Drivers of survival in a small mammal of conservation concern: An assessment using extensive genetic non-invasive sampling in fragmented farmland

Highlights

• Assessing the demography of rare and elusive species is often problematic.

• We evaluated population size and survival of a small-mammal using large-scale gNIS.

• Population size increased during the wet season and decreased during the dry season.

• Survival was higher near road-verges and water-bodies, and lower near farmed areas.

• Our approach provides key information for conserving elusive small mammals.

Abstract

Although important to guide conservation management, detailed demographic studies on rare or elusive species inhabiting fragmented, human-dominated landscapes are often hampered by the species' low densities, and the logistic and ethical constraints in obtaining reliable information covering large areas. Genetic non-invasive sampling (gNIS) provides cost-effective access to demographic information, though its application to small mammals is still scarce. We used gNIS to infer on the demography of an endemic small mammal, the Cabrera vole (Microtus cabrerae), occurring as a spatially-structured population in a 462-ha Mediterranean farmland landscape. We intensively sampled fresh vole feces in four seasons, extracted the DNA, and performed individual identification based on genotypes built using nine microsatellites. We then estimated population size and individual survival relative to environmental variables, controlling for heterogeneity in capture probabilities using capture-mark-recapture modelling. Population size increased during the wet season and decreased during the dry season, while survival remained constant across the study period. Individuals captured along road-verges and around water-bodies survived longer than those captured near agricultural fields. The use of gNIS on a heterogeneous landscape such as our study area allowed us to demonstrate that human land-use activities affect Cabrera vole demographic parameters in Mediterranean farmland, with implications for conservation planning towards its long-term persistence. Our approach can be widely applied to other elusive small mammals of conservation concern, but for which informative demographic data are still scarce.

Introduction

Estimating large-scale demographic patterns (e.g. abundance, population growth, survival) of animal species in relation to both individual traits (e.g. sex, age, weight) and environmental factors (e.g. climate or land-use change) is a difficult but necessary goal to understand species ecology and sustain conservation policies (Smallwood and Schonewald, 1998; Williams et al., 2002). This is particularly true for species occurring in agricultural landscapes where major declines in biodiversity due to agricultural intensification have been reported worldwide (Tscharntke et al., 2012). However, achieving these goals is often difficult due to a number of technical, ethical, and logistic constraints in data collection, particularly for species that are rare, elusive, or otherwise hard to capture or observe.

Capture-Mark-Recapture (CMR) is one of the most popular methods to assess demographic parameters in animal populations (Lebreton et al., 1992), and hence to understand species' biology and ecology in different environments (Smallwood and Schonewald, 1998). Traditional CMR studies have been mostly based on live-trapping techniques, which are usually logistically difficult to implement over large spatial and temporal scales and often expensive (Cheng et al., 2017). In addition, because live-trapping implies both physical confinement and handling of animals, it often involves behavioural and physiological responses due to trapping-induced stress (Beja-Pereira et al., 2009; De Bondi et al., 2010). Stress responses can be reduced with the use of minimally invasive techniques such as camera-trapping, which is expected to be more time-efficient than live-trapping and does not require physical capturing and handling of animals (De Bondi et al., 2010; Mondol et al., 2009). However, camera trapping is unsuitable for CMR studies in species that are difficult to morphologically identify at the individual level, which is the case of most small mammal species (Glen et al., 2013). Furthermore, in the case for rare and elusive species, both live- and camera-trapping often yield insufficient data to be used in CMR models, thus hampering proper evaluation of their population status and trends (Burgar et al., 2018; Mondol et al., 2009).

Genetic non-invasive sampling (gNIS) has been increasingly used to estimate demographic parameters of species that are difficult to trap, mainly due to decreased field sampling effort, ever decreasing lab costs, and increasing DNA amplification success (Beja-Pereira et al., 2009; Marucco et al., 2011). Despite its limitations in retrieving information on relevant individual traits like age, body mass, or reproductive condition, gNIS can provide a more cost-effective solution than traditional live-trapping (Cheng et al., 2017; Ferreira et al., 2018). DNA extracted from non-invasive samples (e.g. feces, hairs, feathers) allows the identification of individuals, providing data that can be easily combined with CMR methods to obtain population parameters that otherwise would be difficult to obtain over large spatial scales (Cheng et al., 2017; Petit and Valiere, 2006). However, to date, applications of gNIS in CMR studies have mostly focused on large and medium-sized mammal species, and often provide snapshots of population size estimates rather than variations over time (but see Brøseth et al., 2010 for an example). Furthermore, very few studies have used gNIS to estimate other important population parameters such as survival (Lampa et al., 2015; Marucco et al., 2012; Zielinski et al., 2013). In the case of small mammals, while some recent studies have used gNIS to estimate population density (DeMay et al., 2017; Gillet, 2016; Sabino-Marques et al., 2018) or to infer dispersal (Ferreira et al., 2018; Gillet, 2016), to our knowledge no study has yet explored the application of this method to understand how demographic parameters relate to large-scale environmental variation.

In this study, we combined gNIS and CMR methods to assess the seasonal variations in abundance, and to evaluate factors affecting survival probability of an elusive small mammal species in a Mediterranean farmland landscape. We focused on the ‘near-threatened’, Iberian endemic Cabrera vole (Microtus cabrerae, Thomas 1906), for which genotyping protocols based on fecal samples have been recently optimized (Barbosa et al., 2013; Ferreira et al., 2018). Additionally, previous studies have also shown the ability of gNIS to provide reliable density estimates for this species (Sabino-Marques et al., 2018). Based on repeated surveys of Cabrera vole feces, we explored the potential of gNIS to (i) assess the seasonal variation in population abundance; and (ii) estimate capture and survival probabilities in relation to variables reflecting survey conditions (genotyping success and season), individual traits (sex), and local and landscape environmental features. We considered variables that might affect survival both positively (e.g. patch area and presence of water) and negatively (e.g. isolation, patch persistence, interactions with the competitor Arvicola sapidus, and human disturbances) (Pita et al., 2014) (see Table 1 for a full description and rationale of covariates considered). Overall, our study illustrates the use of gNIS within a CMR framework, demonstrating its application to retrieve demographic data from elusive small mammals, thus enhancing conservation planning in areas that have been highly modified by human activities.