riparian corridor-channel restoration and management in elm creek, minnesota

236 • September 2010 Ecological REstoRation 28:3

Ecological Restoration Vol. 28, No. 3, 2010ISSN 1522-4740 E-ISSN 1543-4079©2010 by the Board of Regents of the University of Wisconsin System.

Restoration NotesRestoration Notes have been a distinguishing feature of Ecological Restoration for more than 25 years. This section is geared toward introducing innovative research, tools, technologies, programs, and ideas, as well as providing short-term research results and updates on ongoing efforts. Please direct submissions and inquiries to the editorial staff ([email protected] and [email protected]).

Microbial Community Indicators of Soil Development in Tropical Secondary Forests (Costa Rica)William D. Eaton (Dept of Biological sciences and the school of Environmental and life sciences, Kean University, 1000 Morris ave, Union, nJ 07083, [email protected]), Emily giles (Dept of Ecology and Evolutionary Biology, University of tennessee, Knoxville, [email protected]) and Dwight Barry (center of Excellence, Peninsula college, Port angeles, Wa [email protected]])

As land management strategies in tropical regions change from intensive agricultural use to reforesta-

tion, the development of secondary forests is becoming a potentially important restoration strategy for increas-ing soil carbon (C) sequestration (e.g., Guo and Gifford 2002). During earlier stages of regeneration, secondary forests in the tropics are thought to have reduced levels of soil organic carbon (SOC), which over time becomes more complex as labile pools of nutrients and organic matter accumulate on the forest floor. This is believed to stimulate complex microbial activities associated with decomposition and mineralization of organic matter (e.g., Feldpausch et al. 2004), followed by an increase of C in microbial biomass (Cmic ) and the ratio of Cmic to SOC, both of which suggest more efficient use of the SOC and more C incorporation into the microbiota (Table 1).

Chazdon and colleagues (2007) discussed the rates of change in vegetation occurring in secondary tropical for-ests. However, little, if any, work has been conducted to identify and compare changes in the belowground microbial community structure, biomass, or potential for C sequestration, how these change when primary forest is converted into pasture, and different age classes of secondary forests. Soil microbial community structure and function are the most rapidly responding biotic com-ponent of any terrestrial habitat, significantly affecting the nutrient cycle components while actually influencing patterns of vegetation distribution. Thus microbial com-munity changes may be the earliest indicators of habitat condition and the rate and direction of recovery following

implementation of ecosystem management, restoration, and conservation strategies in different habitat types (e.g., Anderson 2003, Wardle et al. 2004).

The Northern Zone of Costa Rica has experienced a variety of extraction-based land management activities over the past 30-plus years, resulting in the loss of approximately 70% of its forests. Attempts at remediation have resulted in a variety of strategies, including an extensive array of secondary forests (Schelhas and Sánchez-Azofeifa 2006), providing a unique opportunity to study the effects that such strategies may have on soil structure and function.

We assessed indicators of soil condition based on C cycle metrics (Table 1), as well as microbial activity, biomass, and abundance and diversity, in soils from old growth forests, three age classes of secondary forests, and adjacent pasture within the La Selva Research Station area of Costa Rica (10°26'N, 84°00'W), which is operated by the Organiza-tion for Tropical Studies. Our goals were to determine if soil community complexity changes across these habitats and to identify a suite of parameters that demonstrate trends in soil development and have the potential for use in developing models to predict the recovery rate of tropical forests following disturbance or management.

Sample sites were chosen so that they shared similar topography and soil structure (previously classified as younger alluvial oxisol soils) but had been managed dif-ferently. The secondary forests were previously part of the nearby old growth forest, typical of the area, that had been cleared 30, 25, and 15 years previously and allowed to recover naturally. The pasture was created from the same old growth forest in 1955, but had not been grazed or managed since 1988. Thus we were able to study the four habitat types within close proximity, each area being at least 200 m × 100 m. Soil was randomly collected along four 50 m transects (separated by approximately 20 m) established within each habitat type.

Using the methods reviewed by Anderson (2003), we determined the Cmic , ratio of Cmic to SOC, and microbial metabolic quotients (qCO2, as a ratio of CO2 from res-piration to Cmic ) as indicators of the amount of C being incorporated into the soil as nonplant, noninvertebrate C biomass, efficiency of utilization of SOC, and as possible metrics to determine rates of change and trends in soil development over time following harvesting. The relative percent contribution (RPC) of bacteria, fungi, and the

September 2010 Ecological REstoRation 28:3 • 237

Table 1. Soil metrics, the soil community functions they indicate, examples of supporting references, and the criti-cally different (p < 0.1 and effect size d > 1.0) mean values observed according to habitat type: microbial biomass carbon (Cmic), metabolic quotient (qCO2), Cmic to SOC ratio, relative abundance (RA) of bacterial and fungal rRNA, ratio of fungal to fungal + bacterial rRNA, and relative abundance of laccase genes. Soils were collected from La Selva Research Station in Costa Rica in pasture (P), old-growth primary forest (OG), and secondary forest sites that were originally cleared 15, 25, and 30 years previously (15-y, 25-y, and 30-y, respectively). Habitats that are grouped together did not have critical differences in their mean values.

