Like all other vertebrates, amphibian dispersal occurs primarily during the juvenile stage (Gill 1978, Berven & Grudzien 1990, Gamble et al. 2007). Initial juvenile movement represents the first stage of the multi-phase process of natal or juvenile dispersal (Bowler & Benton 2005, Semlitsch 2008). The first movements of individuals away from their natal site are particularly crucial as they encounter novel habitats and are highly vulnerable to mortality, lack knowledge of terrestrial resource patches, and have limited perceptual range (Rothermel & Semlitsch 2002, Rothermel 2004, Pittman & Semlitsch, unpubl. data) Examples of initial juvenile movement include: post-fledging movements of juvenile birds (e.g. Walls & Kenward 1998, White & Faaborg 2008, Delgado et al. 2009), hatchling turtles moving away from nests (e.g. Pappas et al. 2009), post-weaning juvenile hares leaving their burrow (e.g. Devillard & Bray 2009), and salmon fry moving into the water column after emerging from the streambed (e.g. De Garcia Leaniz et al. 2000, Bujold et al. 2004). For many species, initial juvenile movement represents an ecological bottleneck in which the majority of juveniles suffer high mortality. Yet, understanding what factors contribute to their movement and success at this initial fine-scale will enable us to begin to predict movement over larger scales and under differing land-use contexts (Lima & Zollner 1996, Zollner & Lima 2005, Nathan et al. 2008).
Pond-breeding amphibians are an ideal system for studying initial juvenile dispersal. Because of their complex life cycle (aquatic larval and terrestrial juvenile phase) pond-breeding amphibians are spatiotemporally concentrated at ponds. Often, 1,000s of metamorphs emerge in multiple bouts over the summer. They provide a source of subjects for studying movement behavior at smaller spatial scales than would be possible for most other vertebrates. Also, the acute susceptibility of amphibians to desiccation (Jorgensen 1997) and limited energetic reserves of recent metamorphs (Scott et al. 2007) make them a particularly sensitive group for examining the effects of landscape and habitat heterogeneity on initial juvenile movement (Rothermel & Luhring 2005, Rothermel & Semlitsch 2002). Habitat heterogeneity and quality have been shown to be important drivers of dispersal (e.g. Ferreras et al. 1992, Bowler & Benton 2005, Revilla and Wiegand 2008).
The specific objective of our research is to link habitat, physiology, and behavior to provide an empirical determination of initial dispersal and habitat resistance based on a real landscape. We will then use empirically-derived resistance layers to modify an existing spatially-explicit Individual Based Model (IBM) called ‘SEARCH’ developed by Pat Zollner-Purdue University (Pauli et al. in prep) to predict dispersal among populations in a second novel landscape. The final step is to test dispersal predictions against estimates of gene flow derived from genetic microsatellite markers across that second landscape (via a DOD SERDP project funded in 2011-see below). Our approach is novel in that we will combine measures of microclimate, energy use, risk of desiccation and predation into GIS data layers that will be used to parameterize and scale-up spatially explicit IBMs. A future application and extension is to link these empirically-based and tested IBMs in other landscapes and with other species that vary in life histories and movement abilities. Results from our proposed research will greatly increase our understanding of animal movement, genetic structure, connectivity, source-sink dynamics, and species occupancy and persistence in human-dominated landscapes. It will also clarify whether incorporating individual behavior into models of connectivity yield better results than existing approaches such as Least-Cost Path or Isolation-by-Resistance modeling approaches based on structural landscape features alone.
Metapopulation theory has commonly been used to address conservation questions, especially in view of increasing loss or alteration of habitat and fragmentation due to land use that can jeopardize the persistence of species. The essential demographic processes are that local populations produce enough offspring to replace those lost to mortality, and that surplus dispersing individuals can “rescue” declining populations (called “sinks”, Pulliam 1988) from extinction, reduce inbreeding depression, and colonize new habitats to maintain a species across the region.
However, two processes are critical for effective metapopulation dynamics and species persistence. First, some local populations must exist in high-quality habitat patches (called “sources”; Pulliam 1988) and produce surplus offspring frequently enough to balance losses in sink populations that occupy low-quality habitat. Secondly, connectivity among habitat patches on the landscape must be maintained to allow successful dispersal and rescue of sink patches or for colonization of new habitat patches. This second process is the result of species vagility, distance between ponds, and quality of the matrix between patches. Despite a clear understanding of the theoretical importance of source-sink dynamics, measurement of both these ecological processes in natural populations has seldom been achieved.
