Skip to main content
NCBI home page
Wiley Open Access Collection logoLink to Wiley Open Access Collection
. 2025 Jul 25;34(17):e70037. doi: 10.1111/mec.70037

Relative Effects of Habitat Amount and Fragmentation Per Se on Genetic Diversity of the Glanville Fritillary Butterfly

Lola Fernández Multigner 1,2,✉, Audrey Bras 1,2, Michelle F DiLeo 3, Marjo Saastamoinen 1,2
PMCID: PMC12376959  PMID: 40709589

ABSTRACT

Habitat loss and fragmentation are considered key drivers of biodiversity loss. Understanding their relative roles is difficult as habitat loss and fragmentation tend to co‐occur. It has been proposed that the total habitat amount available in the local landscape mainly drives species richness, while fragmentation per se—the breaking apart of habitat independent of habitat amount—has a negligible or even a positive effect on biodiversity. Several studies support this at the species richness level. Yet, the potential effects of fragmentation per se on genetic diversity at the landscape scale are understudied. Using the Glanville fritillary butterfly metapopulation in the Åland islands, we tested the effects of fragmentation on genetic diversity using a landscape‐based approach and 2610 individuals genotyped at 40 neutral SNP markers. We assessed the independent effect of habitat amount and fragmentation (i.e., number of patches, habitat aggregation) within the local landscape on focal patch genetic diversity. The amount of habitat in the local landscape had a positive effect on genetic diversity. Fragmentation measured as the number of patches within the landscape had a negligible impact on genetic diversity, whereas habitat aggregation had a negative effect on genetic diversity when the available amount of habitat in the landscape was low. Focal patch size increased genetic diversity, whereas focal patch structural connectivity had no impact. Our results highlight that all fragments are important to contribute to the total amount of habitat available, and that the impact of habitat fragmentation matters more when the total amount of habitat available is low.

Keywords: arthropods, habitat amount hypothesis, landscape configuration, landscape genetics, lepidoptera, metapopulation

1. Introduction

Biodiversity loss is among the greatest challenges human societies are currently facing. Predicting the magnitude and drivers affecting biodiversity loss is urgently needed to prioritise conservation efforts. However, biodiversity loss is often measured as the loss of species diversity, with indicators such as genetic diversity often being entirely neglected (Pereira et al. 2013). Genetic diversity is directly linked with the evolutionary potential of a species and its adaptability (Hu et al. 2021), genetic diversity loss is expected to affect a population's or species' long‐term survival. It has recently been estimated that more than 10% of genetic diversity might have already been lost for many species due to human activities such as habitat destruction (Exposito‐Alonso et al. 2022).

Habitat loss and fragmentation are often cited together as main drivers of biodiversity loss (Tilman et al. 1994; Huxel and Hastings 1999). Based on their current increase rates, both factors are expected to continue causing most of the biodiversity loss in the future (Visconti et al. 2016). However, habitat loss and fragmentation are two distinct processes affecting different landscape features, with potentially varying impacts on populations (Fahrig 2003, 2017). Therefore, disentangling the threats each one imposes is needed for effectively mitigating biodiversity loss. Habitat loss refers to the reduction of habitat amount within a landscape (Martin et al. 2021), and can negatively affect various hierarchical levels of biodiversity, such as species richness (i.e., largely due to the general species‐area relationship; Lomolino 2000), or intraspecific genetic diversity (Exposito‐Alonso et al. 2022). Habitat fragmentation, on the other hand, has widely been used as an umbrella term for many ecological processes and spatial patterns accompanying alteration of landscapes by humans (Lindenmayer and Fischer 2007). Its different definitions and effects on biodiversity have been a continuous debate (see, for instance: Fahrig 2003, 2013; Valente et al. 2023). However, it is usually defined as a previously continuous habitat being broken up into smaller habitat fragments. Some habitats are naturally patchy, and thus, fragmentation can also be seen as a characteristic of the landscape that refers to the spatial distribution of a habitat in a smaller or larger number of fragments. As habitat fragmentation usually happens together with habitat loss, assessments of its impacts alone on biodiversity are difficult (Fahrig 2003; Ewers and Didham 2006). While the number of fragments is the typically used measure of fragmentation at the landscape scale, other habitat configuration metrics like habitat aggregation can capture fragmentation effects while reducing correlation with habitat amount (Wang et al. 2014).

This general coupling of habitat loss and habitat fragmentation has challenged the general assumption of the solely negative effect of habitat fragmentation on biodiversity. It has been suggested instead that the configuration of habitat in the landscape generally has little or no effect on species richness in sampled sites, provided that the habitat amount in the landscape stays constant (Fahrig 2003, 2013, 2017). In this context, habitat fragmentation, or fragmentation per se, is independent of the effects of habitat amount, and the fragmentation process does not necessarily imply habitat loss (Martin et al. 2021). Studies assessing the effects of fragmentation per se on biodiversity have principally focused on species richness changes depending on the amount of habitat or degree of fragmentation in a landscape (Haddad et al. 2017; Seibold et al. 2017; Watling et al. 2020; Herrero‐Jáuregui et al. 2022), while impacts on genetic diversity have been largely neglected (but see Jackson and Fahrig 2014; González‐Fernández et al. 2019). Disentangling the effects of habitat amount and fragmentation per se (hereafter fragmentation) on genetic diversity is, however, equally important to determine their respective direction and magnitude.

Landscape genetic studies have largely looked at the effects of landscape features on genetic diversity with a focus on habitat configuration, but the considered scale has predominantly been at the focal patch level (DiLeo and Wagner 2016). Such patch‐scale approaches cannot determine impacts of fragmentation per se on genetic diversity, as measures of the fragmentation of the habitat are confounded with habitat amount (Fahrig 2019). Determining potential effects of fragmentation per se over habitat amount on genetic diversity at the landscape level will allow the development of more informed conservation management strategies that consider species ability to adapt to changing environmental conditions (DiLeo and Wagner 2016; Savary et al. 2022). Indeed, the size of a fragment can limit the local population size it holds, which, when reduced, could cause a loss in genetic variation through genetic drift (Frankham 1996). Thus, a population subject to habitat reduction and coupled with habitat fragmentation can show lower genetic diversity with an increased genetic differentiation among the subpopulations (Keyghobadi et al. 2005). On the other hand, if the amount of habitat is equal within the landscape, and the remaining habitats are well enough connected to each other, populations living in more fragmented habitats could maintain an equal or have even a higher genetic diversity due to spatially varying selection (Hanski et al. 2011; Cheptou et al. 2017).

The Glanville fritillary butterfly (Melitaea cinxia) in the Finnish Åland Islands archipelago has become a model system to study the effect of fragmentation on metapopulation dynamics, and the ecology and genetics of the butterfly have been well studied (Ojanen et al. 2013; Hanski et al. 2017). Within this metapopulation, each suitable habitat (meadow or a pasture with one of both of the larval host plants present) across the landscape has been characterised (Hanski et al. 2017), and since 1993, the set of approximately 4000 habitat patches has been annually systematically monitored for the presence of conspicuous larval families (Hanski et al. 1996; Ojanen et al. 2013). The effect of focal habitat area and connectivity on the metapopulation dynamics (i.e., occupancy, abundance, local extinction, and re‐colonisation) has been extensively studied using a patch‐based approach (Hanski et al. 1994, 1995; Saccheri et al. 1998; Hanski 2011b; Schulz et al. 2020; Bergen et al. 2020). The spatial configuration and habitat quality rather than the pooled habitat area have been shown to predict metapopulation size and persistence in combination with genetic variation in species' dispersal capacity (Hanski et al. 2017; Fountain et al. 2018; DiLeo et al. 2018). A recent study using a landscape‐based approach has also demonstrated the positive effect of fragmentation (i.e., number of habitat patches) on patch occupancy and colonisation (Galán‐Acedo et al. 2024). While previous work on this system has investigated factors influencing genetic diversity at the patch scale (Saccheri et al. 1998; DiLeo et al. 2018, 2024), little is known about the spatial variation of genetic diversity at a landscape scale (but see Hanski et al. 2017), nor the effect of fragmentation per se on genetic diversity, which influences the butterfly's population performance (DiLeo et al. 2024).

