AMPHIBIAN RESPONSES TO FOREST MANAGEMENT PRACTICES IN SOUTHWESTERN GEORGIA

By

DIANE W. BENNETT

A THESIS PRESENTED TO THE GRADUATE SCHOOL OF THE UNIVERSITY OF FLORIDA IN PARTIAL FULFILLMENT

OF THE REQUIREMENTS FOR THE DEGREE OF MASTER OF ENGINEERING

UNIVERSITY OF FLORIDA

2008

1

© 2008 Diane W. Bennett

2

To the little things.

3

ACKNOWLEDGMENTS

I thank the chair and members of my supervisory committee for their mentoring and

patience, the students and professors of the Environmental Engineering Department for their

inspiration, and International Paper, National Council for Air and Stream Improvement, and

National Fish and Wildlife Foundation for their generous support. I also thank my parents for

encouraging me to take the necessary leaps of faith that separate happiness from complacency,

and my friends for their continuous support and endurance throughout this process.

4

TABLE OF CONTENTS page

ACKNOWLEDGMENTS ...............................................................................................................4

LIST OF TABLES...........................................................................................................................7

LIST OF FIGURES .........................................................................................................................8

LIST OF ABBREVIATIONS..........................................................................................................9

ABSTRACT...................................................................................................................................10

CHAPTER

1 INTRODUCTION ..................................................................................................................11

2 EVALUATION OF THE EFFECTS OF COVERBOARD AGE IN TERRESTRIAL SALAMANDER MONITORING PROGRAMS...................................................................14

Introduction.............................................................................................................................14 Materials and Methods ...........................................................................................................16 Data Analysis..........................................................................................................................17 Results.....................................................................................................................................18

Acclimation Period..........................................................................................................18 Salamander Encounters Under Old versus New Coverboards ........................................18 Sizes.................................................................................................................................19 Coverboard Microhabitat.................................................................................................19

Discussion...............................................................................................................................19

3 AMPHIBIAN RESPONSES TO FOREST HARVEST AND STREAMSIDE MANAGEMENT ZONE PRACTICES IN SOUTHWESTERN GEORGIA........................26

Introduction.............................................................................................................................26 Amphibians as Biological Indicators...............................................................................26 Forestry BMPs in the Southeast ......................................................................................27 Research Problem............................................................................................................28

Methods ..................................................................................................................................29 Site Description ...............................................................................................................29

Climate .....................................................................................................................29 Steep-head ravines....................................................................................................30 Soils ..........................................................................................................................31

Study Design ...................................................................................................................31 Field Sampling.................................................................................................................32

Terrestrial adult salamander monitoring ..................................................................32 Aquatic larval salamander monitoring .....................................................................33 Adult treefrog monitoring ........................................................................................33

5

Data Analysis...................................................................................................................33 Results and Discussion ...........................................................................................................34

Effects of Harvest on Amphibian Presence and Distributions within Watersheds .........34 Results: Adult salamander presence.........................................................................34 Results: Adult salamander distributions...................................................................35 Results: Treefrog presence .......................................................................................35 Results: Treefrog distributions .................................................................................36 Discussion: Effects of harvest on amphibian presence and distributions ................36

Effects of Harvest on Amphibian Abundances within SMZs .........................................38 Results: Adult salamanders ......................................................................................38 Results: Treefrogs ....................................................................................................39 Discussion: Amphibian abundance within SMZs ....................................................40

Effects of SMZ Thinning on Amphibian Abundance .....................................................43 Results: Adult salamanders ......................................................................................43 Results: Larval salamanders.....................................................................................43 Results: Treefrogs ....................................................................................................44 Discussion: SMZ thinning........................................................................................44

Recommendations...................................................................................................................46 SMZ Width......................................................................................................................46 SMZ Thinning .................................................................................................................46

4 CONCLUSIONS ....................................................................................................................56

LIST OF REFERENCES...............................................................................................................58

BIOGRAPHICAL SKETCH .........................................................................................................64

6

LIST OF TABLES

Table page 2-1 Distribution of salamanders under both old and new coverboards....................................23

2-2 Average salamander snout-to-vent length under old and new coverboards ......................23

3-3 Timetable for the Dry Creek Study, Southlands Forest, Bainbridge, GA. ........................48

3-1 SMZ widths by slope class and stream type. .....................................................................48

3-2 Flow statistics in study watersheds during 27-month pre-treatment survey......................48

3-4 Salamander species presence in four sub-watersheds........................................................49

3-5 Treefrog species presence in four sub-watersheds.............................................................50

7

LIST OF FIGURES

Figure page 2-1 Number of salamanders captured relative to number of months since coverboards

were installed .....................................................................................................................24

2-2 Number of salamander captures for new coverboards compared to old coverboards .......24

2-3 Coverboard mean soil temperature and percent moisture compared to natural cover.......25

3-1 Location of study site in relation to physiographic regions...............................................51

3-2 Mean salamander count in relation to distance from streams............................................52

3-3 Mean treefrog count in relation to distance from streams .................................................52

3-4 Mean annual terrestrial salamander counts in four watersheds .........................................53

3-5 Mean annual treefrog counts in forested and harvested watersheds..................................53

3-6 Mean annual terrestrial salamander counts within upstream/downstream portions of forested watersheds compared to thinned and intact SMZs of harvested watersheds .......54

3-7 Mean annual in-stream salamander larvae counts for reference and harvested watersheds..........................................................................................................................54

3-8 Mean annual treefrog counts for reference and harvested watersheds ..............................55

8

LIST OF ABBREVIATIONS

Β Detection probability

BACI-P Before-after-control-impact-paired

BMPs Best management practices

C Count

D Population density

Dry Creek Study Streamside management zone effectiveness of hydrology, water quality, and aquatic habitats in southwestern Georgia headwater streams

IPSF International Paper’s Southlands Forest

NCASI National Council for Air and Stream Improvement

NFWF National Fish and Wildlife Foundation

PVC Polyvinyl chloride

SMZs Streamside management zones

9

Abstract of Thesis Presented to the Graduate School of the University of Florida in Partial Fulfillment of the Requirements for the Degree of Master of Engineering

AMPHIBIAN RESPONSES TO FOREST MANAGEMENT PRACTICES IN SOUTHWESTERN GEORGIA

By

Diane W. Bennett May 2008

Chair: Thomas L. Crisman Major: Environmental Engineering Sciences

Amphibians (frogs and salamanders) were monitored monthly since December 2002 as

part of a study examining the impact of forest harvest and Streamside Management Zone (SMZ)

practices. The study encompassed four adjacent subwatersheds of the Dry Creek Watershed at

the Southlands Experimental Forest of International Paper, Bainbridge, GA. Two watersheds

were left intact, while two were harvested. The SMZ was left intact in the upstream reach of each

treatment stream, while in the downstream, 50% of basal area was removed from the SMZ

(thinned). Terrestrial salamander numbers were assessed using plywood coverboards at fixed

stations throughout the watersheds. Salamander numbers were greatest closer to the streams,

within the width covered by the SMZ, and thinning of SMZs did not affect salamander counts.

Comparison of concurrent old and new coverboard data for one year suggested that board

replacement had an effect on salamander captures, with more encounters occurring under old

boards. Treefrog numbers were assessed using PVC pipes driven vertically into the substrate as

habitat attractants. Capture likelihood was reduced in harvested areas, as well as thinned SMZs.

However, all species of amphibians recorded during the pre-harvest survey period remained

present following harvest. This study suggests that current SMZ widths are adequate for

maintaining amphibian presence. However, thinning in this region may be inappropriate.

10

CHAPTER 1 INTRODUCTION

Forest managers are challenged to balance production of forest products with maintenance

of environmental quality, management of wildlife habitat, and conservation of biodiversity

(Hartley, 2002; Sharitz et al., 1992). Best management practices (BMPs) are sustainable forestry

guidelines developed to inform sivilculture operators of practices to minimize nonpoint source

pollution (e.g., soil erosion and stream sedimentation) and thermal pollution, thus reducing

environmental degradation (Georgia Forestry Commission, 1999). Impacts to environmental

quality as a result of forestry practices have the ability to alter habitats beyond thresholds of

certain species. These indicator species can be monitored, evaluated, and used to assess the

condition of the environment either to provide an early warning of changes in the environment,

or to diagnose the cause of an environmental problem (Dale and Beyeler, 2001).

Amphibians are valuable biological indicators for southeastern forestry due to life history

traits, abundance, and sensitivity to environmental perturbations (Vitt et al., 1990; Welsh and

Droege, 2001). Evaluation of amphibian responses to forestry practices can provide insight into

the effects of forestry management practices on environmental quality. In a literature review on

amphibian responses to forestry (Russell et al., 2004) conflicting results necessitated further

research.

Previous research has concentrated on responses to clear-cut harvesting, with little

attention to forestry BMPs (Ash, 1988, 1997; Clawson et al., 1997; Knapp et al., 2003). Many

studies that reported declines in amphibians occurred in the Pacific Northwest or Appalachians

(Biek et al., 2002; Petranka, 1994, 1998; Harpole and Haas, 1999), whereas studies in the

Southeastern United States reported increased numbers of amphibians following forestry

operations (O’Neill, 1995; Clawson et al., 1997). These contradictions emphasize that amphibian

11

populations are difficult to characterize and that more research is necessary to understand

discrepancies regarding responses to forest management.

Russell et al. (2004) reported only six studies investigating effects of forest management

on southeastern herpetofauna that employed manipulative designs with pre-treatment and post-

treatment data, treatment replication, or true spatial and temporal references (Ash, 1997; Chazal

and Niewiarowski, 1998; Clawson et al., 1997; Harpole and Haas, 1999; Knapp et al., 2003;

Russell et al., 2002). Among these studies, a limited number contained pre-treatment data,

spanned multiple years, or included treefrog monitoring. Furthermore, few studies have

evaluated management practices within SMZs (Streamside Management Zones), making the

appropriateness of these applications uncertain (Grialou et al., 2000; Committee on Riparian

Zone Functioning and Strategies for Management, Water Science and Technology Board

National Research Council, 2002). Contradictory findings and lack of quantifiable evidence on

BMP effectiveness in protecting amphibian biodiversity fail to support presumptions that current

forest management is successful at balancing production with conservation. More standardized,

controlled manipulation experiments are needed.

To meet these research needs, International Paper, along with partners University of

Florida, University of Georgia, Clemson University, Jones Ecological Research Center, National

Council for Air and Stream Improvement (NCASI), Georgia Forestry Commission, and National

Fish and Wildlife Foundation (NFWF) developed the Dry Creek Study (Streamside Management

Zone Effectiveness of Hydrology, Water Quality, and Aquatic Habitats in Southwestern Georgia

Headwater Streams) in 2000 to determine how upland and streamside forest management affects

stream hydrology, water quality and biological indicators. As a component of the study,

amphibians were monitored.

