Research Project No. 2003-18
FHWA/NC/2004-09
Final Report
Ecological Assessment of a Wetlands
Mitigation Bank (Phase III: Restoration
Efforts)
Prepared By:
Kevin K. Moorhead, Irene M. Rossell,
Barbara C. Reynolds, C. Reed Rossell, Jr.
Department of Environmental Studies
University of North Carolina at Asheville
Asheville, NC 28804
and
James W. Petranka
Department of Biology
University of North Carolina at Asheville
Asheville, NC 28804
August 2004
The contents of this report reflect the views of the author(s),
who are responsible for the facts and the accuracy of the
information presented herein. This document is disseminated
under the sponsorship of the U.S. Department of Transportation,
University Transportation Centers Program, in the interest of
information exchange. The U.S. Government assumes no
liability for the contents or use thereof.
2
Technical Report Documentation Page
1. Report No.
FHWA/NC/2004-09
2. Government Accession No. 3. Recipient’s Catalog No.
4. Title and Subtitle
Ecological Assessment of a Wetlands Mitigation Bank
(Phase III: Restoration Efforts)
5. Report Date
August 2004
6. Performing Organization Code
7. Author(s)
Kevin K. Moorhead, Irene M. Rossell, Barbara C. Reynolds,
C. Reed Rossell, Jr., and James W. Petranka
8. Performing Organization Report No.
9. Performing Organization Name and Address
Departments of Environmental Studies and Biology
10. Work Unit No. (TRAIS)
University of North Carolina at Asheville
Asheville, NC 28804
11. Contract or Grant No.
12. Sponsoring Agency Name and Address
US Department of Transportation,
Research and Special Programs Administration
13. Type of Report and Period Covered
Final Report
July 2002 – June 2004
400 7th Street, SW
Washington, DC 20590-0001
14. Sponsoring Agency Code
2003-18
Supplementary Notes:
Supported by a grant from the US Department of Transportation and the North Carolina Department of
Transportation through the Center for Transportation and the Environment, NC State University.
16. Abstract
The overall objective for the Tulula Wetlands Mitigation Bank has been to restore the functional and structural characteristics of a mountain stream and the
adjacent alluvial wetlands. Specific restoration objectives of this study included: 1) determining the success of stream realignment by evaluating the
geomorphology of a new channel before and after water release, 2) evaluating changes in ecosystem structure and function associated with plant community
succession in planted and unplanted portions of the floodplain in response to restored hydrology, and 3) evaluating wildlife use of the site in response to
changing hydrologic conditions (amphibians) and plant community succession (birds). A meandering channel (8,500 linear feet in length) was constructed
across the floodplain and water was released into the new channel in 2001 and 2002. Eight random channel segments were used for measurements of
stream geomorphology and after two years of water flow few differences were noted for channel pattern, although changes were observed for cross-sectional
areas of riffles and pools. Isolated areas of bank and bed erosion were noted. The hydrology of Tulula has been influenced by the stream restoration, with
most notable differences occurring for water-table wells located near the channel. Although the hydrology of Tulula fen was not influenced by stream
restoration, the composition of wetland plant communities in the fen was influenced by natural succession. Restoration did influence the composition of
some plant communities. For example, restored wetland areas contained fewer species than unrestored areas or restored dry areas, and the species that
dominated the restored wet areas were OBL and FACW plants. In addition, production of both vegetative and reproductive stems of a common rush was
influenced by restoration and hydrologic change. Recently disturbed areas at Tulula had lower decomposition rates and fewer litter microarthropods
compared to older plant communities. Data collected from 1996-2004 indicate that constructed ponds are of higher quality than reference ponds based on
physiochemical characteristics, seasonal hydroperiod, and use by resident amphibians. Amphibians rapidly colonized constructed vernal ponds, and the
number of species that utilize these as breeding sites averaged about 50% higher than that of reference ponds. The survivorship and output of wood frog
and spotted salamander juveniles have declined since pond construction, in part due to the accumulation of predators in ponds, the outbreak of a virus
pathogen, and premature pond drying associated with drought. In 2004, bird species richness and relative bird abundance decreased significantly from 2002
levels. Bird species richness decreased 15% and relative bird abundance decreased 52%. Generalist species, such as Song Sparrow and Rufous-Sided
Towhee, continued to be the most abundant species, while many Neotropical migrants of conservation concern, including the Golden-winged Warbler,
Chestnut-sided Warbler, Hooded Warbler, and Yellow-breasted Chat, declined substantially. The significant declines in bird species richness and relative
bird abundance are attributed to habitat changes associated with the flooding of a large proportion of the floodplain as a result of beaver activity.
Management intervention is recommended to control the flooding caused by beaver, and to maintain a variety of early-successional habitat types throughout
the site.
17. Key Words
Wetlands, wetland conservation, mitigation measures, restoration
ecology, site surveys, geomorphology, hydrology, water table,
plant location, amphibians, birds
18. Distribution Statement
19. Security Classif. (of this report)
Unclassified
20. Security Classif. (of this page)
Unclassified
21. No. of Pages
88
22. Price
Form DOT F 1700.7 (8-72) Reproduction of completed page authorized
3
DISCLAIMER
The contents of this report reflect the views of the authors, who are responsible for the facts and the
accuracy of the data presented herein. This document is disseminated under the sponsorship of the
U.S. Department of Transportation and North Carolina Department of Transportation in the interest of
information exchange. This report does not constitute a standard, specification, or regulation. The US
Government assumes no liability for the contents or use thereof.
ACKNOWLEDGMENTS
Support for this project was provided by the U. S. Department of Transportation and the North
Carolina Department of Transportation through the Center for Transportation and the Environment, NC
State University. The authors thank Victor Agraz, Robert Warren, Duncan Quinn, and Dr. Dan Pittillo
for their contributions to this research. We also thank the numerous undergraduate students of UNCA
for their efforts.
4
TABLE OF CONTENTS
LIST OF TABLES………………………………………………………………………. 5
LIST OF FIGURES……………………………………………………………………… 6
EXECUTIVE SUMMARY......................................................................................……. 7
I. INTRODUCTION................................................................................................. 9
II. RESEARCH METHODS AND RESULTS.......................................................…. 10
A. Stream Restoration and Hydrology…………………………………………… 10
B. Vegetation Responses to Restoration…………………………………………. 23
C. Decomposition and Soil Microfauna………………………………………….. 34
D. Amphibian Use of Tulula……………………………………………………… 42
E. Bird Use of Tulula……………………………………………………………. 54
III. DISCUSSION.................................................................................................……. 60
IV. RECOMMENDATIONS.................................................................................…… 62
V. LITERATURE CITED.....................................................................................…… 63
APPENDIX A. (Cross sections of riffles and pools in eight stream segments)…………… 67
APPENDIX B. (Pre- and post-restoration water-table data from electronic wells)……….. 71
APPENDIX C. (Pre- and post-restoration water-table data from manual wells).………… 80
APPENDIX D. (Amphibian and reptile species of Tulula)……………………………….. 85
APPENDIX G. (Bird Species at Tulula Wetland (1994-2004)…………………………… 86
5
LIST OF TABLES
Table 1. Design criteria for the restored Tulula Creek…………………………………………. 12
Table 2. Bankfull width and cross-sectional area of riffles and pools………………………… 15
Table 3. Percent change in cross-sectional area of riffles and pools………………………… 17
Table 4. Sinuosity and slope of the water surface over time…………………………………... 17
Table 5. Width/depth ratio and maximum depth of riffles and pools………………………… 18
Table 6. Other physical characteristics of selected meanders in each stream segment……….. 18
Table 7. Erosion of channel banks after two years of water flow…………………………….. 20
Table 8. Taxa and wetland indicator status of plants occurring in in four study areas……….. 26
Table 9. Contribution of each wetland indicator status in four study areas at Tulula………… 28
Table 10. Effects of restoration on vegetative growth and reproduction of Juncus effusus….. 29
Table 11. Effects of restoration on biomass of plants occurring with Juncus effusus………… 29
Table 12. Effects of hydrology on vegetative growth and reproduction of Juncus effusus…... 29
Table 13. Effects of hydrology on biomass of plants occurring with Juncus effuses…………. 30
Table 14. Total number of overstory trees of each species in 20, 10x10-m2 plots…………….. 31
Table 15. Importance values for overstory trees in 10x10-m2 plots…………………………… 31
Table 16. Total number of understory trees of each species in 20, 4x4-m2 plots……………… 32
Table 17. Importance values for understory trees in 4x4-m2 plots……………………………. 33
Table 18. Mean percent cover of each plant type in 1x1-m2 quadrats…………………………. 33
Table 19. Importance values for plant types in 1x1-m2 quadrats……………………………… 34
Table 20. Survival of commercial red maple seedlings planted in 1995………………………. 35
Table 21. Microarthropod responses to date and site………………………………………….. 37
Table 22. Relative abundance and migratory status of birds………………………………….. 57
Table 23. Means of bird richness, relative bird abundance, and habitat variables……………. 59
6
LIST OF FIGURES
Fig. 1. Restored channels sections of Tulula Creek…………………………………………… 14
Fig. 2. Approximate locations of stream segments used for channel evaluations…………….. 15
Fig. 3. Cumulative pebble counts of seven stream segments………………………………….. 19
Fig 4. Transects and individual electronic wells used to assess site hydrology……………….. 21
Fig. 5. Location of manual wells at Tulula……………………………………………………. 22
Fig. 6. The daily water table and monthly averages for electronic well X1…………………… 24
Fig. 7. Percent litter remaining in litterbags after 17 months in the field……………………… 37
Fig. 8. Average number of microarthropods for three collection dates……………………….. 39
Fig. 9. Average number of total microarthropods for March, 2003…………………………… 40
Fig. 10. Average percent organic carbon for soil from five plant communities……………… 41
Fig. 11. Average pH for soil from five plant communities……………………………………. 42
Fig. 12. Location of standing water habitats within the study site (spring 2004)...…………… 44
Fig. 13. Physiochemical characteristics of reference and constructed ponds………………….. 45
Fig. 14. Mean number of species that bred in reference and constructed ponds………………. 47
Fig. 15. Response of female wood frog and spotted salamanders to pond construction……… 49
Fig. 16. Estimates of the percentage of ponds that produced juveniles, and total juvenile recruitment
from constructed and reference ponds during 1996-2003………………………… 50
Fig. 17. Annual variation in the percentage of constructed and reference ponds that either did not fill or
that dried before larvae could initiate metamorphosis…………………………………… 51
Fig. 18. Changes in the percentage of reference and constructed ponds in which catastrophic die-offs
of larvae occurred from Ranavirus infections…………………………………………….. 52
Fig. 19. Changes in adult breeding population size on the eastern sector based on annual egg mass
counts in all breeding sites………………………………………………………………... 53
Fig. 20. Yearly changes in the proportion of the ten constructed ponds that contained fish… 53
Fig. 21. Location of bird survey and habitat plots…………………………………………… 56
7
EXECUTIVE SUMMARY
Our goal is to document the ecological success of the wetlands at the Tulula Wetlands
Mitigation Bank (Graham County) in response to restored hydrology, soils, and vegetation. Our data
should provide NCDOT an ecological assessment that may be useful for evaluating other wetland
restoration projects located throughout the state. The following objectives provide the framework for a
comprehensive ecological assessment of the restored wetlands of Tulula: 1) determine the success of
stream realignment by evaluating the geomorphology of the new channel before and after water is
introduced, 2) evaluate changes in ecosystem structure and function associated with plant community
succession in planted and unplanted portions of the floodplain in response to a higher water table and
overbank flooding, and 3) evaluate wildlife use of the site in response to changing hydrologic conditions
(amphibians) and plant community succession (birds).
A primary focus of restoration at Tulula was to improve site hydrology. A meandering channel
(8,500 linear feet in length) was constructed across the floodplain in five separate sections that were
connected in fall 2001 and summer 2002. Eight random channel segments were used for measurements
of stream geomorphology, including sinuosity, cross-sectional areas of riffles and pools, bank slope,
slope of the water surface, and overall channel configuration. After two years of water flow, differences
were noted in certain aspects of channel morphology, and localized areas of erosion were noted with
erosion control pins and through increases in the cross-sectional areas of some riffles and pools.
However, the overall configuration of the channel was maintained over the two-year period.
The restoration of hydrology at Tulula was evaluated primarily by changes in water-table depth
as recorded with a series of electronic and manual wells. Our assumption was that the overall water
table of the site would rise after the channel was restored and the drainage ditches were plugged. We
found that the hydrology of Tulula was influenced by these restoration efforts, with most changes
occurring in water-table wells located near the stream channel. Restoration appeared to have little
influence on the hydrology of the fen or of areas located farther from the channel.
Natural succession continues to change the composition of wetland plant communities across
Tulula. In 2003, overstory-sized trees were present in a fen that had been characterized by an open
canopy in 1994, and there was a dramatic increase in the number of understory-sized trees. The
ground layer in this part of the fen also showed an increase in woody species, and a decrease in the
importance of plants that require sunlight, such as rushes.
Soil disturbance attributed to restoration activities increased the taxonomic richness in dry areas.
In wet areas, restoration combined with a high water table led to colonization by almost almost
exclusively OBL and FACW species. Both restoration and the higher water table increased the number
and biomass of vegetative stems of Juncus effusus (soft rush), and the higher water table increased the
number of reproductive stems of this species.
8
Ten ponds were constructed in 1995-1996 to replace natural breeding sites that were
destroyed during golf course construction. Data collected from 1996-2004 indicate that constructed
ponds are of higher quality than reference ponds based on physiochemical characteristics, seasonal
hydroperiod, and use by resident amphibians. The reference ponds have progressively deteriorated
between 1996-2002 with respect to seasonal hydroperiod. In 2002 the majority either did not fill or
dried prematurely, resulting in catastrophic mortality of pond populations. In contrast, the hydroperiod
of most constructed ponds appears to be ideal for most vernal pond breeders. Seven of 10 ponds
underwent seasonal drying in most years, typically in late summer or fall after larvae had
metamorphosed. Fish have colonized many ponds since 2002 in association with above normal rainfall,
beaver activity, and completion of the final phase of reconstruction.
Amphibians rapidly colonized the constructed ponds, and the number of species that utilize
these as breeding sites averaged about 50% higher than that of reference ponds. The survivorship and
output of juveniles of two focal species (wood frog; spotted salamander) have declined since pond
construction, in part due to the accumulation of predators in ponds, the outbreak of a virus pathogen,
and premature pond drying associated with drought. Nonetheless, a small percentage of ponds on site
have successfully produced juveniles annually, and populations of both species are being maintained at
viable levels.
