Research Abstract.--In the context of increasingly fragmented natural habitats and global declines in biodiversity, there is an urgent need to better understand the spatial complexities of rare wildlife populations and the scales at which they operate. Metapopulation theory, originally based on a binary view of habitat patches in a non-habitat matrix, provides one conceptual framework which has been widely accepted for its potential relevance to fragmented populations. A large body of theoretical work has followed in this arena, but remains largely unsubstantiated by empirical studies of real populations. We are addressing this problem with an unprecedented landscape-level investigation into the population dynamics of marbled salamanders (Ambystoma opacum) among clusters of vernal pools in western Massachusetts. In the preliminary phase of this study (1998-2001), we identified optimal study locations, built an extensive field infrastructure, and compiled baseline data on amphibian populations at these sites. In a continuation of this work (2002 to present), we are building and analyzing a multi-year data set for several "breeding populations" in order to address the following objectives:

1) Characterize the range of demographic variation and temporal synchrony among local marbled salamander populations.

2) Quantify dispersal rates among breeding sites and subsequently, the degree of effective isolation or interaction among these sub-populations.

3) Build spatially-explicit and empirically-parameterized population models to generate informed hypotheses about the scale(s) at which marbled salamander populations operate and the significance of sub-population interactions to overall population viability.

Toward these objectives, we are continuously monitoring drift fence arrays which completely encircle 14 breeding ponds in a 1-km-radius study area. Amphibians are captured, recorded, and released as they enter or leave pond basins during breeding migrations and post-metamorphosis emergence. Photographic-recapture methods are being used (in collaboration with Dr. Sai Ravela at MIT; see http://web.mit.edu/ravela) to uniquely identify all adult marbled salamanders and mark-recapture methods associate metamorph salamanders with their natal ponds. A range of vital rates and demographics will be calculated from capture-recapture data and skeletochronology analyses, including local population sizes, sex ratios, age at maturity, reproductive success, and longevity. These rates will be supplemented by abundance data for other amphibians and environmental data characterizing pond and terrestrial habitat features.

We hope to greatly augment our understanding of local dynamics in amphibian populations and to provide data to rigorously quantify dispersal rates in marbled salamanders. In addition, this work will provide a thoroughly documented empirical example of how amphibian populations operate at a landscape-scale, offering to assess the relevance of metapopulation theory to pond-breeding amphibians and potentially to provide alternative conceptual frameworks. To the conservation and natural resource policy communities, this work will provide insights regarding the potential effectiveness or shortfalls of existing regulations targeting ephemeral wetlands and will provide much-needed contextual information to better direct conservation strategies for vernal pool amphibians. While most current efforts remain centered on individual breeding sites and surrounding "buffer zones", this work may provide compelling evidence that larger-scale conservation planning is requisite to ensure the persistence of spatially dynamic populations.