The long‐term monitoring survey of the demographics of the Glanville fritillary butterfly, together with spatio‐temporal genetic data over the Åland islands, offers a comprehensive dataset to assess the separate effects of habitat amount and fragmentation on the species' genetic diversity (Hanski 2015). We used the extensive characterisation of the butterfly's habitat in the Åland archipelago (Ojanen et al. 2013), and an intensive genetic sampling conducted in this system (DiLeo et al. 2018). Using a landscape‐based approach, we first defined the scale of effect which corresponds to a buffer radius for which landscape variables best predict local population responses, following Martin et al. (2021). We then disentangled the effects of habitat amount from fragmentation (number of patches and aggregation) on the neutral genetic diversity of the focal patches. As the amount of habitat available within the landscape is likely to impact the genetic diversity greatly, we predicted higher levels of inbreeding and genetic differentiation, and decreasing levels of heterozygosity with a decreasing amount of habitat available in the landscape. For a constant amount of habitat within the landscape, we further expected a neutral effect of fragmentation on the genetic diversity indices.

2. Material and Methods

2.1. Study System and Landscape Characterisation

The Glanville fritillary butterfly is distributed across Europe up to the Åland archipelago in southern Finland. Within the Åland islands, the butterfly lives in a highly fragmented landscape and persists as a classic metapopulation (Hanski et al. 1994). The butterfly has one generation per year and overwinters as larvae in a winter nest. It has two larval host plants, Plantago lanceolata and Veronica spicata , growing on dry meadows and pastures, which often represent identifiable and discrete habitat patches (Nieminen et al. 2004).

In the Åland islands, each suitable habitat patch for Melitaea cinxia has been characterised and georeferenced based on the availability of one or both host plant species the larvae specialise on, corresponding to over 4000 well‐defined habitat patches (Ojanen et al. 2013). The habitat patches vary in size and are spread out within a heterogeneous landscape that includes agricultural lands, rocky areas, woodland clearings, and urbanised areas. Each potential habitat patch has been systematically visited every fall since 1993 to assess the presence of winter nests and their abundance. This information is used to evaluate the butterfly's population demography, patch occupancy, patch extinctions, and patch recolonisations (Nieminen et al. 2004). Using this extensive dataset, we classified the local landscapes based on the spatial distribution of all these discrete habitat patches (Figure 1).

FIGURE 1.

FIGURE 1

Open in a new tab

Schematic representation of the landscape categorization into focal habitat patches (dark green), non‐focal habitat patches (light green) and local landscapes (orange) with varying fragmentation (x‐axis) and habitat amount (y‐axis).

2.2. Genetic Diversity Indices

To compute genetic diversity indices, we utilised an existing dataset from the years 2011 and 2012, when up to three larvae per nest across the Åland patch network were collected during the fall survey from 4862 nests from 449 habitat patches. These larvae were then genotyped for 45 SNPs, among which 40 were putatively neutral markers (Fountain et al. 2018; DiLeo et al. 2018) and were used thereafter for our genetic analyses. These published datasets are openly available at (DiLeo et al. 2019). To select our focal patches, we chose patches with at least 10 individuals genotyped, to ensure that the measured genetic diversity was consistent and not a result of genetic drift due to recent colonisation events. We randomly sampled across nests a set of 10 genotyped individuals in focal patches to measure genetic diversity from more than one family and to avoid sample size effects. We confirmed that the estimated genetic diversity using 10 individuals was enough and did not significantly change using a larger number of individuals (Appendix: Figure S1). We considered the 2 years separately, even in cases when the same patch was sampled in both years.

Four genetic diversity indices were calculated at every focal patch using the 40 SNPs: the observed heterozygosity H o, the expected heterozygosity H e, the inbreeding coefficient F IS, and patch‐specific F ST (i.e., population‐specific F ST, as defined in Weir and Goudet 2017). The heterozygosity coefficients provide a direct measurement of genetic diversity at the patch level, while the inbreeding coefficient and F ST give information about the genetic consequences of the population dynamics and the differentiation of the subpopulations (Weir and Goudet 2017; Kitada et al. 2021). All analyses and data filtering, as well as the graphs and results presentation, were performed in R 4.4.0 (R Core Team, 2021) using RStudio 2024.09.1+394 software (RStudio Team, 2020). The genetic indices were calculated with the hWierfstat package (Goudet 2005).

2.3. Definition of the Study Scale

We first determined the appropriate landscape scale in which the effect of habitat amount and fragmentation on genetic diversity should be assessed, as proposed by Jackson and Fahrig (2015) and Martin et al. (2021). We studied the response of the four genetic diversity indices measured at the focal patches to the number of suitable habitat fragments in the surrounding landscapes. For this, circular landscapes with radii ranging from 0.5 to 5 km by steps of 0.5 km were defined around the centroid of each focal patch (Figure 2). For each of these landscape radii, we calculated both the total amount of suitable habitat (i.e., the total area of habitat contained in the habitat fragments) and the total number of fragments. We then performed simple linear regressions using the number of fragments as the independent variable, and one of the genetic diversity indices as the response variable for each radius. We assessed the proportion of variation explained by each model r‐squares (r 2) obtained from these regressions at all the different radii and for each of the genetic diversity indices. As suggested by Jackson and Fahrig (2015) and Martin et al. (2021), the radius for which the r 2 was the highest was selected as our final scale of effect and used to define the local landscape surrounding each focal patch. The linear models were performed with lme4 (Bates et al. 2015) and the R stats package. The calculations of the number of fragments and total habitat amount in the surrounding landscape of the focal patches at different scales were calculated using ArcGIS v. 10.8 (ESRI, Redlands, CA, USA, 2019) and QGIS v. 3.20 (QGIS, Open Source Geospatial Foundation Project, Chicago, USA, 2021).

FIGURE 2.

FIGURE 2

Open in a new tab

Estimation of the spatial scale with the strongest relationship between habitat amount and the genetic diversity of the Glanville fritillary butterfly in the Åland islands. Around the centroid of the focal patch, circular landscapes of radius ranging from 0.5 to 5 km by steps of 0.5 km were defined. Each green polygon corresponds to a habitat fragment of the Glanville fritillary butterfly, and the surroundings have been characterised as non‐habitat.

2.4. Testing the Effects of Habitat Amount and Fragmentation on Genetic Diversity

We first tested for a potential response of genetic diversity to habitat amount and fragmentation per se—i.e. measured as the number of fragments in the local landscape—at the landscape level with linear mixed‐effects models. We built a model separately for each of the genetic diversity indices calculated for the focal patches as the response variable. In these models, we included the total habitat amount within the local landscape, the number of fragments within the local landscape, and their interaction.

We additionally included two patch‐level metrics, focal patch size and focal patch connectivity, as explanatory variables, as both of these variables have been shown to predict occupancy, extinction, and colonisation in this system in a large number of previous studies (Hanski et al. 1994; Thornton et al. 2011; Schulz et al. 2020). We calculated the connectivity index S i for each focal patch by adapting the incidence function model described in (Hanski et al. 1994):

Si=∑i≠jexp−adij

where dij is the distance between focal patch i and source patch j, and α is a constant scaling that reflects average dispersal distance, which was set to 1 km as it has been done for this system in previous works (Hanski et al. 1994; DiLeo et al. 2024). We excluded from the original formula the population size, to maintain a measure that excludes habitat amount. Finally, the year was included as a fixed factor and the focal patch as a random effect.

As our aim was to dissociate the effects of fragmentation from the total habitat amount, we followed the protocol by Martin et al. (2021) to deal with the correlation between these two variables. First, we selected a subset of local landscapes from the entire dataset in which the correlation between the total habitat amount and the number of fragments was low. As our data presented a bimodal distribution for the total habitat amount, two different subsets were defined: one including local landscapes with a small total habitat amount (0.135–0.214 km2), and another with a large total habitat amount (0.449–0.521 km2) in the local landscapes (Appendix: Figure S2). The effect of fragmentation on the genetic diversity indices was tested separately for these two sub‐datasets. For each subset, we performed, similarly to the whole dataset, four linear mixed‐effects models with each of the genetic diversity indices as the response variable. We included as explanatory variables the focal patch size, the habitat amount, the number of fragments and their interaction; as a fixed effect, the year; and as a random effect, the patch. We checked that neither of the response variables presented spatial autocorrelation by plotting spline correlograms (BjØrnstad and Falck 2001) of each of the variables and confirming a lack of correlation among nearby observations.