Metric Function References Habitat DifferencesCmic soil complexity, SOC utilization

efficiency, C sequestration potentialGuo and Gifford (2002), Anderson (2003)

25-y = 30-y = OG > P > 15-y

qCO2 soil complexity, SOC utilization efficiency, C sequestration potential

Guo and Gifford (2002), Anderson (2003)

P = 15-y = 25-y > 30-y = OG

Cmic/SOC soil complexity, SOC utilization efficiency, C sequestration potential

Guo and Gifford (2002), Anderson (2003)

25-y = 30-y = OG > 15-y = P

RA Bacteria younger, disturbed, or recovering soils

Guggenberger and Zech (1999), Anderson (2003)

P > OG > 15-y = 25-y = 30-y

RA Fungi restored, recovered and more complex soils

Guggenberger and Zech (1999), Anderson (2003)

25-y = 30-y = OG > 15-y > P

RA [Fungi/(Fungi + Bacteria)]

fungal dominance, soil complexity, C sequestration potential

Guggenberger and Zech (1999), Anderson (2003

OG > 25-y = 30-y > 15-y > P

RA Laccase lignin decomposition, SOC complexity

Rabinovich et al. (2004) OG > 25-y = 30-y > 15-y > P

laccase gene (from the more specialized basidiomycete fungi) were determined by PCR methods, and the diver-sity of the microbes determined by restriction fragment length polymorphisms (RFLP), all using the methods of Eaton and others (Forthcoming). These were used as indicators of the microbial community composition and complexity, and were examined in connection with the C indicators of soil complexity. We used t-tests to compare mean values between habitat types for each metric and also tested effect size.

The pasture soil microbial community appears to be dominated by bacteria, with fewer fungi than the forest soils, whereas the forested soils become more fungally dominated over time as they progress from young second-ary to old growth forests. Also, the amount of the laccase gene similarly increases with age in the different soils. The qCO2 levels are lower and the Cmic /SOC ratios and Cmic levels higher in more complex and evolved soils (Ander-son 2003). This was the case in the old growth and older secondary forest soils relative to the pasture and younger secondary forest soils (Table 1). We also found that as the forests age, the shift in fungal and laccase gene abundance corresponds to an increase in fungal community complex-ity, as the RFLP-based fungal DNA diversity, richness, and dominance increases. The bacterial community complexity increases with age in the forest habitats, but little difference was observed in bacterial community complexity between the old growth forest and pasture soils (Table 2).

Our six metrics of soil condition demonstrated clear trends of increasing soil microbial community complex-ity, increased SOC use efficiency, and greater potential for C sequestration along a habitat gradient from pasture to

younger secondary forest to older secondary forest to old growth forest that may reflect the overall trend of increasing ecosystem complexity observed in the aboveground vegeta-tion in these habitats. The data suggest that these metrics are valuable tools for monitoring and trend projection of belowground biological and C cycle recovery of managed tropical lands as they transition from early to later stages of forest development.

Now that the management emphasis in this Northern Zone of Costa Rica is on establishment of secondary forests (Schelhas and Sánchez-Azofeifa 2006), it may be possible to regionally monitor soil ecosystem condition in concert with vegetation structure. A potential first step would be to establish indices of biotic integrity (IBI), using below- and aboveground metrics to monitor recovery of these secondary forests. The IBI method is an iterative, multimetric approach for monitoring ecosystem condi-tion that has been used around the world (Karr and Chu 1997). This approach requires the establishment of refer-ence sites and disturbed sites along a disturbance gradient, data collection on indicator characteristics from each site, and the examination of specific target metrics, which are analyzed through an iterative statistical process testing each variable’s performance in detecting change along a known disturbance gradient. The conditions in harvested and regenerating forest areas often provide these types of sites, making them ideal for an attempt at developing an IBI monitoring strategy. We encourage this analytical approach and use of the both the below- and aboveground communities to monitor managed and mitigated lands, and we will be implementing it in the future in Costa Rican forests.


Table 2. Diversity indices from RFLP analysis of PCR products from the amplification of soil DNA using the universal bacterial and fungal rRNA primers. The DNA was extracted from soils collected from forests that had been cleared 15, 25, and 30 years previously, and adja-cent pasture and old growth forest within the La Selva Research Station site in Costa Rica.

Richness H Index EvennessBacterial rRNA RFLPPasture 26 2.98 0.59215 y 23 2.74 0.57125 y 24 2.74 0.5630 y 25 2.98 0.597Old Growth 27 3.02 0.635

Fungal rRNA RFLP Pasture 4 1.31 0.65615 y 4 1.38 0.6525 y 4 1.37 0.68930 y 5 1.56 0.687Old Growth 6 1.61 0.698

AcknowledgmentsThis project was supported by the National Science Foundation REU grants DBI-0452328 and DBI 0453504 and was conducted under the Costa Rican Government Permit #063–2008–SINAC The authors wish to thank Dr. Deedra McClearn (Director of the La Selva Biological Research Station) for her support and encourage-ment. We would also like to thank all the OTS, La Selva Research Station staff for their support. In particular, we wish to thank Bérnal Matarrita, Leslie Ragde Sánchez, Cynthia Rossi, Danilo Brenes, Enrique Castro, Orlando Vargas, Brenda Campbell, Shea Mcdonald, and Melanie Roed. We would also like to thank Patricia Leandro at CATIE for her assistance in laboratory analysis.

ReferencesAnderson, T.H. 2003. Microbial eco-physiological indicators

to assess soil quality. agriculture, Ecosystems & Environment 98:285–293.

Chazdon, R.L., S.G. Letcher, M. van Breugel, M. Martínez-Ramos, F. Bongers and B. Finegan. 2007. Rates of change in tree communities of secondary neotropical forests following major disturbances. Philosophical transactions of the Royal society B 362:273–289.

Eaton, W.D., S. McDonald, M. Roed, K.L. Vandecar, J.B. Hauge and D. Barry. Forthcoming. Differences in nutrient dynamics and microbial community characteristics across seasons and forest in Costa Rica: Potential indicators for effect of decreasing soil moisture. tropical Ecology 52(1).

Feldpausch, T.R., M.A. Rondon, E.C.M. Fernandes, S.J. Riha and E. Wandelli. 2004. Carbon and nutrient accumulation in secondary forests regenerating on pastures in central Amazonia. Ecological applications 14:S164–S176.

Guggenberger, G. and W. Zech. 1999. Soil organic matter composition under primary forest, pasture, and secondary forest succession, Región Huetar Norte, Costa Rica. Forest Ecology and Management 124:93–104.