A clear understanding of source-sink dynamics and connectivity of habitat patches is essential for conservation and management as natural habitats become lost, altered, and fragmented due to land use. This is absolutely required for on-the-ground management of species at risk to ensure long-term persistence at a regional level. However, the spatial and temporal dynamics of most metapopulations are not well understood for most species, and especially for globally imperiled amphibians. Thus, the overall objectives of our research are to: 1) identify key environmental conditions associated with varying levels of reproductive success of ringed salamanders to define source habitat patches, 2) identify habitat features associated with effective dispersal of ringed salamanders among populations to define habitat connectivity, 3) examine the effects of a competitor, Ambystoma maculatum (see egg masses below), on ringed salamander larval success, and 4) develop models based on circuit and graph theory that can be used to predict and manage population connectivity in heterogeneous landscapes. Our research seeks to understand the impact of landscape heterogeneity on the source-sink dynamics of a species of management concern (ringed salamander, Ambystoma annulatum) at Fort Leonard Wood, Missouri.
Our approach is multi-faceted in that we will use ecological, genetic, and modeling techniques that allow us to sample and assess populations at multiple spatial and temporal scales. This multi-scale approach is powerful and essential to understand the dynamic nature of source-sink populations. We will sample the abundance of ringed salamanders at local populations across seasons to understand the environmental features associated with reproductive success or failure of ponds-as-patches. We will sample populations across the landscape to understand connectivity and occupancy in relation to dispersal habitat quality and landscape heterogeneity. We will sample gene flow with polymorphic microsatellite markers over multiple generations among nearly 200 populations across the landscape to understand dispersal routes and which populations are providing emigrants. We will develop connectivity models based on actual occupancy, abundance, GIS habitat analysis, and dispersal inferred from genetic analyses that can be used to manage the ringed salamander. This will result in a clear understanding of the source-sink dynamics of a representative species that can be used to direct management at Fort Leonard Wood to ensure species persistence, link management at this site to regional dynamics within the adjacent U.S. Forest Service, Mark Twain National Forest, and possibly be applied to other ambystomatid salamanders of concern at other DOD facilities or public lands.
Disturbance is an important element of forest ecosystems. Both natural and human-caused disturbances can have substantial impacts on forests, modifying community structure and causing extensive tree mortality.
Fire disturbance acts as a driver of ecosystem composition, and fire regimes have changed rapidly over the last several decades due to human population growth and changes in land management practices. Catastrophic wildland fires in 2000 drew national attention to the effects of large fires, which prompted the development of the National Fire Plan (NFP). Fire suppression causes the buildup of combustible forest material (fuel) and creates ideal conditions for intense wildland fires. Wildlife responses to fire vary considerably among species. For instance, many vertebrate species seek refuge or flee from approaching fire, but some vertebrates are attracted to burning areas. Fire seasonality and severity have substantial impacts on wildlife responses and the extent of direct mortality of animals. A few reports found fire-related injury to herpetofauna, but suggest that negative indirect responses occur in fossorial species that require cover objects that may be consumed during fires. Amphibians are ideal for investigating these forest management practices because they are especially sensitive to changes in microhabitat.
Timber harvest is another common disturbance with substantial effects on forest ecosystems. Unlike prescription burns, timber harvest removes at least a portion of the forest canopy. It is generally accepted that timber harvest adversely affects forest-dependent wildlife and that species and life stages can be affected differently (e.g., Semlitsch et al. 2009). We will study the effects of both prescribed fire and timber harvest experimentally in order to explore how various types of disturbances affect amphibians. Specifically, we address the following questions: 1) How do amphibians respond to prescribed burns, timber harvest, and their interaction at the population level? 2) How do amphibians respond behaviorally to prescribed burns and microhabitat changes due to fire?
The population-level study will take place in the Sinkin Experimental Forest, located within the Mark Twain National Forest in the Ozark Highlands region of southeastern Missouri. Our focus of study will be the southern redback salamander, Plethodon serratus. The study site consists of mature (80-100 year old) fully-stocked stands. Oaks (Quercus spp.) dominate the overstory (70% BA), with shortleaf pine (Pinus echinata), hickory (Carya spp.), and maple (Acer spp.) comprising most of the understory. None of the sites have experienced major disturbance within the last 15-20 years. Additional studies will take place at the Daniel Boone Conservation Area in Warren County, Missouri. These studies will focus on the western slimy salamander, Plethodon albagula. The site has an oak (Quercus spp.) and hickory (Carya spp.) dominated overstory, with an understory of sugar maples (Acer saccharum).
Understanding large-scale responses of wildlife to prescribed fire and timber harvest and the mechanistic factors underlying those responses is important for developing more effective forest management practices. Our results will enhance our ability to improve wildlife habitat and protect forest biodiversity.