To complement these analyses with a measure of habitat fragmentation that is less correlated with habitat amount, we performed linear models to test for the effect of habitat aggregation on each genetic diversity index. The Clumpy index (CLUMPY) is an aggregation metric that equals the proportional deviation of the proportion of like adjacencies of the corresponding habitat from that expected under a spatially random distribution (McGarigal et al. 2002). We calculated CLUMPY using the R package landscapemetrics (Hesselbarth et al. 2019) for each local landscape of 3.5 km radius, with a resolution of 10 m because at least 20 m of non‐suitable habitat defines two discrete habitat patches within this system (Ojanen et al. 2013). We tested the effect of fragmentation using CLUMPY instead of the number of fragments on the genetic diversity indices, as described above, for the full dataset.

3. Results

In total, 261 out of the 449 patches had 10 or more individuals genotyped, for which we computed the genetic indices. The final dataset included 189 different patches, as some were occupied during both years.

The analysis of the scale of effect showed the spatial scale at which the four genetic diversity indices present a maximum correlation with the number of fragments. The correlation between the number of fragments and F ST or H e was highest at a 3.5 km radius (r 2  = 0.073 and r 2  = 0.069 respectively; Appendix: Figure S3). The highest correlation for F IS was observed at 2 km (r 2  = 0.029), and at 2.5 km for H o (r 2  = 0.069). As F IS showed the highest correlation with the number of fragments at a smaller scale (i.e., radius), we ran the models at this scale (2 km) and observed the same trends as for a 3.5 km scale (Table S1). Taking a more conservative approach, we thus chose a 3.5 km radius as our study scale for testing the habitat amount hypothesis.

Within our landscapes of 3.5 km radius, the number of fragments ranged from 30 to 291, and the total habitat amount from 0.053 to 0.613 km2. We first looked at the influence of our variables, including habitat amount and fragmentation, on the four genetic indices using the whole dataset. The sampling year did not affect most of the genetic diversity estimates, only H e, with a relatively small coefficient (0.21, SE = 0.1, p = 0.035; Table 1). The focal patch size showed a significant negative effect on F ST (−0.15, SE = 0.07, p = 0.048; Table 1). Patch connectivity did not show any significant effects. The habitat amount within the local landscape had a positive effect on H e (0.56, SE = 0.19, p = 0.004) and F IS (0.32, SE = 0.18, p = 0.011) while it had a negative effect on F ST (−0.54, SE = 0.19, p = 0.005) and a non‐significant positive effect on H o. In comparison, the number of fragments had a weak influence on the genetic indices. The increasing number of fragments tended to increase F ST and decrease H e and F IS at higher habitat amounts; however, main effects of the number of fragments and its interaction with the total habitat amount were non‐significant in models (Figure 3). The number of fragments and the total habitat amount within the local landscape were highly correlated (Pearson's correlation = 0.934, p < 0.001; Appendix: Figure S4). The correlation between the two variables was reduced to 0.258 for the small total habitat amount data subset, and −0.026 for the large total habitat amount, with each subset containing 98 and 57 focal patches, respectively. We repeated the same models using each of the two subsets. We only found a significant positive effect of the habitat amount on H e (0.23, SE = 0.11, p = 0.047) with the small habitat amount subset (Table 2). We found no significant effect of fragmentation on the genetic diversity indices with either of the subsets. In the models performed with the small total habitat amount subset, there was a tendency for habitat amount to reduce F ST and increase H o and H e (Figure 4). The tendencies for the effect of fragmentation were generally weaker. Higher fragmentation tended to be associated with lower F IS and higher H e at lower values of habitat amount for the small total habitat amount subset (Figure 4), and with lower F ST and higher H o for the large total habitat amount subset (Appendix: Figure S5, Table S2).

TABLE 1.

Results of the fitted linear mixed‐effects models relating each of the four genetic diversity indices (F IS, F ST, H e and H o) in the Glanville fritillary butterfly to year of larval collection, focal patch size, logarithm of the connectivity index S i , habitat amount, and number of fragments within a landscape of 3.5 km radius.

F ST Predictors std. beta Standardised std. error Standardised CI p‐value df
Intercept −0.03 0.12 −0.26 to 0.20 0.806 183
Area −0.54 0.19 −0.92 to −0.17 0.005 183
Fragments 0.24 0.19 −0.13 to 0.62 0.201 183
Focal patch area −0.15 0.07 −0.29 to −0.00 0.048 183
log(S i ) −0.08 0.07 −0.21 to 0.06 0.264 183
Year −0.13 0.09 −0.32 to 0.06 0.181 71
Area × Fragments 0.14 0.08 −0.02 to 0.30 0.081 183
N 189 patch
Observations 261
Marginal R 2/Conditional R 2 0.128/0.599
F IS Predictors Std. beta Standardised std. error Standardised CI p‐value df
Intercept 0.19 0.13 −0.06 to 0.44 0.564 183
Area 0.32 0.18 −0.04 to 0.67 0.011 183
Fragments −0.31 0.18 −0.67 to 0.05 0.961 183
Focal patch area −0.00 0.07 −0.13 to 0.13 0.969 183
log(S i ) 0.03 0.06 −0.10 to 0.15 0.652 183
Year −0.07 0.13 −0.32 to 0.18 0.561 71
Area × Fragments −0.15 0.08 −0.31 to 0.00 0.055 183
N 189 patch
Observations 261
Marginal R 2/Conditional R 2 0.035/0.090
H e Predictors Std. beta Standardised std. error Standardised CI p‐value df
Intercept −0.01 0.12 −0.25 to 0.22 0.925 183
Area 0.56 0.19 0.19 to 0.93 0.004 183
Fragments −0.26 0.19 −0.64 to 0.11 0.170 183
Focal patch area 0.14 0.07 −0.00 to 0.29 0.051 183
log(S i ) 0.08 0.07 −0.06 to 0.21 0.267 183
Year 0.21 0.10 0.01 to 0.40 0.035 71
Area × Fragments −0.15 0.08 −0.31 to 0.01 0.067 183
N 189 patch
Observations 261
Marginal R 2/Conditional R 2 0.131/0.594
H o Predictors Std. beta Standardised std. error Standardised CI p‐value df
Intercept −0.13 0.12 −0.38 to 0.11 0.287 183
Area 0.11 0.19 −0.26 to 0.48 0.546 183
Fragments 0.11 0.19 −0.27 to 0.48 0.576 183
Focal patch area 0.10 0.07 −0.04 to 0.24 0.147 183
log(S i ) 0.05 0.07 −0.09 to 0.18 0.501 183
Year 0.20 0.11 −0.02 to 0.42 0.079 71
Area × Fragments 0.00 0.08 −0.16 to 0.16 0.999 183
N 189 patch
Observations 261
Marginal R 2/Conditional R 2 0.069/0.358
Open in a new tab

Note: The bold values are those with a p‐value < 0.05.

FIGURE 3.

FIGURE 3

Open in a new tab

Predicted responses of the Glanville fritillary butterfly's genetic diversity (A) F ST; (B) F IS; (C) H e; and (D) H o of focal patches to habitat amount and fragmentation per se (three groups of hypothetical number of patches), according to the models developed with the whole dataset (N = 261).

TABLE 2.

Results of the fitted linear mixed‐effects models relating each of the four genetic diversity indices (F IS, F ST, H e and H o) in the Glanville fritillary butterfly to year of larval collection, focal patch size, and logarithm of the connectivity index S i , number of fragments, habitat amount within a landscape of 3.5 km radius, for the subset of landscapes with a small habitat amount.