Results of breeding bird surveys in 2004 indicated that species richness and relative abundance
decreased significantly from 2002 levels. Species richness decreased 15%, with 33 species recorded.
American Woodcock, Common Grackle, and Eastern Wood-pewee were new species recorded
during surveys. Relative bird abundance decreased 52%, with a total of 166 observations. Generalist
species, such as Song Sparrow and Rufous-Sided Towhee, continued to be the most abundant species
breeding at Tulula, but their numbers decreased dramatically from 2002 levels. The Red-winged
Blackbird also continued to be one of the most abundant species, but its numbers held steady relative to
2002 levels. Many Neotropical migrants of conservation concern declined substantially in 2004
including the Golden-winged Warbler, Chestnut-sided Warbler, Hooded Warbler, and Yellow-breasted
Chat.
The significant declines in bird species richness and abundance in 2004 are attributed to habitat
changes associated with the flooding of a large proportion of the floodplain as a result of beaver activity.
Productivity of the habitat for birds at Tulula has decreased and correlates with an increase in the large
amounts of area covered with standing water and dominated by rushes and sedges. Management
intervention is needed in order to restore the productivity of the habitat for birds. Management
objectives should include taking appropriate actions to control the flooding caused by beaver, and
maintaining a variety of early-successional habitat types throughout the site.
9
I. INTRODUCTION
Surface transportation projects such as highway construction often impact wetland resources
and cause unavoidable losses of small wetland areas. Increasingly, wetland losses are being mitigated
by the creation of "banks" of restored or natural wetlands that are protected from future disturbance.
Mitigation banks allow the consolidation of efforts to mitigate for small wetland losses, facilitate
advanced planning, and enhance the monitoring and evaluation of mitigation projects (Short 1988).
The Tulula Wetland Mitigation Bank was created to offset impacts of highway projects in western
North Carolina, particularly in the Little Tennessee River basin (1,158,883 ac) located in Macon,
Swain, Graham, Jackson, Clay, and Transylvania Counties. The site was ideal for establishing a
mitigation bank in the mountains of North Carolina because of its relatively large size (235 ac) and its
need for large-scale restoration.
The Tulula Wetland Mitigation Bank (Tulula) (35o17'N, 83o41'W) is located in Graham
County, NC in the floodplain of Tulula Creek, 7.7 miles west of Topton. The site covers approximately
235 ac at an elevation ranging from 2500 to 2800 ft. It is characterized by a relatively large, level
floodplain along Tulula Creek, and is bordered by forested uplands and infrequent seepage communities
on adjacent slopes. A complete description of vegetative communities at Tulula is found in Moorhead
et al. (2001a). Tulula was part of the Nantahala National Forest and owned by the U.S. Forest Service
until the mid-1980's, when it was traded to a group of developers for commercial development of a golf
course. During construction of the golf course, the bed of Tulula Creek was dredged and channelized
and several drainage ditches were dug. Spoil from the drainage ditches and from 11 small golf ponds
was spread over portions of the floodplain. A large portion of the floodplain forest was removed during
the construction of 18 fairways. About 40% of the wetlands were disturbed by drainage and timber
harvest during golf course construction.
Tulula was purchased in 1994 by the North Carolina Department of Transportation (NCDOT)
to develop a wetlands mitigation bank. We have collected information on baseline ecological conditions
(soils, hydrology, flora, and fauna) and have evaluated restoration activities at the site since 1994 (see
www.unca.edu/tulula for details and species lists).
Assessing the success of wetland restoration projects requires an evaluation of ecosystem
structure and function. Long-term success is rarely documented, and failure is common for a variety of
reasons. Our goal was to document the ecological success of the wetlands at Tulula in response to
restored hydrology, soils, and vegetation. Our data should provide NCDOT an ecological assessment
that may be useful for evaluating other wetland restoration projects located throughout the state.
The following objectives provide the framework for a comprehensive ecological assessment of
the restored wetlands of Tulula: 1) determine the success of stream realignment by evaluating the
geomorphology of the new channel before and after water is introduced, 2) following restoration of site
hydrology, evaluate changes in ecosystem structure and function associated with plant community
succession in the floodplain in response to a higher water table and overbank flooding, and 3) evaluate
wildlife use of the site in response to changing hydrologic conditions (amphibians) and plant community
succession (birds).
10
II. RESEARCH METHODS AND RESULTS
Ecological conditions at Tulula have been documented for over ten years by UNCA (see
www.unca.edu/tulula, North Carolina Department of Transportation 1997, Rossell et al. 1999,
Moorhead et al. 2001a, Moorhead et al. 2001b). Ecological success of wetlands restoration at Tulula
has been evaluated by comparing the extensive pre-restoration database to the post-restoration data.
A. Stream Restoration and Hydrology
1. Stream Restoration
A primary focus of restoration at Tulula was to improve site hydrology. A meandering channel
(8,500 linear feet in length) was constructed across the floodplain during the winter of 1999/2000. The
design of the new channel was based partially on the physical characteristics of a relic channel found
primarily at the lower end of the site. The relic channel was used, when practical, as part of the new
meandering channel. The channel was re-constructed in 2001/2002 to correct problems associated
with longitudinal grade. Common streambank erosion techniques, such as fiber matting, coir fiber rolls,
root wads, and live stakes of willow (Salix spp.) and silky dogwood (Cornus amomum), were
installed to improve the short-term stability of the new channel. Four sections of the constructed
channel, in the upper and middle portions of the site, were joined together by crossing the dredged
channel of Tulula Creek in fall 2001. The fifth section was connected in two stages in May (Section V)
and June (Section Va) 2002. The design criteria used to construct the channel are shown in Table 1.
Table 1. Design criteria* for the restored Tulula Creek.
_______________________________________________________
Parameter Proposed Average Value Range
_______________________________________________________
Cross-sectional area 18 ft2 15 – 20 ft2
Bankfull Width 8.5 ft 8 – 10 ft
Average Depth 2.2 ft 1.6 – 2.9 ft
Maximum Depth 3.6 ft 2.2 – 5.3 ft
Width/Depth Ratio 4 3.1 – 6.3
Meander Wavelength 70 – 80 ft 60 – 100 ft
Sinuosity 1.62 1.44-1.93
Arc Length 50 ft 40 – 70 ft
Radius of Curvature 15 ft 10 – 25 ft
Channel Slope 0.0020 0.0017-0.0022
Rosgen Stream Type** E5
_______________________________________________________
*North Carolina Department of Transportation (1997)
**Rosgen (1996)
11
Methods
A primary objective for restoration efforts at Tulula was to determine the success of stream
realignment by evaluating the geomorphology of the new channel before and after water introduction.
Eight random channel segments were chosen in the five stream sections that were restored in
2001/2002. Each segment included four to six riffle-pool sequences varying in length from 120 to 180
ft. Each segment began and ended at the top of a riffle and the origin and end were permanently staked
with PVC pipe and rebar. These two points served as reference to partially describe the channel
geomorphology. A 300-ft measuring tape was secured between the origin pin and the end pin.
Beginning at 0 ft (the origin pin), the orthogonal distance from the tape to the left bank, thalweg, and
right bank was measured every 6 ft on the 300-ft tape. The data were used to develop overall channel
configuration (planview) and to determine sinuosity of channel segments. Data derived from this work
included meander wavelength, arc length, belt width, and the radius of curvature.
In each of the eight segments, two riffles and two pools (defined as the middle of a meander)
were chosen to establish permanent cross-sections. Bankfull width was determined and channel cross-sections
were determined by taking depth measurements every 8 in along a tape that was stretched from
the two bank pins of a riffle or pool at the top of each bank. Bank inclination was determined with a
clinometer. The cross-section data were used to calculate cross-sectional area, average depth,
maximum depth, and the width/depth ratio. Erosion bank pins were installed at the toe or middle of a
channel bank at a few riffle and pool cross-sections. The erosion pins were hammered 21 in into the
bank walls with 3 in exposed in the channel. Pebble counts, using a modified Wolman method (Rosgen
1996), were conducted for each of the eight stream segments although consistent methodology and
results were only available at year 2 of water flow. Pebble counts are used to determine the particle
size distribution of channel materials.
The slope of the water surface was surveyed using standard surveying equipment. A 300-ft
tape was placed in the channel along the thalweg, with a start point in the channel by the origin pin. The
features of each segment (each pool and riffle) were surveyed at the top of the left and right banks and
for the thalweg. The water depth was also noted for the thalweg. The top, middle, and bottom of each
riffle were surveyed as well as the middle of a meander. The distance of these features were noted from
the 300-ft tape lying in the thalweg of the channel. The permanent riffle or pool cross-section pins were
also surveyed. Benchmarks for each segment were chosen by using established NCDOT surveying
points or by placing a nail in a nearby tree (benchmarks were established throughout the Tulula
floodplain by NCDOT during channel construction). Overall slope of the water surface was calculated
by dividing the difference in water surface elevation from the origin to the end of the segment (both
points representing the top of a riffle) by the total stream distance.
The planview was evaluated before water release and after one year of water flow. The
methods used to determine the planview (as described above) are destructive of floodplain vegetation
and annual evaluations are not warranted. The other geomorphic characteristics were evaluated before
water release and after one and two years of water flow. The goal was to evaluate the geomorphology
of the channel annually after the date of water release.
12
Results and Discussion
The restored channel was constructed as five separate sections (Fig. 1). Eight random channel
segments were chosen in the five sections (Fig. 2) to evaluate stream geomorphology over time. Water
release began in Section 1 of the restored channel in September 2001. We placed two segments for
channel evaluation in Section 1, one each in Sections 2 and 3, two in Section 4, and one each in
Sections 5 and 5a. The initial bankfull width and changes in the cross-sectional areas of riffles and pools
of the channel segments are listed in Table 1. There was essentially no change in the bankfull widths
after two years of water flow and therefore, only the initial bankfull widths are reported in Table 1.
Section II Section I
Section III
Section IV
Date of water release:
Section I - September, 2001
Section II and III - October, 2001
Section IV - November, 2001
Section V - May, 2002
Section Va - July 2002
Restored Tulula Creek
Section V
Section Va
Fig. 1. Restored channel sections of Tulula Creek.
I
II Ia
III
IVa
IV
Va
V
Fig. 2. Approximate locations of stream segments used for channel evaluations.
13
As anticipated, riffles typically had lower cross-sectional areas and shorter bankfull widths
compared to pools (Table 2). Although bankfull widths did not change after two years of water flow,
changes in cross-sectional areas were noted for both riffles and pools. The cross-sectional areas of
riffles increased after two years of water flow. Nine of 16 riffles had > 10 % increase in cross-sectional
area after two years of water flow (Table 3). Ten of 16 pools increased in cross-sectional area but six
other pools decreased in cross-sectional area, typically at locations where point bars were forming. The
cross section of a stream changes much more rapidly and frequently in meander bends and, therefore,
there is more variability in pool cross sections than in riffle cross sections (FISRWG, 1998). A visual
representation of riffle and pool cross sections is shown in Appendix A. Changes in cross-sectional
area are often used as an indicator of stream channel stability. Increases in cross-sectional area
represent areas of stream degradation (sediment erosion) while increases indicate aggradation (sediment
deposition) of a stream channel. Changes at Tulula probably represent adjustments of a constructed
channel to various flow regimes over the past two years.
Table 2. Bankfull width (ft) and cross-sectional area (ft2) and of riffles and pools
in eight stream segments.
______________________________________________________________
Bank Full ----------Cross-Sectional Area-------------
Width Initial One Year Two Years
____________________________________________________________________
Segment I
Riffle 1 13.58 20.10 18.80 21.93
Pool 1 15.42 33.27 27.93 24.21
Riffle 2 11.81 14.59 13.99 15.69
Pool 2 15.42 26.71 27.57 28.92
Segment IA
Riffle 1 10.50 13.84 14.42 16.36
Pool 1 10.27 19.07 18.96 19.76
Riffle 2 12.96 19.50 19.86 22.12
Pool 2 12.57 18.94 17.97 18.40
Segment II
Riffle 1 16.34 19.67 20.34 21.93
Pool 1 16.01 30.26 25.03 27.80
Riffle 2 12.80 13.69 14.81 16.36
Pool 2 14.31 20.29 23.35 24.7
Segment III
Riffle 1 13.29 18.55 18.25 20.06
Pool 1 18.87 31.27 30.82 32.99
Riffle 2 16.90 23.89 25.44 24.70
Pool 2 17.88 26.88 21.28 22.49
14
Segment IV
Riffle 1 12.53 16.14 17.15 17.50
Pool 1 14.08 21.35 24.70 23.33
Riffle 2 12.73 18.91 23.34 22.57
Pool 2 14.57 26.38 27.33 27.59
Segment IVa
Riffle 1 12.40 12.22 14.66 15.39
Pool 1 13.58 22.29 19.50 21.15
Riffle 2 15.13 19.22 21.89 21.50
Pool 1 13.52 19.71 19.17 21.74
Segment V
Riffle 1 14.76 17.13 20.51 19.58
Pool 1 16.24 24.08 27.03 24.72
Riffle 2 13.78 15.45 16.70 16.66
Pool 2 16.33 28.32 32.97 33.33
Segment Va
Riffle 1 9.68 15.24 ---- 16.98
Pool 1 11.65 18.14 ---- 19.60
Riffle 2 15.26 18.57 ---- 19.43
Pool 2 10.04 16.68 ---- 18.12
Average
Riffle 1 12.89 16.61 17.73 18.72
Pool 1 14.53 24.97 24.85 24.20
Riffle 2 13.91 17.98 19.43 19.88
Pool 2 14.31 22.99 24.23 24.41
__________________________________________________________________
Table 3. Percent change in cross-sectional area of riffles and pools after
two years of water flow. Numbers in brackets represent a decrease in
cross-sectional area.
_______________________________________________________
Segment Riffle 1 Pool 1 Riffle 2 Pool 2
_______________________________________________________
I 9.1 (27.2) 7.5 8.3
Ia 18.2 3.6 13.5 (2.8)
II 11.5 (8.1) 19.5 22.0
III 8.1 5.5 3.4 (17.1)
IV 8.3 9.3 19.3 4.6
IVa 25.9 (5.1) 11.9 10.3
V 14.3 2.7 7.8 17.7
Va 11.4 8.1 4.6 8.6
Average 13.4 (1.4) 10.9 6.4
_______________________________________________________
15
The average sinuosity of the restored channel was 1.32 (Table 4), compared to the design
sinuosity of 1.62. The slope of the water surface varies for the stream segments and has decreased
over two years in four of seven stream segments (Table 4).