12

This study evaluates the effectiveness of forestry BMPs, specifically SMZs and partial

timber harvesting on amphibian presence, distribution, and relative abundance. This experiment

was conducted using the manipulative Before-After-Control-Impact-Paired (BACI-P) design.

Initial and short-term responses to forestry were evaluated over a four-year period, with one year

serving as pre-harvest data. Amphibians monitored in this study included adult and juvenile

salamanders and treefrogs. Passive trapping techniques (i.e., coverboards and PVC pipes) were

elected for terrestrial amphibian monitoring, while active trapping (i.e., dipnet) was performed

for larval salamander sampling. Salamander size distributions were included to assess impacts on

reproduction.

Chapter 2 is dedicated to evaluation of sampling methodology used for salamanders in this

study. Chapter 3 then covers amphibian responses to forestry management and the final

discussion includes recommendations for SMZs. Finally, Chapter 4 presents conclusions from

both studies, providing insight for future studies and forest management.

13

CHAPTER 2 EVALUATION OF THE EFFECTS OF COVERBOARD AGE IN TERRESTRIAL

SALAMANDER MONITORING PROGRAMS

Introduction

Salamanders are valuable biological indicators of ecosystem health (Welsh and Droege,

2001). Reports of global declines (e.g., Alford and Richards, 1999; Blaustein et al., 1994;

Houlahan, 2000) support the need for research on anthropogenic effects on populations. One

facet of this research is to understand large-scale dynamics of terrestrial salamander populations

and their responses to forest harvest (Petranka, 1994; Ash, 1997). Due to extensive spatial and

temporal scales of these studies, it is often too impractical or costly to determine absolute

abundance; instead populations are estimated with indices of relative (observed) abundance.

However, some assumptions made when equating indices with populations may be invalid when

making comparisons across large spatial and temporal scales. In such cases, indices may produce

results unrepresentative of actual populations, eventually affecting forestry management

decisions.

Several sampling techniques are available to determine relative abundance of

salamanders (Heyer et al., 1994), each able to detect different subsets of populations (Parris,

1999; Dodd and Dorazio, 2004). Coverboards are often used to sample surface dwelling,

terrestrial salamanders, and their application has been evaluated relative to other techniques

(Monti et al., 2000; Houze and Chandler, 2002; Ryan et al., 2002). They are normally non-

treated plywood shingles placed on the ground to mimic natural cover objects. Periodically these

boards are lifted, and the area beneath searched for salamanders. A population index is

developed by the count (C) of salamanders encountered under the coverboards and assumed to

be directly proportional to the actual population density (D) and the probability of ‘detecting’ the

animal in the survey (β) (Pollock et al., 2002; Schmidt, 2003). Salamander detection probability,

14

β, is assumed to be constant over space and time; a critical assumption when assessing

population changes. If this assumption is violated, then changes in both detection probability and

population size are confounded, and conclusions about responses to anthropogenic effects cannot

be made.

Salamander detection probability is dependent on biotic (e.g., weather, rainfall,

microhabitat quality, breeding patterns) and abiotic (e.g., observer ability, enumeration method)

factors that are susceptible to variation (Dodd and Dorazio, 2004). Coverboards can help control

some of the variability in β, making them useful for large-scale studies (Parris, 1999). Detection

probability is stabilized across space when coverboard sampling occurs concurrently, since large-

scale environmental factors such as weather and time of day are controlled. Coverboards

standardize sampling ability and between-observer bias (Smith and Petranka, 2000), and

destruction to the natural habitat is minimized (Marsh and Goicochea, 2003) thus preventing

changes to the microhabitat over time. Additional benefits of the use of coverboards for large-

scale studies include time and labor efficiency (Monti et al., 2000) and low maintenance and cost

relative to other methods (Ryan et al., 2002).

Despite advantages of using coverboards, salamander detection is still imperfect, and its

probability is subject to change over time. Salamanders may require an acclimation period before

species are detected under newly installed coverboards, indicating increasing β. Over time,

plywood coverboards weather and deteriorate, potentially affecting microhabitat quality

underneath. Studies have suggested that salamander detection is related to microhabitat variables

and have postulated that aging of coverboards may create habitat of different quality over time

(Houze and Chandler, 2002). In contrast, Hyde and Simons (2001) suggested that large-scale

factors (i.e., disturbance history, proximity to stream, elevation) have more effect on detection

15

probability than microhabitat, implying that aging of coverboards will not affect population

indices. There is need for understanding factors that influence probability of detection to ensure

that relative abundance comparisons over space and time represent actual population dynamics

accurately.

In a watershed-scale study of terrestrial amphibian response to forestry BMPs, non-

treated plywood coverboards were used to monitor surface dwelling populations of terrestrial

salamanders. After three years of monthly monitoring, the boards had begun to rot and

deteriorate, aging to the point of affecting routine sampling. New coverboards were then

installed directly adjacent to old coverboards. Both old and new boards were checked monthly

for one year until old coverboards deteriorated to the point that the area underneath could no

longer be adequately searched. Such monitoring permitted evaluation of sampling bias attributed

to coverboard age. Acclimation periods for new boards were examined, as were microhabitats to

determine whether coverboards adequately mimicked the natural environment.

Materials and Methods

This study was conducted in the coastal plains physiographic region of southwestern

Georgia. The study site was in the Dry Creek Watershed located in the Southlands Forest of

International Paper at Bainbridge, GA. Four sub-watersheds (A, B, C, D) were selected for study

sites, ranging in area from 25.8 to 48.0 ha. All exhibited steep slopes (35°–45°) towards

groundwater-fed streams. Timber in watersheds B and C was clear-cut harvested in Fall 2003

following Georgia BMPs.

Each sub-watershed was divided into four rectangular sampling grids located on opposite

banks of upstream and downstream reaches of the stream course, for a total of sixteen grids for

the study. Each grid contained three transects (streamside, riparian, midslope) that ran parallel to

the stream for roughly 40 meters. In December 2002, plywood coverboards (60 x 60 x 2 cm)

16

were placed at four stations 5 meters apart along each transect in the sampling grids. The areas

where coverboards were installed were cleared of leaf litter and woody debris so boards had

direct contact with the soil. Grids were then searched monthly for twelve months for

salamanders. By May 2005, the boards had weathered and begun to deteriorate, and new

coverboards of the same materials and dimensions were placed adjacent to the old boards. Both

old and new coverboards were then checked monthly (except December 2005) for salamander

presence from June 2005 through June 2006, when old coverboards had deteriorated to the point

that the area underneath could not be adequately searched.

When searching grids, each coverboard was lifted, and the entire area was examined

underneath for salamanders. All salamanders encountered were identified to species and

measured (snout-vent length, SVL). Individual salamanders were not marked, and analyses were

based on the number of salamanders encountered per search from new versus old coverboards on

the same grid. To compare aged coverboard microclimate to that of the natural environment, soil

temperature and moisture readings were recorded both under old coverboards and in the natural

environment for each transect on each sampling occasion using an Aquaterr M-300 digital meter.

Data Analysis

Salamander capture data from sub-watersheds were combined to examine coverboard

detections. To examine acclimation periods, Spearman’s rank correlations were used to analyze

for temporal trends since coverboards were installed to monitor salamander captures over a

twelve-month period. Variance was evaluated using Levene’s test statistic to assess whether new

and old coverboards had equal variance in number of captures. For each grid, mean salamander

encounter rate was calculated as the number of salamanders encountered per grid search over

twelve searches for both old and new coverboards. These data were then analyzed using the

Mann-Whitney test statistic to evaluate the null hypothesis that two independent samples come

17

from the same population. Total numbers of salamander encounters were analyzed by species for

both old and new coverboards, and a t-test was performed to test for differences in body sizes of

the two most commonly encountered species. Finally, a t-test compared soil temperature and

moisture data of coverboards to the natural environment. All statistical analyses were conducted

using SPSS (vers. 12.0, SPSS Inc.) at a significance level of P<0.05.

Results

Acclimation Period

Because watersheds B and C were clear-cut in 2003, they were eliminated from this

portion of the analysis, and data from watersheds A and D were combined. A total of 107 and

118 salamanders were recorded during a 12-month inspection of newly installed coverboards for

the 2002-03 and 2005-06 sampling periods, respectively. Salamanders were found in the initial

sampling period following a one-month acclimation period for both 2003 and 2005 installations.

There was no relationship between number of salamanders captured and time elapsed since

coverboards were installed (2003, Spearman’s rho = 0.27, P = 0.42; 2005 Spearman’s rho = -

0.05, P = 0.89) over a twelve-month period (Figure 2-1).

Salamander Encounters Under Old versus New Coverboards

A total of 349 salamanders were captured between June 2005 and June 2006 in all four

sub-watersheds. Old and new coverboards accounted for 205 and 144 captures, respectively.

Less than 1% of total coverboards recorded more than one salamander underneath (old = 1.17%,

new = 0.78%). A mean of 1.06 salamanders (+ 0.10 SE; median = 1) was encountered per search

under old coverboards, while a mean value of 0.75 salamanders (+ 0.09 SE; median = 0) was

encountered per search under new coverboards. The number of monthly salamander detections

with both old and new coverboards varied similarly throughout the year, (Levene’s F = 2.03, df =

18

22, P = 0.169), with peaks occurring simultaneously during winter (Figure 2-2). There were

significantly more salamanders encountered under old than new coverboards (P = 0.01).

Sizes

Five salamander species from three genera were encountered under both old and new

coverboards (Table 2-1). Eurycea cirrigera, Eurycea longicauda guttolineata, and Plethodon

grobmani were the most common (45.6%, 29.8%, and 16.6% of total individuals encountered,

respectively) with Pseudotriton ruber vioscai and Desmognathus apalachicolae infrequently

encountered. The number of salamander species varied among watersheds. E. cirrigera, E.

longicauda guttolineata, and P. grobmani were detected in all watersheds. P. ruber vioscai was

encountered in watersheds A and C only, while D. apalachicolae was recorded in only C and D.

SVLs were measured for 154 E. cirrigera and 98 E. longicauda guttolineata found under

old and new coverboards. Mean SVL values of individuals found under old versus new

coverboards were not significantly different, suggesting that coverboard age has no effect on

salamander length for these two species (Table 2-2).

Coverboard Microhabitat

Coverboards and natural cover displayed mean temperatures of 65.04 + 6.28 and 64.52 +

6.53, respectively (Figure 2-3). There were no statistical differences in temperature between

coverboard and natural cover (P = 0.39). Furthermore, percent moisture under coverboards

(mean = 61.14 + 35.74) was not significantly different from natural cover objects (mean = 59.37

+ 35.75; P = 0.61).

Discussion

Using indices of relative abundance to evaluate population dynamics is common in

terrestrial salamander studies. However, inconsistent responses to forestry (Russell et al., 2004)

and other anthropogenic influences (Collins and Storfer, 2003) illuminate the fact that

19

salamander populations are difficult to characterize, suggesting flaws in monitoring and/or

statistical techniques. To develop a more accurate depiction of absolute abundance, underlying

assumptions in relative abundance indices must be scrutinized. This study assessed the

assumption of constant detection probability (β); specifically that detection of salamanders under

coverboards is constant over time.

