Ecological assessment of a wetlands mitigation bank |
Previous | 1 of 2 | Next |
|
small (250x250 max)
medium (500x500 max)
Large
Extra Large
large ( > 500x500)
Full Resolution
|
This page
All
|
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
Object Description
Description
Title | Ecological assessment of a wetlands mitigation bank |
Other Title | Final report: ecological assessment of a wetlands mitigation bank |
Contributor |
Moorhead, Kevin K. CTE/NCDOT Joint Environmental Research Program. North Carolina. Department of Transportation. United States. Department of Transportation. |
Date | 2004-08 |
Subjects |
Wetland mitigation banking--North Carolina--Evaluation Wetland conservation--North Carolina Restoration ecology--North Carolina |
Place |
North Carolina, United States |
Description | Phase III: Restoration efforts |
Table Of Contents | Phase I. Baseline ecological conditions and initial restoration efforts -- Phase II. Continued restoration efforts -- Phase III. Restoration efforts -- Phase IV. Post-restoration |
Publisher | CTE/NCDOT Joint Environmental Research Program |
Agency-Current |
North Carolina Department of Transportation |
Rights | State Document see http://digital.ncdcr.gov/u?/p249901coll22,63754 |
Physical Characteristics | v. : ill. ; 28 cm |
Collection |
North Carolina State Documents Collection. State Library of North Carolina |
Type | Text |
Language |
English |
Format |
Reports |
Digital Characteristics-A | 588 KB; 88 p. |
Digital Collection |
North Carolina Digital State Documents Collection |
Digital Format |
application/pdf |
Related Items | CD-ROMs contains PDF version of report.; http://worldcat.org/oclc/156974365/viewonline |
Audience |
All |
Pres File Name-M | pubs_finalreportecological200408.pdf |
Pres Local File Path-M | \Preservation_content\StatePubs\pubs_borndigital\images_master\ |
Full Text | 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 |
OCLC number | 156974365 |