods

In 2009, we conducted the fourth UK national survey of H. comma’s distribution (for previous surveys in 1982, 1991 and 2000 see Thomas et al. 1986; Thomas & Jones 1993; Davies et al. 2005). As H. comma is univoltine, this represents nine generations since the previous survey (Davies et al. 2005). The search was restricted to five main habitat networks in south-east England (in the counties of Kent, Sussex, Surrey, Hampshire, and the Chiltern hills), which encompass the majority of H. comma populations in Britain (Davies et al. 2005). As in previous surveys, suitable habitat patches were defined as any unimproved chalk grassland containing more than an estimated 5% cover of F. ovina plants <10 cm tall, and neighbouring patches were defined as separate if their nearest perimeter points were divided by at least 25 m of unsuitable grassland, or a woodland or scrub barrier. All habitat patches meeting these criteria within a 30 km radius of known populations were surveyed; this radius is considered sufficient to detect all new populations because it exceeds the maximum recorded colonization distance over an 18-year period (1982–2000; Davies et al. 2005), and because the species was not recorded between 2001 and 2008 more than 13 km from populations known in 2000 by the British ‘Butterflies for the New Millennium’ recording scheme (R. Fox, pers. comm.).

Once located, the perimeter of each patch was mapped using a hand-held Global Positioning System (GPS; accuracy <±10 m) and later digitized using ArcMap software (ESRI 2009). Habitat characteristics that could affect H. comma occupancy were recorded, including the total area of the patch in hectares, the percentage cover of bare ground and vegetation <10 cm tall and proportional host plant cover in vegetation <10 cm (Table 1). Patches were then searched for the presence of H. comma, based on either observation of adults or timed egg searches.

Table 1.   Explanatory variables and the analyses they are used in: survival (S), colonization (C), and/or population density (P)
Variable nameSymbolDescriptionSpatial scaleAnalyses
Area <10 cmAREffective area (ha) of turf <10 cm tallLocalS, C
Host plant coverHOProportional coverage (%) of Festuca ovina in turf <10 cmLocalS, C, P
Bare ground coverBGProportional coverage (%) of bare ground in patchLocalS, C, P
Solar indexINIncident solar radiation at midday in mid-AugustLocalS, C, P
MacroclimateMCMean August maximum temperature over 9-year periodLocalS, C, P
Direct connectivityCDConnectivity to occupied sitesLandscapeS, C
Indirect connectivityCIConnectivity to initially unoccupied sites, weighted by connectivity to occupied sitesLandscapeS, C
<10 cm coverTEProportional coverage (%) of turf <10 cm tall in patchLocalP
Population ageAGECategorical variable indicating whether population was founded after 2000 (‘young’) or prior to 2000 (‘old’)LocalP

For patches visited in favourable weather during the flight period, adult densities were estimated using transect walks (Pollard & Yates 1993). Using weekly H. comma counts from UK Butterfly Monitoring Scheme (BMS) transects, we calculated the proportion of peak abundance on the day each transect was walked. We then divided observed abundance on survey transects by proportion of regional peak to estimate peak density at each site, setting a minimum proportion of 20% to avoid excessively large density adjustments.

Variable Collation

The area of each patch (ha) was calculated using digitized patch perimeters in ArcGIS. By multiplying patch area by proportion of the patch with vegetation <10 cm tall, an estimate of effective breeding area was obtained (Table 1).

To assess landscape-scale gradients in macroclimate during the adult flight season, we calculated the mean August daily maximum temperature for each site for two separate periods (1982–1991 and 2000–2009), using the UKCP09 5 km resolution gridded observation data set from the UK Meteorological Office (Perry & Hollis 2005). To assess patch-scale differences in microclimate, we calculated incoming solar radiation as a function of aspect and slope: we applied the ‘hillshade’ function to a 5-m resolution digital terrain model in ArcMap (Intermap Technologies 2007; vertical accuracy ±60 cm), using solar azimuth of 180° and altitude 60° (equivalent to the maximum solar radiation in south-eastern UK during mid-August), and extracted the median solar index for each patch using the spatial analyst tool (ESRI 2009).