Guo, L.B. and R.M. Gifford. 2002. Soil carbon stocks and land use change: A meta analysis. global change Biology 8:345–360.

Karr, J.R. and E.W. Chu. 1997. Biological monitoring: Essential foundation for ecological risk assessment. Human and Ecological Risk assessment 3:933–1004.

Rabinovich, M.L., A.V. Bolobova and L.G. Vasil’chenko. 2004. Fungal decomposition of natural aromatic structures and xenobiotics: A review. applied Biochemical Microbiology 40:1–17.

Schelhas, J. and G.A. Sánchez-Azofeifa. 2006. Post-frontier forest change adjacent to Braulio Carrillo National Park, Costa Rica. Human Ecology 34:407–431.

Wardle, D.A., R.D. Bardgett, J.N. Klironomos, H. Setala, W.H. van der Putten and D.H. Wall. 2004. Ecological linkages between aboveground and belowground biota. science 304:1629–1633.

Salt Tolerance of Invasive Phalaris arundinacea Exceeds That of Native Carex stricta (Wisconsin)nick Prasser and Joy B. Zedler (Botany Dept and arbore-tum, University of Wisconsin–Madison, 430 lincoln Dr, Madison, Wi 53706)

Invasive plants that are more salt tolerant than their native competitors would be favored in wetlands that

receive inflows of road salt (the most common being NaCl). Although sodium and chlorine ions are essential in minute quantities for plant growth, high concentra-tions are stressful to non-salt-tolerant plants. While some inland wetlands are naturally saline, wetlands of the Upper Midwestern USA tend to experience saline runoff follow-ing winter application of deicing salt. And since more than 70% of U.S. roads are in snowy regions (FHWA 2009), considerable salt-enriched runoff makes its way into ditches and downstream wetlands.

Salt causes most plants to wilt, which triggers the pro-duction of abscisic acid, a hormone that mediates stomatal closure, thus reducing carbon dioxide intake and photosyn-thesis (Redondo-Gómez et al. 2007). In addition, sodium can replace potassium in leaf tissues, causing necrosis; chlorine and sodium ions can disrupt amino acid bonding in proteins; and affected plants divert energy from growth to export toxic ions from cells. The visible effects of salt exposure include leaf necrosis, decreased productivity, and mortality.

If invasive species are more salt tolerant than natives, road salts could contribute to their establishment and abundance in affected wetlands. Reed canarygrass (Phalaris arundinacea) has been called a model invader (Lavergne and Molofsky 2004). It has replaced the native tussock


sedge (carex stricta) over large areas of sedge meadow (Zedler 2009). Here, we tested the hypothesis that reed canarygrass is more salt tolerant than tussock sedge.

We tested salt tolerance of similar-size individuals by growing them in a University of Wisconsin–Madison greenhouse under seven treatments (watering with 0.0, 0.5, 1.0, 2.0, 4.0, 8.0, and 16.0 ppt NaCl) and assessing four response variables: new biomass production, necrosis severity, mortality, and ash weight (mineral content, which would be high where salt accumulated). We collected propagules of each species from a rural, groundwater-fed wetland with no record of salt influx in the Town of Dunn, Dane County. On October 1, 2008, we took 35 individu-als of reed canarygrass from a shallow coldwater spring near the headwater. We had previously collected seeds of tussock sedge growing in the adjacent sedge meadow and germinated them in May 2008 in a greenhouse to produce 35 individuals.

On October 3, we pruned all 70 individuals to 10 cm height and weighed them wet, after washing and blotting, then planted each in its own 1 L container with moist vermiculite. One replicate per treatment was randomly assigned to each of the five blocks within a 1 × 1.25 m glass enclosure in the greenhouse. We watered the plants weekly; controls received 200 mL of deionized water, and NaCl treatments received 200 mL of solution with either 0.5 g, 1 g, 4 g, 8 g, or 16 g of NaCl per liter of deionized water. On October 3, October 21, and November 11, 2008, we mixed 14.7 g of Miracle-Gro water-soluble all-purpose plant food with 4 L of deionized water and watered all 70 containers equally to provide good growing condi-tions. We illuminated plants with a 430 W high-pressure sodium lamp on a 24-hour schedule. A HOBO data logger recorded temperatures during the treatment period (mini-mum 15.5°C, maximum 31.6°C). The experiment ended on December 15, 2008.

Intermittently (Oct. 27 and 29; Nov. 6, 11, and 18; Dec. 15), we visually rated necrosis using classes that ranged

from 1 (= none to minimal) to 5 (= complete). We pooled classes 1–2 as low levels and 3–5 as high in comparing the two species to create a 2 × 2 contingency table with pooled NaCl treatments (0–4 vs. 8–16 ppt), which we analyzed with the chi-square test. We also recorded the number of plants that died. To determine net biomass production, we rinsed each individual (including dead mass), blotted it, obtained wet weight, dried it at 60°C, and reweighed. We used the wet/dry ratios (65.2% for reed canarygrass and 72.4% for tussock sedge) to estimate initial dry weights. Because of low individual weights, we pooled replicates for biomass and compared results using a paired t-test. We compared ash content for the 0 ppt and 16 ppt NaCl treatments by pooling the dry biomass, grinding in a Wiley mill, combusting in a muffle furnace at 500°C for nine hours, and reweighing.

In our greenhouse experiment, the invasive reed canarygrass experienced necrosis later, produced more new biomass, and had higher survival than the native tussock sedge. For both reed canarygrass and tussock sedge, necrosis increased with greater NaCl levels, but tussock sedge was first to show class 3 necrosis in the 16 ppt NaCl treatment (on Nov. 6 vs. Nov. 11 for reed canarygrass), class 4 necrosis (on Nov. 11 vs. Dec. 15), and complete necrosis (on Dec. 15). The pattern was significant (p < 0.01).