This study found that coverboards did not require an extended acclimation period before

detection occurred. Salamanders colonized coverboards quickly following both 2002 and 2005

installations. Coverboards attracted salamanders after only a 1-month acclimation period, despite

being installed during different times of year (December, 2002; June, 2005). This was consistent

with a study by Houze and Chandler (2002) that detected salamanders within the first week

following installation.

Salamander captures did not increase as coverboards aged, when examined over a one-year

period. This agrees with a study by Monti et al. (2000) evaluating capture bias due to 1-year old

coverboards. When 3-year old boards were examined next to new boards, however, aged boards

captured more salamanders. Within the first year of coverboard usage there was no correlation

between numbers of salamander captures and coverboard age; however, in the third year a

relationship existed. Development of this relationship probably occurred in the second year of

coverboard usage, for which this study does not have data.

Despite increasing salamander counts with coverboard age, both new and old coverboards

did a good job of detecting the same species. Furthermore, size distribution of E. cirrigera and E.

longicauda guttolineata did not differ with board age, and there was no evidence that newly

installed coverboards biased sampling with respect to size, at least for the most common species.

20

Differences in physical characteristics between new and old coverboards may contribute to

salamander capture biases. As coverboards decay, they may more effectively reproduce

microclimates of natural cover and attract more salamanders than new coverboards. In this study,

the microclimate of aged coverboards was similar to the natural environment suggesting that as

coverboards age, they more closely resemble natural microhabitats of salamanders. However, it

still remains unclear whether new coverboards adequately simulate salamander microclimates.

Houze and Chandler (2002) recorded daily temperatures under 4-month old coverboards and

compared them to natural cover objects and found that although there was no difference in mean

daily temperatures, there was significantly more variability in coverboard temperatures than

natural cover.

Knowing that capture biases due to coverboard age occur within the third year of

coverboard usage, but not the first year, it is difficult to draw definitive conclusions about

relative abundance from temporal comparisons of more than two years. Therefore, it is suggested

that coverboard replacement occur within the second year of monitoring to prevent changes in

salamander detection probability due to coverboard age; thus avoiding violation of assumptions

inherent in relative abundance indices. Because only a short acclimation period is required,

frequent coverboard replacement should not require extensive planning. New boards should be

placed directly next to old boards to keep sampling areas consistent over time. Since salamander

detection with coverboards may be more dependent on time of year than on acclimation periods,

it may be important to consider the timing of new coverboard installation. Once coverboards are

installed, side-by-side monitoring of new and old boards should continue until new boards

become colonized. However, due to natural salamander population fluctuations and the inability

21

to standardize β completely, statistical estimates of detection probability should be considered to

ensure accurate representation of salamander population dynamics (MacKenzie et al., 2002).

22

Table 2-1. Distribution of salamanders under both old and new coverboards, Dry Creek Watershed, Georgia.

Number of encounters Species Old coverboards New coverboards Total Eurycea cirrigera 85 (53.5%) 74 (46.5%) 159Eurycea longicauda guttolineata 66 (63.5%) 38 (36.5%) 104Plethodon grobmani 38 (65.5%) 20 (34.5%) 58Pseudotriton ruber vioscai 10 (58.8%) 7 (41.2%) 17Desmognathus apalachicolae 6 (54.5%) 5 (45.5%) 11TOTAL 205 (58.7%) 144 (41.3%) 349

Table 2-2. Average salamander snout-to-vent length (SVL) (+ 1 SD) under old and new coverboards. P-values indicate no significant difference in salamander size. Average SVL (cm) Old Coverboards New Coverboards P E. cirrigera 3.40 (+0.07) 3.34 (+0.07) 0.55 E. longicauda guttolineata 4.34 (+0.13) 4.48 (+0.17) 0.50

23

0

10

20

30

1 2 3 4 5 6 7 8 9 10 11 12Months after Installation

Sala

man

ders

cou

nted

20032005

Installation Year

Figure 2-1. Number of salamanders captured relative to number of months since coverboards

were installed for both 2003 and 2005

0

10

20

30

40

JUN

JUL

AUGSEP

OCTNOV

JAN

FEBMAR

APRMAY

JUN

Month

Sala

man

ders

cou

nted

Old coverboardsNew coverboards

Figure 2-2. Number of salamander captures for new coverboards installed in June 2005

compared to old coverboards

24

Figure 2-3. Coverboard mean soil temperature and percent moisture compared to natural cover. Error bars represent 95% confidence intervals. Coverboard microhabitat did not differ significantly from natural microhabitat (temperature, P = 0.39; moisture, P = 0.58).

25

CHAPTER 3 AMPHIBIAN RESPONSES TO FOREST HARVEST AND STREAMSIDE MANAGEMENT

ZONE PRACTICES IN SOUTHWESTERN GEORGIA

Introduction

Amphibians as Biological Indicators

Amphibians represent much of the faunal biomass within riparian and in-stream habitats of

southeastern United States ecosystems (Burton and Likens, 1975b; Stewart and Woolbright,

1996; Petranka and Murray, 2001). Approximately 80 of the 245 species nationwide occur in

Georgia (New Georgia Encyclopedia, 2007). Amphibians have unique biphasic lifecycles,

occupying both terrestrial and aquatic habitats during their lives. As a result, they have evolved

as an integral component of energy transfer and nutrient cycling in riparian forest ecosystems

(Burton and Likens, 1975a; Petranka and Murray, 2001). They are an important constituent of

forest food webs and may be keystone species in habitats where they have a disproportionately

large effect on ecosystem structure (Holomuzki et al., 1994; Wissinger et al., 1999).

Forest amphibians (especially salamanders) have unique qualities that make them suitable

as biological indicators (Vitt et al., 1990). They have narrow tolerance ranges for several

environmental variables, are sensitive to changes in microclimates, have permeable skin and gills

that are sensitive to sedimentation, are located in mid-trophic levels within food webs, are

numerous in southeastern forests, and can be easily and cheaply sampled (Dale and Beyeler,

2001; Welsh and Droege, 2001). These qualities make amphibians valuable biological indicators

for sustainable forestry in the southeastern U.S.

Reports of global amphibian declines (Stuart et al., 2004) spur the necessity for collecting

baseline data on populations and their distributions, as well as understanding their ecological

niches. Amphibian population declines have been attributed to anthropogenic habitat loss, global

climate change, chemical contamination, UVB radiation, overexploitation, both singly and in

26

combination (Wake, 1991; Blaustein and Kiesecker, 2002). Forestry operations are a major

contributor to habitat loss. Of the 9.9 million ha of forestland in Georgia, nearly 98% is available

for timber production (Georgia Forestry Commission, 2004). This amounts to nearly 60% of the

total land area of the state being available for harvest, with likely amphibian impacts.

Forestry BMPs in the Southeast

BMPs are sustainable forestry guidelines developed to inform sivilculture operators of

practices to minimize nonpoint (e.g., soil erosion and stream sedimentation) and thermal

pollution (Georgia Forestry Commission, 1999). SMZs are a specific type of BMP designed to

protect stream channel and riparian ecosystems from potential impacts related to forestry and

other operations (United States Environmental Protection Agency, 2005). They are vegetated

riparian buffer strips located parallel and adjacent to streams that are left intact during

sivilculture operations.

Riparian vegetation is beneficial to water quality and stream habitat. It can protect stream

water quality from sedimentation by filtering runoff from the watershed as it flows over

disturbed, harvested land toward the stream (Georgia Forestry Commission, 1999; Sun et al.,

2004; United States Environmental Protection Agency, 2005). Canopy species remaining within

the SMZ provide an additional benefit by shading surface water, thereby moderating water

temperatures (Georgia Forestry Commission, 1999; Sun et al., 2004; United States

Environmental Protection Agency, 2005). Trees within an SMZ also supply organic matter vital

to the stream ecosystem and maintain landscape connectivity to adjacent watersheds by serving

as habitat and wildlife corridors (Georgia Forestry Commission, 1999; United States

Environmental Protection Agency, 2005).

27

The State of Georgia Forestry Commission has developed guidelines for delineating the

width of SMZs. Instead of providing a formula, principles related to the stream type, slope

stability, and soil erosion potential of the harvested land are used (Table 3-1).

Research Problem

Given that amphibians occupy riparian-forest habitats, are biological indicators of

ecosystem quality, and are species of global concern, it is surprising that they are not considered

in forest management plans more often. Timber harvesting has several impacts on amphibian

habitat. Air, soil, and stream temperatures may increase, affecting microhabitat quality. Streams

may experience increased sedimentation, making in-stream habitats inhospitable to larvae. Many

studies in the southeastern U.S. have demonstrated that amphibian populations are negatively

affected by timber harvest (Ash, 1997; Petranka, 1994); however, there is debate about the

degree of impact, time required for recovery, and validity of sampling methods (Ash and Bruce,

1994; Petranka, 1994).

Amphibians are dependent on the same water quality characteristics that SMZs are

intended to protect; therefore, inferences could be made about the effectiveness of SMZs by

monitoring amphibian responses to harvesting. Though use of SMZs is a widely accepted

practice in southeastern states, little is known about the amphibian response to harvest practices

within SMZs (Russell et al., 2004). To address these concerns, development of monitoring

programs that span several years and employ non-biased, reliable, and efficient surveying

techniques appropriate for multiple year comparisons has become paramount.

To meet these research needs, International Paper, along with partners University of

Florida, University of Georgia, Clemson University, Jones Ecological Research Center, NCASI,

Georgia Forestry Commission, and NFWF established the Dry Creek Study in 2000 to determine

28

how upland and streamside forest management affects stream hydrology, water quality and

biological indicators. As a component of the study, amphibians were monitored.

The objective of this research was to evaluate amphibian responses to forest harvest and

SMZ management techniques. Major questions that were addressed included impacts of clear-cut

harvest on amphibian presence, distribution, and abundance within intact and thinned SMZs. By

learning how sensitive species such as amphibians respond to current forest management

practices, an important contribution can be made to developing sustainable forestry practices.

This study will provide forest managers with recommendations for forest management,

specifically regarding development of streamside management zone practices that are compatible

with maintaining healthy populations of forest amphibians.

Methods

Site Description

International Paper’s Southlands Forest (IPSF), an experimental and working forest since

1947, is located in the Coastal Plain physiographic province, approximately 16 km south of

Bainbridge, Georgia (Figure 3-1).

Climate

Climate of the region is characterized by warm, humid summers, and mild winters.

Temperatures in January range from an average maximum of 16.3 °C to a minimum of 2.8 °C.