As a measure of potential immigration into each patch, we calculated functional connectivity using Hanski’s connectivity index:

  • image

Where i is the focal patch and j all other patches, which have area Aj and are separated from i by distance dij (Hanski 1999). Here, Aj is effective area <10 cm (ha) of patch j, and dij edge-to-edge distances between patches i and j (km), both based on 2009 data. α (a negative exponential dispersal kernel) and b (a scaling function for patch emigration) are estimated from a previous study (Wilson, Davies & Thomas 2010). In the original formulation, pj denotes the presence or absence of the butterfly at site j and is here used in two different ways. ‘Direct connectivity’ calculates connectivity to occupied patches (pj = 1), discounting unoccupied patches (pj = 0). ‘Indirect connectivity’ calculates connectivity to unoccupied patches, weighting those patches by their direct connectivity scores (i.e. taking area of patch j as AjSj). Therefore, whilst direct connectivity is a measure of the probability of colonization in a single dispersal event from currently occupied patches, indirect connectivity is a measure of the probability of colonization in two dispersal events using currently unoccupied habitat as a ‘stepping stone’. Both connectivity measures were calculated for occupancy patterns in 3 years: 1982, 2000 and 2009.

Analyses

Analyses were conducted in two stages. First, we examined whether establishment, survival and density of H. comma populations from 2000 to 2009 could be explained by local and regional factors (Table 1). Second, we used the best models to predict the probabilities of survival and colonization across all patches for two time periods (1982–1991 and 2000–2009). All models were fitted using generalized linear modelling (Crawley 2007) in r 2.12.2 (R Development Core Team 2011).

Survival models considered existing populations in 2000, with a binary response of ‘survival’ or ‘extinction’ by 2009; colonization models considered unoccupied patches in 2000, with the response as ‘colonized’ or ‘uncolonized’. In addition to present information from the 2009 survey, colonization analyses included temporary colonizations based on additional information from surveys conducted by the authors in 2001 and 2002, and records from 2001 to 2008 from Butterfly Conservation, the organization which coordinates the UK butterfly distribution monitoring scheme. We considered the butterfly as absent from any site not surveyed in 2000 that was more than 5 km from an occupied patch. Population density analyses focussed on density data from the 2009 survey.

We assumed a binomial error distribution for survival and colonization models. Population density models assumed a Tweedie error distribution, a compound poisson-gamma distribution family suited to analysing positive continuous data with exact zeroes; parameters were estimated using the ‘tweedie’ package (Dunn 2010).

Explanatory factors included local and landscape variables (Table 1). Local variables comprised solar index and habitat characteristics (Table 1; see also Survey methods). A squared bare ground term was included because intermediate levels of bare ground are likely to be optimal for this species. We used habitat data gathered in 2000 in survival analyses, but because these data were not available for all unoccupied patches in 2000, we used 2009 habitat data in colonization analyses. Landscape variables included macroclimate, and direct and indirect connectivity (Table 1).

Population density analyses considered only local variables from 2009 and macroclimate (Table 1). Density analyses used proportion of turf <10 cm instead of effective area <10 cm, because the former investigates how density increases with breeding habitat independently of patch size. An additional variable, ‘age’, was also included, which indicated whether a population was present in 2000 (‘old’) or had been founded after 2000 (‘young’).