Table 4. Sinuosity and slope of the water surface over time.
________________________________________________________
Segment Sinuosity Initial slope At 1 year At 2 years
________________________________________________________
I 1.23 0.0030 0.0036 ---
Ia 1.22 0.0024 0.0010 0.0006
II 1.26 0.0022 0.0019 0.0018
III 1.43 0.0028 0.0026 0.0016
IV 1.29 0.0044 0.0047 0.0059
IVa 1.22 0.0022 0.0025 beaver
V 1.32 0.0024 0.0014 0.0018
Va 1.58 --- --- ---
Average 1.32 0.0028 0.0025 0.0020
_______________________________________________________
The width/depth (W/D) ratio of riffles was slightly higher than for pools and decreased after two
years of water flow (Table 5). The decrease in W/D was a result of slightly higher average and
maximum depths of the channel with no increase in bankfull width. A W/D ratio of 12 is a high end
value for “E” stream types (Rosgen 1996). The W/D ratio is used to understand the distribution of
energy within a channel. If the W/D ratio increases, the hydraulic stress against the banks also increases
and bank erosion is accelerated (Rosgen 1996).
Table 5. Width/depth (W/D) ratio and maximum depth (ft) of
riffles and pools (represents the average of seven stream segments).
_____________________________________________________
Time Riffle 1 Pool 1 Riffle 2 Pool 2
_____________________________________________________
Initial W/D 11.4 9.2 11.3 10.0
Two Years W/D 10.0 9.6 10.2 9.5
Initial max depth 2.06 2.97 2.21 2.72
Two years 2.74 3.07 2.88 3.24
_____________________________________________________
16
Other physical characteristics of the stream segments suggest that the restored channel
was not as sinuous as designed. This was reflected in the higher meander wavelengths and radius of
curvature and lower belt widths of channel segments (Table 6) as compared with design criteria (Table
1). However, channel configuration has not changed after two years of water flow, suggesting that the
geometry of the restored channel was suitable for the various flow conditions that occur in Tulula Creek.
Table 6. Other physical characteristics of selected meanders in each stream segment.
_________________________________________________________________
Section Meander Arc Belt Radius of
Wavelength (ft) Length (ft) Width (ft) Curvature (ft)
_________________________________________________________________
I 65.6 45.3 42.7 19.4
Ia 68.9 24.3 43.6 10.2
II 95.1 55.8 55.8 23.3
III 98.4 66.3 57.4 21.0
IV 137.8 61.4 77.1 21.3
IVa 75.5 42.7 22.9 24.3
V 75.5 59.1 57.1 22.3
Average 88.3 50.5 50.9 20.3
_________________________________________________________________
The cumulative pebble counts of the eight stream segments are shown in Fig. 3. With the
exception of stream segment Va, 40 to 70 % of the cumulative pebble counts were found in the silt/clay
fraction. Segment Va is the closest representation of the relic channel of Tulula, with minor adjustments
made to small portions of the stream bank during stream re-construction. Roughly 10 % of the pebble
count was silt/clay in this segment. With the addition of sands, 80 to 98 % of the pebble counts were
accounted for in the eight stream segments. The additional stream bed materials consisted of gravel.
17
Particle Size (in)
0.001 0.003 0.005 0.010 0.020 0.040 0.080 0.160 0.2400.310 0.470 0.630 0.940 1.260 1.9002.500
Percent (Cummulative - Finer Than)
0
20
40
60
80
100
Segment I
Segment Ia
Segment II
Segment III
Segment IV
Segment V
Segment Va
Sands Gravels Silt/
Clay
Fig. 3. Cumulative pebble counts of seven stream segments.
Bank inclinations of riffles and pools created for the restored channel were commonly between
20 and 30 degrees (data not shown). Although significant erosion was noted at the bottom of the banks
(toe of the bank slope) of riffles and pools (Table 7), overall bank inclinations did not change
appreciably after two years of water flow because of the lack of erosion in the middle and upper
portions of stream banks. The erosion noted at the bottom of channel banks through erosion control
pins can be used to evaluate the lateral stability of a channel. Several points along the re-constructed
Tulula channel are at risk of instability based on lateral erosion, most notably the riffle/pool sequence of
Section Ia, and to a lesser extent Riffle 2 of Section III and Pool 1 of Section IV. The meander width
ratio (meander belt width divided by bankfull channel width) is another indicator of lateral stability.
Given the lack of changes in meander belt or bankfull width after two years of water flow, the ratio has
not changed, suggesting that the re-constructed channel is fairly stable.
The overall channel configuration has not changed substantially after two years of water flow.
However, changes in channel depth have altered the cross-sectional areas of riffles and pools and
changed the W/D ratio. Desirable features have formed in the channel, most notably point bars on
inside banks of many meanders. Changes in cross section and bank erosion at certain locations suggest
that the channel is still adjusting to the flow regimes of Tulula Creek. Minor adjustments can be made
for areas that appear to have unstable banks or stream bed conditions.
18
Table 7. Erosion of channel banks after two years of water flow,
based on erosion control pins.
_________________________________________________
Segment Feature Location Erosion (inches)
_________________________________________________
I Pool 1 Toe 2.02
I Riffle 2 Toe 5.26
Ia Riffle 1 Toe 11.26
Ia Pool 1 Toe 11.98
Ia Riffle 2 Toe 3.85
II Pool 1 Toe 1.62
II Pool 1 Middle 2.02
II Riffle 2 Toe 0.16
III Riffle 1 Toe 0.40
III Riffle 2 Toe 5.66
IV Riffle 1 Toe 0.40
IV Pool 1 Toe 4.86
IV Pool 1 Middle 0.81
IV Riffle 2 Toe 0.40
__________________________________________________
2. Hydrology
Concurrent with construction of the new channel, drainage ditches were blocked and filled. The
expectation was that re-constructing a meandering channel would decrease water velocity, which, when
coupled with blocked drainage ditches, would raise the level of the water table across the floodplain and
allow for more frequent overbank flooding. One of our objectives was to determine if site restoration
improved the overall site hydrology. Electronic water table wells were installed in July 2000 along
transects that were perpendicular to the new channel (Fig 4). In addition, site hydrology has been
monitored for over ten years with a series of manual water table wells and piezometers (Fig. 5). Many
of the manual wells and all of the piezometers are located in a 4-ha floodplain/fen complex that serves as
a reference area for several UNCA research projects. We have documented seasonal patterns of
water-table elevation and vertical hydraulic gradient in this area and determined the influence of
hillslopes and drought on fen hydrology (Moorhead 2001, Moorhead 2003).
19
A1
A2
A3
B1
B2
B3
B4
D1
D2
E1
E2
E3
E4
F1
F2
F3
X1
G1 G2
I1
H1 H2
H3
B5
A4
A5
D3
C1 D4
C2
Approximate locations of wells (no GPS data).
Electronic Wells at Tulula
Fig 4. Transects and individual electronic wells used to assess site hydrology of the restored stream
channel. See Appendix A for daily water-table levels of wells.
Methods
Both electronic and manual water-table wells were used to determine if the floodplain water
table was higher because of the new channel and blocked drainage ditches. Methods of installation are
described in Moorhead et al. (2001a). The manual wells were read two to four times a month. The
electronic wells were programmed to record the water-table depth on a daily basis. The data for both
types of wells were converted to monthly averages to compare the pre- and post-restoration conditions.
The monthly data were then used to construct hydrographs over a one-year period that coincided with
the release of water in the various stream sections. For example, the months of September through the
following August were used for developing hydrographs for electronic or manual wells in stream section
I (water release in September, 2001). Differences between the average monthly pre- and post-restoration
water-table levels were analyzed with a Student’s t-test in Microsoft Excel.
Results and Discussion
The success of hydrology restoration at Tulula, like many wetland sites, will be determined
primarily by changes in water-table depth. The assumption was that after the channel was restored and
the drainage ditches were plugged, the overall water table of the site would rise. The electronic wells
were also used by NCDOT to determine the success of wetland hydrology as determined by the
Section 404 permitting system of the U.S. Army Corps of Engineers (at least 12 consecutive days of
inundation or saturation during the growing season; North Carolina Department of Transportation,
2003).
20
# #
#
#
# #
#
#
#
#
#
#
#
#
#
#
#
#
#
# #
#
# #
#
#
##
# #
# F1
W1
T11
T10
T9
T8
T7
T6
T5
T4
T14
T13
8C 3C
3F
9I
6I
7F T12
T1
T3
T2
I1
I2
II1
II2
III1
III2 IV1 IV2
Manual Wells at Tulula: Eastern Side of Site
Tulula Fen
Restored Tulula Creek
Old Tulula Creek
#
#
# #
#
# #
#
#
#
#
#
# # #
#
#
#
D20
D60
D125
C-50
C-20
A20
A60
B20
B60
M3 M2
B120
B160
B200
A120
A160
C20
M1
Manual Wells at Tulula: Western Side of Site
Fig. 5. Location of manual wells at Tulula. Wells A160, A120, B200, B160, B120, and D20 were destroyed during site
restoration and were not replaced.
The electronic wells were installed in July 2000 and one or two years of pre-restoration data
were compared to two years of post-restoration data, depending on the date of water release into the
various stream sections. The data from individual wells are organized by stream section. As an
example, the daily water-table graphs of electronic well X1 (in stream section I) and the monthly
averages are shown in Fig.6. A comparison of the pre- and post-restoration monthly averages provides
an easier visual interpretation of changes in water table depth due to restoration. The remaining monthly
averages of water-table graphs of electronic wells are found in Appendix B.
21
A rise in the water table was viewed as an improvement in site hydrology. For example, the
restoration of the stream channel improved the hydrology at X1. In stream section I, the water table
increased in the following electronic wells: H3, G1, G2, and X1 (Appendix B1). It was not as
consistent at H2, and although there appeared to be an overall raise at I1, the restoration of Tulula
Creek and hydrology did not improve site hydrology at I1 to meet the requirements of wetland
hydrology for the permitting process. Water-table graphs from electronic wells in stream section II and
III showed a consistent raise in the water table after restoration for electronic wells E1, E2, E3, and F2
(Appendix B2). There was no consistent water table rise for D-transect electronic wells associated
with stream section IV (Appendix B3). In section V and Va, the water rose after restoration for
electronic wells C1, C2, B1, B3, B4, B5, and A3. However, several of these wells were influenced by
the flooding of the lower end of the site by beaver dams. In particular, C1, C2, B4, B5, and A3 are
located near or in areas of flooded conditions from beaver dams.
Data from some of the manual wells have been collected since 1994 (locations of wells shown
in Fig. 3). Monthly averages of water-table depth were calculated for seven years of pre-restoration
data and two years of post-restoration data. The figures illustrating the pre- and post-restoration water-table
data from individual manual wells are found in Appendix C. There are seven years of pre-restoration
data including three years of drought conditions (July 1998 through fall 2001 (Moorhead
2003). The data from manual wells provide a more comprehensive view of site hydrology, given the
varied conditions of annual precipitation before restoration, given the three drought years and the higher
than average annual precipitation during June 1994 through 1997.
At the eastern side of the site, the depth of the water table of Tulula fen (wells 3C, 3F, 6I, 9I,
7F, and 8C; Appendix C1a) and the floodplain adjacent to it (wells II1 and 2, III1 and 2, IV1 and 2;
Appendix C1b) showed few statistical differences before and after restoration of site hydrology. The
statistical differences were noted more often in summer months, during periods of plant transpiration.
Based on manual wells, the water table of Tulula was improved (higher) for wells located near the
stream channel (F1, T13, T14), with little or no improvement documented for wells located farther from
the channel (T1, T2, T3, T4, T5, T6, T7, T8, T9, T10, T11, T12; Appendix C1c). Collecting water-table
data over the next few years and comparing pre- and post-restoration water-table levels will
provide a more comprehensive view of how site restoration has changed the hydrology of Tulula
wetlands.
22
X 2000/2001 pre
-48
-36
-24
-12
0
12
9/1/2000
10/1/2000
11/1/2000
12/1/2000
1/1/2001
2/1/2001
3/1/2001
4/1/2001
5/1/2001
6/1/2001
7/1/2001
8/1/2001
depth (in)
X 2001/2002 post
-48
-36
-24
-12
0
12
9/1/2001
10/1/2001
11/1/2001
12/1/2001
1/1/2002
2/1/2002
3/1/2002
4/1/2002
5/1/2002
6/1/2002
7/1/2002
8/1/2002
depth (in)
X 2002/2003 post
-48
-36
-24
-12
0
12
9/1/2002
10/1/2002
11/1/2002
12/1/2002
1/1/2003
2/1/2003
3/1/2003
4/1/2003
5/1/2003
6/1/2003
7/1/2003
8/1/2003
depth(in)
monthly avg pre and post X
-48
-36
-24
-12
0
12
S O N D J F M A M J J A
month
depth (in)
1 YR PRE 2 YR POST
Fig. 6. The daily water table and monthly averages for electronic well X1. Statistical differences (P <
0.05) were noted for the monthly averages of all months except June. Depth of “0” represents the
surface of the soil.
23
The main concern of NCDOT will be whether the wetlands of the Tulula floodplain have the
appropriate hydrology to meet permit conditions. The data required for this determination are collected
with the electronic wells and analyzed on a yearly basis (see North Carolina Department of
Transportation, 2003 for examples). A more interesting ecological question is how the overall
hydrology has changed at Tulula with site restoration. The manual wells will provide more information
for this question since they were installed in 1994.
B. Vegetation responses to restoration
One of our objectives for restoring wetlands in the Tulula floodplain has been to monitor the
response of native wetland plant communities. We have been monitoring the community composition of
an intact fen since 1994, and during this funding cycle, we were able to examine the community post-restoration.
We also sought to gain a better understanding of the relationship between wetland plants
and environmental factors such as hydrology. We used Juncus effusus L. (soft rush), which is an easily
recognizable and widespread species in the Tulula floodplain (and elsewhere), as an indicator species to
evaluate the effects of hydrology and restoration on plant growth and reproduction. In a previous seed
bank study at Tulula, Rossell and Wells (1999) reported that Juncus spp. dominated the wetland seed
bank, especially in an early successional area of the fen. Our objectives were to determine whether the
growth and reproduction of Juncus effusus were enhanced by wetland restoration, and how overall
species richness responded to restoration.
