INTRODUCTION
Habitat (patch) size is a fundamental determining factor of an organism’s population size, as it is inversely related to local extinction probability (Rosenzweig, 1995; Tjørve & Tjørve, 2017). Likewise, the degree of isolation influences a species’ colonization rate as it is directly related to the degree of connectivity across a landscape (Thrall et al., 2000; Wang & Altermatt, 2019). Both Island biogeography theory (IBT) (MacArthur & Wilson, 1967; Whittaker et al., 2008) and the metapopulation theory (Levins, 1969; Hanski & Gaggiotti, 2004) predict population dynamics across the spatial variation of patch size and connectivity. In metapopulation regional survivability depends on threshold levels of migration events among subpopulations of spatially isolated habitat patches as unoccupied patches are constantly (re)colonized from dispersal events (Levins, 1969; Hanski & Gaggiotti, 2004). Adopting the metapopulation concepts offers a fundamental framework from which to study population demographic and evolutionary processes in fragmented habitats (Hanski & Gaggiotti, 2004; Hanski, 2012; Hanski et al., 2017) and for landscape conservation planning (Thrall et al., 2000; Fahrig, 2019). Indeed, preserving the effective dispersal of organism individuals and their genes among habitat patches (i.e., plant functional connectivity ) is crucial for long-term metapopulation persistence across a fragmented landscape (Auffret et al., 2017; Hanski et al., 2017; Vellend et al., 2017).
Anthropogenic deforestation is radically transforming landscape configurations worldwide (Haddad et al., 2015; Taubert et al., 2018; Fischer et al., 2021). In biodiverse hotspots such as the Amazon, projections point out that deforestation is increasing at a 0.5% rate per year, thus dramatically augmenting the density of small tropical forest patches (Taubert et al., 2018). To date, the Amazon Forest has lost an estimated 17% of its original forest in the past half century, while more than 50% of the remaining forests are characterized as degraded (Lovejoy & Nobre, 2018; Matricardi et al., 2020; Lapola et al., 2023; Albert et al., 2023). Such rapid ecosystem-wide changes continue to preclude species’ ability to adapt to such spatial disturbances over short time scales (Albert et al., 2023) resulting in impaired global biodiversity and ecosystem functions (Sala et al., 2000; Morris, 2010; Gomes et al., 2019; Daskalova et al., 2020). The effects of habitat fragmentation are separated into three components, outright habitat loss, reduced remnant size, and increased isolation, all of which effectively drive metapopulations into a non-equilibrium state (Thrall et al., 2000). Ecosystem decay in fragments demonstrably accelerates biodiversity loss (Hanski et al., 2013; Chase et al., 2020) as predicted by the species-area relationship (SAR): smaller habitats harbor fewer species (Hanski et al., 2013). However, the compounding effects of forest fragmentation (such as edge effects, neighborhood density, lowered dispersal, or demographic stochasticity) further exacerbate demographic stability by increasing extinction risk and/or decreasing recolonization potential (Lovejoy et al., 1984; Chase et al., 2020; Scott et al., 2021).
The demographic dynamics (colonization/extinction) among isolated metapopulation patches are fundamental in determining standing genetic diversity (Aycrigg & Garton, 2014; Buza et al., 2000; Vellend & Geber, 2005; Wang & Altermatt, 2019) but it remains difficult to predict (Vranckx et al., 2012). For oceanic island species, genetic diversity generally declines with few exceptions (see Laenen et al., 2011) as patch size decreases and isolation increases (Costanzi & Steifetten, 2019; Hill et al., 2017; Whittaker et al., 2017; Hamabata et al., 2019). These studies assume that processes influencing island species diversity are reflected in the genetic structure allowing for opportunities to predict micro-evolutionary outcomes in light of community-level patterns (Losos & Ricklefs, 2009; Vellend, 2003; Vellend & Geber, 2005). Conservation genetic theory posits that stochastic extinction events, disproportionately experienced by smaller-sized populations, will lead to an increased loss of alleles, and genetic differentiation through genetic drift (Lowe et al., 2005; Aguilar et al., 2008). Interruptions to migration subsequently impede (re)colonization of unoccupied patches precluding the renewal of locally extinct genotypes (Lowe et al., 2005; Auffret et al., 2017).