July is the hottest month with an average maximum temperature of 33.5 °C and minimum of

21.5 °C (SERCC, 2004). Mean annual precipitation is 1412 mm. June has the highest mean

rainfall (152.1 mm) and October the lowest (77.5 mm) (SERCC, 2004). Summer rains are

usually short, with high intensity events giving way to low intensity frontal events from late fall

to early spring. Due to proximity of the Gulf of Mexico, heavy rainfall associated with

hurricanes and tropical storms in late summer is not unusual.

29

Steep-head ravines

The steeply sloping Pelham Escarpment on which the site is situated forms the boundary or

surface-water divide between the Flint River basin to the west and Ochlockonee River basin to

the east (Couch et al., 1996). Streams originating from the Pelham Escarpment are characterized

by perennial headwaters that become intermittent downstream or drain directly into the Flint

River. Bluffs and deep ravines, or steep-head ravines, also characterize this transitional area

(Entrekin et al., 1999).

Means (1985, 2000) defined steep-head ravines as landscape features formed by large

rivers that cut deep through porous sands, intersect with perched groundwater tables, then

infiltrate laterally atop impermeable layers of silty marl, clay, or limestone. Pore spaces within

the karst limestone endemic to the region serve as underground conduits for groundwater to

travel laterally, eventually seeping out at points of low topography and creating groundwater

dominated headwater streams. Steep-head ravine systems are characterized by undulating waves

of topographical relief. The eroding churn of surface water overflow, rise and fall of the

groundwater table, and dissolution of karst limestone, result in solution caverns formed within

the substrate, weakening upper soil layers and causing strain on soil strata. The combination of

these forces creates highly erodable landscapes with diverse topographical features. Low-

gradient seepage streams at the base of valleys may be as much as 35 m below the upland

habitat, with walls sloping more than 45° (Means, 1985). Steep slopes serve as physical barriers

that distinguish watersheds. Within these watersheds are cool microclimates formed by the

constant seeping of groundwater of relatively uniform temperature. This provides the humidity

necessary to buffer air temperatures, providing adequate habitat to support rare plant and animal

species (Enge, 2002).

30

Soils

Soils of this area are dominated by Ultisols. Summer et al. (2003) described the riparian

soils as Chiefland and Esto series that are well-drained, fine sands over clay loams. The lower

slopes feature Eustis series soils, which are loamy sands over sandy loams and are classified as

somewhat excessively well drained. The upland soils are comprised of Wagram, Norfolk,

Lakeland, Orangeburg, and Lucy series, which are generally well-drained, loamy sands over

sandy clay loams, with the exception of the Lakeland Unit, which has a sandy texture throughout

and is characterized as excessively well drained.

Study Design

The overall Dry Creek Study design follows the BACI-P experimental design. The streams

in this study drain four adjacent watersheds with similar aspect, size, shape, soils and vegetative

cover type. Sub-watersheds were paired according to valley floor geomorphologic differences

into what was initially believed to be most optimal groups (A+B and C+D). Watersheds A and B

have broad, flat valleys with riparian wetlands, while watersheds C and D have more channelized

streams running through steeper, v-shaped valleys (Jones et al., 2003). However, pre-harvest

hydrological survey results showed that stream flow characteristics are more similar between

watersheds B+C and watersheds A+D (Summer et al., 2005) (Table 3-2).

Watersheds A and D were designated as controls, thus left undisturbed throughout the

entire study period. Watersheds B and C were selected as treatment watersheds and were

harvested during fall 2003. A timetable of study components is provided in Table 3-3. Each

watershed was harvested according to minimum requirements set forth in the State of Georgia

BMP Manual. SMZ widths ranged from 30–40 m perpendicular to the stream edge on either

side. Watersheds B and C each received two sivilculture treatments; mechanical clear-cut upland

harvesting and partial harvesting of downstream SMZ. Half of the basal area was removed

31

within the downstream portion of harvested sites SMZ, while the upstream SMZ was left intact

in order to evaluate the effects of harvesting within the SMZ.

Field Sampling

Monthly amphibian monitoring began in December 2002, allowing for ten months of base-

line data collection before the harvest period began in September 2003. The experimental layout

of semi-aquatic salamander and treefrog monitoring for each watershed was a grid consisting of

four transects running parallel to each side of the stream and perpendicularly upslope along a

habitat gradient. The four transects were designated as habitat zones: (1) stream (2) riparian (3)

midslope (4) upslope, representing increasing distance from stream. Sampling techniques

employed to capture amphibians included coverboard shelter attractants (for adult salamanders),

vertical PVC pipe shelter attractants (for treefrogs), and dipnet sweeps (for larval salamanders).

Due to lack of reliable, efficient techniques available for marking amphibians, animals were not

mark-recaptured (Grialou et al., 2000; Monti et al., 2000).

Terrestrial adult salamander monitoring

Experimental grids contained transects with four passive sampling locations for which

semi-aquatic salamanders and treefrogs could be monitored. Coverboards were used as shelter

attractants for adult, terrestrial salamanders (Houze and Chandler, 2002). Boards were cut from

2.0-cm untreated plywood sheets into 60 x 60 cm squares (Grant et al., 1992) and placed along

transects perpendicular to stream channels toward uplands. Eight coverboards were placed in

designated habitat zones for a given sample reach (four coverboards on either side of the stream,

256 total coverboards). Salamanders found under coverboards were identified to species,

counted, and measured for snout-to-vent length (SVL, cm).

32

Aquatic larval salamander monitoring

Active sampling was used for in-stream larval salamander monitoring. To sample all

potential microhabitats within the stream, the flat surface of a standard D-frame dipnet (V ≅

0.02-m3; dimensions: 0.3-m2 opening, 0.5-m length, 1,000-μm mesh) was swept along the

bottom of the stream and under incised banks. For each sample reach, 20 dipnet sweeps were

performed, each ~1-m long. Captured larvae were counted, identified to species, and released

into the stream reach where captured.

Adult treefrog monitoring

Vertical polyvinyl chloride (PVC) pipes (5.1-cm diameter, 60-cm height above ground)

were used for treefrog monitoring (Boughton et al., 2000). PVC pipes act as shelter attractants by

shielding inhabitants from extreme wind and temperature, thereby providing moist refuge (Wyatt

and Forys, 2004). One sampling pipe was installed at each coverboard location (256 total pipes).

Frogs inhabiting the artificial habitat were identified to species, counted, and PVC pipe location

was noted.

Data Analysis

Amphibians were grouped as salamanders and treefrogs. Low monthly captures resulted in

multiple zeros causing amphibian detection probability to be less than 1 (MacKenzie et al., 2002;

Royle, 2004). Hence data were compiled annually to strengthen statistical analyses. Non-

parametric tests for multiple independent samples were then used due to small number of

monthly samplings (N = 12 for year 0, 1, 2; N = 11 for year 3).

Amphibian presence was noted for each year of the study, and distributions were

calculated as percentage of total captures within habitat zones. To test for the effects of clear-cut

harvesting on amphibian capture rates within SMZs, comparisons of amphibian counts within

riparian areas were made between watersheds and among years. To examine differences in

33

amphibian captures among all watersheds during the pre-harvest survey, Kruskal-Wallis rank

sum tests were performed to evaluate the null hypothesis that multiple independent samples

come from the same population. When tests yielded a significant H statistic, Wilcoxon rank sum

tests were then used to determine which watershed pairs contributed to the overall differences.

To evaluate for changes in number of captures among years, the non-parametric analog of a

Repeated Measures ANOVA (Friedman’s F-Test Statistic) was used for each watershed. Post-

hoc comparisons were performed using Wilcoxon’s Matched Pairs to test for differences among

consecutive years. Analysis of salamander SVLs was performed by combining data from all

years following harvest for the most common species encountered for harvested and forested

sites. T-tests were then performed to analyze for differences in size due to harvest. The effect of

partial harvesting of SMZs was analyzed by combining data from harvested and forested reaches

(because of the low total number of catches for each reach). Mann Whitney tests were then used

to compare the capture means of upstream/downstream and thinned/intact reaches.

Analyses were performed at significance levels of P<0.05.

Results and Discussion

Effects of Harvest on Amphibian Presence and Distributions within Watersheds

Results: Adult salamander presence

A total of 993 salamanders, six species from five genera, were captured during the period

December 2002 – December 2006 (Table 3-4). All watersheds were inhabited by E. cirrigera, E.

longicauda guttolineata, P. grobmani, and P. ruber vioscai. D. apalachicolae was found only at

watersheds C and D, and N. viridescens was detected only in watershed D.

For the first two years of the study, E. longicauda guttolineata and D. apalachicolae were

not detected in forested watersheds A and D, respectively. However, both species were found in

their respective watersheds during the final two years of monitoring. In the first year following

34

harvest, P. ruber vioscai was not detected at harvested site C, but was found again during years 2

and 3. In the final year of sampling, P. grobmani was not detected in harvested watershed B, and

P. ruber vioscai was not detected at watershed D.

Results: Adult salamander distributions

Salamanders were monitored monthly for one year prior to harvest to establish baseline

conditions. During the pre-harvest survey period (Year 0), the majority of salamander captures

occurred within boundaries of SMZs (90.8% of total captures, Figure 3-2). Only one salamander

capture occurred on the upslope transect for the entire study (watershed C, Feb 2003, P.

grobmani). In the years following harvest, forested watersheds continued to maintain the

majority of salamanders within boundaries of SMZs (Year 1 = 85.5%; Year 2 = 91.7%; Year 3 =

91.2%) with just two salamander species being recorded in the midslope and upslope regions (P.

grobmani and E. cirrigera). There were no captures outside of the SMZ in watershed B

following harvest, and in watershed C there were 5 total captures of P. grobmani on the

midslope.

Results: Treefrog presence

Six species of treefrogs from two genera were captured during the study totaling 2236

captures (Table 3-5). All watersheds recorded Hyla squirrella, Hyla cinerea, Hyla chrysoscelis,

and Pseudacris. crucifer. Watersheds A, C, and D reported captures of Hyla femoralis, while

Hyla avivoca was detected only at watersheds A and D.

Treefrog presence varied for all watersheds during all years of study. Forested watersheds

did not record H. chrysoscelis or H. femoralis until the second and third year of study,

respectively. Harvested watersheds did not record presence of H. femoralis following harvest.

35

Results: Treefrog distributions

The results of the pre-harvest survey reveal that 76.1% of total treefrogs captured were

located within boundaries of SMZs (Figure 3-3). Watersheds B and C continued to capture the

majority of treefrogs within SMZs following harvest (Year 1 = 93%; Year 2 = 90%; Year 3 =

85%), while forested watersheds captures decreased within SMZs (Year 1 = 81%; Year 2 = 61%;

Year 3 = 62%) and increased outside the SMZs. Captures in regions outside SMZs were

significantly greater for forested watersheds compared to harvested (forested N = 591, harvested

N = 29).

Discussion: Effects of harvest on amphibian presence and distributions

Watersheds displayed similar species assemblages during pre- and post-harvest surveys.

