NATURAL ENVIRONMENTAL GRADIENTS PREDICT THE MICROHABITAT USAGE, FINE-SCALE DISTRIBUTION, AND ABUNDANCE OF THREE WOODLAND SALAMANDERS IN AN OLD-GROWTH FOREST

Introduction:�Biological gradients

• Biological patterns along physical gradients
• Preservation of biodiversity
• Ecology of fauna across environmental gradients
• Niche requirements
• Population dynamics
• Biotic interactions
• Strong responses
• Limited dispersal (Cushman 2006)
• Low reproductive output (Elton 2000)
• Acute sensitivity to environmental conditions (Buckley & Jetz 2007)

​

​

​

BD Todd, UC Berkeley

(Hutchinson 1957,

MacArthur & Wilson 1976, Simberloff 1974)

Introduction:�Amphibians

• Acute sensitivity to…
• Chemical environment
• Thermal and hydrologic regimes
• Microbiome (i.e. emerging disease)
• Carnivorous ectotherms
• Prey availability
• Landscape structure
• Global population declines (Houlahan et al. 2000, Stuart et al. 2004)
• Effective ecological indicator (Welsh & Ollivier 1998, Welsh & Droege 2001)
• Tremendous component of ecosystem biomass
• Aquatic (Gibbons et al. 2006)
• Terrestrial (Burton & Likens 1975)
• Riparian (Peterman et al. 2008)

​

Capps et al. 2014, Luhring et al. 2017

Introduction:�Woodland salamanders

• Terrestrial woodland salamanders

(direct-developing, lungless salamanders)

• Most abundant vertebrates in eastern deciduous forests^†

• Key energetic intermediary
• Detrital communities (Burton and Likens 1975)
• High-order predators (Semlitsch 2014)
• Top-down regulation of energy loss^‡
• Predation of detrital food webs
• Reduce decomposition, increase organic matter retention
• Bottom-up augmentation of energy accumulation^§
• Transformation of detrital nutrients
• Transportation of energy to forest consumers

​

^‡ Wyman 1998, Walton 2013

^§ Burton & Likens 1975, Semlitsch 2014

^†Burton & Likens 1975, Petranka & Murray 2001, Semlitsch et al. 2014

Introduction:�Role of woodland salamanders

• Wyman 1998—top-down effects (seminal paper)
• Woodland salamanders reduce detrital processing rates 11–17%
• Walton 2005, 2013—environmental variability
• Change the sign and magnitude of top-down effects
• Hickerson et al. 2017—predator abundance
• Increased salamander abundance amplifies top-down effects
• The role of woodland salamanders is contingent upon:
• Distribution of their biomass in the system
• Gradients of environmental conditions
• Population dynamics?

Introduction:�Population Ecology

• Population dynamics of single-species
• Environmental gradients (Peterman & Semlitsch 2013)
• Presence of heterospecifics influence
• Distribution (Hairston 1950, Jaeger 1970, 1971)
• Abundance (Hairston 1951)
• Microhabitat usage (Keen 1982, Farallo & Miles 2016)
• Diversity & endemism
• USA: 188 species
• Geographically nuanced communities
• >100 years of ecological studies of woodland salamanders
• Complex natural history & ecology
• Museum collections, occurrence data, or ‘count indices’
• Solid data for inferring population dynamics?

​

​

​

Introduction:�Challenges & Objectives

• Presence-only data—confounded by…
• Subsurface migration (temporary emigration)
• Observer ability
• Non-detection doesn’t imply absence
• Count indices—confounded by…
• Season/weather (surface activity)
• Observer ability
• Counts underestimate true abundance
• Imperfect detection

Objectives

Determine if natural environmental gradients influence the 1.) microhabitat use, fine-scale distribution, and abundance of terrestrial salamanders; and 2.) determine if those patterns are species-specific.

​

​

​

​

​

​

They’ll never find me in here

Methods:�Study Location

• Lilley Cornett Woods Appalachian Ecological Research Station
• 102 ha of old-growth forest
• Virtually no anthropogenic disturbance

​

​

• Canopy gaps: a natural disturbance regime
• Increase solar exposure
• Evapotranspiration
• Understory veg
• Environmental gradients
• Dissected topography
• modifies local climate
• edaphic variation

​

Methods:�Sampling plots:

• “Shop Hollow”
• 44.25 ha of old-growth forest
• Experiences little disturbance from guided hiking
• Circular, 0.08 ha sampling plots (N=40)

Shop Hollow

0.08 ha

ca. 7% upland area

Kelley Hoefer

Methods: �Salamander species

• Caudata:Plethodontidae:Plethodon
• Woodland salamanders

Methods:�Salamander Surveys

• Sampling events
• Four 2-day events, 5 days apart
• 15 Oct – 13 Nov 2016
• Visual Encounter Surveys
• 40 circular 0.08-ha sample plots
• Randomized transect placement
• Natural cover (woody & rocky)
• Animal Measurements (IACUC# 05-2015)
• Mass:
• 10- & 20-gram PESOLA scales
• Length:
• Snout-to-vent length (green)
• Tail length (purple)