At the beginning of the experiment, the two species had similar pooled dry biomass (4.16 g, reed canarygrass and 4.25 g, tussock sedge). Growth was considerably greater for reed canarygrass at all NaCl concentrations, but decreased with NaCl addition for both species (Figure 1).

Despite experiencing necrosis, only one individual of reed canarygrass died (at 16 ppt NaCl, recorded on Dec. 15), whereas mortality of tussock sedge was first recorded at 0.5 ppt NaCl on November 6 and increased with time and NaCl addition. By December 15, eight plants had died (all 5 at 16 ppt NaCl). We found no change in ash weight of plants watered with 0 vs. 16 ppt NaCl for either species.

The invasive reed canarygrass appears to be more salt tolerant than the native tussock sedge. Several mechanisms of salt tolerance have been identified for plants that occur in naturally saline wetlands. For example, the well-studied saltbush atriplex portulacoides (Redondo-Gómez et al. 2007) keeps its stomata open instead of wilting when exposed to salt. Alternatively, reed canarygrass might deal with toxic ions or their accumulation effectively because of its higher growth rate or by dropping leaves that contain toxic ions. Tussock sedge, by contrast, retains leaves as they senesce from the tip to the base. However, we found no evidence for salt accumulation in tissues of either species (ash content remained low).

The individuals we collected from an unpolluted stream might have been preadapted to salt, since genetic selection occurs rapidly in reed canarygrass (Lavergne and Molofsky 2004). Elsewhere, within a single rangeland, Maeda and others (2006) found genetic selection for increased NaCl

Figure 1. Growth (net new biomass) of reed canarygrass and tussock sedge with increasing NaCl additions.


tolerance by reed canarygrass and that roots of tolerant individuals had increased ATPase and ability to sustain water potential and potassium uptake under NaCl stress.

Greater NaCl tolerance would help explain reed canarygrass abundance in salt-laden roadside ditches and urban wetlands. Where road salts flow into wetlands, invaders could colonize canopy gaps and outcompete stressed native plants. Salt-mediated invasion would not be unique to this species. In Massachusetts, road salt enhanced invasion of a fen by another salt-tolerator, common reed (Phragmites australis) (Richburg et al. 2001). Throughout Wisconsin, common reed, cattails (typha angustifolia, t. × glauca), and reed canarygrass are common in roadside ditches where water and road salt likely accumulate. While their distributions might be due to other disturbances (grading, sedimentation, nutrients), the precautionary principle leads us to suggest that icy streets and highways be kept safe using less toxic alternatives to road salt ( Jull 2009).

ReferencesFederal Highway Administration (FHWA). 2009. Snow and ice.

USDOT FHWA Road Weather Management Program. ops .fhwa.dot.gov/weather/weather_events/snow_ice.htm

Jull, L.G. 2009. Winter salt injury and salt-tolerant landscape plants. University of Wisconsin Extension Publication No. A3877. learningstore.uwex.edu/Assets/pdfs/A3877.pdf

Lavergne, S. and J. Molofsky. 2004. Reed canary grass (Phalaris arundinacea) as a biological model in the study of plant invasions. critical Reviews in Plant sciences 23:415–429.

Maeda, Y., S. Hirano, M. Yoshiba and T. Tadano. 2006. Variations in salt tolerance of reed canarygrass (Phalaris arundinacea L.) plants grown at sites with different degrees of cattle urine contamination. soil science and Plant nutrition 52:13–20.

Redondo-Gómez, S., E. Mateos-Naranjo, A.J. Davy, F. Fernández-Muñoz, E.M. Castellanos et al. 2007. Growth and photosynthetic responses to salinity of the salt-marsh shrub atriplex portulacoides. annals of Botany 100:555–563.

Richburg, J.A., W.A. Patterson and F. Lowenstein. 2001. Effects of road salt and Phragmites australis invasion on the vegetation of a western Massachusetts calcareous lake-basin fen. Wetlands 21:247–255. Abstract available at www.sgnis .org/publicat/richja.htm.

Zedler, J.B. 2009. Feedbacks that might sustain natural, invaded and restored states in herbaceous wetlands. Pages 236–258 in R. Hobbs and K.N. Suding (eds), new Models for Ecosystem Dynamics and Restoration. Washington DC: Island Press.

Riparian Corridor-Channel Restoration and Management in Elm Creek, Minnesotachristian lenhart (Research associate, Dept of Bioproducts and Biosystems Engineering, University of Minnesota, st. Paul Mn, 612/269-8475, [email protected]), Britta suppes (Research assistant, Dept of Forest Resources, Uni-versity of Minnesota), Kenneth Brooks (Professor, Dept of Forest Resources, University of Minnesota) and Joseph Magner (senior Hydrologist, Minnesota Pollution control agency)

The University of Minnesota’s Center for Integrated Natural Resources Management, with funding sup-

port from the Minnesota Pollution Control Agency and Martin County Soil and Water Conservation District, established a multipurpose stream restoration on an impaired reach of Elm Creek, a 700 km2 subwatershed of the Minnesota River Basin in south central Minnesota. The purpose of the project was to demonstrate cost-effec-tive stream restoration techniques within an economically productive agroecosystem to enhance channel stability, reduce sediment loads, and improve aquatic and ripar-ian habitats. Along with its ecological and agricultural benefits, the project site serves as a unique opportunity for public education and outreach, featuring affordable alternatives for stream restoration, agroforestry practices, and cattle grazing management in southern Minnesota.