Harvest did not affect species presence within SMZs of harvested watersheds.

The pre-harvest survey indicated that the majority of salamanders inhabit the region within

the SMZ, with few occurrences on midslope, and zero occurrences on upslope reaches of

watersheds. After harvest, salamanders were not found in clearcut areas outside SMZs; however,

forested watersheds continued to yield captured salamanders on midslope transects occasionally.

SMZ widths extended 30 m from each side of the stream edge, much less than suggested buffer

widths necessary to protect 50% of salamander populations (i.e., 125 m; Semlitsch, 1998).

Despite the small buffer area, nearly 95% of salamanders captured in forested watersheds

occurred within this zone.

The majority of salamander species disperse parallel to aquatic habitats (Maxcy, 2000).

However, dispersal patterns can be species specific depending on body-size and breeding

strategy (Grover, 2000). In this study, the majority of salamanders were Plethodontid stream

salamanders. Although migratory at some stage in their life cycle, these salamanders remain

close to the edges of streams, seldom moving more than 20 – 30 m from aquatic habitats

36

(Grover, 1998). However, P. grobmani, a terrestrial breeding species abundant in this study,

accounted for the majority of captures at midslope transects that were ~ 50-m from the stream

edge. These results may be indicative of the breeding strategy of P. grobmani, which lays eggs in

leaf mats in terrestrial habitats away from stream edges. The larger body size enables it to resist

desiccation (Grover, 2000), allowing travel farther from humid environments compared to

smaller semi-aquatic species. Dispersal patterns of other species may be inhibited by steeply

sloping ravines characteristic of this region. Perpendicularly from the stream, elevation increases

rapidly, resulting in a shift in microclimate. Sharp gradients in moisture-temperature regimes

may cause species to remain close to stream edges where cool, humid, microclimates are

protected by ravine slopes.

Harvest did not affect presence of treefrogs. During the pre-harvest survey, the majority of

treefrogs were captured within riparian regions of watersheds, with about 25% of captures

occurring on mid- and upslope transects. Following harvest, most captures occurred within

SMZs of watersheds B and C, while in forested watersheds, the proportion of population outside

of SMZ grew over consecutive years. Anurans have larger dispersal distances than salamanders

with some species traveling as far as 1000 – 1600 m (Semlitsch et al., 2003). Only 29 individuals

were recorded in uphill regions of clear-cut areas, compared to forested areas, with 591

detections. Nearly 40% of all treefrog captures were recorded in uphill regions of forested

watersheds by the fourth year of study, suggesting large dispersal ranges for treefrogs in this

region. Impacts of clear-cutting may interrupt anuran movements resulting in isolation of

metapopulations.

Studies have shown positive responses by frogs to forest harvest (O’Neill, 1995; Russell et

al., 2002). Anuran resilience has been attributed to non-selective breeding strategies (often

37

depositing eggs in skidder ruts) and the ability to resist desiccation resulting from habitat

alterations (Russell et al., 2004). However, in those studies, Bufos, Ranids and Hylids were

grouped together, with few reports detailing information regarding just treefrog responses. Bufos

and Ranids occupy different landscape patches, with select species being negatively associated

with forest area (Guerry, 2002). Because of the increased difference in treefrog encounters in

upland areas compared to clear-cut over consecutive years, this study suggests that treefrogs may

respond negatively to clear-cut operations.

Effects of Harvest on Amphibian Abundances within SMZs

Results: Adult salamanders

Because so few salamanders were captured outside of SMZs, they were eliminated from

this portion of the analysis. Instead, salamander abundance inside SMZs was analyzed. Kruskal

Wallis test results were marginally significant in detecting similarities in salamander abundance

among all sub-watersheds during the pre-harvest survey period (H = 7.62, df = 3, P = 0.06).

Following harvest, there were significant differences in salamander captures between watersheds

for all years (2004 H = 13.75, df = 3, P = 0.00; 2005 H = 13.06, df = 3, P = 0.01; 2006 H = 9.22,

df = 3, P = 0.03).

Watershed pairs were similar in salamander captures during the pre-harvest survey (A+B

W = -1.13, P = 0.26; C+D W = -0.77, P = 0.44) (Figure 3-4). Following harvest of watershed C,

there were significantly fewer salamander captures compared to D for the first two years (2004

W = -2.77, P = 0.01; 2005 W = -2.12, P = 0.03), but the watershed pair was similar again the

third year (2006 W = -1.73, P = 0.08). For the first two years after harvest of watershed B, there

were no differences in salamander captures when compared to forested watershed A (2004 W = -

1.59, P = 0.11; 2005, W = -0.49, P = 0.62); however, in the third year, the pair was different

(2006 W =-2.96, P = 0.00). Comparisons across years for each watershed resulted in no

38

significant differences in salamander abundance for either forested (A F = 3.95, df = 3, P = 0.

27; D F = 3.38, df =3, P= 0.34) or harvested watersheds (B F = 2.72, df =3, P = 0. 44; C, F =

5.97, df = 3, P = 0.11).

Tests for differences in salamander SVLs (cm) were performed for the most abundant

salamander species captured at all watersheds. No statistical differences were found between

salamander SVLs of forested and harvested sites for E. cirrigera (forested mean =3.4+0.0 cm;

harvested mean =3.4+0.0-cm; t = 0.45, df = 309, P = 0.66) or P. grobmani (forested mean

4.1+0.1-cm; harvested mean = 4.1+0.1-cm; t = 0.43, df = 314, P = 0.67). However, SVLs of E.

guttolineata were significantly greater at harvested than forested sites (forested mean =4.0+ 0.9-

cm; harvested mean = 4.3+0.1-cm; t = -2.84, df = 304, P = 0.01).

Results: Treefrogs

Kruskal Wallis tests were significantly different for treefrog captures between watersheds

for all years (2003 H =10.43, P = 0.02; 2004 H =16.22, P = 0.00; 2005 H =12.78, P = 0.01; 2006

H =29.81, P = 0.00). Watersheds A+B were similar in treefrog captures during the pre-harvest

survey (2003 W =-0.27, P = 0.79) and first year following harvest (2004 W =-0.10, P =0.92);

however, in the second and third year, they were different (2005 W =-2.14, P = 0.03; 2006 W =-

3.07, P = 0.00). Watersheds C+D were different in the pre-harvest survey (2003 W =-2.07, P =

0.04) and all years following harvest (2004 W =-3.06, P = 0.00; 2005 W =-2.94, P = 0.00; 2006

W =-3.06, P = 0.00) (Figure 3-5).

Friedman F-statistical tests were significantly different for treefrog captures among years

for forested watersheds (A F = 27.00, df = 3, F = 0.00; D F = 20.45, df = 3, P = 0.00) but not

harvested watersheds (B F =6.98, df = 3, P = 0.07; C F = 3.31, df = 3, F = 0.35). Post hoc

comparisons of watershed A revealed an increase in annual captures among consecutive years

(Year 0-1, W = -3.08, P = 0.00; Year 1-2, W = -1.81, P = 0.07; Year 2-3, W = -2.50, P = 0.01)

39

Watershed D experienced an increase in the first year (Year 0-1, W = -2.54, P = 0.01), stayed

steady in the second (Year 1-2; W = -0.59, P = 0.56), and increased again in the third year (Year

2-3, W = -2.85, P = 0.00).

Discussion: Amphibian abundance within SMZs

Observed amphibian abundance (C) is a function of population density (D) and detection

probability (β). Differences in abundances can be attributed to natural fluctuations in amphibian

populations, their response to harvest, or changes in detection ability (Chapter 3).

All watersheds were similar in salamander captures during the pre-harvest survey period,

and watershed pairs were well matched. However, differences in salamander capture following

harvest between watershed pairs and among years were confounded, making definitive

conclusions regarding salamander responses difficult. Responses to harvest may have been

delayed or the number of captures may have been influenced by con-specific attraction, inter-

annual breeding cycles, stream flows, or local meteorology.

Adult salamander counts may reflect a delayed response to harvest due to faster re-

hydration rates than juveniles (Grover, 2000). Adult body size is larger and able to retain more

moisture than smaller body sizes, thus enabling them to be more tolerant to environmental

disturbances. Clear-cut harvest can alter microclimates within SMZs beyond thresholds of

juveniles. If there is high juvenile mortality, a delayed response in decreasing adult populations

may occur. This, coupled with an increase in captures of larger adults due to smaller individuals

burrowing or desiccating, may suggest a negative response of salamanders within SMZs to forest

harvest. These results may be exemplified in the third year following harvest of watershed B

when captures decreased relative to A, and SVLs of E. longicauda guttolineata were smaller in

harvested than forested watersheds. However, body sizes of the more abundant salamander

40

species such as E. cirrigera remained similar between forested and harvested watersheds.

Responses to harvest may be taxon specific.

Harvested watersheds contained more fallen trees and large woody debris than that of

forested. Con-specific attraction of adult salamanders to natural cover in harvested watersheds

over coverboards may result in decreased counts in harvested compared to forested watersheds.

Grover (2002) showed a high correlation between salamander abundance and large woody

debris. Moreover, coverboard microhabitat can be more variable than natural cover (Houze and

Chandler, 2002). Salamanders may seek larger cover objects in harvested watersheds due to

increased temperatures, making coverboards less attractive habitat.

Harvest may have resulted in a different subset of salamander populations being sampled

relative to forested watersheds. Variability in captures increased in watershed C compared to D

with most captures at site C occurring during winter breeding months. Salamanders at harvested

sites were infrequently encountered during summer when they burrow underground seeking

refuge from elevated temperatures. Captures at site D were more consistent year-round, with

peaks still occurring during winter. Harvest may result in only breeding salamanders being

sampled, thus violating assumptions of constant detection probability between watershed pairs.

Large-scale habitat variables such stream flow and meteorology were highly variable and

could not be standardized. Amphibians have been linked to hydrologic cycles, with breeding

events occurring sporadically in response to meteorology. Several hurricanes occurred during

the 2004 season, and amphibians may have responded with increasing populations. Watershed A

experienced a high number of no-flow days in the stream during 2003, 2004, but not during 2005

or 2006. Terrestrial salamander captures did not indicate any differences among years for any

watersheds. However, controlling factors for amphibian captures become tangled by reactions to

41

environmental fluctuations resulting in too much variability to detect small changes in

populations with few sample sizes (N = 12). Abundance of salamanders is a function of many

variables, and increases or decreases may be attributed to factors other than forest management.

Pre-harvest survey data indicated that watershed D was different from all other watersheds

in number of treefrog captures. There were disproportionately more treefrogs recorded in this

watershed at the beginning of the study, and by the final year, six times more treefrogs were

captured at this watershed relative to others. Due to this large difference, comparisons of treefrog

captures of watershed D relative to C resulted in differences during all 3 years following harvest,

rendering this watershed pair a poor comparison.

