Fidelity and dispersal in the pond-breeding amphibian, Ambystoma opacum: Implications for spatio-temporal population dynamics and conservation

Abstract

The spatio-temporal dynamics of amphibian populations and the models that describe them are largely influenced by the frequency of dispersal among breeding sites; however, dispersal has rarely been addressed rigorously in empirical studies. In a 7-year landscape-level investigation, we monitored breeding populations of marbled salamanders (Ambystoma opacum) among 14 seasonal ponds in western Massachusetts, USA, to quantify dispersal probabilities and distances. Emerging juveniles (n = 11,168) received cohort marks and adults (n = 5560 capture events) were photographed for individual identification using computer-aided dorsal pattern analysis. We found that 91.0% of first-time breeders returned to natal ponds to breed and 96.4% of experienced breeders maintained breeding site fidelity through multiple seasons. These findings confirm a high level of philopatry in this species and the prominence of local factors in determining local population trends. However, the remaining survivors dispersed to other ponds, with several individuals exceeding distances of 1000 m. Though breeding populations were clearly subdivided, dispersal at these rates may offset effects of genetic drift and inbreeding depression by increasing effective population size (through the aggregation of breeding populations). Outward dispersal probabilities were higher at ponds with small breeding populations and inward dispersal was biased toward larger populations, suggesting that salamanders were cueing to the presence of other individuals and/or to unmeasured habitat characteristics. Our findings suggest that small and dynamic local populations may operate inter dependently in a metapopulation context. Effective conservation strategies targeting these and similarly structured amphibian populations must address landscape-level processes (e.g., dispersal) as well as local demographic factors.

Introduction

Much debate has centered on whether the metapopulation paradigm accurately describes many amphibian populations (Semlitsch, 2000, Marsh and Trenham, 2001, Smith and Green, 2005). Though the terminology has been applied broadly, metapopulations are generally defined as groups of populations that experience routine local extinctions but may persist regionally as a result of dispersal and recolonization (Levins, 1969, Levins, 1970, Hanski and Gilpin, 1997). Broader interpretations include source-sink populations (Brown and Kodric-Brown, 1977), rescue-effect populations (Stacey et al., 1997), or virtually any spatially subdivided populations connected by occasional dispersal. Where the paradigm applies, it carries clear conservation implications, such as the need to focus on landscape-level connectivity (e.g., see Cushman, 2006) and to maintain both occupied and potential habitat patches. However, misapplication of the paradigm can be risky, as in cases where a fragmented population in decline may be viewed as a metapopulation at equilibrium or where deterministic local factors may be more important to population persistence than regional connectivity (Marsh and Trenham, 2001).

At face value, the metapopulation paradigm appears well suited to many amphibians. For example, pond-breeding amphibians: (1) are associated with discrete aquatic breeding sites, (2) often experience wide population fluctuations (Semlitsch et al., 1996, Green, 2003) that may vary temporally among proximate sites (Trenham et al., 2003) and (3) often appear to have high fidelity to natal ponds (i.e., philopatry) (Pechmann et al., 1991). The degree to which populations exhibit philopatry is critical in determining the scales at which populations operate; however, this is often the condition we know least about in real populations. For example, how often do individuals successfully disperse in a true ecological sense – that is, to leave a natal site and breed at a new location – and what factors regulate this behavior? Is dispersal a density-dependent response to local conditions or does it occur at more constant, predictable rates? How is dispersal success mediated by distance, landscape permeability, and/or active habitat selection? Addressing these types of questions is necessary to assess the applicability of the metapopulation paradigm and its conservation implications for pond-breeding amphibians (Semlitsch, 2002).

Though a number of studies have concluded that pond-breeding amphibians exhibit philopatry, most of these studies have been limited primarily to observations of returning individuals at a single breeding site. For example, both Husting, 1965, Whitford and Vinegar, 1966 conducted multi-year studies on spotted salamanders (Ambystoma maculatum) that were focused on single breeding sites with limited sampling at nearby ponds. In both cases, no marked animals from the primary sites were observed elsewhere and it was concluded that little or no dispersal occurred. Findings have differed in studies that have incorporated numerous breeding sites. For example, in a study of the California tiger salamander (Ambystoma californiense), Trenham et al. (2001) directly observed interpond dispersal at probabilities exceeding 20% in both first-time and experienced breeders. Gill (1978) documented perfect breeding site fidelity among adult red-spotted newts (Notophthalmus viridescens) but inferred dispersal by juveniles based on the persistence of several apparent sink populations. Multi-site studies of anurans have documented abundant dispersal in some cases (Breden, 1987) and infrequent, age-specific dispersal in others (Berven and Grudzien, 1990). To our knowledge, no published studies have directly measured dispersal in any Ambystomatids in eastern North America.

Here we present results from a 7-year field study in which we continuously monitored breeding populations of the marbled salamander (Ambystoma opacum) at 14 seasonal ponds in a western Massachusetts landscape. Our primary objectives in this paper are to:

• address pond fidelity in marbled salamanders by measuring successful return and dispersal probabilities among first-time and experienced breeders,

• use these empirical data to fit a distance-dispersal function that can (a) help to generalize our findings to different landscapes (with several important limitations) and (b) be used to set parameters for spatial population and connectivity models,

• describe several observations that reveal elements of dispersal behavior and may be used to frame future experimental work,

• discuss the implications of our findings for spatio-temporal population dynamics in this species and for conservation approaches to pond-breeding amphibians in general.