F ST Predictors Std. beta Standardised std. error Standardised CI p‐value df
Intercept 0.19 0.18 −0.16 to 0.55 0.279 64
Area −0.23 0.11 −0.45 to 0.00 0.051 64
Fragments 0.03 0.14 −0.25 to 0.30 0.854 64
Focal patch area −0.11 0.11 −0.34 to 0.11 0.321 64
S i (log) −0.14 0.11 −0.35 to 0.08 0.203 64
Year −0.26 0.20 −0.66 to 0.14 0.195 27
Area × Fragments −0.01 0.16 −0.34 to 0.31 0.937 64
N 70 patch
Observations 98
Marginal R 2/Conditional R 2 0.076/0.322
F IS Predictors Std. beta Standardised std. error Standardised CI p‐value df
Intercept 0.01 0.18 −0.34 to 0.37 0.937 64
Area 0.08 0.11 −0.14 to 0.30 0.479 64
Fragments −0.17 0.13 −0.44 to 0.10 0.202 64
Focal patch area −0.02 0.11 −0.24 to 0.21 0.877 64
S i (log) 0.10 0.11 −0.12 to 0.31 0.367 64
Year −0.06 0.20 −0.48 to 0.36 0.770 27
Area × Fragments 0.10 0.16 −0.22 to 0.42 0.535 64
N 70 patch
Observations 98
Marginal R 2/Conditional R 2 0.057/0.244
H e Predictors Std. beta Standardised std. error Standardised CI p‐value df
Intercept −0.24 0.18 −0.60 to 0.11 0.171 64
Area 0.23 0.11 0.00 to 0.45 0.047 64
Fragments −0.04 0.14 −0.31 to 0.23 0.784 64
Focal patch area 0.11 0.11 −0.11 to 0.33 0.334 64
S i (log) 0.14 0.11 −0.07 to 0.35 0.187 64
Year 0.34 0.20 −0.07 to 0.74 0.097 27
Area × Fragments 0.02 0.16 −0.30 to 0.34 0.915 64
N 70 patch
Observations 98
Marginal R 2/Conditional R 2 0.085/0.312
H o Predictors Std. beta Standardised std. error Standardised CI p‐value df
Intercept −0.21 0.17 −0.55 to 0.14 0.243 64
Area 0.14 0.12 −0.10 to 0.37 0.251 64
Fragments 0.11 0.14 −0.18 to 0.39 0.450 64
Focal patch area 0.10 0.12 −0.13 to 0.34 0.392 64
S i (log) 0.02 0.11 −0.20 to 0.23 0.879 64
Year 0.31 0.18 −0.07 to 0.68 0.107 27
Area × Fragments −0.05 0.17 −0.39 to 0.28 0.745 64
N 70 patch
Observations 98
Marginal R 2/Conditional R 2 0.073/0.445
Open in a new tab

Note: The bold values are those with a p‐value < 0.05.

FIGURE 4.

FIGURE 4

Open in a new tab

Predicted responses of the Glanville fritillary butterfly's genetic diversity (A) F ST; (B) F IS; (C) H e; and (D) H o of focal patches to habitat amount and fragmentation per se, from the models developed with the subset of small total habitat amount within local landscapes (0.134–0.214 km2, N = 98).

In addition to assessing the impacts of fragmentation as a number of patches within the landscape, we also tested the effect of fragmentation using CLUMPY, which describes the degree of habitat aggregation in a landscape. The aggregation of habitats showed low levels of correlation with habitat amount (Pearson's correlation = 0.439). Overall, the studied landscapes had a nonrandom spatial aggregation (0.595–0.776). Habitat aggregation within the landscape had a negative, significant effect on F ST (−0.2, SE = 0.07, p = 0.009), and a significant positive effect on F IS (0.16, SE = 0.07, p = 0.024) and H e (0.2, SE = 0.07, p = 0.024). The habitat amount within the local landscape had a positive effect on H o (0.20, SE = 0.07, p = 0.008; Table 3), while it had a negative effect on F ST (−0.15, SE = 0.07, p = 0.049). Even though the effects were not significant, we found a positive trend of the habitat amount on H e and a negative trend on F IS (Table 3). The interaction between the habitat amount and habitat aggregation had a significant positive effect on F ST (0.13, SE = 0.06, p = 0.032) and a significant negative effect on H e (−0.13, SE = 0.06, p = 0.031), so that the impacts of habitat aggregation on genetic diversity were more evident when the total amount of landscape available was low. The interaction between habitat amount and habitat aggregation was negative but non‐significant on F IS and H o (Figure 5). Also in this model, the focal patch size showed a significant negative effect on F ST (−0.16, SE = 0.07, p = 0.03; Table 3) and a significant positive effect on H e (0.15, SE = 0.07, p = 0.032), whereas patch connectivity showed no significant effects. The sampling year also did not affect any of the genetic diversity estimates.

TABLE 3.

Results of the fitted linear mixed‐effects models relating each of the four genetic diversity indices (F IS, F ST, H e and H o) in the Glanville fritillary butterfly to year of larval collection, focal patch size, logarithm of the connectivity index S i , habitat amount, and CLUMPY index within a landscape of 3.5 km radius.

F ST Predictors Std. beta Standardised std. error Standardised CI p‐value df
Intercept 0.03 0.09 −0.16 to 0.22 0.749 183
Area −0.15 0.07 −0.29 to −0.00 0.049 183
CLUMPY −0.20 0.07 −0.34 to −0.05 0.009 183
Focal patch area −0.16 0.07 −0.30 to −0.02 0.030 183
S i (log) −0.09 0.07 −0.22 to 0.04 0.185 183
Year −0.11 0.10 −0.30 to 0.08 0.269 71
Area × CLUMPY 0.13 0.06 0.01 to 0.25 0.032 183
N 189 patch
Observations 261
Marginal R 2/Conditional R 2 0.160/0.596
F IS Predictors Std. beta Standardised std. error Standardised CI p‐value df
Intercept 0.09 0.11 −0.12 to 0.30 0.412 183
Area −0.12 0.07 −0.26 to 0.02 0.086 183
CLUMPY 0.16 0.07 0.02 to 0.30 0.024 183
Focal patch area 0.01 0.06 −0.12 to 0.14 0.881 183
S i (log) 0.03 0.06 −0.09 to 0.16 0.584 183
Year −0.09 0.13 −0.35 to 0.16 0.468 71
Area × CLUMPY −0.07 0.06 −0.18 to 0.05 0.246 183
N 189 patch
Observations 261
Marginal R 2/Conditional R 2 0.032/0.042
He Predictors Std. beta Standardised std. error Standardised CI p‐value df
Intercept −0.08 0.09 −0.26 to 0.11 0.418 183
Area 0.14 0.07 −0.01 to 0.28 0.059 183
CLUMPY 0.20 0.07 0.06 to 0.35 0.007 183
Focal patch area 0.15 0.07 0.01 to 0.29 0.032 183
S i (log) 0.09 0.07 −0.04 to 0.22 0.187 183
Year 0.18 0.10 −0.01 to 0.37 0.061 71
Area × CLUMPY −0.13 0.06 −0.25 to −0.01 0.031 183
N 189 patch
Observations 261
Marginal R 2/Conditional R 2 0.163/0.589
Ho Predictors Std. beta Standardised std. error Standardised CI p‐value df
Intercept −0.09 0.10 −0.30 to 0.11 0.355 183
Area 0.20 0.07 0.05–0.34 0.008 183
CLUMPY 0.02 0.07 −0.13 to 0.17 0.788 183
Focal patch area 0.10 0.07 −0.04 to 0.24 0.154 183
S i (log) 0.05 0.07 −0.08 to 0.19 0.437 183
Year 0.19 0.11 −0.04 to 0.41 0.098 71
Area × CLUMPY −0.07 0.06 −0.19 to 0.05 0.262 183
N 189 patch
Observations 261
Marginal R 2/Conditional R 2 0.074/0.367
Open in a new tab

Note: The bold values are those with a p‐value < 0.05.

FIGURE 5.

FIGURE 5

Open in a new tab

Predicted responses of the Glanville fritillary butterfly's genetic diversity (A) F ST; (B) F IS; (C) H e; and (D) H o of focal patches to habitat amount and clumpiness (three groups of hypothetical levels of clumpiness), according to the models developed with the whole dataset (N = 261).

4. Discussion

Habitat amount and fragmentation are intimately linked landscape features that often change in parallel (Huxel and Hastings 1999; Fahrig 2003; Wagner et al. 2021), hindering the ability to attribute their independent effects on biodiversity. In our study, we specifically aimed to assess the independent effects of habitat amount and fragmentation on different genetic diversity indices. Overall, we found that habitat amount had a positive effect on genetic diversity both at the landscape and at the focal patch level. In contrast, the effects of fragmentation were generally weaker, but the strength and direction of effects varied depending on the fragmentation measure and genetic diversity index used and the total amount of habitat in the landscape. While most effects of fragmentation tended to be negligible or even slightly positive, negative effects on genetic diversity were found in landscapes with low total habitat amount when an aggregation index was used in place of the number of fragments in models. We therefore conclude that habitat fragmentation impacts genetic diversity less than the total amount of habitat available, but its negative impacts become relevant when the available habitat amount is limited.