Most interestingly though is the increase in treefrog captures among consecutive years in

forested watersheds compared to harvested. PVC pipes appear to accumulate treefrogs over time

in forest watersheds, but not in harvested. Treefrogs within SMZs may be affected by harvest

and may not have the population densities available to increase numbers in PVC pipes. Treefrog

habitat is directly affected by removal of canopy trees. Since treefrogs require larger habitat

areas and have greater dispersal distances, it is possible that upland habitat removal affected

treefrogs inhabiting SMZs. However, as in the case of salamanders, definitive conclusions are

difficult to make due to other confounding variables. Other habitats may be available in

harvested watersheds (e.g., downed logs, ruts, pools) that attract treefrogs over PVC pipes. PVC

pipes located in harvested watersheds were inhabited by other species (e.g., wasps) more often

than forested watersheds, which may also have interfered with detections.

Due to greater dispersal distances, treefrogs seemingly are more impacted by harvest than

salamanders due to upland habitat loss. However, because treefrogs are presumably more

resilient to disturbances than salamanders (compared with salamanders, anurans have relatively

42

high operating and tolerance temperatures and have the ability to store and reabsorb large

quantities of water in the bladder) (Duellman and Trueb, 1994), populations may be expected to

recover. However, there is concern that upland habitat alterations may interrupt metapopulation

dynamics, resulting in regional decline of treefrogs. SMZs presumably provide wildlife corridors

to adjacent forested watersheds, but Niemela (2001) expressed apprehension about the belief that

animals isolated within corridors utilize them for movement between landscapes. These concerns

highlight the need for quantitative evidence on the utility of SMZs as movement corridors and

their role in metapopulation dynamics.

Effects of SMZ Thinning on Amphibian Abundance

Results: Adult salamanders

For this portion of the analysis, watersheds A, B, C, and D were divided into

upstream/downstream segments, and captures were combined for respective segments. Results of

Mann Whitney comparisons found no difference between salamander counts downstream

compared to upstream portions of any watersheds during pre-harvest survey (Down-Up, S = -

0.35, P = 0.73; Thinned-Intact, S =-0.24, P = 0.81) (Figure 3-6). After thinning the downstream

SMZ of watersheds B and C, no significant differences were found between thinned and intact

SMZs for the duration of the study (Year 1 P = 0.19; Year 2 P = 0.13; Year 3 P = 0.65). Forested

watersheds recorded significantly higher captures in the upstream reaches of watersheds during

year 1 (S =-0.347, P = 0.03), were marginally higher in year 2 (S = -1.78, P = 0.07) and were

significantly higher in year 3 (S =-2.61, P = 0.01).

Results: Larval salamanders

There were significant differences in upstream-downstream segments during the pre-

harvest survey (Year 0) for all watersheds (Figure 3-7). Watersheds A and D had significantly

more larvae in their upstream sections (S = -3.67, P = 0.00) compared to downstream, while

43

watersheds B and C had more in their downstream sections (S = -2.57, P = 0.01). Following

harvest of watersheds B and C, there were no significant differences in number of larvae

captured in the streams of thinned versus intact SMZs for all years (Year 1 S = -0.18, P = 0.86;

Year 2 S = -1.44, P = 0.15; Year 3 S = -0.05, P = 0.96). Reference watersheds showed

significantly more larvae in the upstream segments for all years (Year 1, S = -4.38, P = 0.00;

Year 2 S = -2.29, P = 0.02; Year 3 S = -2.15, P = 0.03).

Results: Treefrogs

Mann Whitney tests revealed no differences in upstream or downstream treefrog counts

during the pre-harvest survey (Down-Up, S = -0.48, P = 0.63; Thinned-Intact, S = -1.00, P =

0.32) (Figure 3-8). Following harvest, watersheds B and C recorded significantly fewer treefrogs

in thinned SMZs than intact for years 1 (S = -2.12, P = 0.03) and 3 (S = -3.22, P = 0.00) but not

year 2 (S = -0.92, P = 0.36). Forested watersheds A and D did not show any significant

differences in the number of treefrog captures for the upstream-downstream segments for all

years (Year 1, S = -1.40, P = 0.16; Year 2, S = -1.28, P = 0.20; Year 3, S = -0.78, P = 0.44).

Discussion: SMZ thinning

Grialou et al. (2000) evaluated redbacked salamander responses to thinning within SMZs

in the Pacific Northwest and found that forest thinning stimulated salamanders. No statistical

differences were found between thinned and intact reaches of harvested watersheds for this

study, though slightly more salamanders were captured in intact segments over consecutive

years. Adult salamander counts were greater in upstream segments of forested watersheds.

Means (2000) found that salamanders of this region prefer headwater habitats of upstream

reaches that contain small seeps, no fish competition, and little variability in flow regimes. One

similarity of all watersheds was that the upstream areas where salamanders were sampled were

44

broad and flat compared to downstream regions. Hydrology in the upstream segments may have

been less variable, and more suitable for salamander breeding.

Larval salamanders showed the same trend as adults, with significantly greater numbers in

upstream sections of forested watersheds. Harvested watersheds showed different trends with

more larvae in their downstream segments during the pre-harvest survey. After harvest, there

were no statistical differences. This may not be a result of thinning, but rather a sampling bias in

reference watershed A. Disproportionately larger numbers of juveniles were captured in the

upstream segments of watershed A for all years of the study. During the first two years, this

particular site had many no flow days in the downstream portion, making it impossible to dipnet

the stream. However, although dry in the downstream portion, the upstream portion contained

big, deep pools of quiescent water, enabling large volumes to be sampled with a dipnet. Juvenile

salamanders were concentrated in these pools, enabling easy detection. Comparatively, upstream

segments of all other watersheds were shallow and flowing throughout the study. Smaller

volumes were sampled, and juveniles were not concentrated. This scenario is reflected in the

variability of the upstream sections during the first two years of study. Variability in the

detection probability of juvenile captures precluded an analysis of the effects of thinning. Size

class descriptions may be more informative.

More treefrogs were captured in intact SMZs for all years following harvest, although the

second year did not show a significant difference. Reference watersheds did not indicate a

preference for upstream or downstream reaches. Differences in treefrog abundances may be

attributed directly to thinning and associated habitat loss.

45

Recommendations

SMZ Width

SMZs are important habitat for feeding, overwintering, and breeding of amphibians. They

not only buffer aquatic habitat but also provide an area for the biological interdependence

between aquatic and terrestrial habitats that is essential for persistence of populations. Harrison

and Voller (1998) suggested that corridor location and design should reflect the ecology of an

area and that riparian buffer widths be adjusted proportionally with stream width, intensity of

adjacent harvest, and slope. Individual taxa should also be considered in buffer width design.

SMZs have multiple purposes (e.g., prevent stream sedimentation, thermal pollution,

habitat), and widths should be optimized for all management goals (e.g., production,

conservation of biodiversity). Microclimates presumably decrease in humidity along elevation

gradients perpendicular to streams, and the intensity of this gradient may be a good determinant

of SMZ width for this region. SMZ boundaries may occur where microclimates exceed

thresholds of target organisms such as salamanders.

Wider SMZs that incorporate midslope reaches may offer necessary habitat for terrestrial

breeding amphibians and provide additional buffer for semi-aquatic salamander and treefrog

territories. However, this may be unnecessary for species survival and recovery, and will result

in more costs and loss of revenue for forestry operations. For this study, a 30 m buffer width

appeared adequate for species’ existence.

SMZ Thinning

Natural preference of headwater habitats by salamanders suggests that thinning of

downstream segments be more appropriate to protect amphibian-breeding areas. However,

reports of salamanders responding positively to thinning in the Pacific Northwest (Grialou et al.,

2000) imply that thinning may not adversely affect salamander populations. Treefrogs may be

46

more affected by thinning than salamanders due to habitat loss, but because they are presumably

more resilient, this may not be a concern. It is important to note, however, that thinning resulted

in substantially more canopy loss due to windthrow. Hurricanes frequently pass through this

region, and thinned canopies are susceptible to hurricane force winds. Additional canopy loss

due to wind-throw may result in cost benefit ratios greater than one. Managers should weigh

decisions and optimize production with maintenance of biodiversity. Thinning of SMZs in this

region may not be appropriate.

47

Table 3-3. Timetable for the Dry Creek Study, Southlands Forest, Bainbridge, GA. Date Action

December 2000 Study established May 2001 Initiation of continuous stream flow gauging June 2001 Initiation of water quality data collection December 2002 Initiation of terrestrial amphibian monitoring September–November 2003 Upland harvesting and SMZ partial harvesting May 2004 Initiation of soil moisture-temperature data collectionSeptember 2004 Site preparation herbicide November 2004 Site preparation burn December 2004 Site replanting September 2006 End soil moisture-temperature data collection March 2007 End of amphibian monitoring

Table 3-1. SMZ widths by slope class and stream type.

Minimum width (ft) SMZ on each side Slope Class Perennial Intermittent Trout Slight (<20 %) 40 20 100 Moderate (21-40%) 70 35 100 Steep (>40%) 100 50 100 (Georgia Forestry Commission, 1999)

Table 3-2. Flow statistics in study watersheds during 27-month pre-treatment survey. Site Area Mean Q (L/s/ha) Max Q

(L/s/ha) Zero Flow Days (/822)

A 25.8 .055 8.37 163 (20%)B 34.7 .074 14.08 6 (.7%)C 42.7 .073 9.91 2 (.02%)D 48.0 .042 7.17 206 (25%)Summer, W.B., Jackson, R.C., Jones, D., Golladay, S.W., & Miwa, M., (2005). Hydrologic and Sediment Transport Response to Forestry; Southwest Georgia Headwater Streams. (In: Proceedings of the 2005 Georgia Water Resources Conference, held April 25-27, 2005. The Institute of Ecology: The University of Georgia, Athens, GA.)

48

Table 3-4. Salamander species presence in four sub-watersheds of the Dry Creek Basin, Georgia. Watersheds B and C were harvested in September 2003; watersheds A and D were left intact for the entire study.

Watershed A: Forested Watershed B: Harvested Year 0

(Pre-harvest)

Year 1

Year 2

Year 3

Year 0 (Pre-

harvest)

Year 1

Year 2

Year 3

Eurycea cirrigera X X X X X X X X Eurycea longicauda guttolineata

- - X X X X X X

Plethodon grobmani X X X X X X X - Pseudotriton ruber vioscai

X X X X X X X -

Desmognathus apalachicoloae

- - - - - - - -

Notophthalmus viridescens

- - - - - - - -

Watershed D: Forested Watershed C: Harvested Eurycea cirrigera X X X X X X X X Eurycea longicauda guttolineata

X X X X X X X X

Plethodon grobmani X X X X X X X X Pseudotriton ruber vioscai

X X X X X - X X

Desmognathus apalachicoloae

- - X - X X X X

Notophthalmus viridescens

- X X - - - - -

49

Table 3-5. Treefrog species presence in four sub-watersheds of the Dry Creek Basin, Georgia. Watershed A: Forested Watershed B: Harvested Year 0

(Pre-harvest)

Year 1

Year 2

Year 3

Year 0 (Pre-

harvest)

Year 1

Year 2

Year 3

Hyla squirella X X X X X X X X Hyla cinerea X X X X - X X X Hyla chrysoscelis - X X X - X X - Hyla femoralis - - - X - - - - Hyla avivoca - - - X - - - - Pseudacris crucifer

X X X X X X - X

Watershed D: Forested Watershed C: Harvested Hyla squirella X X X X X X X X Hyla cinerea X X X X X X X X Hyla chrysoscelis - - X X - X X X Hyla femoralis - - X X X - - - Hyla avivoca - - - X - - - - Pseudacris crucifer

X X X X X X - -

50

Figure 3-1. Location of study site in relation to physiographic regions (modified from US Forest

Service, 1969).