​

​

​

​

​

Methods:�Microhabitat surveys

0.08 ha

• Survey of cover items
• Woody cover: N = 491
• Rocky cover: N = 800
• Capture Rate
• Woody cover: 14% inhabitance
• Rocky cover: 4% inhabitance
• Inhabited (N = 108)
• Soil moisture (Decagon Devices Inc., Pro Check)
• Surface temperature (Kintrex, IRT0421)
• Uninhabited (N = 1191)
• Surface temperature
• Ambient (inhabited + uninhabited) (N = 1297)
• Surface temperature

​

3 m

Methods:�Site Covariates

Quantifying site-level variation in environmental conditions using:

I) In situ measurements

• Soil moisture
• Volumetric soil content (Pro Check, Decagon Devices, Inc.)
• Measured at 5 equidistant points during each survey (grand mean)
• Canopy openness
• Hemispherical canopy photography (Fall 2016, prior to leaf off)
• 24 megapixel DSLR camera w/ 10.5 mm fisheye lens
• Leveled with tripod; angled at underside of canopy
• Analyzed with ImageJ (Abramoff et al. 2004)

II) Geospatially derived…

Methods:�Site Covariates—Geospatial

0.08 ha

• Geospatial covariates
• Primary layers
• Aspect
• Slope
• Elevation

​

​

• Secondary layers
• Topographic Position Index
• Beers Aspect (linearized)
• Direct Solar Radiation

​

3 m

Study Location

Methods:�Site Covariates—Geospatial

• Topographic Position Index
• Ravine Ridge-top
• Measure of topographic convexity

​

3 m

Methods:�Site Covariates—Geospatial

• Topographic Position Index (“Topographic Convexity”)
• Beers Aspect (Beers et al. 1996)
• Linearizes aspect (circular data)
• Measure of northeasterliness

​

​

​

​

Methods:�Site Covariates—Geospatial

• Topographic Position Index (“Topographic convexity”)
• Beers Aspect (“Northeasterliness”)
• Direct Solar Radiation (“Solar radiation”)
• = total solar radiation – [reflected solar radiation + diffuse solar radiation]
• Watts/hour/m²
• During specified sampling period

​

Methods:�Site Covariates—Geospatial

• Topographic Position Index (“Topographic convexity”)
• Beers Aspect (“Northeasterliness”)
• Direct Solar Radiation (“Solar radiation”)
• Normalized Difference Vegetation Index (NDVI)
• Barren Heavily vegetated
• Measure of canopy density

​

​

​

Methods:�Site Covariates—Geospatial

• Topographic Position Index (“Topographic convexity”)
• Beers Aspect (“Northeasterliness”)
• Direct Solar Radiation (“Solar radiation”)
• Normalized Difference Vegetation Index (“Canopy Density”)

​

​

Methods:�Sampling covariates

Image: BD Todd, UC Berkeley

• Quantifying variation in sample event conditions
• Quantity of coarse woody debris (diameter < 30 mm)
• Quantity of rocky cover items (area ≥ cobble area)
• Depth of leaf litter measured at 5 equidistant points along transect
• Time of sample event (24 hrs)
• Date of sample event (Julian date)
• Solar conditions (luminous flux)
• Digital illuminance meter (TekPower, LX1330B); measured at breast height

Methods:�Data analysis—microhabitat & body size

Image: BD Todd, UC Berkeley

Does body size predict thermal microhabitat preference?

• Ordinary least square regression (α = 0.05)
• Compare body sizes among species: two-way ANOVA; Tukey’s honest significant difference (HSD) test (α = 0.05)

Do thermal microhabitat preferences vary among species?

• Two-way ANOVA; Tukey’s HSD test

Do thermal microhabitat preferences differ from ambient microhabitat temperature?

• Multiple independent t-tests

​

Statistical procedures performed in R programming environment (v. 3.4.1., R Core Team 2017)

Methods:�Data analysis—hierarchical modeling

Image: BD Todd, UC Berkeley

Do environmental gradients influence the occupancy and abundance of Plethodon salamanders in LCW?

• Detection probabilities assumed <1 (Bailey et al. 2004)
• Hierarchical models estimate occupancy & abundance (MacKenzie & Royle 2005)
• Occupancy models estimated probability a species occupies a given site, ψ (MacKenzie et al. 2002)
• N-mixture models estimated species’ true abundance, λ (Royle 2004)
• “unmarked” & “AICcmodavg”, R Programming Environment (v. 3.4.1.)

Do patterns of species co-occurrence vary along natural environmental gradients?

• “Conditional Two-species occupancy model” (Richmond et al. 2010)
• Complex parameterization scheme precluded the inclusion of P. glutinosus, perhaps due to sparse detections
• Model co-occurrence of P. richmondi and P. kentucki
• program “PRESENCE” (v. 11.7) (Hines 2006)

​

Results:�Microhabitat usage

Image: BD Todd, UC Berkeley

Does body size predict thermal microhabitat preference?