4.1. Landscape Size and Glanville Fritillary Butterfly's Biology

The scale of effect is defined as the spatial extent at which the landscape structure best predicts the response of interest (Jackson and Fahrig 2015; Martin et al. 2021). It ensures that the landscape is measured at the appropriate scale (Jackson and Fahrig 2012). In the Åland islands study system, the scale at which the genetic diversity of the Glanville fritillary butterfly was responding the most to fragmentation was at a 3.5 km radius. This is identical to what was found in a recent study in this system testing specifically the effects of fragmentation per se on patch occupancy using 20 years of demographic data (Galán‐Acedo et al. 2024). The multiscale analysis of the local landscape for the four different genetic indices, however, showed that the relationship between genetic diversity and the landscape extent varied depending on the genetic index considered. The scale of effect was the smallest for the inbreeding coefficient, F IS, while the largest scale of effect was found for the expected heterozygosity, H e, and the genetic differentiation coefficient, F ST. When simulating population responses to habitat amount and fragmentation per se, Jackson and Fahrig (2014) also obtained different scales of effect depending on the response of interest, with genetic diversity presenting the largest scale. Genetic diversity indices based on allele frequencies might change faster, making their responses more apparent at smaller spatial scales (Crow and Aoki 1984; Keyghobadi et al. 2005). Our results do not match this expectation, and it is unclear why F ST and H e, which are based on allele frequencies, showed the largest scale of effect. It has been argued that the scale of effect is primarily a function of species mobility and that dispersal distance would be its strongest predictor (Jackson and Fahrig 2012), with the scale of effect being at least equal to or higher than a species average distances (Jackson and Fahrig 2015). The observed scales of effect in our study are in line with the reported average dispersal distances of the Glanville fritillary butterfly (Hanski et al. 1995; Fountain et al. 2018; DiLeo et al. 2018). The average dispersal distance for the butterfly is 1.5 km (DiLeo et al. 2018, 2022), with the highest breeding dispersal distances reaching even 10 km based on genetic assignment tests (DiLeo et al. 2022). This suggests that we are capturing the appropriate landscape size to disentangle the role of fragmentation per se from habitat amount.

4.2. Effects of Habitat Amount and Fragmentation on Genetic Diversity

As suggested in the habitat amount hypothesis (Fahrig 2013) and metapopulation persistence theory (With 2004), the total habitat amount within the landscape seemed to be the main factor affecting the genetic diversity of the Glanville fritillary butterfly. Jackson and Fahrig (2014) found a similar trend in genetic diversity in their theoretical study. We found that with increasing habitat amount, expected and observed heterozygosity increased, while population differentiation decreased. Unexpectedly, we also found that inbreeding increased with habitat amount; however, this result might be explained by a confounding of habitat amount and fragmentation in our full model. For example, when the correlation between habitat amount and fragmentation was reduced by subsetting the data or using an aggregation index in place of the number of fragments, we found a neutral or the expected negative relationship between habitat amount and inbreeding. An increase in the total habitat amount in the local landscape can thus reduce the rate of genetic drift, improve gene flow between patches (Keyghobadi et al. 2005; Lebigre et al. 2022) and in turn, favour the evolutionary potential of the local population (Cheptou et al. 2017).

After accounting for the correlation between habitat amount and the number of fragments within local landscapes (Martin et al. 2021), the association between habitat amount and genetic diversity was less clear. Within local landscapes with small total habitat amounts, the genetic diversity of the Glanville fritillary butterfly seemed to respond positively to habitat amount, independently from fragmentation. However, within local landscapes with a large habitat amount, the genetic diversity was not affected by habitat amount (Appendix: Figure S4). This might suggest that genetic diversity saturates after a certain level of habitat is achieved in a landscape, as is often seen in the shape of species‐area relationships (Lomolino 2000; Exposito‐Alonso et al. 2022).

The independent effects of fragmentation, measured as the number of fragments, on genetic diversity of the Glanville fritillary butterfly were generally weak and non‐significant. However, we acknowledge that due to a very high correlation between habitat amount and fragmentation, standard errors of estimated effects of fragmentation from the model using the complete dataset are likely inflated and unstable, and thus must be interpreted with caution (Dormann et al. 2013). After controlling for the correlation with habitat amount by subsetting the data, no clear effects of fragmentation on genetic diversity were evident, despite our attempt to maintain as much variability in the number of fragments as possible and to reduce the variability in the total habitat amount. In contrast, we detected negative impacts of habitat fragmentation in landscapes with low total habitat amount when assessing fragmentation using an aggregation index. At low total habitat amounts, less aggregated (i.e., more fragmented) landscapes had higher population differentiation (F ST) and lower expected heterozygosity (H e). This suggests that when habitat is scarce, less aggregated landscapes hold smaller populations and have lower immigration and emigration rates, which could all increase the effect of genetic drift and accelerate the loss of genetic diversity. Counterintuitively, we also found that less aggregated landscapes had lower levels of inbreeding (F IS) compared to more aggregated landscapes, when total habitat amount was low. This could be due to a Wahlund effect (Wahlund 1928) in aggregated landscapes at low habitat amount. Higher population turnover is expected in landscapes with less habitat, leading to more opportunities for re‐colonisation following local patch extinction (Hanski 2011b). Larger, or more aggregated patches are more likely to be recolonised by multiple females, which could lead to Wahlund effects if colonising genotypes are distinct. Together, these results support the idea that, although habitat amount has a much bigger effect on biodiversity, fragmentation becomes relevant when there is very little habitat available (Fahrig 1998; Hanski and Gaggiotti 2004).

From the patch level parameters and in line with previous ecological observations on the role of increased habitat patch area in the Glanville fritillary butterfly (Hanski 1999), we found that the size of the focal patch affected F ST and showed a clear trend with H e. When the size of the focal patch increased, patch genetic differentiation decreased and expected heterozygosity H e showed a trend to increase, indicating that more immigrants from the surrounding patches are colonising larger focal patches. Larger and well‐connected habitat patches tend to support higher occupancy and population abundance (Schulz et al. 2020). On the other hand, small habitat fragment patches do not often have a very high occupancy rate, high population abundance, nor do they stay occupied for many consecutive years (Dallas et al. 2020), and instead are more likely to work as stepping stone patches. Smaller focal patches will likely contain smaller subpopulations, show a lower immigration rate, and increase genetic drift, which can lead to genetically more differentiated subpopulations.

The focal patch connectivity reflecting patch‐level structural connectivity, on the other hand, showed no significant effect on genetic diversity patterns in any of the analyses. However, genetic diversity is expected to depend on species dispersibility between patches, favouring sufficient gene flow between habitat patches, reducing genetic drift and thus preventing the negative effects of fragmentation (Amos and Harwood 1998). The patch connectivity index Si has been associated with population differentiation in some but not all cases in previous studies in this system (DiLeo et al. 2018). The lack of effect of connectivity could imply that all the studied focal patches were connected enough to allow gene flow and prevent genetic diversity loss, or that actual connectivity depends on other factors (e.g., matrix permeability) not captured in our model.

4.3. Disentangling Habitat Amount from Fragmentation Effects on Genetic Diversity of Wild Populations

Characteristics and quality of local habitat patches may be important determinants of effective population size (Ne) and, consequently, genetic diversity (Crawford and Keyghobadi 2018). Studies have shown the importance of patch quality and patch connectivity in butterflies on population density (Turlure et al. 2010; Crawford and Keyghobadi 2018), successful dispersal within the landscape (With 2004; DiLeo et al. 2022), patch occupancy (Dennis and Eales 1997; Hanski et al. 2017; Galán‐Acedo et al. 2024) as well as genetic diversity (Lebigre et al. 2022). When only taking into account the habitat amount available within the local landscape, patch quality is not considered, while patch connectivity is assumed to increase with the total habitat amount but is not directly measured (Fahrig 2013). For small organisms like insects, direct measures of patch connectivity (e.g., using the incidence function model) have been found to be good estimators of both dispersibility and maintenance in the landscape, while patch quality has been found to be a better estimator of individual settlement (DiLeo et al. 2022). If individuals cannot colonise some patches within the landscape due to their isolation or poor quality, it will decrease the effective population size the landscape can hold, potentially exposing the population to a higher risk of extinction. In turn, the population is likely to be more sensitive to stochasticity and random events, which could lead to a reduced genetic diversity. This might become even more important as extreme events (i.e., droughts, flooding) are increasing (Calvin et al. 2023), directly affecting insects (Wagner et al. 2021), including the Glanville fritillary butterfly (Bergen et al. 2020). Measuring the genetic diversity of a species at the landscape level might then give better insight into the overall viability of the population by providing more representative information about the maintenance of genetic diversity while the patch scale approach would provide information on species movement, barriers (Holderegger and Wagner 2008) and density fluctuations between patches. Genetic studies combining patch‐based and landscape‐based approaches will thus permit more accurate predictions of the responses of a species' genetic diversity to landscape characteristics. Such studies will be especially relevant as Riva and Fahrig (2022) found opposite patterns in species diversity when using either a landscape‐scale or patch‐scale approach.