Data were checked for spatial autocorrelation (Beale et al. 2010) using Mantel tests with the ‘vegan’ package (Oksanen et al. 2011). There was no evidence that population density estimates were spatially correlated (r = −0·043, P > 0·05). We expected some spatial autocorrelation in survival and colonization data because of dispersal between neighbouring patches, as explicitly modelled by connectivity terms. Therefore, to test for autocorrelation over and above that accounted for by these covariates, we conducted mantel tests of the residuals from models including solely the two connectivity measures. Although there was no significant autocorrelation in colonization model residuals (r = −0·0080, P > 0·05), there was some sign of autocorrelation remaining in survival residuals (r = 0·24, < 0·01). Consequently, we fitted mixed models including a random effect of ‘grid square’, classifying patches into spatial blocks of (i) 5 km and (ii) 10 km squares, using the ‘lme4’ package (Bates & Maechler 2010). However, top model sets were similar and effect sizes were within standard error bounds of those in the original models, indicating that spatial autocorrelation was not seriously affecting model parameters.

For each response variable, we fitted all possible combinations of linear terms, with no interactions, and used the Akaike Information Criterion, adjusted for small sample size (AICc; Burnham & Anderson 2002) to rank models. To obtain our best model sets, we selected models that were within six AICc units of the top-ranked model (Richards 2005), excluding models with simpler, higher-ranking nested variants (Richards 2008). This procedure guards against the selection of over-parameterized models whilst maintaining a high probability of selecting the true best model (Richards 2008).

To further explore which variables best explained survival and colonization, and to examine how variance was partitioned among them, we conducted hierarchical partitioning (Mac Nally 2002), implemented with the ‘hier.part’ package (Mac Nally & Walsh 2004). Hierarchical partitioning investigates the average change in fit (here, log-likelihood) between equivalent models with and without a given variable X, to assess the explanatory power of X independently of other terms (Ix). By randomizing the data set 1000 times and recalculating Ix, we estimated the probability of obtaining a value of Ix equal to or greater than that observed by chance, allowing us to assign statistical significance to the explanatory power of each variable (Mac Nally 2002). In addition, the average effect of other variables on the explanatory power of X (‘joint’ effects, JX) can be calculated; their direction indicates whether other variables are acting additively (positive), increasing the variation explained by X, or suppressively (negative), sharing variation with X (Mac Nally & Walsh 2004).

Following model selection, we applied our best model sets to predict the probabilities of colonization and survival across the metapopulations. Predictions were based on distribution data from either 1982 or 2009; in both cases, local variables (Table 1) were based on 2009 data because habitat information was not available for all patches in 1982. We used model-averaging (Burnham & Anderson 2002) implemented by the ‘AICcmodavg’ package (Mazerolle 2010) to obtain a single predicted survival and colonization probability for each patch. Doing so incorporates model selection uncertainty whilst weighting the influence of each model by the strength of its supporting evidence (Burnham & Anderson 2002).

To identify how survival and colonization limitations varied across the British distribution of H. comma, we classified each patch into one of four categories, based on whether they were primarily limited by survival, colonization, both (‘marginal’ habitat), or neither (‘supported’ habitat).

Results

  1. Top of page
  2. Summary
  3. Introduction
  4. Materials and methods
  5. Results
  6. Discussion
  7. Conclusion
  8. Acknowledgements
  9. References

Between 2000 and 2009, 67 new patches were colonized, 168 populations survived, whilst 48 populations suffered local extinctions (Fig. 1). In addition, records from the intervening years detected 29 temporary colonizations that subsequently became extinct before 2009, giving a minimum of 96 colonizations and 77 local extinction events (Fig. 1).

Population survival was closely associated with climatic factors, being more likely in warmer regions and on south-facing slopes (macroclimate: Z = 5·03, < 0·0001; solar index: Z = 3·13, < 0·0001; Fig. 2a; Table 2a). Solar index did not influence the chance of colonization (Z = −0·26, P > 0·05; Fig. 2b). The relationship between macroclimate and colonization was equivocal; although hierarchical partitioning indicated a significant relationship (Z = 2·75, < 0·01; Fig. 2b), the direction of this effect varied, much of its explanatory power was shared with other variables (large ‘joint’ effect, Fig. 2b), and macroclimate did not appear in any of the top-ranked colonization models (Table 2b).

image

Figure 2.  Independent (black) and joint (grey) contributions of explanatory variables (Table 1) in (a) survival and (b) colonization models, based on hierarchical partitioning analyses and expressed as mean per cent change in log-likelihood. Asterisks indicate significance of independent effects (< 0·05, 0·01, 0·001 and 0·0001 denoted by one, two, three and four asterisks, respectively).