1. Plant growth responses to restoration and hydrologic regime
Methods
We used data from groundwater wells to select four sites at Tulula: an undisturbed wet area, a
nearby undisturbed drier area, a restored wet area, and a nearby restored drier area. At each site, we
delineated a 50m x 10m study area in relatively uniform plant communities. Within each of the four study
areas, we established 20, 0.25-m2 quadrats at randomly selected points (using a table of random
numbers). Our only criterion was that all quadrats contained Juncus effusus. If a randomly selected
quadrat did not contain J. effusus, it was rejected, and another random quadrat was selected.
In July 2003, we surveyed the plant associates of Juncus effusus in each study area. All plants
occurring in all 80 quadrats were identified to species, and coverage within the quadrat was visually
estimated. We obtained the Region 2 (southeastern United States) wetland indicator status for each
species by consulting U.S. Fish and Wildlife Service (1996) and U.S. Department of Agriculture
(2001). Wetland indicator status categories describe wetland affinities as follows: obligate wetland
plants (OBL) occur in wetlands >99% of the time, facultative wetland plants (FACW) occur in
wetlands 67-99% of the time, facultative plants (FAC) occur in wetlands 34-66% of the time,
facultative upland plants (FACU) occur in wetlands 1-33% of the time, and upland plants (UPL) occur
in wetlands <1% of the time.
24
In early September 2003, we used shears to harvest all aboveground plant material within each
quadrat. Plant material was placed on tarps, then sorted into four categories: vegetative stems of
Juncus effusus, reproductive stems of Juncus effusus, non-Juncus effusus herbaceous plants, and
woody plants. All plant material was placed into paper bags, air-dried to constant weight in a warm dry
building, and weighed. The numbers of Juncus effusus vegetative and reproductive stems were
counted. All Juncus effusus inflorescences were clipped off of reproductive stems, and weighed
separately.
We performed an analysis of variance (ANOVA) to determine the effects of restoration status
(restored vs. unrestored) and hydrology (wet vs. dry) on the following variables: number and biomass
of vegetative Juncus effusus stems, number and biomass of reproductive Juncus effusus stems,
biomass of Juncus effusus inflorescences, biomass of non-Juncus effusus vegetation, and biomass of
woody vegetation. Statistical Analysis Systems was used for all analyses (SAS 2001).
Results and Discussion
Juncus effusus (a FACW species), although present in all quadrats, never occupied more than
25% of the area of any one quadrat. In half to three-fourths of all quadrats, Juncus effusus occupied
<5% of the area of the quadrat. Clearly, although a consistent presence in all of our quadrats, Juncus
effusus was not a dominant species overall. It had many associate species that were reflected in our
calculations of taxonomic richness.
Taxonomic richness was greatest in the restored dry area (48 taxa), and lowest in the restored
wet area (17 taxa). Richness was intermediate in the unrestored dry (33 taxa) and unrestored wet (37
taxa) areas (Table 8). OBL and FACW species made up the greatest percentage of the flora in the
restored wet area (93.3%), and the smallest percentage in the unrestored dry area (55.5%) (Table 9).
The percentage of OBL and FACW species in the unrestored wet (67.7%) and in the restored dry
areas (60.5%) were similar.
25
Table 8. Taxa and wetland indicator status of plants occurring in 0.25-m2 quadrats
in four study areas at Tulula.
Area
Wetland
Taxon Unrest.
Dry
Unrest.
Wet
Restored
Dry
Restored
Wet
indicator
status
Acalypha rhomboidea Raf. x FAC
Acer rubrum L. X x FAC
Agalinis purpurea (L.) Pennell x OBL
Agrimonia parviflora Ait. X FAC
Agrostis sp. x FACW
Alnus serrulata (Ait.) Willd. X FACW
Ambrosia artemisiifolia L. x X x FACU
Ambrosia trifida L. X FAC
Andropogon virginicus L. x FAC
Apios americana Medicus x X x FACW
Aster novae-angliae L. x X x NA
Aster pilosus Willd. x x NA
Bidens frondosa L. x FACW
Boehmeria cylindrica (L .) Sw. X FACW
Campanula aparinoides Pursh. x OBL
Carex annectens (Bickn.) Bickn. X FACW
Carex debilis Michx. x FACW
Carex festucacea Willd. X FACW
Carex lurida Wahl. x X x X OBL
Carex scoparia Schkuhr ex. Willd. x X x FACW
Carex sp. 1 x NA
Carex sp. 2 X NA
Cassia fasciculata Michx. x NA
Clematis virginiana L. x X x FAC
Cuscuta campestris Yuncker X NA
Cyperus strigosus L. x FACW
Desmodium cuspidatum (Willd.) Loudon x NA
Dicanthelium clandestinum (L.) Gould x X x FACW
Dicanthelium ensifolium x X x NA
Eleocharis obtusa (Willd.) Schultes X OBL
Eleocharis tenuis (Willd.) Schultes X FACW
Epilobium ciliatum Raf. X NA
Erigeron annuus (L.) Pers. x FACU
Erigeron philadelphicus L. x FAC
Eupatorium fistulosum Barratt x FAC
Eupatorium perfoliatum L. x FACW
Galium tinctorium L. X x X FACW
Grass sp.1 x NA
Grass sp.3 x NA
Grass sp.4 x NA
26
Holcus lanatus L. x X FACU
Hypericum mutilum L. x X x X FACW
Impatiens capensis Meerb. x X x X FACW
Juncus acuminatus Michx. X x X OBL
Juncus brevicaudatus (Engelm.)Fern. X OBL
Juncus effusus L. x X x X FACW
Juncus tenuis Willd. x X x FAC
Lespedeza cuneata (Dumont) G.Don x NA
Liriodendron tulipifera L. x FAC
Lobelia puberula Michx. x FACW
Ludwigia alternifolia L. X x OBL
Mimulus ringens L. X x X OBL
Onoclea sensibilis L. X FACW
Osmunda cinnamomea L. X FACW
Oxalis sp. x x UPL
Oxalis stricta L. X UPL
Panicum virgatum L. X FAC
Persicaria hydropiper L. x X x X OBL
Persicaria sagittatum L. x X x X OBL
Persicaria spp. x NA
Potentilla simplex Michx. x x FACU
Prunella vulgaris L. x FAC
Pycnanthemum verticillatum (Michx.)
Pers.
x X UPL
Rhynchospora glomerata (L.) Vahl. x OBL
Rosa palustris Marsh. x X OBL
Rubus argutus Link x X x X FACU
Rubus hispidus L. x FACW
Sagittaria latifolia Willd. X OBL
Sambucus canadensis L. x x FACW
Scirpus expansus Fern. X X OBL
Scirpus polyphyllus Vahl. x OBL
Solidago gigantea Aiton x X x FACW
Solidago rugosa Miller x x FAC
Sparganium americanum Nutt. X OBL
Trifolium campestre Schreb. x NA
Trifolium repens L. x FACU
Vernonia noveboracensis (L.) Michx. x x FAC
Viola sp. x NA
27
Table 9. Contribution of each wetland indicator status (as a percent of all vegetation in 0.25-m2
quadrats) in four study areas at Tulula (for plants with a known indicator status).
__________________________________________________________________
Unrestored area Restored area
Wetland Indicator Status Dry Wet Dry Wet
OBL 18.5 26.5 23.7 60.0
FACW 37.0 41.2 36.8 33.3
FAC 22.2 17.6 23.7 0
FACU 14.8 8.8 13.2 6.7
UPL 7.4 5.9 2.6 0
__________________________________________________________________
The results of our ANOVA showed that vegetative Juncus stems were more numerous
(P<0.0001), as well as heavier (P=0.004) in unrestored areas (Table 10). Similarly, non-Juncus herbs
were heavier in unrestored areas (P=0.002) (Table 11). Neither the number nor the biomass of
reproductive Juncus stems were influenced by restoration (P>0.05).
When water table was considered, vegetative Juncus stems were more numerous (P=0.007) as
well as heavier (P=0.002) in wet areas (Table 12). Reproductive Juncus stems were more numerous
(P=0.015), but not heavier (P=0.064), in wet areas. The biomass of non-Juncus herbs was lower in
wet areas than in dry areas (P<0.0001)(Table 13).
Table 10. Effects of restoration on vegetative growth and reproduction of Juncus effusus. Within
columns, means followed by the same letter do not differ significantly (P>0.05).
_________________________________________________________________
Vegetative Reproductive Juncus
Juncus stems Juncus stems inflorescences
Treatment No. Biomass (g) No. Biomass (g) Biomass (g)
_________________________________________________________________
Unrestored 190.0a 23.4a 15.5a 8.4a 1.6a
Restored 69.0b 13.6b 11.5a 6.2a 1.2a
_________________________________________________________________
28
Table 11. Effects of restoration on biomass of plants occurring with Juncus effusus in 0.25-m2
quadrats. Within columns, means followed by the same letter do not differ significantly (P>0.05).
___________________________________________________
Herbaceous plants
Treatment (non-Juncus ) (g) Woody plants
_____________________________________________________________
Unrestored 50.3b 78.5a
Restored 82.4a 39.7a
___________________________________________________
Table 12. Effects of hydrology on vegetative growth and reproduction of Juncus effusus. Within
columns, means followed by the same letter do not differ significantly (P>0.05).
__________________________________________________________________
Vegetative Reproductive Juncus
Juncus stems Juncus stems inflorescences
Treatment No. Biomass (g) No. Biomass (g) Biomass (g)
__________________________________________________________________
Wet 156.2a 23.8a 19.7a 9.5a 1.8a
Dry 102.8b 13.1b 7.4b 4.3a 0.8a
__________________________________________________________________
In summary, the disturbance that is inherently part of restoration activities clearly benefited the
growth of non-Juncus herbaceous plants, perhaps by opening up the canopy and minimizing
competition for light. In contrast, Juncus effusus was more numerous and heavier in undisturbed areas,
perhaps because it is less competitive than the associated flora. A high water table benefited Juncus
effusus (a FACW species) more than the associated flora, however, and stimulated the production of
reproductive stems, ensuring the continued presence of Juncus effusus in the seed bank over the long
term. Overall, plant taxonomic richness was greatest in restored dry areas, but lowest in restored wet
areas, implying that a high water table inhibited many species and favored the establishment of OBL and
FACW plants.
29
Table 13. Effects of hydrology on biomass of plants occurring with Juncus effusus in 0.25-m2
quadrats. Within columns, means followed by the same letter do not differ significantly (P>0.05).
__________________________________________________
Herbaceous plants
Treatment (non-Juncus ) (g) Woody plants (g)
__________________________________________________
Wet 41.0b 98.5a
Dry 92.1a 36.3a
__________________________________________________
2. Vegetation dynamics in Tulula Fen and adjacent floodplain
To determine the effects of wetland restoration on plant communities in an intact fen at Tulula,
we examined the community composition of open and closed canopy areas of fen and adjacent
disturbed floodplain. The vegetation in these areas was inventoried twice prior to restoration (1994 and
2001). We repeated the inventory of each area in July 2003, in order to evaluate any changes that
might have arisen as a result of the altered hydrology at the site.
Methods
We inventoried vegetation using the protocol established in 1994, and a grid of 120 yd2 plots
that was laid out throughout the fen in 1994. Within this grid, 20 plots were randomly selected in an
area with a closed canopy, and 20 plots in an area with an open canopy. In each 32.8 ft x 32.8 ft plot,
we identified all overstory trees with a DBH > 4 in, and measured its DBH. In nested 13.1 ft x 13.1 ft
plots, we identified all understory trees and shrubs with a DBH of 0.8 – 4.0 in, and measured their
DBH. In nested 3.3 ft x 3.3 ft quadrats, we identified all herbaceous plants and woody seedlings (DBH
< 0.8 in), and visually estimated their percent cover. In an adjacent floodplain that was disturbed by the
golf course developers for the purpose of creating a golf fairway, 6, 65.6 ft x 98.4 ft plots were
established in 1994. Within each of these 6 plots, overstory trees were inventoried in an 59 ft x 59 ft
plot, and understory trees were inventoried in a 23 ft x 23 ft plot (these plot sizes were selected so that
the total area inventoried in the floodplain was consistent with the total area inventoried in each area of
the fen). Within each of the 6 plots, we inventoried herbaceous and woody vegetation in 4, 3.3 ft x 3.3
ft quadrats (N=24).
We obtained the Region 2 (southeastern United States) wetland indicator status for all woody
species by consulting U.S. Fish and Wildlife Service (1996) and U.S. Department of Agriculture
(2001). Importance values (IV’s) were calculated for all overstory and understory woody species, and
for six groups of herbs/woody seedlings (ferns, forbs, grasses, rushes, sedges, and woody seedlings).
For the overstory and understory species, IV’s were calculated based on density, basal area, and
frequency of occurrence. For the six groups of herbs/woody seedlings, IV’s were calculated based on
percent cover and frequency of occurrence.
30
Results and Discussion
During the nine years of this study, the number of red maples in the overstory of the closed
canopy fen increased from 104 in 1994, to 123 in 2003 (Table 14). However, the overall importance
of red maple (Acer rubrum L.) declined, from IV=93 in 1994, to IV=86 in 2003 (Table 15). At the
same time, the importance of white pine (Pinus strobus L.) increased, from IV=3 in 1994, to IV=11 in
2003. Conversely, in the open canopy area of the fen, the importance of red maple increased during
this 9-year period. In 1994 there were no overstory trees (DBH > 4.0 in) in the open canopy area of
the fen. By 2003, 14 overstory-sized red maples were present in this area.
Table 14. Total number of overstory trees of each species in 20, 10x10-m2 plots.
________________________________________________________________________
Closed Fen Open Fen Floodplain
Species 1994 2001 2003 1994 2001 2003 1994 2001 2003
________________________________________________________________________
Acer rubrum L. 104 109 123 - 4 14 - - -
Amelanchier sp. 1 - - - - - - - -
Ilex opaca Ait. 1 1 1 - - - - - -
Malus angustifolia - 2 1 - - - - - -
(Ait.) Michx.
Pinus strobus L. 2 5 11 - - - - - -
Total 108 117 136 - 4 14 - - -
________________________________________________________________________
Table 15. Importance values for overstory trees in 10x10-m2 plots.