51

0

20

40

60

streamside riparian midslope upslope

Distance from stream

Sala

man

der c

ount

(Pre

-har

vest

)

SMZ

boun

dary

Figure 3-2. Mean salamander count in relation to distance from streams in four sub-watersheds

of the Dry Creek Basin, Georgia. Captures occurred during pre-harvest survey period December 2002–November 2003. Line designates generalized location of SMZ. Bars represent standard deviation.

0

20

40

60

streamside riparian midslope upslope

Distance from stream

Tree

frog

cou

nt(P

re-h

arve

st)

SMZ

boun

dary

Figure 3-3. Mean treefrog count in relation to distance from streams in four sub-watersheds of

the Dry Creek Basin, Georgia. Captures occurred during pre-harvest survey period December 2002–November 2003. Line designates hypothetical boundary where SMZ would occur. Bars represent standard deviation.

52

Figure 3-4. Mean (+1 SE) annual terrestrial salamander counts in four watersheds (A, B, C, D).

Harvested (H) watersheds were paired with forested (F) watersheds during the pre-harvest survey (Year 0) and monthly monitoring continued three years following clear-cut harvesting. Significant differences in counts among watershed pairs (A+B, C+D) are denoted (P<0.05 = *, P< 0.01 = **).

Figure 3-5. Mean (+1 SE) annual treefrog counts in forested (F) and harvested (H) watersheds of

the Dry Creek Basin, GA. Significant differences in paired watersheds A+B and C+D are denoted (P<0.05 = *; P<0.01 = **).

53

Figure 3-6. Mean (+1 SE) annual terrestrial salamander counts within upstream/downstream

portions of forested watersheds compared to thinned and intact SMZs of harvested watersheds. Significant differences between upstream-downstream and thinned-intact stream segments are denoted (P<0.05 = *).

Figure 3-7. Mean (+1 SE) annual in-stream salamander larvae counts for reference and harvested

watersheds. Significant differences between upstream-downstream and thinned-intact stream segments are denoted (P<0.05 = *, P<0.01 = **).

54

Figure 3-8. Mean (+1 SE) annual treefrog counts for reference and harvested watersheds.

Significant differences between upstream-downstream and thinned-intact stream segments are denoted (P<0.05 = *, P<0.01 = **).

55

CHAPTER 4 CONCLUSIONS

The objective of this research was to determine amphibian responses to forest harvest and

SMZ management techniques applied in southwestern Georgia. Conflicting evidence on

responses of amphibians to forest harvest across the U.S. warranted additional research on this

subject. Amphibians are abundant in southeastern ecosystems and are reliant on the same habitat

variables that SMZs are intended to protect. Therefore inferences on how amphibians respond to

forest harvest and SMZ practices can aid in determining if these practices are effective in

protecting stream and riparian ecosystems.

Amphibian populations are difficult to characterize due to fluctuations in detection

probability. When developing relative (observed) abundance indexes, it is assumed that detection

probability is constant across space and time. However, when salamander detections with

coverboards were evaluated (Chapter 2) it was found that detection probability changes as

coverboards age. Salamanders were detected more often with three-year old coverboards than

new coverboards. This suggests that temporal comparisons of abundance indices made over 3-

years violate assumptions of constant detection probability. Therefore it is suggested that

coverboards be replaced every two years to maintain similar detection probability. To alleviate

bias in relative abundance indices, it is suggested that statistical estimates of detection

probability be considered.

This study also evaluated amphibian responses to forest harvest and management practices

in SMZs (Chapter 3). All species of amphibians detected during the pre-harvest survey period

were found present following clear-cut timber harvesting of watersheds. Counts of terrestrial

salamanders within SMZs were similar before and after harvest. There was a decline in treefrog

counts within SMZs of harvested relative to forested watersheds. SMZs in downstream segments

56

were thinned and compared to upstream intact SMZs. There was no effect of thinning SMZs on

terrestrial or larval salamanders, however, thinning adversely affected treefrogs, resulting in

lower counts compared to intact SMZs. Treefrogs may be more susceptible to forest harvest than

salamanders due to their need for large dispersal ranges. Whether treefrogs use SMZs as

corridors for movement and metapopulation dynamics remains uncertain. Furthermore, thinning

resulted in more canopy loss due to windthrow than intact SMZs. Thinning in this region may

not be appropriate due to high frequency of hurricanes and tropical storms.

57

LIST OF REFERENCES

Alford, R. A. & Richards, S. J. (1999). Global amphibian declines: A problem in applied ecology. Annual Review of Ecology and Systematics, 30, 133-165

Ash, A. N. (1997). Disappearance and return of plethodontid salamanders to clearcut plots in the southern blue ridge mountains. Conservation Biology, 11, 983-989

Biek, R., Mills, S. L., & Bury, B. R. (2002). Terrestrial and stream amphibians across clearcut-forest interfaces in the Siskiyou Mountains, Oregon. Northwest Science, 76, 129-140

Blaustein A. R & Kiesecker J. M. (2002). Complexity in conservation: lessons from the global decline of amphibian populations. Ecology Letters, 5, 597–608

Blaustein, A. R., Wake, D. B. & Sousa, W. P. (1994). Amphibian declines: Judging stability, persistence, and susceptibility of populations to local and global extinctions. Conservation Biology, 8, 60-71

Boughton, R. G., Staiger, J. & Franz, R. (2000). Use of PVC pipe refugia as a sampling technique for hylid treefrogs. The American Midland Naturalist, 144, 168-177

Burton T. M., & Likens G. E. (1975a). Energy flow and nutrient cycling in salamander populations in the Hubbard Brook experimental forest, New Hampshire. Ecology 56, 1068–1080

Burton, T. M. & Likens, G. E. (1975b). Salamander populations and biomass in the Hubbard Brook experimental forest, New Hampshire. Copeia, 1975, 541-546

Clawson, R. G., Lockaby, B. G. & Jones, R. H. (1997). Amphibian responses to helicopter harvesting in forested floodplains of low order, blackwater streams. Forest Ecology and Management, 90, 225-235

Collins, J. P. & Storfer, A. (2003). Global amphibian declines: Sorting the hypotheses. Diversity and Distributions, 9, 89-98

Committee on Riparian Zone Functioning and Strategies for Management, Water Science and Technology Board, National Research Council. (2002). Riparian Areas: Functions and Strategies for Management [Electronic version]. Retrieved December 13, 2007, from http://www.nap.edu/catalog.php?record_id=10327

Couch, C. A, Hopkins, E. H., & Hardy, P. S. (1996). Influences of environmental settings on aquatic ecosystems in the Apalachicola-Chattahoochee-Flint river basin. U.S. Geological Survey National Water-Quality Assessment Program. Water-Resources Investigations Report 95-4278. 58 pp.

Dale, V. H., & Beyeler, S. C. (2001). Challenges in development and use of ecological indicators. Ecological Indicators, 1, 3-10

58

Dodd, C. K. & Dorazio, R. M. (2004). Using counts to simultaneously estimate abundance and detection probabilities in a salamander community. Herpetologica, 60, 468-478

Duellman, W. E., & Trueb, L. (1994). Biology of Amphibians. (The Johns Hopkins University Press)

Entrekin, S., Golladay, S., Ruhlman, M., & Hedman, C. (1999). Unique steephead stream segments in Southwest Georgia: invertebrate diversity and biomonitoring. Proceedings of the 1999 Georgia Water Resources Conference, held March 30-31, 1999, at University of Georgia. Kathryn J. Hatcher, editor, Institute of Ecology, The University of Georgia, Athens, Georgia.

Enge, K. M. (2002). Herpetofaunal drift-fence survey of steephead ravines and seepage bogs in the western Florida Panhandle. Final Performance Report. Florida Fish and Wildlife Conservation Commission, Tallahassee, Florida, USA.

Georgia Forestry Commission (1999). Georgia’s Best Management Practices for Forestry. Retrieved December 13, 2007, from http://www.gfc.state.ga.us/ForestManagement/documents/GeorgiaForestryBMPManual.pdf

Grant, B. W., Tucker, A., D., Lovich, J. E., Mills, A. M., Dixon, P. M., & Gibbons, J. W. (1992). The use of coverboards in estimating patterns of reptile and amphibian biodiversity. (In D. R. McCullough, & R. H. Barrett (Eds.), Wildlife 2001: Populations (pp. 379–403). Elsevier Science Publ. Inc., London, England.)

Grialou, J. A., West, S. D. & Wilkins, N. R. (2000). The effects of forest clearcut harvesting and thinning on terrestrial salamanders. The Journal of Wildlife Management, 64, 105-113

Grover, M. C. (2000). Determinants of salamander distributions along moisture gradients. Copeia, 2000, 156-168

Grover, M. C. (1998). Influence of cover and moisture on abundances of the terrestrial salamanders plethodon cinereus and plethodon glutinosus. Journal of Herpetology, 32, 489-497

Guerry, A. D., & Hunter, M. L. (2002). Amphibian distributions in a landscape of forests and agriculture: an examination of landscape composition and configuration. Conservation Biology, 16, 745-754

Harrison, S., & Voller, J. (1998). Conservation biology principles for forested landscapes. (University of British Columbia Press)

Harpole, D. N., & Haas, C. A. (1999). Effects of seven silvicultural treatments on terrestrial salamanders. Forest Ecology and Management, 114, 349-356

Hartley, M. J. (2002). Rationale and methods for conserving biodiversity in plantation forests. Forest Ecology and Management, 155, 81-95

59

Heyer, W. R., Donnelly M. A., McDiarmid, R. W., Hayek, L. C., & Foster, M. S. (1994). Standard Techniques. (In W. R. Heyer, M. A. Donnelly, R. W. McDiarmid, L. C. Hayek & M. S. Foster (Eds), Measuring and Monitoring Biological Diversity: Standard Methods for Amphibians (pp. 75-78) Smithsonian Inst. Press, Washington, D. C.)

Holomuzki, J. R., Collins J. P., & Brunkow P. E. (1994). Trophic control of fishless ponds by tiger salamander larvae. Oikos, 71, 55–64.