Download figure to PowerPoint

Table 2.   Top-ranking models for (a) survival, (b) colonization and (c) population density analyses
RankKΔAICc%LLFormula
  1. K, number of parameters; ΔAICc, difference in Akaike Information Criterion (AICc) between current and best model; %LL, per cent change in log-likelihood from null model.

  2. Formula indicates which terms were included in the model (Table 1) and the direction of their coefficients (±).

(a) Survival (n = 214)
170·000·25+ AR + IN + CD + MC + BG − BG2
272·840·24+ AR + IN + CD + BG − BG2 − CI
362·920·23+ AR + IN + CD + BG − BG2
463·200·23+ AR + IN + CD + MC + HO
554·260·21+ AR + IN + CD + MC
(b) Colonization (n = 510)
140·000·27+ CD + CI + HO
252·720·27+ CD + CI + BG − BG2
333·940·26+ CD + CI
(c) Population density (n = 142)
160·000·04+ AGE + IN + TE + HO
250·350·04+ AGE + IN + TE
351·050·03+ AGE + IN + HO
441·590·03+ AGE + TE
552·020·03+ AGE + IN + BG
642·830·03+ AGE + IN
743·200·03+ AGE + HO
843·840·03+ AGE + BG
934·530·02+ AGE

Survival was also closely linked to local management variables (Fig. 2a; Table 2a). There was strong evidence for an effect of patch area (Z = 3·81, < 0·001), with populations in larger patches more likely to survive (Table 2a). Survival was also more likely in higher quality patches with increased host plant availability (Z = 2·31, < 0·05). The top-ranked survival models supported a squared relationship with bare ground (Table 2a), indicating highest survival at sites with intermediate (around 15%) bare ground cover.

Colonization was also more likely for patches with greater host plant cover and an intermediate proportion of bare ground (Host plant cover: Z = 8·65, < 0·0001; Bare ground: Z = 1·80, < 0·05; Fig. 2b, Table 2b), with around 22% bare ground as the predicted optimum. However, the areal extent of a patch did not affect its chance of colonization (Z = −0·38, P > 0·05; Fig. 2b).

Survival and colonization increased with direct connectivity to occupied sites (Survival: Z = 16·05, < 0·0001; Colonization: Z = 26·75, < 0·0001; Fig. 2). Indirect connectivity improved colonization chances (Z = 55·14, < 0·0001; Table 2b), but had little impact on survival of pre-existing populations (Z = 0·17, P > 0·05; Fig. 2a). The high correlation between direct and indirect connectivity means that they share explanatory power (large ‘joint’ effect, Fig. 2b).

We obtained density estimates for 142 populations. Population density models corresponded well with survival analyses: density was highest at south-facing sites with high proportions of short vegetation (<10 cm), F. ovina, and bare ground (Table 2c). Density was lower in more recently established populations (Table 2c).

To illustrate spatial and temporal variation in factors limiting the persistence and expansion of H. comma, Fig. 3b displays model predictions for (i) 1982–1991 and (ii) 2000–2009 for the Sussex network. Close to the expanding front are clusters of ‘survival-limited’ sites which have a high chance of colonization, but a low probability of supporting a population in the long term. Further beyond the leading edge, there are scattered ‘colonization-limited’ patches, which are relatively unlikely to be colonized over a 9-year period, but, if colonized, have a high chance of supporting H. comma populations because of their large size and/or high quality. With the expansion of H. comma, many formerly isolated patches now have closer sources of migrants available and hence improved chances of colonization and/or survival (‘supported’ patches, Fig. 3).

image

Figure 3.  Map of patches in Sussex network, classified according to predicted probabilities of colonization and survival for (a) 1982–1991 and (b) 2000–2009. Hesperia comma occupancy at the start of each period is indicated by filled (present) and open (absent) circles.