Elm Creek is a substantial contributor of sediment to the Blue Earth River, the largest supplier of streamflow (46%) and total suspended solids (55%) to the Minnesota River (MPCA 2005). Draining approximately one-fifth of the state of Minnesota, the Minnesota River is of particular concern because it transports a disproportionate amount (88%) of the total sediment entering the Upper Mississippi River Basin (Engstrom et al. 2009). Accelerated sedimenta-tion causes turbidity levels to exceed total maximum daily load (TMDL) standards and is also detrimental to aquatic biota. Engstrom and others (2009) found that the majority of sediment entering the Upper Mississippi River in recent decades originates from stream channel, ravine, and bluff erosion in the Minnesota River Basin (MRB). Increased rates of channel erosion have been caused by recent tile drainage expansion, land cover change, and precipitation increases. Cumulatively, native prairie conversion, extensive annual cropping, wetland drainage, ditch construction, and artificial subsurface drainage have increased stream flow throughout the MRB (Lenhart 2008), with contin-ued soybean and corn expansion and subsurface drainage increases in the past 30–40 years. A 10%–15% increase in annual precipitation in southern Minnesota over the last three decades has likely contributed to increased streamflow as well.

Though less widely recognized, the straightening of streams at road crossings has also exacerbated channel


Table 1. Timeline of restoration activities at the Elm Creek restoration site in the Blue Earth River watershed of Minnesota.

Year Time Milestone Activity2006 Fall Site selection Initial meeting with landowners to discuss project plans and objectives

Summer Baseline dataCollect geomorphic channel data (cross-sections, longitudinal survey,

pebble counts)

Restoration plan Draft plan developed2007 Fall Permit approval Finalized and approved by Minnesota DNR

November ConstructionContractor completes all earthwork and restoration structures during low-

flow conditions

Seeding Seed mixes hand spread and erosion control fabric placed2008 June Planting Willow stakes along 600 m of streambank

July FundingGrant to Martin County SWCD to extend activities directly upstream of

initial project reach

September Construction Contractor completes additional earthwork and restoration structures

during low-flow conditions; bluff stabilized by moving channel away from escarpment

Planting Seed mixes, erosion control fabric, and willow stakes in upstream addition

2009 August Data collectionChannel dimensions resurveyed; bank pins and clay pads emplaced;

monitoring plan developed

2010–2012 Ongoing monitoringChannel resurveys, sediment budget measurements, vegetation surveys,

and aquatic biota assessments

Planting Installation of high-value tree speciesGrazing Low-maintenance rotational grazing plan

instability by decreasing channel sinuosity and instigating channel evolution. Elm Creek has lost approximately 15% of its length since 1938. Increases in slope and streamflow promote channel downcutting and subsequent channel adjustment until a new equilibrium state is reached (e.g., Schumm et al. 1984). However, within the MRB, altered flow regimes and channelization have kept streams in dis-equilibrium and reduced floodplain connectivity.

The watershed-scale disequilibrium in the MRB makes intensive streambank stabilization projects infeasible. While watershed management has helped reduce sedi-ment loads, reductions of in-stream sediment sufficient to meet turbidity standards may also require extensive reduction of sediment from stream erosion, which is not economically viable or sustainable. Therefore, targeting channel reaches exceeding some threshold rate of lateral migration is needed.

Since row-crop agriculture is the major economic driver in the MRB, covering 86% of the watershed, channel stabilization projects must address agroeconomic con-siderations. Multipurpose riparian management systems such as the Elm Creek restoration site are an attractive option because they provide a tool to help meet sediment reduction goals while maintaining agricultural productiv-ity through perennial crop systems and managed grazing. Successful implementation depends on the acceptance by participating landowners, which is strongly influenced by economic incentives.

The Elm Creek watershed consists of flat to rolling topography typical of the glacial Des Moines Lobe till

plain, which contains fine-textured, loamy soils. Histori-cally, land cover was predominantly prairie with numerous small lakes and wetlands, or prairie potholes. The project site was about 11 ha of riparian area located in a broad alluvial valley about 6.5 km from the stream terminus at the Blue Earth River, near Winnebago, Minnesota, and close to a county road for easy public access and high visibility. The Elm Creek site exemplified problems common to streams of southern Minnesota, including channel entrenchment, loss of channel sinuosity, floodplain disconnection, and high rates of bank erosion. At this particular site, channel entrenchment was initiated decades ago, following channel alterations during bridge construction at the county road, in conjunction with basinwide changes in stream flow. Since then, the channel has been actively migrating laterally and downcutting vertically. The outer bend was migrating at a rate of 1.8 meters laterally per year, in excess of typi-cal erosion rates measured in MRB tributaries. Pervasive soil compaction, riparian vegetation loss, and streambank erosion from 30 cattle confined to the riparian corridor contributed to the ecological degradation of the site. In-stream cattle crossings contributed to channel widening and increased deposition of fine sediments.

The goals were to reduce sediment loads by expedit-ing the channel evolution process to a more stable state and to reestablish floodplain functionality. The channel evolution model (CEM) describes the sequence of five channel adjustment stages that occur following a change in watershed equilibrium, based on the concept that a stream adjusts to the new energy forces acting upon it in order to


reach a dynamic equilibrium and maintain channel dimen-sion, pattern, and profile (Schumm et al. 1984). To mimic these CEM equilibrium processes, it was necessary to counteract stream channel widening, facilitate deposition on the active floodplain, revegetate the streambanks and riparian corridor, and introduce a rotational grazing plan.

To increase floodplain connectivity, an oxbow cutoff was reconnected (at high flows only) using a cross vane constructed with locally available rock. Cottonwood logs were used asa cost-saving measure because few on-site rocks were large enough. and To reduce the highest rates of bank erosion, revetment logs were placed along the outer bend into the direction of flow and secured to the bank with duckbill anchors to divert flow away from the bank. Additionally, an actively flooded bankfull “bench” was constructed within the incised portion of the channel at an elevation determined by the 1.5 year flow (a flood that occurs, on average, every one to two years), calculated from data from a stream gauge located near the site and field indicators, such as the occurrence of a flat surface with recently deposited sediment.