______________________________________________________________________________________
Wetland Closed Fen Open Fen Floodplain
Species Status 1994 2001 2003 1994 2001 2003 1994 2001 2003
______________________________________________________________________________________
Acer rubrum FAC 93.2 88.6 86.0 - 100 100 - - -
Amelanchier sp. - 1.7 - - - - - - - -
Ilex opaca FAC 1.7 1.8 1.4 - - - - - -
Malus angustifolia - - 2.1 1.4 - - - - - -
Pinus strobus FACU 3.4 7.5 11.2 - - - - - -
________________________________________________________________________
31
In the understory of the closed canopy fen, neither the total number of stems (Table 16) nor the
importance of any species changed appreciably between 1994 and 2003 (Table 17). However, in the open
canopy fen, the total number of stems of understory-sized trees (DBH 0.8 - 4.0 in) increased dramatically
from 121 stems in 1994, to 234 stems in 2003 (Table 16). Most of this increase was due to the number of
red maple and tag alder (Alnus serrulata (Ait.) Willd.) stems that entered this size class. The overall
importance of red maple declined from IV=89 in 1994, to IV=71 in 2003, primarily because the taxonomic
richness in this area increased from 3 understory-sized species in 1994, to 12 understory-sized species in
2003 (Table 17).
Table 16. Total number of understory trees of each species in 20, 4x4-m2 plots.
____________________________________________________________________________________
Closed Fen Open Fen Floodplain
Species 1994 2001 2003 1994 2001 2003 1994 2001 2003
______________________________________________________________________________________________
Acer rubrum L. 73 77 72 111 157 174 - 6 1
Alnus serrulata (Ait.) Willd. 3 4 1 8 23 33 - 6 15
Amelanchier laevis Wiegand - - - - 2 2 - - -
Aralia spinosa L. - - - - 1 1 - - -
Aronia arbutifolia (L.) Ell. - 1 - - 7 3 - - -
Aronia melanocarpa (Michx.)Ell.
- - - - 5 - - - -
Ilex opaca Ait. - 5 4 - - 1 - - -
Ilex verticillata (L.)Gray 3 5 6 - 1 - - - -
Liriodendron tulipifera L. - - - - 1 1 - - -
Malus angustifolia (Ait.)Michx.
2 1 - - - - - - -
Nyssa sylvatica Marsh. 5 5 6 - 2 4 - - -
Oxydendrum arboreum (L.)DC
1 2 2 - - - - - -
Pinus strobus L. 4 7 7 2 4 1 - - 1
Prunus serotina Ehrhart - - - - 1 1 - - -
Rhus copallina L. - - - - - - - 1 -
Rosa palustris Marsh. - - - - 1 - - - -
Salix sericea Marsh. - - - - 2 12 - - -
Sambucus canadensis L. 1 - - - 4 1 - 2 -
Vaccinium corymbosum L. - 1 1 - - - - - -
Viburnum cassinoides L. 9 6 6 - - - - - -
Total 101 114 105 121 211 234 - 15 17
______________________________________________________________________________________________
In the ground-layer community, the closed canopy fen showed an increasing dominance by ferns
(mostly cinnamon fern, Osmunda cinnamomea L.) over the nine years of this study (IV=32.5 in 1994,
IV=48.5 in 2003) (Tables 18 and 19). The open canopy fen showed a decline in the importance of rushes
(IV=10.2 in 1994, IV= 0 in 2003) and an increase in the importance of woody plants (IV=19.5 in 1994,
IV=29.2 in 2003). Since many rushes flourish in open, sunny areas or those with only partial shade
(Thunhorst 1993), it is likely that the shading created during natural succession at Tulula will largely eliminate
rushes from this area of the fen. Woody plants increased even more in the adjacent disturbed floodplain
(IV=19.3 in 1994, IV=40.2 in 2003).
32
Table 17. Importance values for understory trees in 4x4-m2 plots.
______________________________________________________________________________________
Wetland Closed Fen Open Fen Floodplain
Species Status 1994 2001 2003 1994 2001 2003 1994 2001 2003
________________________________________________________________________________________________
Acer rubrum FAC 69.8 65.3 65.6 89.0 67.9 70.9 - 41.6 13.4
Alnus serrulata FACW 3.4 3.0 1.4 10.6 7.8 9.2 - 34.2 73.3
Amelanchier laevis - - - - - 1.2 1.5 - - -
Aralia spinosa FAC - - - - 1.0 1.4 - - -
Aronia arbutifolia FACW - 1.2 - - 5.5 2.9 - - -
Aronia melanocarpa FAC - - - - 2.6 - - - -
Ilex opaca FAC - 4.2 4.6 - - 1.4 - - -
Ilex verticillata FACW 3.4 4.2 5.3 - 1.0 - - - -
Liriodendron tulipifera FAC - - - - 1.0 1.4 - - -
Malus angustifolia - 1.9 1.2 - - - - - - -
Nyssa sylvatica FAC 5.3 4.2 6.4 - 2.1 4.3 - - -
Oxydendrum arbreum UPL 1.5 2.4 2.9 - - - - - -
Pinus strobus FACU 6.3 8.5 8.0 3.3 3.3 1.4 - - 13.3
Prunus serotina FACU - - - - 1.0 1.4 - - -
Rhus copallina FACU - - - - - - - 9.9 -
Rosa palustris OBL - - - - 1.0 - - - -
Salix sericea OBL - - - - 1.2 3.0 - - -
Sambucus canadensis
FACW 1.5 - - - 3.3 1.4 - 14.3 -
Vaccinium corymbosum
FACW - 1.2 1.4 - - - - - -
Viburnum cassinoides
FACW 6.9 4.6 4.2 - - - - - -
________________________________________________________________________
Table 18. Mean percent cover of each plant type in 1x1-m2 quadrats.
_______________________________________________________________________
Closed Fen Open Fen Floodplain
Plant type 1994 2001 2003 1994 2001 2003 1994 2001 2003
_____________________________________________________________________________________
Fern 21.8 12.2 15.2 8.9 11.0 6.4 0.3 1.8 3.1
Forb 0.9 0.5 0.2 4.8 4.3 2.7 29.7 10.4 14.7
Grass 0.4 0.2 0.1 12.6 4.5 2.8 23.3 3.0 6.5
Rush 0.2 0 0 4.6 0.1 0 6.2 1.0 0.8
Sedge 12.0 6.4 2.9 32.9 21.0 18.8 2.9 20.4 3.1
Woody 19.5 13.6 4.9 15.6 23.3 17.1 14.7 41.8 42.5
____________________________________________________________________________________
The number of dead trees in the closed canopy region of the fen increased somewhat during our study for
understory-sized trees. In 1994, we recorded 7 dead stems in this size class, compared with 13 dead stems in 2001 (we
counted 11 dead stems in 2003, but some of those could have been standing since 2001). Because the site restoration
was not complete by 2001 and the hydrology had not been altered in this part of the floodplain, the most likely
causes of death for these stems are dry conditions at the site during the 1990’s, shading, and/or disease.
33
Table 19. Importance values for plant types in 1x1-m2 quadrats.
____________________________________________________________________________________
Closed Fen Open Fen Floodplain
Plant type 1994 2001 2003 1994 2001 2003 1994 2001 2003
____________________________________________________________________________________
Fern 32.5 33.8 48.5 11.4 14.9 13.6 1.9 3.4 5.6
Forb 7.4 3.8 3.6 11.3 12.1 11.5 29.4 17.0 20.6
Grass 3.6 3.3 2.6 17.2 13.9 13.6 25.3 11.4 14.3
Rush 1.5 0 0 10.2 1.2 0 12.9 7.4 8.3
Sedge 24.1 23.3 18.9 30.4 28.0 32.1 11.2 23.4 11.0
Woody 31.0 35.8 26.5 19.5 29.8 29.2 19.3 37.5 40.2
_______________________________________________________________________
In summary, the changes in the fen reflect what might be expected due to natural succession, but
not to changes in hydrology due to site restoration. The closed canopy fen continues to be dominated
by red maple, although the overstory shows a small increase in white pine. Given the overall lack of
disturbance in recent years, the open fen is reverting to a forested canopy, and is dominated by red
maple. Heliophytic herbaceous plants like rushes are decreasing throughout the fen, while shade-tolerant
herbs such as ferns are increasing.
3. Survival of commercial red maple stock
Methods
During the winter of 1995, we planted 77 red maple seedlings in each of three of the 65.6 ft x
98.4 ft plots in the disturbed floodplain (N=231). We re-inventoried these red maple saplings during
fall 2003, so that we could compare their survival to that determined in several previous (pre-restoration)
years.
Results and Discussion
Survival of the commercial red maple seedlings appears to have declined somewhat during 2003
(Table 20). The number of surviving saplings had been relatively steady from 1995 until 2002 (some of the
discrepancies in the results of each year’s survey are likely due to the fact that there are now thousands of
naturally-regenerating red maple saplings in this floodplain, and it is sometimes difficult to determine whether
a saplings was planted, or has regenerated on its own). Survival during 2003 was 10% less than it has been
since 2000.
It is premature to pinpoint the factor(s) that are responsible for the decreased survival of these
saplings during 2003, but one of the most significant is likely competition from the aggressive growth of
blackberries (Rubus argutus Link) and other tall shrubs in some areas of this floodplain. In these areas, the
planted red maple seedlings have been overtopped by other vegetation. Other factors that have influenced the
survival of the planted saplings over the last few years have included browsing by deer, and the spraying of
herbicides in the vicinity of a large powerline that crosses the floodplain (the herbicides were sprayed by the
local power company, in an effort to control vegetation under the powerline).
34
Table 20. Survival of 231 commercial red maple seedlings planted in Tulula floodplain during winter
1995.
______________________
Year Survival (%)
__________________________
1995 77
1996 71
2000 76
2001 81
2002 76
2003 66
______________________
Despite the reduced survival of planted saplings, naturally-regenerating red maple saplings
continue to flourish in the Tulula floodplain. This trend is documented by Warren et al. (2004), who
conducted comprehensive surveys of red maple regeneration across this floodplain in 1994 and 2001.
They reported that red maple readily colonized wetland habitats, with a post-disturbance recruitment
window lasting at least twice as long as that reported for terrestrial habitats.
C. Effects of Restoration on Decomposition and Soil Microfauna
Decomposition is a primary ecosystem function in the recycling of nutrients (Swift et al. 1979,
Seastedt 1984), and is influenced by factors such as soil nutrients, temperature, composition of plant
material, and composition and activity of soil fauna. Although many studies have examined
decomposition in upland hardwood communities in the southern Appalachians (see Reynolds et al.
2003), and some research has focused on decomposition in cypress-gum wetlands (Battle and Golliday
2001) and playa wetlands in the southern Great Plains (Anderson and Smith 2002), little is known
about decomposition in wetlands of the southern Appalachians.
The vital role of microarthropods in decomposition and nutrient cycling has been long
established (Swift et al. 1979), but research in wetland systems appears to be minimal. Braccia and
Batzer (2001) examined invertebrates associated with woody debris in a southeastern floodplain
wetland, but their study did not include decomposition. Indeed, these authors emphasized that terrestrial
wetland fauna have been overlooked, and they found that non-aquatic (including Acari and Collembola)
rather than aquatic arthropods, were the most significant component of overall community structure. We
conclude, therefore, that the present research, combining decomposition studies with microarthropod
data, is not only useful but ground-breaking. In this section, we report on decomposition and
microarthropod studies conducted in five plant communities at Tulula, and relate these data to soil pH
and organic matter.
35
1. Decomposition
Methods
Six plots, co-located with water table wells, were established in each of five plant community
types at Tulula. Plant communities used were the red maple forest (RM), open (OF) and closed fen
(CF), floodplain(FP), and the former fairway – a disturbed alluvial bottomland forest (DA). Twelve
fiber-glass screen litter bags, 6 x 6” with mesh size of 1/16 “, containing known weights of air-dried
Acer rubrum (red maple) leaves were placed in each plot in a 4 x 3 grid. The fresh-fallen leaves were
collected in October, 2002, and the litter bags placed in the field in January, 2003. Each litterbag was
anchored with a survey flag and lightly covered with surrounding litter. One litterbag was removed from
each plot every other month, beginning in March, 2003 and continuing through May of 2004. Bags
were transported in zip-loc bags to the lab, and the litter content weighed after microarthropod
extraction. Percent mass of the remaining litter was calculated. Tukey’s Studentized Range (HSD)
Test, (SAS version 8), was used for statistical analysis.
Results and Discussion
After 17 months in the field, the percent mass of litter remaining averaged 50% for the RM site
to 54.8% for the FP site (Fig. 7). The percent mass remaining in RM, CF, and OF was not significantly
different, nor was the mass remaining in FP, DA, OF, and CF. However, decomposition was
significantly greater in RM than in DA and FP. This indicates to us that in the least disturbed site, which
has an intact canopy and least disturbed soil, the important process of decomposition occurs most
readily. Therefore, one would assume that the ensuing process of nutrient cycling would also occur
most readily in the least disturbed, RM site. Since there is no significant difference in percent mass
remaining between the FP and the OF, we conclude that the presence of the closed canopy in the RM
site is not as important in determining decomposition rate as the intact soil, although the differences in
moisture may be a factor in decomposition.
2. Litter Microarthropods
Methods
Microarthropods were extracted from litterbags using a modified Tullgren funnel apparatus
(Mallow and Crossley 1984). Litterbags were left on the funnels for 3 to 4 days; the extracted
microarthropods were preserved in 70% ETOH. Microarthropods were sorted under a
stereomicroscope into the following categories: oribatid, prostigmatid, and mesostigmatid mites,
Collembola, and others. Microarthropod abundances were determined as the mean number of
animals/15.43gr litter. Since the abundance values were not normally distributed, the data were
analyzed using a Generalized Linear Model (Proc Genmod SAS version 8e, 2000) (Crawley 1993).
Standard errors in graphs are provided for comparison purposes, but aren’t statistically rigorous
because the data do not conform with the assumptions of normality.
36
Fig. 7. Percent litter remaining in litterbags after 17 months in the field. Plant communities are RM=red
maple, CF=closed fen, OF=open fen, FP=flood plain, and DA=disturbed alluvial bottomland forest.
Each bar is the average of 37 to 48 litterbags. Bars with the same letter are not significantly different;
error bars are ± 1 SE.
Results and Discussion
As expected, microarthropod numbers varied significantly among the three dates analyzed
(Table 21, Fig. 8). Similar seasonal variations have been reported for upland hardwood forests in the
southern Appalachians (Reynolds et al. 2003). We also found differences in litter microarthropod
numbers by site for total microarthropods and all individual taxa counted except for prostigmatida,
which were not abundant enough for statistical analysis. However, due to significant date*site
interactions, the interpretation of significant site differences for mesostigmatida and collembola is unclear
(Table 21).