Download figure to PowerPoint

Discussion

  1. Top of page
  2. Summary
  3. Introduction
  4. Materials and methods
  5. Results
  6. Discussion
  7. Conclusion
  8. Acknowledgements
  9. References

The empirical evidence base to inform management of species’ range shifts is currently lacking (Heller & Zavaleta 2009). Here, we present a case study of a thermally constrained habitat specialist, the silver-spotted skipper butterfly Hesperia comma, which is expanding its northern range limit in Britain under rising temperatures (Brohan et al. 2006). The results allow us to draw conclusions regarding the roles of local- and landscape-scale processes in adapting conservation management to climate change.

Determinants of Extinction and Colonization

Although H. comma continued to expand its distribution in Britain between 2000 and 2009, there were also a number of setbacks, with 48 local extinctions offsetting the 67 recorded colonization events. Populations in habitats of lower climatic suitability – in cooler regions and on north-facing slopes – were most likely to suffer extinction. Two non-mutually exclusive explanations are that (i) populations in less suitable climates were smaller (Hanski 1999) or (ii) the dynamics of populations at the range edge showed greater fluctuations with weather (Thomas, Moss & Pollard 1994; Opdam & Wascher 2004). Although we cannot examine inter-annual variability in population size based on the data presented here, we found that population densities were smaller in both cooler microclimates and in more recently founded populations. Thus, although warming temperatures permitted expansion into habitat that was previously too cool (Thomas et al. 2001), the vulnerability of these new populations remained higher than those in core habitats.

Several previous studies have suggested that range edge populations may face heightened extinction risk in comparison with those in core habitats. Indirect evidence is provided by analyses of population density across species’ ranges: in many but not all cases, it has been shown that population densities decline towards range margins (Sagarin & Gaines 2002). A few studies have also gathered direct evidence that extinction risk increases in climatically marginal habitats at leading margins. For example, a study on ringlet butterflies Aphantopus hyperantus indicated a retraction to core (partially shaded) habitats following a period of drought (Sutcliffe et al. 1997), whilst Mehlman (1997) found that populations of three North American passerine bird species were more likely to go extinct at sites which experienced more severe winters or were closer to the range edge. Our study shows that populations in climatically marginal habitats remain vulnerable when a species is expanding its distribution under increasingly favourable climates.

As population survival, and not just establishment, limits range expansions under climate change, species’ range shifts could be enhanced through management to improve population persistence. Local measures such as increasing patch area have been proposed as an option for building resistance to climate change at contracting margins (Pearce-Higgins et al. 2011), but are rarely considered in the context of expanding margins. We found that site-scale factors promoted the survival of populations at the expanding front. Survival was higher in larger patches, either because larger patches support larger populations (Hanski 1999) or perhaps because larger patches capture greater environmental heterogeneity which buffers populations against climatic variability (Opdam & Wascher 2004; Oliver et al. 2010). Moreover, survival was also improved in higher quality patches (those with higher host plant availability and optimal bare ground cover). These results indicate that local management can be important for securing species’ footholds in habitats at the range edge and thus, help to facilitate range expansions.

In contrast to population survival, the chance of a patch being colonized was not strongly related to local management, although we did find evidence that higher host plant availability increased colonization chances. Instead, there was an overriding positive effect on colonization probability of direct and indirect connectivity. The latter indicates that well-connected networks of habitat, even if initially unoccupied, can facilitate ‘stepping stone’ colonizations, either within generations (connected habitat patches are used as migration routes) or between generations (unoccupied patches are colonized and serve as migrant sources in subsequent years), or both (Lawler 2009). This provides empirical evidence, in the context of an expanding range margin, that beliefs about connectivity enhancing range expansions are well-founded (see also Melles et al. 2011).