Finally, we lost substantial statistical power by subsetting our data to control for high correlations between habitat amount and fragmentation when assessing fragmentation via number of fragments. Future studies could avoid this limitation by selecting a priori and sampling in landscapes with more contrasting characteristics (i.e., including more landscapes with high habitat amount and low fragmentation, and landscapes with low habitat amount and high fragmentation). Considering the urgency to protect the remaining biodiversity (Cowie et al. 2022), such an approach would allow us to determine if measuring only the effect of habitat amount is a more efficient method to develop appropriate conservation measures than trying to consider the full complexity of the landscape (Seibold et al. 2017).

5. Conclusion

Studies testing the effect of fragmentation per se on species' genetic diversity using wild populations are still scarce (but see Connor et al. 2022), and to our knowledge, do not exist for insects. Our results suggest that the total habitat amount available within the landscape affects genetic diversity positively, while the effects of fragmentation per se (i.e., number of fragments) seem to have a neutral effect on genetic diversity. However, it is noteworthy that fragmentation, when measured as habitat aggregation, seems to have a relative negative impact on genetic diversity when habitat is scarce. These outcomes thus provide support for the benefit of first prioritising the conservation of large enough habitat availability within the local landscape and in a second step, its habitat configuration (Riva et al. 2024). Overall, our study demonstrates the complexity of assessing the impacts of fragmentation per se on genetic diversity estimates, even in a species for which extensive landscape and genetic datasets are available. Future studies testing the habitat amount hypothesis at both neutral and functional genetic diversity in more contrasted landscapes would give more insight into the separate role of habitat amount and fragmentation per se on species.

Author Contributions

Lola Fernández Multigner: data analysis, manuscript drafting. Audrey Bras: study design, data analysis, manuscript drafting. Michelle F. DiLeo: study design, data analysis, manuscript drafting. Marjo Saastamoinen: study design, manuscript drafting.

Conflicts of Interest

The authors declare no conflicts of interest.

Supporting information

Appendix S1.

MEC-34-e70037-s001.docx (461.8KB, docx)

Acknowledgements

We acknowledge the Biodata analytics unit at the University of Helsinki for providing helpful inputs on the statistical analyses, all the people who have participated in the Åland survey over the years, the University of Helsinki and all funders of this long‐term survey: Waldemar von Frenckel Foundation, Nesslingin Foundation, Weisell Foundation, Ella and Georg Ehrnrooth Foundation, Novo Nordisk Challenge Programme and Vuokko Conservation Foundation. Open access publishing facilitated by Helsingin yliopisto, as part of the Wiley ‐ FinELib agreement.

Handling Editor: Jeremy B Yoder

Funding: This study was performed in EcoGenetics—Centre for Ecological Genetics funded by the Novo Nordisk Challenge Programme grant number NNF20OC0060118.

Lola Fernández Multigner and Audrey Bras authors contributed equally to this work.

Data Availability Statement

The data used for this study are publicly available, published in: DiLeo MF, Husby A, Saastamoinen M. 2019, November 16. Data from: Landscape permeability and individual variation in a dispersal‐linked gene jointly determine genetic structure in the Glanville fritillary butterfly. Dryad. Available from https://datadryad.org/dataset/https://doi.org/10.5061/dryad.mp25s15 (accessed June 9, 2025). The code of the analysis is available in: https://zenodo.org/records/14218771?token=eyJhbGciOiJIUzUxMiJ9.eyJpZCI6IjVjMDY4MDdlLTRhYzMtNGFhZi1hMDY4LWRkODRmZWZmM2NiOCIsImRhdGEiOnt9LCJyYW5kb20iOiIzYTgxOTk5ODY2ZjE1YjUwYTIwZTk2MGRmMDgyZmUwMiJ9.xr7sdfs1WVDXVZ2Oc_N6cYlpkXrRpUMcjmZfSzzO_HjjYJGFe51ydhBaas‐rNKooX2GNvbFNtXdXTINF3I40wQ