In all sites, oribatid mites were by far the most common microarthropod (Fig. 8) and they were
most abundant in the RM community, followed by CF. Abundances of oribatids (and total
microarthropods) appear to be significantly lower in OF, FP, and DA. These findings could be related
to the presence of a canopy in RM and CF, protecting litter-dwelling arthropods from extremes in
temperature and from dessication when exposed to solar radiation. However, the low numbers of total
microarthropods in DA and FP, compared to sites with more canopy (RM, CF, and OF), before leaves
are present in March (Fig. 9), indicates that other factors are involved. We posit that soil disturbance,
once again, plays a major role in a critical ecosystem factor – the abundance of litter microarthropods.
P e r c e n t L i t t e r R e m a i n i n g a f t e r 1 7 M o n t h s
P l a n t C o m m u n i t i e s
R M C F O F F P D A
Percent litter remaining
0
10
20
30
40
50
60
A
B B
A B
A B
37
Table 21. Microarthropod responses to date and site. Data analyzed were average numbers of
microarthropods per 15.43 grains of red maple litter from litterbags collected on each of three dates
(March, May, and July of 2003).
_________________________________________________________________________________________
___
Organism Log-Likelihood Terms Chi-square df P
_________________________________________________________________________________________
___
Total 487.56 Date 40.72 2 <0.0001
Site 16.08 4 0.0029
Date*Site 11.31 8 0.1850
Oribatida 400.99 Date 33.20 2 <.0001
Site 14.76 4 0.0052
Date*Site 10.57 8 0.2275
Mesostigmatida 157.29 Date 54.54 2 <0.0001
Site 11.09 4 0.0256
Date*Site 17.17 8 0.0284
Collembola 55.26 Date 9.40 2 0.0091
Site 23.38 4 0.0001
Date*Site 19.25 8 0.0136
_________________________________________________________________________________________
__
38
PlantCommunities
RM CF OF FP DA
0
20
40
60
80
100
120
140
160
180
Total
mesostigs
col
Total
ori
mesostigs
col
oribatids
mesostigs
collembola
Average number of microarthropods for
March, May, and July, 2003 in five plant communities
average # microarths/15.4 gr dry litter Fig. 8. Average number of microarthropods/15.43 gr dry red maple litter for all three collection dates,
March, May and July, 2003. Plant communities are RM=red maple, CF=closed fen, OF=open fen,
FP=flood plain, and DA=disturbed alluvial bottomland forest. Each bar is the average of 15 to 18
litterbags; error bars are ± 1 SE.
39
Average number of total microarthropods
for March, 2003
Plant Communities
RM CF OF FP DA
Average # microarths/15.4 gr dry litter
0
5
10
15
20
25
30
35
Fig. 9. Average number of total microarthropods/15.43 gr dry red maple litter for March, 2003. Plant
communities are RM=red maple, CF=closed fen, OF=open fen, FP=flood plain, and DA=disturbed alluvial
bottomland forest. Each bar is the average of 6 litterbags; error bars are ± 1 SE.
3. Soil Properties
Methods
Eight samples were collected from the top 2” of soil from each litterbag plot with a soil probe in July
2003. Those samples were then composited into one sample/plot, giving 6 samples per plant community.
Percent organic carbon (OC) content was determined by the Walkley-Black method (Nelson and Sommers
1982); pH was measured on a 1:1 slurry of soil:distilled water using a Fisher Accumet pH meter and standard
electrodes. Average values of pH and OC were calculated for each plant community and comparisons among
the 5 sites were done using Tukey’s Studentized Range (HSD) test, SAS version 8e (2000).
Results and Discussion
Average organic carbon varied from 11.79% to 2.8%, and was highest in soils from the open fen
(11.79%), with OC decreasing in this order: closed fen (11.11%) > red maple forest (9.52%) > flood plain
(8.53%) > disturbed alluvial forest (2.80%) (Fig. 10). The significantly lower OC for DA is probably the
result of bulldozing the area for a fairway. Sites with the least disturbance, the fens and red maple forest,
have the highest OC in the soil. Since soil organic matter is known to be strongly influenced by soil fauna
(Coleman and Crossley 1996), these results appear to be correlated with the distribution of microarthropod
abundances, especially for RM and OF (Fig. 8).
40
Fig. 10. Average percent organic carbon for soil from five plant communities: RM=red maple,
CF=closed fen, OF=open fen, FP=flood plain, DA=disturbed alluvial bottomland forest. Each bar is
the average of 48 soil samples. Bars with the same letter are not significantly different; error bars are ±
1 SE.
Average soil pH values ranged from 4.36 to 3.72, with the DA having the highest pH (4.36),
followed by FP (4.2), OF (4.04), CF (3.98), and RM (3.72) (Fig. 11). pH is significantly lower for the
red maple forest.
Organic carbon in soil
Plant Communities
RM CF OF FP DA
% organic carbon
0
2
4
6
8
10
12
14
AB
AB
A
B
C
41
Fig. 11. Average pH for soil from five plant communities: RM=red maple, CF=closed fen, OF=open
fen, FP=flood plain, and DA=disturbed alluvial bottomland forest. Each bar is the average of 48 soil
samples. Bars with the same letter are not significantly different; error bars are ± 1 SE.
Summary of Decomposition and Soil Fauna
We found that the least disturbed plant communities, red maple in particular, have the quickest
decomposition, the greatest amount of litter microarthropods, the most soil organic carbon, and the
lowest soil pH. We conclude that soil characteristics, related to less disturbance, rather than the
presence of a closed canopy, are probably the main influences on decomposition and litter
microarthropods. Therefore, the most intact ecosystems appear to be functioning at the healthiest
levels.
Soil pH
Plant Communities
RM CF OF FP DA
pH
0.0
0.5
1.0
1.5
2.0
2.5
3.0
3.5
4.0
4.5
5.0
A
B BC CD D
42
D. Amphibian Use of Tulula
Introduction
Amphibians are increasingly being used as indicator species in restoration projects for small
freshwater wetlands (e.g., Pechmann et al. 2001) because they are often community dominants, are
sensitive to site hydrology, and can be easily monitored to assess ecosystem function. Amphibians play
key ecological roles in wetlands in the southern Appalachian Mountains, and are the dominant
vertebrate group in standing water habitats at Tulula. Because a major goal of wetlands restoration is to
restore ecosystem integrity (e.g., to create functional ecosystems where all major community elements
are sustained at viable levels), the response of amphibians to site restoration is a useful indicator of
ecosystem function.
Because of their strong reliance on seasonal wetlands for breeding, the reproductive success of
many amphibian species is strongly influenced by hydroperiod (seasonal duration of ponds). The
hydroperiod affects the likelihood of amphibian larvae reaching a minimum developmental stage to
complete metamorphosis. It also influences the distribution and abundance of predators such as fish and
aquatic insects that feed on amphibian eggs and larvae. Short hydroperiods during periods of drought
can result in catastrophic mortality of larvae due to premature pond drying, but also reduce or eliminate
aquatic predators. Long hydroperiods during wet years provide ample time for amphibian larvae to
complete metamorphosis, but may result in heavy mortality from predators such as dragonfly larvae that
prefer semi-permanent ponds.
At the initiation of the study in 1994, the site contained aquatic habitats that varied from highly
ephemeral to permanent ponds. Most natural breeding sites were filled during golf course construction.
During a detailed survey of the site during 1994-1995, we located 155 standing-water habitats that
included 11 permanent ponds that were constructed as golf course obstacles. Permanent ponds
contained predatory fish (bluegills, largemouth bass) and were not used as breeding sites by most
resident amphibians. The remaining 144 sites were fish-free, seasonal habitats that were mostly small,
shallow depressions. These included mud puddles, water-filled tire ruts, test wells for pond sites,
sluggish ditches, and stream cut-offs associated with the channelization of Tulula Creek.
Monitoring of seasonal habitats during 1994-1995 indicated that most breeding sites were of
very low quality because of altered site hydrology associated with stream channelization, ditching, and
the filling of low-lying areas. All species of vernal pond-breeders suffered high larval mortality during
1994 and 1995 because most breeding sites dried prematurely before tadpoles or salamander larvae
could complete their larval stages. Despite heavy rains in late winter and early spring, about 75% of the
breeding sites dried prematurely in 1994 and 60-70% in 1995. These observations indicated a need to
construct larger and deeper ponds to replace natural breeding sites that were destroyed during golf
course construction.
43
Ten vernal ponds were constructed between October 1995 and January 1996 to replace
natural breeding habitats. Depth and contour were manipulated to create seven temporary and three
permanent fish-free ponds that provide suitable habitat for all pond-breeding amphibians at Tulula. At
seven sites small standing water habitats existed prior to the construction of ponds. We selected 10 of
the largest existing breeding sites as reference ponds to compare hydrological, physiochemical, and
biotic characteristics. One reference pond was destroyed in 2001 in conjunction with reconstruction of
the stream channel. Two others did not fill in 2001-2002 due to construction activity, but were
functional in 2003 and 2004.
Thirteen new breeding sites were also created in the fall of 1999 when golf course ponds were
either filled or partially filled to create shallow ponds. Most of these were stream-fed, and now exist as
shallow, permanent sites that contain small fish. In others, fish were eliminated and the sites were
converted into temporary ponds. Sections of the restored stream channel also were temporarily
blocked with check dams to allow channel re-vegetation prior to restoring stream flow. Small pools
formed in the deepest sections of these channel segments and were used as breeding sites by resident
amphibians in 2001. Additional pools were formed in conjunction with stream and site restoration in
2001-2003. In February 2004 the site had over 60 breeding sites (Fig. 12).
Fig. 12. Location of standing water habitats within the study site (spring 2004).
constructed ponds
reference ponds
other breeding sites
44
Methods
The 10 constructed and 10 reference ponds were sampled 3-19 times annually to obtained data on pond pH,
temperature, conductivity, and oxygen saturation. Samples were taken during the day (900-1700 hrs) and all
constructed and reference ponds were sampled haphazardly during the same day. Three subsamples of water were
taken from each pond at approximately equidistant points along the center of the long axis and approximately 10 cm
below the water’s surface. Subsamples were pooled and readings were taken from the pooled sample. Samples were
placed on ice during warm weather and dissolved oxygen was measured in the field < 3 hours after samples were
collected using Corning Check-mate meters. Conductivity and pH were measured using Corning Check-mate and
Corning 430 bench meters, respectively. We used the yearly mean for all seasonal samples in statistical comparisons
of reference and constructed ponds.
Results
Reference ponds were smaller and shallower than constructed ponds, which could influence physiochemical
characteristics. At full capacity, surface areas of reference ponds averaged 888 ft2 (range = 145-2367 ft2) versus 5165
ft2 (range = 2421-9931 ft2) for constructed ponds. Respective values for maximum depths were 13.4 inches (range =
5.1-23.6 inches) and 24.4 inches (range = 15-34 inches). Comparisons of physiochemical characteristics of
constructed and reference ponds from 1996-2004 are in Fig. 13.
Fig. 13. Physiochemical characteristics of reference and constructed ponds. Symbols are annual means based on 3-
19 seasonal samples per year. Vertical bars are 1 SE. Asterisks indicate means that differed significantly within years.
0
5
10
15
20
25
1996 1997 1998 1999 2000 2001 2003 2004
*
* * *
0
10
20
30
40
50
60
70
1996 1997 1998 1999 2000 2001 2002 2003 2004
reference constructed
20
30
40
50
60
70
80
90
100
1996 1997 1998 1999 2000 2001 2002 2003 2004
*
Conductivity (mS/cm) Temperature (C)
Oxygen saturation (%)
4.8
5
5.2
5.4
5.6
5.8
6
6.2
1996 1997 1998 1999 2000 2001 2002 2003 2004
pH
* *
* *
*
*
*
*
*
45
Respective grand means (+ 1 SE) based on annual averages for reference versus constructed
ponds were 5.46 (0.08) versus 5.60 (0.05) for pH, 14.4oC (0.53) versus 17.1oC (0.86) for
temperature, 42.1 (2.09) versus 38.3 (1.60) dS/cm for conductivity, and 58.8 (3.3) versus 77.8 (3.6)
for percent O2 saturation. T-tests (alpha = 0.05) indicate that means for pH differed only in 2002 and
2004, while conductivity did not differ significantly for any year (conductivity: P > 0.19). However,
constructed ponds were significantly warmer in five of seven years and had significantly higher oxygen
saturation levels in all but two years.
2. Use of constructed and reference ponds by amphibians.
Methods
All constructed ponds filled with water before amphibians began breeding in February 1996.
We monitored all constructed and reference ponds annually to determine patterns of use by resident
species. We visited ponds every 1 to 3 weeks between January-August and searched for amplexed
adults, eggs, or larvae. Larvae were collected when conducting open-bottom sampling to estimate
survival (see below) and when ponds were dip-netted periodically during the spring and summer to
sample resident amphibians.
Results
Resident amphibians rapidly colonized constructed ponds that first filled in 1996 (Fig. 14). Eight
species of amphibians bred in the constructed ponds within 1 year of construction and 10 species have
used the ponds through 2004. These are the wood frog, green frog, bullfrog, gray treefrog, spring
peeper, American toad, spotted salamander, red salamander, three-lined salamander, and the red-spotted
newt (Appendix F). The only species unique to constructed ponds was the bullfrog, which
prefers permanent or semipermanent habitats. Reference ponds were also used by 10 species of
amphibians and only one, the two-lined salamander, was unique to reference ponds (breeding in 1 of 10
reference ponds).
Overall, constructed ponds contained a significantly greater number of breeding species (mean
+ 1 SE = 4.21 + 0.24 species) than reference ponds (2.74 + 0.16 species) during the 8-year period
(paired t-test; P = 0.0002). For individual years, the mean number of species per pond was
significantly higher in constructed ponds for five of eight years and approached significance (P < 0.10)
for two other years (Fig. 14). Regression analysis indicates that the mean number of species using
ponds annually did not increase between 1996-2003 (P values for reference and constructed ponds =
0.92 and 0.19, respectively). The latter suggests that constructed ponds quickly reached saturation
levels within one year of construction. A more detailed analysis of pond colonization and community
turnover is in Petranka (2000a).
46
0
1
2
3
4
5
6
1996 1997 1998 1999 2000 2001 2002 2003
reference constructed
*
*
*
* *
Fig. 14. Mean number of species that bred in reference and constructed ponds. Symbols are means
and bars are + 1 SE. Years with asterisks are significantly different.