Population survival was also improved by increased connectivity, indicating a possible role for rescue effects (Hanski 1999), or re-colonizations within the 9-year survey interval. We found little evidence that indirect connectivity influenced survival, perhaps because most patches colonized between 2000 and 2009 were too remote from most pre-existing populations to provide rescue effects. Few discussions on managing range shifts have explicitly considered how increasing connectivity could improve persistence of populations at the expanding front. Our data, using a fine enough spatial resolution to identify individual patches and populations, and a temporal resolution of nine generations, show that increased connectivity enhances both colonization and survival.

As both patch area and proximity are incorporated in our connectivity terms, the results do not allow us to differentiate among measures that increase per capita dispersal success (e.g. corridors or matrix management; Lovejoy & Hannah 2005; Krosby et al. 2010) and measures that augment the number of migrants by increasing the number or size of extant populations (Hodgson et al. 2009). However, because our results show that both population density and persistence, and hence migrant numbers, increased with patch size and quality (Table 2a,c), it follows - assuming a causal relationship between these variables - that local management also increases landscape-scale connectivity. Thus, management to improve patch size or quality brings both direct local and indirect regional benefits (Hodgson et al. 2009).

Implications for Managing a Range Expansion

Current management recommendations for H. comma focus on local management, with measures increasing the effective size of habitat patches through scrub removal, and grazing to control sward height and bare ground availability (Barnett & Warren 1995; Davies et al. 2005). Landscape-scale management of connectivity is less straightforward to achieve, potentially involving coordinated management among several landowners, although voluntary agri-environment schemes such as Environmental Stewardship have successfully improved connectivity for some taxa (Donald & Evans 2006). Local prioritization of management strategies will depend on the current chances of colonization by and survival of H. comma (Table 3).

Table 3.   Site classification scheme and management recommendations
CategoryColonization probabilitySurvival probabilityRecommended management actions
SupportedHigh (>50%)High (>50%)Continue existing management Monitor for changes and adapt management if necessary
Colonization-limitedLow (<50%)High (>50%)Increase connectivity of surrounding habitat through matrix softening or creation of corridors/stepping stones Increase patch quality through grazing management Assisted colonizations where viable Await further spread
Survival-limitedHigh (>50%)Low (<50%)Increase patch size through habitat restoration (e.g. scrub removal) Increase patch quality through grazing
MarginalLow (<50%)Low (<50%)Low priority May become more important as species spreads

Applying our models to predict colonization and survival across the British distribution of H. comma for 2009–2018 showed that colonization and survival are differentially limiting across the species’ range margin, because of among-patch differences in size and quality (influenced by local management) and connectivity to existing populations (influenced by landscape management and the species’ current distribution; Fig. 3b). Because connectivity (direct and indirect) is the primary determinant of colonization probability, colonization becomes increasingly limiting as distance beyond the expanding front increases (Davies et al. 2005), whereas survival is strongly influenced by local factors and is therefore less constrained by proximity to H. comma’s current distribution. Practically, this means that whilst survival limitation can be remedied through site-specific actions (Table 3; survival-limited and marginal patches, Fig. 3b), there is less that local management can do to directly encourage colonization of a patch. Nonetheless, knowing the chance of colonization can usefully inform site-scale management decisions (Table 3). For large, south-facing, but isolated patches (colonization-limited patches, Fig. 3), artificial introductions might be considered, because their success rate is likely to be high (McLachlan, Hellmann & Schwartz 2007; Hoegh-Guldberg et al. 2008; Richardson et al. 2009). Alternatively, as managers must balance decisions for many species with different requirements and distributions, local management for species that are currently unlikely to reach a site may represent a low priority.