References

  1. Amos, W. , and Harwood J.. 1998. “Factors Affecting Levels of Genetic Diversity in Natural Populations.” Philosophical Transactions of the Royal Society of London. Series B, Biological Sciences 353: 177–186. [DOI] [PMC free article] [PubMed] [Google Scholar]
  2. Bates, D. , Mächler M., Bolker B., and Walker S.. 2015. “Fitting Linear Mixed‐Effects Models Using lme4.” Journal of Statistical Software 67, no. 1: 1–48. 10.18637/jss.v067.i01. [DOI] [Google Scholar]
  3. Bergen, E. , Dallas T., DiLeo M. F., et al. 2020. “The Effect of Summer Drought on the Predictability of Local Extinctions in a Butterfly Metapopulation.” Conservation Biology 34: 1503–1511. [DOI] [PubMed] [Google Scholar]
  4. BjØrnstad, O. N. , and Falck W.. 2001. “Nonparametric Spatial Covariance Functions: Estimation and Testing.” Environmental and Ecological Statistics 8: 53–70. [Google Scholar]
  5. Calvin, K. , Dipak D., Gerhard K., et al. 2023. IPCC, 2023: Climate Change 2023: Synthesis Report. Contribution of Working Groups I, II and III to the Sixth Assessment Report of the Intergovernmental Panel on Climate Change [Core Writing Team, H. Lee and J. Romero (eds.)]. Intergovernmental Panel on Climate Change (IPCC). [Google Scholar]
  6. Cheptou, P.‐O. , Hargreaves A. L., Bonte D., and Jacquemyn H.. 2017. “Adaptation to Fragmentation: Evolutionary Dynamics Driven by Human Influences.” Philosophical Transactions of the Royal Society, B: Biological Sciences 372: 20160037. [DOI] [PMC free article] [PubMed] [Google Scholar]
  7. Connor, T. , Qiao M., Scribner K., et al. 2022. “Complex Effects of Habitat Amount and Fragmentation on Functional Connectivity and Inbreeding in a Giant Panda Population.” Conservation Biology 36, no. 2: e13828. [Google Scholar]
  8. Cowie, R. H. , Bouchet P., and Fontaine B.. 2022. “The Sixth Mass Extinction: Fact, Fiction or Speculation?” Biological Reviews 97: 640–663. [DOI] [PMC free article] [PubMed] [Google Scholar]
  9. Crawford, L. A. , and Keyghobadi N.. 2018. “Analysis of Genetic Diversity in a Peatland Specialist Butterfly Suggests an Important Role for Habitat Quality and Small Habitat Patches.” Conservation Genetics 19: 1109–1121. [Google Scholar]
  10. Crow, J. F. , and Aoki K.. 1984. “Group Selection for a Polygenic Behavioral Trait: Estimating the Degree of Population Subdivision.” Proceedings of the National Academy of Sciences 81: 6073–6077. [DOI] [PMC free article] [PubMed] [Google Scholar]
  11. Dallas, T. A. , Saastamoinen M., Schulz T., and Ovaskainen O.. 2020. “The Relative Importance of Local and Regional Processes to Metapopulation Dynamics.” Journal of Animal Ecology 89: 884–896. [DOI] [PubMed] [Google Scholar]
  12. Dennis, R. L. H. , and Eales H. T.. 1997. “Patch Occupancy in Coenonympha tullia (Muller, 1764) (Lepidoptera: Satyrinae): Habitat Quality Matters as Much as Patch Size and Isolation.” Journal of Insect Conservation 1: 167–176. [Google Scholar]
  13. DiLeo, M. F. , Husby A., and Saastamoinen M.. 2018. “Landscape Permeability and Individual Variation in a Dispersal‐Linked Gene Jointly Determine Genetic Structure in the Glanville Fritillary Butterfly.” Evolution Letters 2: 544–556. [DOI] [PMC free article] [PubMed] [Google Scholar]
  14. DiLeo, M. F. , Husby A., and Saastamoinen M.. 2019. “Data From: Landscape Permeability and Individual Variation in a Dispersal‐Linked Gene Jointly Determine Genetic Structure in the Glanville Fritillary Butterfly.” Dryad. https://datadryad.org/dataset/doi:10.5061/dryad.mp25s15. [DOI] [PMC free article] [PubMed]
  15. DiLeo, M. F. , Nair A., Kardos M., Husby A., and Saastamoinen M.. 2024. “Demography and Environment Modulate the Effects of Genetic Diversity on Extinction Risk in a Butterfly Metapopulation.” Proceedings of the National Academy of Sciences of the United States of America 121: e2309455121. [DOI] [PMC free article] [PubMed] [Google Scholar]
  16. DiLeo, M. F. , Nonaka E., Husby A., and Saastamoinen M.. 2022. “Effects of Environment and Genotype on Dispersal Differ Across Departure, Transfer and Settlement in a Butterfly Metapopulation.” Proceedings of the Royal Society B: Biological Sciences 289: 20220322. [DOI] [PMC free article] [PubMed] [Google Scholar]
  17. DiLeo, M. F. , and Wagner H. H.. 2016. “A Landscape Ecologist's Agenda for Landscape Genetics.” Current Landscape Ecology Reports 1: 115–126. [Google Scholar]
  18. Dormann, C. F. , Elith J., Bacher S., et al. 2013. “Collinearity: A Review of Methods to Deal With It and a Simulation Study Evaluating Their Performance.” Ecography 36: 27–46. [Google Scholar]
  19. Ewers, R. M. , and Didham R. K.. 2006. “Confounding Factors in the Detection of Species Responses to Habitat Fragmentation.” Biological Reviews 81: 117–142. [DOI] [PubMed] [Google Scholar]
  20. Exposito‐Alonso, M. , Booker T. R., Czech L., et al. 2022. “Genetic Diversity Loss in the Anthropocene.” Science 377: 1431–1435. [DOI] [PubMed] [Google Scholar]
  21. Fahrig, L. 1998. “When Does Fragmentation of Breeding Habitat Affect Population Survival?” Ecological Modelling 105: 273–292. [Google Scholar]
  22. Fahrig, L. 2003. “Effects of Habitat Fragmentation on Biodiversity.” Annual Review of Ecology, Evolution, and Systematics 34: 487–515. [Google Scholar]
  23. Fahrig, L. 2013. “Rethinking Patch Size and Isolation Effects: The Habitat Amount Hypothesis.” Journal of Biogeography 40: 1649–1663. [Google Scholar]
  24. Fahrig, L. 2017. “Ecological Responses to Habitat Fragmentation Per Se.” Annual Review of Ecology, Evolution, and Systematics 48: 1–23. [Google Scholar]
  25. Fahrig, L. 2019. “Habitat Fragmentation: A Long and Tangled Tale.” Global Ecology and Biogeography 28: 33–41. [Google Scholar]
  26. Fountain, T. , Husby A., Nonaka E., et al. 2018. “Inferring Dispersal Across a Fragmented Landscape Using Reconstructed Families in the Glanville Fritillary Butterfly.” Evolutionary Applications 11: 287–297. [Google Scholar]
  27. Frankham, R. 1996. “Relationship of Genetic Variation to Population Size in Wildlife.” Conservation Biology 10: 1500–1508. [Google Scholar]
  28. Galán‐Acedo, C. , Fahrig L., Riva F., and Schulz T.. 2024. “Positive Effects of Fragmentation Per Se on the Most Iconic Metapopulation.” Conservation Letters 17: e13017. [Google Scholar]
  29. González‐Fernández, A. , Arroyo‐Rodríguez V., Ramírez‐Corona F., Manjarrez J., Aguilera‐Hernández A., and Sunny A.. 2019. “Local and Landscape Drivers of the Number of Individuals and Genetic Diversity of a Microendemic and Critically Endangered Salamander.” Landscape Ecology 34: 1989–2000. [Google Scholar]
  30. Goudet, J. 2005. “HIERFSTAT, a Package for R to Compute and Test Hierarchical F‐Statistics.” Molecular Ecology Notes 5: 184–186. [Google Scholar]
  31. Haddad, N. M. , Gonzalez A., Brudvig L. A., Burt M. A., Levey D. J., and Damschen E. I.. 2017. “Experimental Evidence Does Not Support the Habitat Amount Hypothesis.” Ecography 40: 48–55. [Google Scholar]
  32. Hanski, I. 2011b. “Eco‐Evolutionary Spatial Dynamics in the Glanville Fritillary Butterfly.” Proceedings of the National Academy of Sciences of the United States of America 108: 14397–14404. [DOI] [PMC free article] [PubMed] [Google Scholar]
  33. Hanski, I. 2015. “Habitat Fragmentation and Species Richness.” Journal of Biogeography 42: 989–993. [Google Scholar]
  34. Hanski, I. , and Gaggiotti O.. 2004. “Preface.” In Ecology, Genetics and Evolution of Metapopulations, edited by Hanski I. and Gaggiotti O. E., xiii–xiv. Academic Press. [Google Scholar]
  35. Hanski, I. , Kuussaari M., and Nieminen M.. 1994. “Metapopulation Structure and Migration in the Butterfly Melitaea cinxia .” Ecology 75: 747–762. [Google Scholar]
  36. Hanski, I. , Moilanen A., Pakkala T., and Kuussaari M.. 1996. “The Quantitative Incidence Function Model and Persistence of an Endangered Butterfly Metapopulation.” Conservation Biology 10: 578–590. [Google Scholar]
  37. Hanski, I. , Mononen T., and Ovaskainen O.. 2011. “Eco‐Evolutionary Metapopulation Dynamics and the Spatial Scale of Adaptation.” American Naturalist 177: 29–43. [DOI] [PubMed] [Google Scholar]
  38. Hanski, I. , Pakkala T., Kuussaari M., and Lei G.. 1995. “Metapopulation Persistence of an Endangered Butterfly in a Fragmented Landscape.” Oikos 72: 21. [Google Scholar]
  39. Hanski, I. , Schulz T., Wong S. C., Ahola V., Ruokolainen A., and Ojanen S. P.. 2017. “Ecological and Genetic Basis of Metapopulation Persistence of the Glanville Fritillary Butterfly in Fragmented Landscapes.” Nature Communications 8: 14504. [DOI] [PMC free article] [PubMed] [Google Scholar]
  40. Hanski, L. 1999. Metapopulation Ecology. Oxford University Press. [Google Scholar]
  41. Herrero‐Jáuregui, C. , Camba G., Andries D. M., Aguiar S., Fahrig L., and Mastrangelo M.. 2022. “Past and Present Effects of Habitat Amount and Fragmentation Per Se on Plant Species Richness, Composition and Traits in a Deforestation Hotspot.” Biological Conservation 276: 109815. [Google Scholar]
  42. Hesselbarth, M. H. K. , Sciaini M., With K. A., Wiegand K., and Nowosad J.. 2019. “Landscapemetrics: An Open‐Source R Tool to Calculate Landscape Metrics.” Ecography 42, no. 10: 1648–1657. [Google Scholar]
  43. Holderegger, R. , and Wagner H. H.. 2008. “Landscape Genetics.” Bioscience 58: 199–207. [Google Scholar]
  44. Hu, Y. , Fan H., Chen Y., et al. 2021. “Spatial Patterns and Conservation of Genetic and Phylogenetic Diversity of Wildlife in China.” Science Advances 7: eabd5725. [DOI] [PMC free article] [PubMed] [Google Scholar]
  45. Huxel, G. R. , and Hastings A.. 1999. “Habitat Loss, Fragmentation, and Restoration.” Restoration Ecology 7: 309–315. [Google Scholar]
  46. Jackson, H. B. , and Fahrig L.. 2012. “What Size Is a Biologically Relevant Landscape?” Landscape Ecology 27: 929–941. [Google Scholar]
  47. Jackson, H. B. , and Fahrig L.. 2015. “Are Ecologists Conducting Research at the Optimal Scale?” Global Ecology and Biogeography 24: 52–63. [Google Scholar]
  48. Jackson, N. D. , and Fahrig L.. 2014. “Landscape Context Affects Genetic Diversity at a Much Larger Spatial Extent Than Population Abundance.” Ecology 95: 871–881. [DOI] [PubMed] [Google Scholar]
  49. Keyghobadi, N. , Roland J., Matter S. F., and Strobeck C.. 2005. “Among‐ and Within‐Patch Components of Genetic Diversity Respond at Different Rates to Habitat Fragmentation: An Empirical Demonstration.” Proceedings of the Royal Society B: Biological Sciences 272: 553–560. [DOI] [PMC free article] [PubMed] [Google Scholar]
  50. Kitada, S. , Nakamichi R., and Kishino H.. 2021. “Understanding Population Structure in an Evolutionary Context: Population‐Specific F ST and Pairwise F ST.” G3: Genes, Genomes, Genetics 11: jkab316. [DOI] [PMC free article] [PubMed] [Google Scholar]
  51. Lebigre, C. , Turlure C., Vandewalle H., Binard F., Habel J. C., and Schtickzelle N.. 2022. “Diverging Effects of Geographic Distance and Local Habitat Quality on the Genetic Characteristics of Three Butterfly Species.” Ecological Entomology 47: 842–854. [Google Scholar]
  52. Lindenmayer, D. B. , and Fischer J.. 2007. “Tackling the Habitat Fragmentation Panchreston.” Trends in Ecology & Evolution 22: 127–132. [DOI] [PubMed] [Google Scholar]
  53. Lomolino, M. V. 2000. “Ecology's Most General, Yet Protean Pattern: The Species‐Area Relationship.” Journal of Biogeography 27: 17–26. [Google Scholar]
  54. Martin, A. E. , Bennett J. R., and Fahrig L.. 2021. “Habitat Fragmentation.” In The Routledge Handbook of Landscape Ecology, 1st ed., 118–139. Routledge. [Google Scholar]
  55. McGarigal, K. , Cushman S. A., Neel M. C., and Ene E.. 2002. FRAGSTATS: Spatial Pattern Analysis Program for Categorical Maps. Computer Software Program Produced by the Authors at the University of Massachusetts. [Google Scholar]
  56. Nieminen, M. , Siljander M., and Hanski I.. 2004. “Structure and Dynamics of Melitaea cinxia Metapopulations.” In On the Wings of Checkerspots: A Model System for Population Biology, edited by Ehrlich P. R. and Hanski I.. Oxford University Press. [Google Scholar]
  57. Ojanen, S. P. , Nieminen M., Meyke E., Pöyry J., and Hanski I.. 2013. “Long‐Term Metapopulation Study of the Glanville Fritillary Butterfly (Melitaea cinxia): Survey Methods, Data Management, and Long‐Term Population Trends.” Ecology and Evolution 3: 3713–3737. [DOI] [PMC free article] [PubMed] [Google Scholar]
  58. Pereira, H. M. , Ferrier S., Walters M., et al. 2013. “Essential Biodiversity Variables.” Science 339: 277–278. [DOI] [PubMed] [Google Scholar]
  59. Riva, F. , and Fahrig L.. 2022. “Landscape‐Scale Habitat Fragmentation Is Positively Related to Biodiversity, Despite Patch‐Scale Ecosystem Decay.” Ecology Letters 26: e14145. [DOI] [PubMed] [Google Scholar]
  60. Riva, F. , Koper N., and Fahrig L.. 2024. “Overcoming Confusion and Stigma in Habitat Fragmentation Research.” Biological Reviews 99: 1411–1424. [DOI] [PubMed] [Google Scholar]
  61. Saccheri, I. , Kuussaari M., Kankare M., Vikman P., Fortelius W., and Hanski I.. 1998. “Inbreeding and Extinction in a Butterfly Metapopulation.” Nature 392: 491–494. [Google Scholar]
  62. Savary, P. , Foltête J.‐C., Moal H., and Garnier S.. 2022. “Combining Landscape and Genetic Graphs to Address Key Issues in Landscape Genetics.” Landscape Ecology 37: 2293–2309. [Google Scholar]
  63. Schulz, T. , Vanhatalo J., and Saastamoinen M.. 2020. “Long‐Term Demographic Surveys Reveal a Consistent Relationship Between Average Occupancy and Abundance Within Local Populations of a Butterfly Metapopulation.” Ecography 43: 306–317. [Google Scholar]
  64. Seibold, S. , Bässler C., Brandl R., et al. 2017. “An Experimental Test of the Habitat‐Amount Hypothesis for Saproxylic Beetles in a Forested Region.” Ecology 98: 1613–1622. [DOI] [PubMed] [Google Scholar]
  65. Thornton, D. H. , Branch L. C., and Sunquist M. E.. 2011. “The Influence of Landscape, Patch, and Within‐Patch Factors on Species Presence and Abundance: A Review of Focal Patch Studies.” Landscape Ecology 26: 7–18. [Google Scholar]
  66. Tilman, D. , May R. M., Lehman C. L., and Nowak M. A.. 1994. “Habitat Destruction and the Extinction Debt.” Nature 371: 65–66. [Google Scholar]
  67. Turlure, C. , Choutt J., Van Dyck H., Baguette M., and Schtickzelle N.. 2010. “Functional Habitat Area as a Reliable Proxy for Population Size: Case Study Using Two Butterfly Species of Conservation Concern.” Journal of Insect Conservation 14: 379–388. [Google Scholar]
  68. Valente, J. J. , Gannon D. G., Hightower J., et al. 2023. “Toward Conciliation in the Habitat Fragmentation and Biodiversity Debate.” Landscape Ecology 38: 2717–2730. 10.1007/s10980-023-01708-9. [DOI] [Google Scholar]
  69. Visconti, P. , Bakkenes M., Baisero D., et al. 2016. “Projecting Global Biodiversity Indicators Under Future Development Scenarios.” Conservation Letters 9: 5–13. [Google Scholar]
  70. Wagner, D. L. , Grames E. M., Forister M. L., Berenbaum M. R., and Stopak D.. 2021. “Insect Decline in the Anthropocene: Death by a Thousand Cuts.” Proceedings of the National Academy of Sciences of the United States of America 118: e2023989118. [DOI] [PMC free article] [PubMed] [Google Scholar]
  71. Wahlund, S. 1928. “Zusammensetzung Von Populationen Und Korrelationserscheinungen Vom Standpunkt Der Vererbungslehre Aus Betrachtet.” Hereditas 11: 65–106. [Google Scholar]
  72. Wang, X. , Blanchet F. G., and Koper N.. 2014. “Measuring Habitat Fragmentation: An Evaluation of Landscape Pattern Metrics.” Methods in Ecology and Evolution 5: 634–646. [Google Scholar]
  73. Watling, J. I. , Arroyo‐Rodríguez V., Pfeifer M., et al. 2020. “Support for the Habitat Amount Hypothesis From a Global Synthesis of Species Density Studies.” Ecology Letters 23: 674–681. [DOI] [PubMed] [Google Scholar]
  74. Weir, B. S. , and Goudet J.. 2017. “A Unified Characterization of Population Structure and Relatedness.” Genetics 206: 2085–2103. [DOI] [PMC free article] [PubMed] [Google Scholar]
  75. With, K. A. 2004. “2—Metapopulation Dynamics: Perspectives From Landscape Ecology.” In Ecology, Genetics and Evolution of Metapopulations, edited by Hanski I. and Gaggiotti O. E., 23–44. Academic Press. [Google Scholar]