3. Response of focal species to constructed ponds.
Methods
We selected the spotted salamander (Ambystoma maculatum) and wood frog (Rana
sylvatica) as focal species for monitoring ecosystem function and restoration success. Both species are
widely distributed across the site and are largely restricted to temporary ponds that predominated prior
to golf course construction. These species lay large egg masses that can be accurately counted, and that
serve as an index of the size of the female breeding population.
To obtain estimates of the overall response of the focal species to restoration efforts, we
conducted a complete count of egg masses on the eastern half of the site beginning in 1995. This census
included the 10 constructed ponds, the reference ponds, and all other breeding sites in the eastern
sector.
To estimate relative changes in embryonic and larval survival across years, we estimated the
total population size of hatchlings and larvae nearing metamorphosis in each pond using open-bottomed
samplers. Populations were sampled using 30 gallon galvanized trashcans with bottoms that were
removed with a blowtorch (approximate area of can bottom = 1.2 ft2). When sampling, the can was
pushed into the pond substrate to trap larvae. Repeated sweeps of the can were made with aquarium
nets until no larvae were captured for five consecutive sweeps.
47
Ponds were sampled by walking a zig-zag transect across the entire area of the pond and taking
samples at approximately equidistant points along the transect. The number of samples per pond
increased with pond size and varied from 15-80. If ponds were not at full capacity, then pond surface
area was estimated at the time of sampling based on 3-5 measurements of length and width using a
meter tape. The total population size of hatchlings or larvae nearing metamorphosis was estimated using
data on the mean number of larvae per sample, the surface area of the sampler, and the surface area of
the pond.
We obtained an initial sample of hatchlings within 1-3 weeks after > 95% of the egg masses
were estimated to have hatched in a pond. We intensively dip-netted ponds as larvae approached
metamorphosis, and obtained a final sample immediately after the first metamorphosing larva was
observed in each pond. Criteria used to recognize metamorphosing larvae were the emergence of both
front legs for wood frog tadpoles and the partial or complete reabsorption of gills and dorsal fins for
spotted salamander larvae. We used this estimate as a relative measure of the number of juveniles that
were recruited into the terrestrial population each year.
Changes in adult population size are the most meaningful measure of the response of amphibians
to site restoration efforts. However, a significant time lag in population responses occurs because of the
prolonged juvenile stage. That is, juveniles that metamorphose and leave ponds may not return for 2-4
years as breeding adults. We used total egg mass censuses of the eastern half of the site to measure the
effects of pond construction and site restoration on breeding populations.
Results
The responses of breeding populations of wood frogs and spotted salamanders to pond
construction are shown in Fig. 15. These data exclude two constructed ponds (7X; 10X) that occurred
on the western end of the site and three small reference ponds that were either destroyed (2C) or were
nonfunctional in 2002 (3C; 4C) and 2003 (4C only) due to construction activities. During 1996 (first
year after pond construction and filling), 71% of the resident wood frogs and 59% of spotted
salamanders bred in the constructed ponds. A corresponding decline in breeding effort occurred in the
remaining small depressions, suggesting that many adults abandoned historical breeding sites in favor of
newly constructed ponds.
The percentage of adult wood frogs that bred in constructed ponds between 1996 and 1999
increased slightly. However, adults decreased use of constructed ponds after 1999 and shifted to other
sites. This reflects a progressive increase in the number of ponds on site in association with stream and
final site reconstruction. In contrast, use of constructed ponds by spotted salamanders was similar
across years, perhaps because adults favor larger, deeper ponds for breeding. In 2004, approximately
48% of wood frogs and 44% of spotted salamanders bred in the constructed ponds, while reference
ponds provided breeding habitat for < 8% of the population.
48
Wood Frog
0
20
40
60
80
100
1995 1997 1999 2001 2003
constructed reference other sites
Spotted Salamander
0
20
40
60
80
1995 1997 1999 2001 2003
constructed reference other sites
Fig. 15. Response of female wood frog and spotted salamanders to pond construction. Symbols are
the number of egg masses laid on the eastern half of the site in constructed ponds, reference ponds, and
all remaining breeding sites. Numbers are expressed as a percentage of all masses laid in the eastern
half of the site. ‘Other” includes all sites other than reference and constructed ponds, including sites that
were created during stream channel restoration. Data for 1995 ‘constructed’ are masses laid in
preexisting sites where ponds were constructed.
Fig. 16 shows annual changes in the percentage of ponds that successfully produced juveniles
(upper graphs) and total yearly output of juveniles from constructed and reference ponds (lower
graphs). The percentage of ponds that successfully produced juveniles has declined annual from 60-
100% in 1996 to < 30% in 2003. The estimated output of terrestrial juveniles from constructed ponds
was exceptionally high during 1996 (N = 253,696 wood frogs; 30,831 spotted salamanders), but
progressively declined in later years (e.g., N = 5,819 and 753 in 2003, respectively). A similar trend
occurred in reference ponds. These trends parallel a general decline in the percentage of ponds that
have successfully produced juveniles each year. Nonetheless, a small percentage of ponds on site have
successfully produced juveniles annually, and viable populations of both species occur on site (see
below).
Comparisons of the number of hatchlings and number of larvae surviving to the initiation of
metamorphosis (see Petranka 2003b for details) indicate that the decline in juvenile output was primarily
due to increased larval mortality rather than increased embryonic mortality. Embryonic survival varied
among years, but there was no evidence of catastrophic mortality for any year. In contrast, overall
juvenile production per egg mass declined markedly during the study period for both species and both
sets of ponds. The reduction in juvenile production is attributable to at least three factors: (1) premature
pond drying and/or the failure of ponds to fill seasonally, (2) outbreaks of a pathogen that caused larval
die-offs, and (3) the accumulation of predators in constructed ponds after 1996.
49
Wood Frog
0
20
40
60
80
100
1996 1997 1998 1999 2000 2001 2002 2003
reference constructed
Spotted Salamander
0
20
40
60
80
100
1996 1997 1998 1999 2000 2001 2002 2003
reference constructed
Wood frog
-50
0
50
100
150
200
250
300
1996 1997 1998 1999 2000 2001 2002 2003
reference constructed
Spotted Salamander
-5
0
5
10
15
20
25
30
35
1996 1997 1998 1999 2000 2001 2002 2003
reference constructed
Fig. 16. Estimates of the percentage of ponds that produced juveniles, and total juvenile recruitment
from 10 constructed and 10 reference ponds during 1996-2003. Symbols for upper panels are the
percentage of ponds that produced juveniles annually, whereas those in the lower panels are the
estimated number of larvae surviving to the initiation of metamorphosis (in thousands).
Fig. 17 shows the percentage of ponds that either did not fill or that filled and dried prematurely
between 1996-2003. Constructed ponds filled annually and usually held water sufficiently long to allow
metamorphosis of both species. An exception is 2001 when 20% of ponds dried prematurely, causing
catastrophic mortality.
The more shallow reference ponds tended to progressively deteriorate with respect to
hydroperiod between 1996-2002. During 2002, 43% and 100% of the reference ponds either did not
fill or dried prematurely for Rana and Ambystoma, respectively. This pattern may in part reflect a
regional drought that occurred from the summer 1998 to fall 2002. The proportion of reference ponds
that dried prematurely decreased after 2001-2002 as the drought ended and rainfall increased to
average or above average levels.
50
Wood Frog
-10
0
10
20
30
40
50
60
70
80
1996 1997 1998 1999 2000 2001 2002 2003
reference constructed
Spotted Salamander
-10
10
30
50
70
90
110
1996 1997 1998 1999 2000 2001 2002 2003
reference constructed
Fig. 17. Annual variation in the percentage of constructed and reference ponds that either did not fill or
that dried before larvae could initiate metamorphosis.
Disease is a second factor that contributed strongly to the decrease in juvenile output between
1996-2003. Outbreaks of a disease that caused catastrophic larval mortality were first observed in
1997. Moribund specimens were sent to the National Wildlife Health Center in Madison, Wisconsin,
and detailed histological and molecular studies revealed that the pathogen is an iridovirus (Ranavirus).
Larvae of both the wood frog and spotted salamander are susceptible to Ranavirus infections.
Infected larvae tend to become lethargic, often float at or near the water surface, and develop
characteristic bloody, hemorrhagic patches on the body and fins. Infected larvae are first noticed
seasonally during the mid- to latter half of the larval stage. Catastrophic mortality typically occurs within
1-2 weeks after the first infected individuals are detected. Typically, outbreaks result in 100% mortality
of larvae in a pond.
The extent to which the disease has impacted local populations in reference and constructed
ponds at Tulula is shown in Fig. 18. Diseased animals and die-offs were not observed prior to 1997, at
which time two die-offs occurred in two ponds. The disease rapidly spread to other ponds on site and
has been a major source of larval mortality since 1998. The smaller percentage of reference ponds with
die-offs between 1998-2002 reflects the fact that many reference ponds dried prematurely (e.g., prior
to the time when the disease normally develops
51
Wood Frog
0
10
20
30
40
50
60
70
80
90
100
1996 1997 1998 1999 2000 2001 2002 2003
reference constructed
Spotted Salamander
0
10
20
30
40
50
60
70
1996 1997 1998 1999 2000 2001 2002 2003
reference constructed
Fig. 18. Changes in the percentage of reference and constructed ponds in which catastrophic die-offs
of larvae occurred from Ranavirus infections.
Egg and larval predation was the third significant source of premetamorphic mortality that
contributed to the decline in juvenile output between 1996-2003. In particular, egg predation by green
frog tadpoles on wood frogs (Petranka and Kennedy 1999), and wood frog tadpoles on spotted
salamanders (Petranka et al. 1998) were significant sources of mortality in certain ponds. Odonates and
other predatory aquatic insects accumulated in constructed ponds after 1996 and presumably
contributed to higher larval mortality.
Despite impacts from drought, disease, and predators, populations of both species have not
suffered severe crashes and remain at viable levels (Fig. 19). The size of the wood frog population
declined from 1995-1998, increased dramatically (366%) through 2000, and declined thereafter. The
population has remained relatively stable since 2002. Female wood frogs require 3-4 years to reach
sexual maturity after metamorphosing (Bervin 1982). Thus, the marked increase in population size in
1999 corresponds to when the large output of juveniles in 1996 first returned to breed as adults. The
decline since 2000 presumably reflects the impact of Ranavirus and premature pond drying on the adult
population.
The population of spotted salamanders has not changed as markedly. The size of the breeding
population slowly increased from 1995(N = 1,265 egg masses) to 2004 (N = 1,831 masses). Females
of this species may require 3-5 years to reach sexual maturity (Petranka 1998), so the gradual increase
in breeding population size may reflect recruitment from the relatively large output of juveniles in 1996
and 1997. The decline in 2002 may reflect the impact of Ranavirus outbreaks that began in 1997-
1998. However, in 2004 the population reached the highest level (1,831 masses), indicating that
recruitment has been sufficient to gradually increase population size.
52
wood frog (east of Mason's)
0
500
1000
1500
2000
2500
3000
1995 1997 1999 2001 2003
reference all ponds constructed
spotted salamander
0
500
1000
1500
2000
1995 1997 1999 2001 2003
constructed reference all ponds
Fig. 19. Changes in adult breeding population size on the eastern sector based on annual egg mass
counts in all breeding sites.
4. Altered site hydrology and emerging concerns. The completion of reconstruction activities,
above average precipitation in 2003-2004, and invasions of the site by beavers have increased the
number of habitats with fish. Damming of Tulula Creek by beavers caused spillover into most of the
nearby wetlands that parallel the stream on the west end of the site (Fig. 12). Almost all of these sites
now contain fish and provide little habitat for seasonal pond breeders. Although reference ponds are
too ephemeral to support fish, fish have invaded many of the constructed ponds since 2002 (Fig. 20).
Amphibians that use fish-free habitats have responded by not ovipositing in ponds with fish; however, it
is uncertain whether adults that avoid ponds with fish are successfully breeding in other habitats on site.
0
10
20
30
40
50
60
70
80
90
1995 1996 1997 1998 1999 2000 2001 2002 2003 2004
Fig. 20. Yearly changes in the percentage of the ten constructed ponds that contained fish.
53
Summary
Data collected from 1996-2004 indicate that constructed ponds are of higher quality than
reference ponds based on physiochemical characteristics, seasonal hydroperiod, and use by resident
amphibians. The constructed ponds tended to be warmer and have higher oxygen levels. Since larval
growth is directly proportional to temperature, and high oxygen levels reduce physiological stress,
physiochemical conditions are judged to be superior to those of reference ponds. Amphibians rapidly
colonized the constructed ponds, and the number of species that utilize these as breeding sites averaged
about 50% higher than that of reference ponds.
Reference ponds progressively deteriorated between 1996 and 2002 with respect to seasonal
hydroperiod. In 2002 the majority either did not fill or dried prematurely, resulting in catastrophic
mortality of pond populations. In contrast, the hydroperiod of most constructed ponds appears to be
adequate for most vernal pond breeders. Seven of 10 ponds normally undergo seasonal drying in late
summer or fall when larvae have metamorphosed. However, fish have colonized many since 2002 in
association with above normal rainfall, beaver activity, and completion of the final phase of
reconstruction.
Outbreaks of Ranavirus have dramatically reduced the output of juveniles from both
constructed and reference ponds. Similar outbreaks of this disease have been reported in several areas
of the United States (Daszak et al. 1999) and have resulted in catastrophic die-offs of larvae.
Amphibians often exhibit boom-and-bust recruitment patterns in which juvenile recruitment may be near
zero in some years and high in others (e.g., Gill 1978, Semlitsch et al. 1996). Local populations are
buffered from these effects since the adults may live many years and metapopulation dynamics allow for
some recruitment annually. Thus, years with complete reproductive failure in local ponds may not
necessarily translate to long-term declines of local populations. We have documented high rates of
reproductive failure in most ponds in most years. However, annual recruitment from a small subset of
ponds annually appears to be sufficient to maintain viable adult populations of wood frogs and spotted
salamanders.