However, as species expand their distributions, connectivity to extant populations, and hence local management priorities, can change rapidly. With the proliferation of H. comma across Sussex, owners of previously isolated sites may consider management for H. comma of higher importance now that colonization chances have improved, whilst formerly survival-limited patches on the periphery of the H. comma distribution are now well-supported by rescue effects, making local management to improve population survival less of a necessity (Fig. 3). This rapid and widespread shift in priorities suggests that, if management is to be conducted on an independent, site-by-site basis, regular monitoring of species shifting distributions (Lepetz et al. 2009) and flexible decision-making (Lovejoy & Hannah 2005) will be required to maximize efficiency. However, there are two reasons why coordinated landscape-scale management might prove preferable to independent local management. First, retrospective adjustments of policy might be too slow for species with particularly dynamic patterns of population turnover or rapid rates of spread. If so, predictive modelling (Hannah et al. 2007; Huntley et al. 2010) could help management to keep pace with species’ shifting distributions. Second, connectivity will only increase if habitat availability provides the opportunity to expand, such as in the Sussex network here. Where habitat patches are more isolated, such as in Hampshire (Fig. 1), colonization may remain limiting pending increased habitat connectivity, artificial introductions, or rare long-distance colonization events (McLachlan, Hellmann & Schwartz 2007).

The simple models employed here are not intended to provide an exact blueprint for the management, because they assume that correlations observed between environmental variables and population survival or colonization are causal, and assess limiting factors for each site based on a single point in time, ignoring the complex interplay among changes in management practices, occupancy patterns and connectivity. Nonetheless, Fig. 3 demonstrates how, for H. comma at least, survival is the primary limiting factor close to the leading edge and remains so as the distribution shifts, potentially slowing the rate of expansion. Our results provide encouragement that local management can help to alleviate survival limitation at an expanding front and provide larger and more stable platforms for future expansion.

Conclusion

  1. Top of page
  2. Summary
  3. Introduction
  4. Materials and methods
  5. Results
  6. Discussion
  7. Conclusion
  8. Acknowledgements
  9. References

Range shifts under climate change are not a continuous march towards pastures new, but a dynamic process resulting from an imbalance in extinction and colonization events. Consequently, founding new populations at the leading edge is only the first step towards a lasting range expansion, because populations at the range edge face heightened extinction risk. As colonization rates may be high close to species’ current distributions, survival can become the primary factor limiting expansion at the range edge. There are thus two complementary strategies for facilitating species’ range shifts: encourage colonization of unoccupied habitat through landscape management for connectivity and/or assisted colonization; and support populations by increasing patch size or quality, or providing nearby migrant sources, wherever new footholds are established. As local management can increase both population persistence and connectivity, it represents an important strategy for facilitating range shifts under climate change.

Acknowledgements

  1. Top of page
  2. Summary
  3. Introduction
  4. Materials and methods
  5. Results
  6. Discussion
  7. Conclusion
  8. Acknowledgements
  9. References

2009 fieldwork was carried out with the help of Alice Daish, Tracy Gray, Bonnie McBride, Crispin Holloway, Tim Bernhard, Phil Budd, Mark Edwards, Rachel Hoyes, Jenna Poole, Helen Silver and Tim Yardley. Additional monitoring and distribution data were provided by Tom Brereton, Richard Fox and Ian Middlebrook (Butterfly Conservation), David Roy (Centre for Ecology and Hydrology) and Zoe Davies and Dan Hoare; we thank the many volunteers who collected these data (including Richard Soulsby), and the many landowners and managers who provided access for the survey. We thank Karen Anderson, John Hopkins, the editor and two anonymous reviewers for helpful comments. The survey was funded by the UK Natural Environment Research Council (NERC) grant NE/G006296/1. CRL is funded by a Climate Change and Sustainable Futures studentship from the University of Exeter.

References

  1. Top of page
  2. Summary
  3. Introduction
  4. Materials and methods
  5. Results
  6. Discussion
  7. Conclusion
  8. Acknowledgements
  9. References