Following grading and in-stream work, a native prairie seed mix obtained from a local supplier was hand-spread on the streambanks and covered with a straw-coconut erosion control fabric. Also, willow (salix spp.) cuttings were collected from a nearby stand and planted along 600 meters of streambanks to increase channel rough-ness and enhance stability through root strength. As an agroforestry demonstration project, a planting scheme was developed with landowner input, including perennial grasses and economically valuable timber species such as black walnut ( Juglans nigra), oak (Quercus spp.), and black cherry (Prunus serotina). Finally, a rotational grazing plan was proposed to initially fence cattle out of the restora-tion area and gradually rotate them back into the riparian zone at low densities and short durations to avoid soil compaction and overgrazing.

The timeline for project completion is detailed in Table 1. Initial bank grading, in-stream work, and hand-seeding of plant mixtures were all completed by November 2008, and willow stakes were planted and established along the restoration reach in June 2008. By summer 2008, addi-tional funding was allocated to continue restoration efforts upstream of the initial restoration reach, allowing further construction to commence in September 2008. Restora-tion activities at the Elm Creek site were fully complete by October 2009 at a cost of approximately $30,000, not including salaries.

A postrestoration monitoring plan was created to observe channel development over time. Several permanent cross-section, longitudinal profile, and bed material survey points were established for annual resurvey. Also, erosion pins were emplaced at various locations throughout the

restoration reach to quantify erosion rate over time. Clay pads were constructed on the floodplain bench, the upper terrace, and reconnected backchannel to measure sedi-ment deposition rates during peak flows and to eventually develop a sediment budget for the site. Lastly, biological monitoring of invertebrates and fish species using the Index of Biotic Integrity as well as the surveying of riparian vegetative species will be done in 2010.

In the first year following restoration (2009), the stream stabilization project withstood overbank flow during spring snowmelt. Preliminary findings suggest that reductions in bank erosion have occurred where logs were placed to divert flow from the outer bend. Sediment deposition within the reconnected oxbow was not measurable yet, since no overbank flows have occurred since 2008. Although the landowner was satisfied with the project, particularly the aesthetic and forestry aspects, the farmer renting the property has not yet begun rotational grazing because of the extra labor required. Since most farmers in the region use cattle grazing as a secondary income source, widespread adoption of rotational or lower-density grazing in riparian corridors is limited by their willingness to put in extra hours of work for small financial gains.

Experimentation with grazing management systems using an adaptive management strategy will be proposed to the farmer, based on availability of funding. The adjacent floodplain, which has been placed under a conservation easement, will be planted with native prairie vegetation, and the agroforestry planting will be completed by 2012.

ReferencesEngstrom, D.R., J.E. Almendinger and J.A. Wolin. 2009.

Historical changes in sediment and phosphorus loading to the upper Mississippi River: Mass-balance reconstructions from the sediments of Lake Pepin. Journal of Paleolimnology 41:563–588.

Lenhart, C.F. 2008. The Influence of watershed hydrology and stream geomorphology on turbidity, sediment and nutrients in tributaries of the Blue Earth River, Minnesota, USA. PhD dissertation, University of Minnesota–Twin Cities.

Minnesota Pollution Control Agency (MPCA). 2005. Minnesota River Basin: Watonwan, Blue Earth, and Le Sueur River Watersheds. Minnesota Pollution Control Agency. proteus .pca.state.mn.us/water/basins/mnriver/watershed-blueearth .pdf

Schumm, S.A., M.D. Harvey and C.C. Watson. 1984. incised channels, Morphology, Dynamic, and control. Littleton CO: Water Resources Publications.


The Stillaguamish Big Trees Project: Watershed-Scale Riparian Restoration (Washington)nate Hough-snee (University of Washington Botanic gardens and school of Forest Resources and snohomish county Public Works), Rodney Pond (University of Washington Botanic gar-dens and school of Forest Resources, Box 354115, seattle, Wa 98195-4115, and snohomish county Public Works, surface Water Management M/s 607, 3000 Rockefeller ave, Everett, Wa 98201-4046, 425/388-3464, [email protected]) and Jake Jacobson (snohomish county Public Works)

“a society grows great when old men plant trees whose shade they know they shall never sit in.”

—Greek proverb

Logging and subsequent agricultural and urban devel-opment have left many Puget Sound rivers and ripar-

ian forests heavily degraded and unable to support Pacific salmon (oncorhynchus spp.) at their historic levels of abun-dance. Given that the root cause of Pacific Northwest riparian and in-stream degradation lies in landscape-scale clearcutting of near-stream mature conifer forests, resto-ration of ecological processes for salmon recovery should address deforestation at the watershed scale. While numer-ous entities approach riparian restoration planning from a process-driven, watershed scale, project implementation largely remains at the subbasin and reach scale, emphasiz-ing either in-stream habitat structures or streamside reveg-etation. This short-term, site-specific implementation insufficiently targets landscape-scale habitat degradation such as lethal in-stream temperatures, loss of large woody debris, and sedimentation. In response to this discon-nect, Snohomish County Surface Water Management has undertaken a watershed-scale project to restore riparian conifer forests around the Stillaguamish River, remnant habitat for Chinook salmon (o. tshawytscha).

To understand why watershed-level forest restoration is essential to restore salmon runs, one must consider the relationship between trees, rivers, and fish and how logging has altered the current landscape. Prior to European settle-ment, frequent flooding scoured and deposited material within forested riparian zones, contributing large woody debris (LWD) to active channels and floodplains, providing nurse logs, in-stream habitat, and structural reinforcement to new landforms (Collins et al. 2003). This disturbance resulted in structurally complex and successionally diverse riparian forests. The component old-growth conifers Sitka spruce (Picea sitchensis), Douglas-fir (Pseudotsuga menziesii), western redcedar (Thuja plicata), and western hemlock (tsuga heterophylla) were located along river terraces, and early seral hardwoods such as big-leaf maple (acer mac-rophyllum), black cottonwood (Populus balsamifera), and

red alder (alnus rubra) were near active channels (Collins et al. 2003).