Houlahan, J. E., Scott, F. C., Benedikt, S. R., Meyer, A. H., & Sergious, K. L. (2000). Quantitative evidence for global amphibian population declines. Nature, 404, 752-755.

Hyde, E. J. & Simons, T. R. (2001). Sampling plethodontid salamanders: Sources of variability. The Journal of Wildlife Management, 65, 624-632

Houze, M. C., & Chandler, C. R. (2002). Evaluation of coverboards for sampling terrestrial salamanders in South Georgia. Journal of Herpetology, 36, 75-81

Jones, D.G., Summer, W.B., Miwa, M., & Jackson, C.R. (2003). Baseline characterization of forested headwater stream hydrology and water chemistry in southwest Georgia. (In: Connor, K.F. (Eds). Proceedings of the 12th biennial southern silviculture research conference. Gen. Tech. Rep. SE-71. Asheville, NC: U.S. Department of Agriculture, Forest Service, Southern Research Station: 161-165)

Knapp, S. M., Haas, C. A., Harpole, D. N., & Kirkpatrick, R. L. (2003). Initial effects of clearcutting and alternative silvicultural practices on terrestrial salamander abundance. Conservation Biology, 17, 752-762

MacKenzie, D. I., Nichols, J. D., Lachman, G. B., Droege, S., Royle, J. A., & Langtimm, C. A. (2002). Estimating site occupancy rates when detection probabilities are less than one. Ecology, 83, 2248-2255

Means, D. B. (1985). The canyonlands of Florida. Nature Conservancy News, September/October, 13-17

Means, D. B. (2000). Southeastern U.S. Coastal Plain habitats of the Plethodontidae: the importance of relief, ravines, and seepage. (In R. C. Bruce, R. J. Jaeger, and L. D. Houck, (Eds.) The biology of plethodontid salamanders. (Pages 287–302) Plenum, New York, New York, USA.)

Marsh, D. M. & Goicochea, M. A. (2003). Monitoring terrestrial salamanders: Biases caused by intense sampling and choice of cover objects. Journal of Herpetology, 37, 460-466

Maxcy, K. A. & Richardson, J. (2000). Abundance and movements of terrestrial salamanders in second-growth forests of southwestern British Columbia. (Proceedings of a conference on the biology and management of species and habitats at risk, Kamloops, B. C.)

60

Monti, L., Hunter, M., & Witham, J. (2000). An evaluation of the artificial cover object (ACO) method for monitoring populations of the redback salamander plethodon cinereus. Journal of Herpetology, 34, 624-629

New Georgia Encyclopedia, 2007. http://www.georgiaencyclopedia.org/nge/Article.jsp?id=h-2188&hl=y. Last accessed March 26, 2007.

Niemela, J. (2001). The utility of movement corridors in forested landscapes. Scandinavian Journal of Forest Research, 16, 70

O’Neill, E. D. (1995). Amphibian and reptile communities of temporary ponds in a managed pine flatwoods. Gainesville, FL: University of Florida. 106 p. M.S. thesis.

Parris, K. M. (1999). Review: Amphibian Surveys in Forests and Woodlands. Contemporary Herpetology, 1, 1-14.

Petranka, J. W. (1994). Response to impact of timber harvesting on salamanders. Conservation Biology, 8, 302-304

Petranka, J. W. (1998). Salamanders of the United States and Canada. Smithsonian Institution Press, Washington DC.

Petranka, J. W., & Murray, S. S. (2001). Effectiveness of removal sampling for determining salamander density and biomass: a case study in an Appalachian streamside community. Journal of Herpetology, 35, 36–44

Pollock, K. H., Nichols, J. D., Simons, T. R., Farnsworth, G. L., Bailey, L. L., & Sauer, J. R. (2002). Large scale wildlife monitoring studies: statistical methods for design and analysis. Environmetrics, 13, 105-119

Royle, J. A. (2004). N-mixture models for estimating population size from spatially replicated counts. Biometrics, 60, 108-115

Russell, K. R., Van Lear, D.H., & Guynn, D.C., Jr. (2002). Responses of isolated wetland herpetofauna to upland forest management. Journal of Wildlife Management. 66, 603-617

Russell, K. R., Wigley, T. B., Baughman, W. M., Hanlin, H. G., & Ford, M. W. (2004). Responses of southeastern amphibians and reptiles to forest management: a review. (In Southern Forest Science. United States Department of Agriculture Forest Service, Southern Research Station. General Technical Reprort SRS-75, pp. 319-334.)

Ryan, T. J., Philippi, T., Leiden, Y. A., Dorcas, M. E., Wigley, T. B. & Gibbons, J. W. (2002). Monitoring herpetofauna in a managed forest landscape: Effects of habitat types and census techniques. Forest Ecology and Management, 167, 83-90

Schmidt, B. R. (2003). Count data, detection probabilities, and the demography, dynamics, distribution, and decline of amphibians. Comptes Rendus Biologies, 326, 119-124

61

Semlitsch, R. D. (1998). Biological delineation of terrestrial buffer zones for pond-breeding salamanders. Conservation Biology, 12, 1113-1119

Semlitsch, R. D. & Bodie, J. R. (2003). Biological criteria for buffer zones around wetlands and riparian habitats for amphibians and reptiles. Conservation Biology, 17, 1219-1228

Sharitz, R. R., Boring, L. R., Van Lear, D. H., & Pinder, J. E. (1992). Integrating ecological concepts with natural resource management of southern forests. Ecological Applications, 2, 226-237

Smith, C. K. & Petranka, J. W. (2000). Monitoring terrestrial salamanders: Repeatability and validity of area-constrained cover object searches. Journal of Herpetology, 34, 547-557

Southeast Regional Climate Center. Bainbridge, Georgia Climate Information. http://climate.engr.uga.edu/bainbridge/index.html, Last accessed August 8, 2004.

Stewart M. M., & Woolbright L. L. (1996). Amphibians. (In: Reagan DP and Waide RP (Eds). The food web of a tropical rainforest. Chicago, IL: University of Chicago Press.)

Stuart, S. N., Chanson, J. S., Cox, N. A., Young, B. E., Rodrigues A. S., Fischman, D. L., & Waller, R. W. (2004). Status and trends of amphibian declines and extinctions worldwide. Science 306, 1783–86

Summer, W.B., Jackson, C. R., Jones, D. G., & Miwa, M. (2003). Characterization of hydrologic and sediment transport behavior of forested headwater streams in southwest Georgia. (In: Proceedings of the 2003 Georgia Water Resources Conference, Athens, GA. 23-24 April 2003. The Institute of Ecology: The University of Georgia, Athens, GA. 157-160.)

Summer, W.B., Jackson, R.C., Jones, D., Golladay, S.W., & Miwa, M., (2005). Hydrologic and Sediment Transport Response to Forestry; Southwest Georgia Headwater Streams. (In: Proceedings of the 2005 Georgia Water Resources Conference, held April 25-27, 2005. The Institute of Ecology: The University of Georgia, Athens, GA.)

Sun, G., Riedel, M., Jackson, R., Kolka, R., Amataya, D., & Shepard, J., (2004). Chapter 19: Influences of Management of Southern Forests on Water Quality and Quantity. United States Department of Agriculture, Forest Service Southern Research Station General Technical Report SRS-75. Southern Forest Science: Past, Present, Future. 408 p.

United States Environmental Protection Agency (2005). National Management Measures to Control Nonpoint Source Pollution from Forestry. http://www.epa.gov/owow/nps/forestrymgmt/. Last accessed March 26, 2007.

United States Forest Service (1969). A Forest Atlas of the South. Southern Forest Experiment Station and Southeastern Forest Experiment Station Publication. 27 pp.

U.S. Geological Survey National Water-Quality Assessment Program. Water-Resources Investigations Report 95-4278. 58 pp.

62

Vitt, L. J., J. P. Caldwell, H. M. Wilbur, & D. C. Smith (1990). Amphibians as harbingers of decay. BioScience 40, 418

Wake D. B. (1991). Declining amphibian populations. Science 253, 860

Welsh, H. H. & Droege, S. (2001). A case for using plethodontid salamanders for monitoring biodiversity and ecosystem integrity of north american forests. Conservation Biology, 15, 558-569

Wissinger S. A., Whiteman, H. H., Sparks, G. B., Rouse, G. L., & Brown, W. S. (1999). Foraging tradeoffs along a predator-permanence gradient in subalpine wetlands. Ecology, 80, 2102–16

Wyatt, J. L. & Forys, E. A. (2004). Conservation implications of predation by Cuban Treefrogs (Osteopilus septentrionalis) on native hylids in Florida. Southeastern Naturalist, 3, 695-700

63

BIOGRAPHICAL SKETCH

Diane W. Bennett spent her childhood years in Nashville, Tennessee as the youngest of

three children. In middle school, she was relocated to the suburbs of Orlando, Florida where she

spent her teenage years attending Lyman High School. After graduating high school in 1997, she

traveled for two years working odd jobs such as meat cutting and cable lineman until learning

she had been awarded the Florida Bright Futures Scholarship, for which she aptly returned to

Florida to enroll in Valencia Community College.

College began with courses in remedial math and English, but hard work and

determination led her to tutoring mathematics, eventually earning over 32 credits in mathematics

alone. She attended leadership workshops and worked closely with students and professors to

improve campus environmental quality through club and community forums. She served as

president for Phi Theta Kappa (International Honors Society) and co-founded The

Environmental Club, which was awarded “Best Environmental Outreach” by Florida Leader

Magazine, May 2002. In May 2003, she graduated with Honors, receiving her Associates Degree

for Pre-Engineering, and was recognized as Valencia Community College Alumni Association’s

Distinguished Graduate, for which she was keynote speaker at her graduation.

In January 2004, Diane enrolled in the Environmental Engineering Sciences Department at

The University of Florida to complete her bachelor’s degree. After attending an applied ecology

course in summer 2004, she began working for Dr. Thomas L. Crisman in the Howard T. Odum

Center for Wetlands. Once there, she branched off from engineering to study ecological

indicators and watershed management. In May 2005 she was granted a monetary award by the

University Scholar’s Program to complete a study with the advisement of Dr. Crisman to

evaluate amphibians along disturbance gradients in urban headwater streams. As part of her

capstone design course for engineering, Diane worked with a team of student engineers to design

64

an expansion to a wastewater treatment facility. This team was awarded First Place at both

Regional and National conferences in 2006. She graduated with her bachelor’s degree in

environmental engineering sciences in May 2006, and immediately enrolled in the master’s

program to continue working under the direction of Dr. Crisman.

Upon completion of her master’s degree, Diane will search for government employment

involving sustainable watershed development. She hopes to apply her knowledge of ecology

with engineering, and to bridge gaps between these sciences. She plans to continue her

education, eventually earning her PhD, with anticipation that one day she may return inspiration,

nurturing, and guidance to students like that she received throughout her academic career.

65