The effects of patch size and isolation on genetic diversity depend on the time since isolation (Young et al., 1996; Aguilar et al., 2008). Due to the long generation times of most plants, integrating the demographic and genetic consequences of the recent exponential increase in forest fragmentation over multiple generations is scarce, particularly in tropical regions harboring the greatest plant diversity (Hamilton, 1999; Aldrich et al., 1998; Côrtes et al., 2013). Since the effects of human-induced habitat loss are relatively recent in an evolutionary context, long-lived species will take decades to reveal extinction debts as characterized by a loss of genetic variability through genetic drift (Morris et al., 2008; Vranckx et al., 2012). Terrestrial plant groups characterized by sufficiently accelerated demographic parameters (such as colonization, extinction, and local growth) are ideal systems to examine habitat fragmentation consequences in a tractable time frame (Pharo & Zartman, 2007; Spagnuolo et al., 2007; Zartman et al., 2006).
Bryophytes stand out as a model system to bridge demography and population genetics in the context of fragmentation (Pharo & Zartman, 2007). Bryophytes are characterized by widely divergent reproductive strategies (Snäll et al., 2004; Wang et al., 2013; Holá et al., 2015; Sierra et al., 2019a; Alonso-García et al., 2021; Lang et al., 2021) thus providing opportunities from which to generate predictions of dispersal trait related long-term gene-flow patterns in metapopulations (Obbard et al., 2006; Patiño et al., 2013). Moreover, habitat-dependent reproductive performance observed in bryophytes (Maciel-Silva et al., 2012) will help in the understanding of the species’ different sensitivity to habitat isolation (Sierra et al., 2019b).
Bryophyte metapopulations inhabiting the leaves of vascular plants (epiphyllous bryophytes) present short generation times (~6 months), where colonization is mainly carried out through the dispersal of microscopic spores and asexual propagules (Sierra et al., 2019a; Mežaka et al., 2019). Local patch occupancy is determined by the immediate environment (Sonnleitner et al., 2009), and neighboring colony densities (Zartman et al., 2012). Epiphyllous bryophytes are ideal bellwethers of biodiversity associated with landscape changes in diverse ecosystems in the Amazon and Atlantic Forests (Zartman, 2003; Alvarenga et al., 2009). At the Biological Dynamics of Forest Fragmentation Project (BDFFP) in the Amazonian basin, populations of epiphyllous bryophytes have persisted at low local abundances for decades (>40 years) in small forest fragments <10-ha (Sierra et al., 2019b) governed by suppressed colonization (Zartman & Shaw, 2006). In these small fragments after 20 years of isolation, Zartman et al., (2006) observed linkage disequilibrium between loci from amplified fragment length polymorphisms (AFLPs) markers, but it was no observed sign of genetic drift. However, a multi-species assessment of Amazonian bryophytes suggested a high genetic structure of populations across a large spatial scale (Ledent et al., 2020; Campos et al., 2022). Such contrasting results suggest a priori that the degree of functional connectivity among Amazonian bryophyte metapopulations is not universal and is dependent on the interaction of landscape configuration and species dispersal capacity (Auffret et al., 2017).
In this study, we focus on the demographic and genetic basis of metapopulation persistence of two epiphyllous bryophyte species distributed across a 10 000-km2 experimentally fragmented Amazonian landscape in the BDFFP. Herein, we combine evidence from long-term censuses of two ephemeral epiphyllous bryophytes to create a comprehensive profile of the ecological genetic impacts of fragmentation at a medium-term (half-century) time scale. Single nucleotide polymorphisms (SNPs) identified from Genotyping by Sequencing (GBS) were used to analyze the genetic diversity and differentiation in fragmented (1-, 10- and 100-ha) and continuous forests. Specifically, we addressed the following questions: Are demographic changes associated with the intensity of fragmentation (patch size and isolation) reflected in their genetic structures? Subsequently, does a species’ mating system confers different sensitivity at the genetic level to the effects of habitat fragmentation? Considering these questions, we hypothesized: (H1) reduced colony densities in smaller fragments over 40 years of isolation leads to population genetic drift (lower diversity and higher differentiation) when compared to continuous forests. Secondly (H2), these fragmentation impacts are less evident in bisexual as compared to unisexual species due to the former’s demonstrably greater reproductive performance and dispersal capacities (Maciel-Silva et al., 2012; Laenen et al., 2016). Consequently (H3), the landscape connectivity network of the bisexual species will present high migration (Figure 1: patchy metapopulation), with all patches exchanging migrants among them, irrespective of size and isolation. While the unisexual species, we expect to observe only migration from the continuous forest towards nearby small forest fragments at a low rate (Figure 1: island-mainland metapopulation).