Scientists currently know very little about the epidemiology of amphibian Ranavirus. For
example, it is unknown how the virus is spread between ponds, whether a subset of larvae are resistant
to the virus, or whether the infections subside after several years of outbreaks. Preliminary studies that
we have conducted suggest that humans and other vertebrates such as raccoons and birds may play a
role in spreading the disease via movement of contaminated mud or water between local ponds. One
scenario for the Tulula populations is that the severity of die-offs will decline with time as local
populations evolve immunity or as the virus undergoes normal erratic patterns of outbreak. A second is
that the virus will consistently produce annual die-offs in most ponds that do not dry prematurely. If the
proportion of ponds that suffer die-offs increases significantly in the future, then the latter could result in
resident amphibian species undergoing population bottlenecks or even local extinctions.
54
The invasion of beavers (Castor canadensis) and the completion of stream restoration are influencing
site hydrology and the dynamics of amphibian populations at Tulula. Beaver invaded the site shortly before
stream channel construction began and were eliminated through trapping. They have since reinvaded and
have significantly altered the landscape. Fish have become far more abundant on site since 2002 and have
invaded most of the constructed ponds. In general, habitat quality for amphibians that use seasonal wetlands
has declined. Monitoring of focal species in future years will document how amphibians respond to altered
hydrology from stream restoration and beaver activity. It will also help resolve the extent to which Ranavirus
infections ultimately impact breeding populations of amphibians.
D. Bird Use of Tulula
Birds are used as a common indicator for assessing changes in habitat attributes that are associated
with many types of restoration projects (Morrison 1986). Since 1994, we have conducted breeding bird
surveys and measured habitat characteristics of the Tulula floodplain (Rossell et al. 1999, Moorhead et al.
2001). Restoration of Tulula Creek was completed during the summer of 2002. Here we report results of
breeding bird surveys and habitat analyses conducted during 2004. These results are the first year of data
evaluating the response of bird populations to post-restoration habitat changes at Tulula.
1. Bird Surveys
Methods
Breeding bird surveys were conducted from 17 May to 29 May 2004, at 65, 25-m radius plots
located across Tulula floodplain (Fig. 21). Thirty-two plots were separated by at least 100 m. An additional
33 plots were separated by at least 50 m and surveyed because habitat data have been collected at these plots
since 1994 (see Bird-Habitat Relations below). Surveys were conducted from sunrise until 1000 hrs. After a
1-min quiet time, all birds heard or seen within 25 m of the plot center were recorded for 3 min. Birds that
flushed within 25 m of the plot center during the approach also were recorded. Plots were sampled three
times during the survey period. Bird richness was defined as the total number of species, and relative bird
abundance was defined as the total number of individuals of a species.
Results and Discussion
Results of breeding bird surveys are presented in Table 22. In 2004, species richness declined
15% from 2002 levels, with 33 species recorded. American Woodcock, Common Grackle, and
Eastern Wood-pewee were new species recorded during surveys (See Appendix C for complete list of
birds and scientific names). Common Grackle and Eastern Wood-pewee are common in the mountains
of North Carolina (Hamel 1992), and both species were likely breeding on site. American Woodcock
are considered rare in the southern Appalachians, although they have no designated conservation status
(Hamel 1992). The American Woodcock is associated with moist woodland thickets and bottomland
forests that have an abundance of dead leaves on the ground (Hamel 1992). American Woodcock
have been observed in past years using the Tulula floodplain for singing grounds; this species likely
breeds in low numbers throughout the site.
55
Fig. 21. Location of bird survey and habitat plots. S = survey plots, H = habitat plots, and B = survey
and habitat plots.
Relative bird abundance in 2004 decreased 52% from 2002 levels, with a 166 total
observations (Table 22). Song Sparrow and Rufous-sided Towhee continued to be the most abundant
species on site, however, their numbers decreased by almost 50% from 2002 levels. Red-winged
Blackbird also continued to be one of the most abundant species on site, but its numbers held steady
relative to 2002 levels. Many species of conservation concern declined substantially in 2004 (Hamel
1992). The most notable declines included the Golden-winged Warbler, Hooded Warbler, and
Yellow-breasted Chat. Golden-winged Warblers and Yellow-breasted Chats have declined steadily
since 1998. Other species that declined in 2004 included Red-eyed Vireo and White-eyed Vireo.
Brown-headed Cowbirds, which were breeding at Tulula in 2002, were conspicuously absent in 2004.
The declines in species richness and relative bird abundance are likely associated with the large
proportion of the floodplain that was inundated with standing water. Beaver have colonized the western
end of Tulula Creek, constructing a series of dams that flooded much of the interior of the site. The site
was so wet during the spring of 2004 that chest waders had to be worn to conduct surveys. Species
associated with standing water, such as Red-winged Blackbirds and Wood Ducks, have generally
increased in abundance, while species associated with early-successional habitats, including many of the
Neotropical migrants of conservation concern, have generally decreased in abundance.
56
The Golden-winged Warbler is the species of highest conservation concern breeding at Tulula.
This species is federally listed as a species of special concern (LeGrand and Hall 2004). Since 1994,
the Golden-winged Warbler has decreased 94% (31 to 2 birds) in breeding bird surveys at Tulula.
Golden-winged Warblers require a variety of seral stages for breeding, including patches of herbaceous
cover, shrub thickets, and a forested edge (Klaus and Buehler 2000, Rossell 2001, Rossell et. al.
2002). As a result of stream construction and backfilling the old stream channel during the spring of
2002, most of the herb and shrub layers were eliminated from the interior of Tulula. This area
encompassed a substantial portion of many Golden-winged Warbler territories (Rossell et al. 2002). In
2004, additional habitat was lost due to the flooding of the site by beaver.
General observations of Golden-winged Warblers at Tulula indicated that 6-8 territories were
established in 2004. The majority of territories were located along the periphery of the floodplain where
conditions were drier and where there was a large shrub component. Areas with large amounts of
standing water were generally not inhabited by Golden-winged Warblers. Interestingly however, all
Golden-winged Warbler territories established in 2004 contained some standing water.
Table 22. Relative abundance and migratory status of birds recorded during breeding bird surveys in
65, 25-m radius (0.2 ha) plots during 1994, 1998, 2000, 2002, and 2004.
_______________________________________________________________________
1994 1998 2000 2002 2004 Migratory
Species Number Status*
_______________________________________________________________________
Acadian Flycatcher 2 14 3 1 5 N
American Goldfinch 19 13 7 5 2 Y
American Robin 0 1 0 12 1 D
American Woodcock 0 0 0 0 1 D
Belted Kingfisher 0 1 0 0 0 Y
Blue-gray Gnatcatcher 11 13 10 9 11 N
Blue-headed Vireo 0 0 0 1 0 N
Brown-headed Cowbird 0 0 0 2 0 D
Brown Thrasher 1 0 0 4 1 D
Black-and-White Warbler 1 3 1 0 3 N
Blue Jay 0 2 0 0 0 Y
Carolina Chickadee 15 4 7 10 8 Y
Carolina Wren 3 6 3 2 7 Y
Common Yellowthroat 7 1 0 2 5 N
Chestnut-sided Warbler 23 2 7 14 3 N
Cedar Waxwing 9 10 4 9 0 D
Common Grackle 0 0 0 0 1 Y
Downy Woodpecker 6 1 2 3 2 Y
Eastern Phoebe 0 0 0 1 0 D
Eastern Wood-Pewee 0 0 0 0 1 N
57
Golden-winged Warbler 31 21 8 6 2 N
Gray Catbird 4 0 0 0 0 Y
Hooded Warbler 11 21 6 12 4 N
Indigo Bunting 83 55 15 17 13 N
Kentucky Warbler 17 9 9 2 9 N
Mourning Dove 0 2 0 1 0 Y
Northern Bobwhite Quail 0 0 2 7 1 Y
Northern Cardinal 8 3 4 12 5 Y
Northern Flicker 1 0 0 1 0 Y
Northern Parula 17 24 10 26 11 N
Northern Rough-winged Swallow 0 2 0 4 0 N
Ovenbird 2 6 2 5 0 N
Pileated Woodpecker 0 2 1 2 1 Y
Red-eyed Vireo 21 28 28 25 10 N
Ruby-throated Hummingbird 6 5 6 7 3 N
Rufous-sided Towhee 22 24 14 26 15 Y
Red-winged Blackbird 0 0 0 13 12 D
Scarlet Tanager 0 1 1 0 0 N
Song Sparrow 4 11 11 31 16 Y
Swainson’s Warbler 1 4 0 0 0 N
Tufted Titmouse 3 5 8 11 5 Y
White-breasted Nuthatch 1 0 1 1 1 Y
White-eyed Vireo 22 26 29 20 3 N
Wood Duck 0 0 0 1 2 D
Wood Thrush 0 1 0 3 1 N
Yellow-breasted Chat 18 23 12 7 1 N
Yellow-throated Vireo 4 1 3 3 0 N
Yellow-throated Warbler 3 4 1 3 0 N
Yellow Warbler 0 1 0 0 0 N
Total Species 31 36 29 39 33
Total Individuals 378 350 215 321 166
_______________________________________________________________________
*Note: Migratory status from Hamel (1992).
N = Neotropical migrant, D = Short-distance migrant, Y = Year-round resident.
58
2. Bird-Habitat Relations
Methods
Habitat data were collected in 41, 25-m radius (0.2 ha) permanent plots from 7 June to 28 June
2004. Bird-habitat plots were selected in 1994 based on the criterion that they had at least one bird
species recorded in two out of three surveys. Within each plot, herbaceous cover, shrub thickness, and
canopy cover were estimated at 16 regularly spaced points along two perpendicular transects.
Understory (2.5-10 cm dbh) and overstory (> 10 cm dbh) tree densities were also estimated in each
plot using the closest individual method (Bonham 1989). Herbaceous cover was estimated for
vegetation < 0.5 m in height using a 0.25-m2 quadrat. Shrub thickness was estimated for vegetation
0.5-2 m tall using a shrub profile board (Hays et al. 1981). Canopy cover was estimated using a
spherical densiometer (Hays et al. 1981).
Bird richness and relative bird abundance were calculated for each plot. Cedar Waxwings and
American Goldfinches were excluded from the analysis because their flocking behavior tended to inflate
estimates. Correlation analysis was used to examine associations between the habitat variables and bird
richness and relative bird abundance. Analysis of variance (ANOVA) tests were used to compare
differences among years for bird richness, relative bird abundance, and the habitat variables. If a
significant difference was found with ANOVA, then Tukey’s Studentized Range test was used to
determine between year differences.
Results and Discussion
Means of bird richness, relative bird abundance, and habitat variables for the 41 habitat plots
are summarized in Table 23. Both bird richness and relative bird abundance were significantly lower in
2004 than in 2002 (P < 0.05). In 2004, herbaceous cover was significantly greater than in 2002 (P <
0.05), while all other habitat variables were similar between the two years (all P > 0.05). There was a
significant negative correlation between relative bird abundance and overstory tree density (r = -0.14, P
= 0.04). A similar relationship was evident between bird richness and overstory tree density, although
the correlation was not statistically significant (r = -0.12, P = 0.09). All other correlations between bird
richness or relative bird abundance and the habitat variables were extremely low (all Pearson r, between
-0.07 and 0.05; all P > 0.05).
59
Table 23. Means (SD) of bird richness, relative bird abundance, and habitat
variables for 41, 25-m radius (0.2 ha) plots during 1994, 1998, 2000, 2002, and 2004.
___________________________________________________________________________________
Year
Variable 1994 1998 2000 2002 2004
___________________________________________________________________________________
Bird Richness 4.6 (2.1)b 4.0 (1.8)b 2.8 (1.9)a 3.7 (2.2)b 1.8 (1.9)a
Rel. Bird Abund. 6.6 (3.0)a 5.2 (2.8)a 3.4 (2.3)ab 4.4 (2.7)a 2.2 (1.6)b
Herb. Cov. (%) 60.0 (17.5)a 53.9 (20.6)a 52.4 (17.9)a 28.1 (15.6)b 48.5 (18.7)a
Shrub Thick. (%) 35.2 (15.9)ab 28.5 (14.7)b 38.9 (17.7)a 25.9 (16.7)ab 32.6 (12.0)b
Canopy Cov. (%) 59.2 (23.8) 45.4 (21.8) 51.7 (25.0) 45.6 (26.5) 47.4 (26.0)
Understory dens.
(no./0.2 ha) 11.5 (15.3) 6.3 (18.8) 21.7 (27.1) 18.5 (30.2) 22.2 (31.0)
Overstory dens.
(no./0.2 ha) 7.1 (13.9)a 7.6 (13.8)a 10.8 (20.5)ab 8.9 (16.0)ab 21.8 (40.1)b
____________________________________________________________________________________
Note: Values followed by the same or no letters within a row are not significantly different (P > 0.05).
The negative trends in bird richness and relative bird abundance in the habitat plots support the
results of the breeding bird surveys. As discussed in the Results and Discussion of the Bird Survey
section of this report, the declines in bird richness and relative bird abundance are related to loss of
habitat due the large proportion of the site with standing water. In addition, the negative correlations
found between bird richness and relative bird abundance and overstory tree density also help to explain
the declines in species that require early-successional habitats. These declines in early-successional
species are likely to continue as succession proceeds and overstory tree densities increase across the
site.
In 2002, significant reductions in herbaceous cover and shrub thickness reflected high levels of
disturbance of the interior of Tulula that occurred during restoration activities. These habitat changes
were accompanied by significant increases in bird richness and relative bird abundance as a result of
generalist species colonizing the site. Many of the generalist species that experienced large increases in
2002, such as the American Robin, Rufous-sided Towhee, and Song Sparrow, declined dramatically in
2004 as a result of the site being flooded by beaver (Table 22). The significant increase in herbaceous
cover in 2004 compared to 2002 reflects the large increase in areas with standing water colonized by
sedges and rushes. Observations during surveys indicated that few bird species use this rush/sedge
dominated habitat, with the exception of a few blackbirds and wood ducks.
Bird surveys and habitat analyses are scheduled for 2006 to continue monitoring the responses
of bird populations to post-restoration habitat changes. Results reported here indicate that some type of
management is needed at Tulula to maintain the productivity of the habitat for birds (especially the
habitat of the interior of the site). Management objectives should include taking appropriate actions to
eradicate beaver or control the flooding caused by beaver, and maintaining a variety of early-successional
habitat types.
60
DISCUSSION
Tulula continues to change as restoration proceeds and as natural processes respond to
changing site conditions. We have developed a fairly comprehensive understanding of annual and
seasonal variability in the structural and functional attributes of this restoration project.
The overall pattern of the restored stream channel has not changed since water was released in
the first restored section in September 2001. We have noticed isolated areas of bank and bed erosion,
but the channel is performing remarkably well after two years of water flow. Most of the notable a