While dam-free rivers still experience flooding, land-scape fragmentation and an absence of postlogging conifer seed sources have allowed deciduous species to dominate floodplain forests. Additionally, the prevalence of shade-tolerant invasive species has led to a novel ecotype familiar to local restorationists: deciduous, even-aged gallery forests with non-native Himalayan blackberry (Rubus discolor) and knotweed (Polygonum cuspidatum, P. sachalinense, P. × bohemicum) dominated understories. When short-lived (ca. 60–100 years), fast-growing alder and cottonwood suc-cumb to competition, flooding, windthrow, or age, black-berry and knotweed rapidly infill canopy gaps, precluding canopy seedling recruitment, thus effectively halting forest succession. In the absence of catastrophic disturbance, the remnant riparian canopy declines and a stalled succes-sional trajectory retrogresses into a stable invasive species community (Figure 1) (Walker and del Moral 2009). This trajectory—widespread conifer loss leading to deciduous canopy dominance and invasive species persistence—fails to regenerate long-lived canopy trees that shade streams, stabilize banks, and provide keystone pieces of large woody debris to the channel. Reduced ecosystem services and habitat value result, a tale told by current salmon return numbers.

To restore conifer-mediated processes that aid salmon recovery at the watershed scale, Snohomish County began process-driven successional management under the banner of the Stillaguamish Big Trees Project (SBTP) in 2007. The Stillaguamish River drains an 181,300 ha water-shed into Puget Sound and supports several salmon runs, including federally endangered Chinook salmon. While the Stillaguamish was once overwhelmingly covered by

Figure 1. An example of the novel hardwood/invasive community type: an early seral red alder/black cottonwood (Alnus rubra/Populus balsam-ifera) riparian forest allows for the development of a dense Himalayan blackberry/knotweed (Rubus discolor/Polygonum spp.) understory that prevents conifer seedling recruitment. A stalled successional trajectory results that is common throughout deforested Puget Sound riparian zones. Photo by Rodney Pond


mature forest, land cover analyses show that within the lower Stillaguamish—the area most impacted by log-ging and agriculture—mature conifer forest comprises only 3%–10% (SIRC 2005). Historic declines in water quality and salmon have been linked to deforestation and land use changes in the watershed (Nehlsen et al. 1991), especially stream temperature, which may drive Chinook survival and reproduction. Chinook spawn as early as late August, when annual stream temperatures peak near the physiological limit for salmon (ca. 18°C), staying on their spawning beds through December when seasonal floods begin. Flood-resilient in-stream LWD and reduced river temperatures via shading, both conifer-mediated processes, are critical components of the SBTP, which designs resto-ration to meet critical salmon habitat requirements. The SBTP works to restore riparian forests on both public and private lands within the North and South Forks, and currently has projects in 16 sites.

The SBTP differs from most regional approaches to riparian forest restoration by managing succession to address both the in-stream and riparian ecological lega-cies of logging at the watershed scale (Figure 1) rather than applying limited resources to small-scale, disconnected rapid revegetation projects. In most projects, low-cost, small bare-root or potted trees and shrubs are planted at high densities following minimal invasive plant treatment to establish a multilayer canopy structure, with an emphasis on species diversity and rapid cover in a single installation. The reasoning is that as diversity increases, so does the likelihood that those species best adapted to site conditions will come to dominate the site quickly. While this approach may provide the desired end result, a vegetated riparian zone, the rapidly establishing hardwoods may suppress concurrently planted, slower growing conifers. The shorter lifetime and stature and rapid decay rate of cottonwood and alder also render them less effective at providing ecosystem functions required by salmon. Additionally, the size of the plantings is small, allowing for reinvasion and suppression by blackberry or knotweed. The SBTP focuses on reintro-ducing larger, longer-lived conifers so that decay-resistant keystone LWD, persistent channel-spanning shade, and mature conifer seed sources will be available over longer ecological time frames.

Using intensive site preparation and larger conifer stock to push succession forward, SBTP intends to complement the shorter-term “diverse portfolio” approach to riparian

restoration. Because these conifers grow more slowly, and the planting stock is less dense than in diverse portfolio restoration, aggressive invasive control with species-appro-priate combinations of brush cutting, mowing, grubbing, and herbicide application must precede the tree instal-lation. Bare-root stock of local genetic origin is secured whenever available and then field grown at the Snohom-ish County native plant nursery or containerized for two to three growing seasons prior to final site installation. Field-grown (4 years and 2 up-transplants) bare-root and containerized (18–38 L) native conifers are planted across a site, with particular attention to keeping plantings at or above active floodplain terraces so that annual floods are less likely to cause mortality. A Washington Conservation Corps crew presently carries out all site preparation, plant production, and planting.

The additional height (1.2–1.8 m) of this larger stock reduces competition with invasives and native shrubs, thus recruiting conifers into the upper canopy more quickly. Sur-vival is also enhanced by the planted trees’ well-developed root systems that are more resilient to summer drought and dormant-season flood events. Planting occurs at reduced densities (Table 1) to accommodate existing stem densities and lower the impacts of site maintenance. To maximize tree survival, plantings are released from competition as needed by herbicide application or mowing of adjacent vegetation.

Restoration project success is carefully measured within the SBTP at the plant and ecosystem scale. The SBTP’s primary goal is to establish riparian conifers over ten river miles (16 km) by supplementing existing stem densities of 0 to 250 stems/ha to between 432 and 503 stems/ha (Table 1) over the next decade. In-stream and air temperatures are concurrently monitored at four stations within each fork, while forest community data are collected within 400 m2 plots across all 16 SBTP sites to assess species com-position, stem density, and invasive species cover; stream cross-sections and LWD are regularly mapped within the watershed as well, so that eventual LWD contributions from the SBTP may be detected. This intensive data col-lection will allow not only for the assessment of plant-based metrics common to restoration, such as tree survival and height and invasive recurrence, but also to make detailed comparisons of vegetation communities and river condi-tions over time. The objectives for SBTP assume that eco-logical functions can be manipulated at the landscape scale

Table 1. Stillaguamish Big Trees Project watershed-level objectives.