• Yes!

Results:�Microhabitat usage

Image: BD Todd, UC Berkeley

Do thermal microhabitat preferences vary among species, and/or differ from ambient microhabitat temperature

• Yes & yes!

Results:�Population Parameters

Image: BD Todd, UC Berkeley

No effect

Do natural environmental gradients influence the fine-scale distribution and abundance of Plethodon salamanders in LCW?

*

Heterogeneous effects

Results:�Population Parameters—P. richmondi

Image: BD Todd, UC Berkeley

Soil moisture

Northeasterliness

Canopy density

Canopy openness

Elevation

Solar Radiation

Topographic convexity

Results:�Population Parameters—P. kentucki

Image: BD Todd, UC Berkeley

Canopy closure

Soil moisture

Elevation

Northeasterliness

Solar radiation

Topographic convexity

Canopy openness

Results:�Population Parameters—P. glutinosus

Image: BD Todd, UC Berkeley

Canopy density

Northeasterliness

Topographic convexity

Elevation

Solar radiation

Soil moisture

Canopy openness

Results:�Fine-scale distribution—P. richmondi

Results:�Fine-scale distribution—P. kentucki

Image: BD Todd, UC Berkeley

(Canopy Density)

Fine-scale distribution:�Fine-scale distribution—P. glutinosus

Image: BD Todd, UC Berkeley

(Canopy Density)

Results:�Abundance—P. richmondi

Results:�Abundance—P. kentucki

Image: BD Todd, UC Berkeley

Fine-scale distribution:�Co-occurrence—P. richmondi & kentucki

Image: BD Todd, UC Berkeley

Do patterns of co-occurrence vary along natural environmental gradients?

• Yes!

Is it resultant of interspecies competition, perhaps due to resource limitation or physical stress?

• …maybe not

​

Results:�Co-occurrence—P. richmondi & kentucki

Image: BD Todd, UC Berkeley

Probability of P. richmondi occupying a site where P. kentucki is known to be present

Discussion:�Fine-scale distribution and abundance

• Plethodon richmondi:
• Distribution
• restricted to forests stands with moist soil and robust canopy cover
• below exposed ridge-tops
• Abundance
• strongly influenced by the presence of mesic habitat
• Plethodon kentucki
• Distribution
• less restricted distribution than P. richmondi
• Abundance
• can be reduced in exposed areas, especially with canopy disturbance
• Plethodon glutinosus
• Distribution and Abundance
• robust body-size may buffer environmental conditions
• may respond to more extreme environmental variation (human disturbance)?

​

​

Discussion:�Microhabitat usage & body size

• Body size predicts thermal microhabitat preferences
• Thermal differentiation in microhabitat preferences between P. glutinosus and P. richmondi & P. kentucki
• Explained by…
• Physiological requirements^†
• Competition—stratification of microhabitats^‡
• Look to the data…

^† Spight 1968, Peterman et al. 2013, Riddell & Sears 2015

^‡Jaeger 1971, Shoener 1974, Farallo & Miles 2016

Discussion:�Microhabitat usage & body size

• Use of subterranean refugia for desiccation avoidance & thermoregulation (Jaeger 1980, Grover 1998)
• P. richmondi and P. kentucki (Nagel 1979, Green & Pauley 1987, Bailey & Pauley 1993, Marvin 1996)
• Not documented in P. glutinosus
• Probability salamanders are alive, but underground

​

​

Discussion:�Conservation implications

• Large standing crop of biomass
• High assimilation-to-production ratio: 60.7% (Burton & Likens 1975)
• Biomass available for forest consumers—1.02 kg۰ha^-2

​

​

—2.77 kg۰ha^-2

Discussion:�Conservation implications

• The biomass of salamanders is non-randomly and non-uniformly distributed
• The nature of woodland salamanders influence on terrestrial ecosystems varies with environmental conditions (Walton 2005, Walton 2013)
• The role salamanders play is contingent upon the inhabitability of the system
• Regions which generate large crops of salamander biomass should have greater conservation value
• mesic forest stands
• north-to-east facing slopes
• dense canopy

​

​

Acknowledgements:

Image: BD Todd, UC Berkeley

• Stephen Richter, Advisor
• Graduate advisory committee
• David Brown
• Brad Ruhfel
• Research Technicians
• Emily Jones
• Jake Hutton
• Kelley Hoefer
• Funding
• Society for the Study of Amphibians and Reptiles
• Kentucky Academy of Sciences
• Division of Natural Areas, EKU
• Department of Biological Sciences, EKU

​

Questions:�and pictures of salamanders

Image: BD Todd, UC Berkeley

Plethodon caddoensis (Caddo Mountain salamander)

Caddo Mountain, Arkansas—2013

Plethodon wehrlei (Wehrle’s salamander)

Lilley Cornett Woods, Kentucky—2016

Plethodon jordani (Jordan’s red-cheeked salamander)

Great Smoky Mtns, North Carolina—2015

Plethodon yonahlossee (Yonahlossee salamander)

Whitetop Mountain, Virginia—2016