Associated Data

This section collects any data citations, data availability statements, or supplementary materials included in this article.

Supplementary Materials

Appendix S1.

MEC-34-e70037-s001.docx (461.8KB, docx)

Data Availability Statement

The data used for this study are publicly available, published in: DiLeo MF, Husby A, Saastamoinen M. 2019, November 16. Data from: Landscape permeability and individual variation in a dispersal‐linked gene jointly determine genetic structure in the Glanville fritillary butterfly. Dryad. Available from https://datadryad.org/dataset/https://doi.org/10.5061/dryad.mp25s15 (accessed June 9, 2025). The code of the analysis is available in: https://zenodo.org/records/14218771?token=eyJhbGciOiJIUzUxMiJ9.eyJpZCI6IjVjMDY4MDdlLTRhYzMtNGFhZi1hMDY4LWRkODRmZWZmM2NiOCIsImRhdGEiOnt9LCJyYW5kb20iOiIzYTgxOTk5ODY2ZjE1YjUwYTIwZTk2MGRmMDgyZmUwMiJ9.xr7sdfs1WVDXVZ2Oc_N6cYlpkXrRpUMcjmZfSzzO_HjjYJGFe51ydhBaas‐rNKooX2GNvbFNtXdXTINF3I40wQ

Articles from Molecular Ecology are provided here courtesy of Wiley

RESOURCES