MetricBig Trees Goal Achieved Year One Anticipated Year Two

S Fork N Fork Both Forks Both ForksArea Planted (ha) 11 13 4 8River Miles 4.6 5.2 1.0 2.0Stocking Density (trees/ha) 503 432 172 172Landowner Partners 10 10 11 12


and that reintroduction of conifers will initiate autogenic processes for long-term ecosystem resilience. At present, the most important metrics are those that assess tree sur-vival and invasive suppression, but as the trees grow, this priority will shift to the abiotic (light, temperature, LWD contributions) and biotic (seedling regeneration, under-story recruitment) functions that dictate project success.

While the SBTP project is in its inception, it is our hope that it sets a precedent for succession and process-based approaches in future watershed restoration projects within the Puget Sound region. By moving away from an instant gratification approach, where densely planted, small, fast-growing trees show immediate results, we hope to also illustrate that the most difficult functions to restore—LWD sources, long-term stream shade and salmon—take time, patience and a novel approach to planting. Additionally, we anticipate that when SBTP site data are paired with fisher-ies and climate data, future generations of practitioners can create new models for ecological succession and in-stream function, improving the efficacy of watershed restoration activities across Puget Sound.

AcknowledgmentsWe wish to thank the many landowners within the watershed who are working cooperatively with us on this project, volunteer Native Plant Stewards from the Washington Native Plant Society for project monitoring support, Snohomish County’s Washington Conservation Corps crews for site preparation and planting support, the Snohomish County Native Plant program for logistical support, and the Washington State Department of Ecology for funding this project under Centennial Clean Water Act agreements #G0700234 and #G0800397. Additionally, we thank Chris Reyes for valuable comments on early drafts of this manuscript.

ReferencesCollins, B.D., D.R. Montgomery and A.J. Sheikh. 2003.

Reconstructing the historical riverine landscape of the Puget Sound lowland. Pages 79–128 in D.R. Montgomery, S. Bolton, D. Booth and L. Wall (eds), Restoration of Puget sound Rivers. Seattle: University of Washington Press.

Nehlsen, W., J.E. Williams and J.A. Lichatowich. 1991. Pacific salmon at the crossroads: Stocks at risk from California, Oregon, Idaho and Washington. Fisheries 16(2):4–21.

Stillaguamish Implementation Review Committee (SIRC). 2005. Stillaguamish watershed chinook salmon recovery plan. Everett WA: Snohomish County Department of Public Works, Surface Water Management Division.

Walker, L.R. and R. del Moral. 2009. Lessons from primary succession for restoration of severely degraded habitats. applied Vegetation science 12:55–67.

Mowing and Herbicide of Scrub Oaks in Pine Barrens: Baseline Data (New York)Jason t. Bried (albany Pine Bush Preserve commission, 195 new Karner Road, albany, nY 12205-4605, [email protected], 518/456-0655 x1221) and neil a. gifford (albany Pine Bush Preserve commission)

The Albany Pine Bush (APB) Preserve, located in densely populated east-central New York State,

protects a globally rare inland pine barrens community dominated by pitch pine (Pinus rigida) and scrub oaks (Quercus ilicifolia, Q. prinoides). Despite its urban context and small area (1,255 ha), the APB supports many rare and declining shrubland fauna (Barnes 2003, Gifford et al. 2010), including the endangered Karner blue butterfly (lycaeides melissa samuelis). Fire suppression has long been the primary threat to this pyrogenic community through-out the northeastern United States (Finton 1998). With-out frequent fire, the unique shrubland ecosystem rapidly becomes tall dense thicket and eventually forest domi-nated by white pine (Pinus strobus), tree oaks, or invasive hardwoods (Finton 1998, Malcolm et al. 2008). Scrub oak density in major shrubland fragments of the APB is currently twice the desired amount, or about 60%–70%. A primary goal of APB ecological management is to reduce scrub oak density by about half to 30%–35%. Meeting this goal should facilitate the restoration of an open bar-rens where the grasses and forbs essential to the Karner blue and many other shrubland species are codominant with scrub oak and other native shrubs.

Scrub oak regenerates immediately after dormant-season prescribed fire or growing-season prescribed fire + mowing; the latter has been used since 2004 to restore APB shrub-land. Posttreatment growth may decelerate within a few years, but open barrens may have already changed to thicket or begun the transition. An alternative restoration approach is needed that creates a less ephemeral open bar-rens. Combined mowing and herbicide was recommended in utility corridors to reduce thicket cover, promote wild blue lupine (lupinus perennis), and thus increase suitable habitat for New York Karner blue populations (Forrester et al. 2005)—the same may be needed for APB scrub oak patches, as the Karner blue has rarely been observed in these areas.

In 2008, a growing-season mow + herbicide treatment was applied to four scrub oak patches (sites) totaling approximately 45 ha of APB preserve. Here we document first-year treatment effects on vegetation as a baseline for monitoring. No further management is scheduled in these areas for at least three years so as not to interfere with assessment of mow + herbicide effects on plant and animal communities.

Mowing and herbicide work was contracted to two private companies and supervised by APB staff. The four


sites were mowed between 2 July and 6 August 2008. All vegetation was cut to 20–25 cm height using a Hydro-Ax, and debris was left on site. Between 2 and 23 September 2008, herbicide applicators equipped with backpack pump sprayers (ultra-low volume) walked five abreast along adja-cent flagged routes, each spraying two scrub oak crowns (usually single stem of Q. ilicifolia or stem cluster of Q. prinoides) and then skipping one in a repeating pattern without regard to species. The herbicide mixture included Krenite “S” (active ingredient fosamine), which is selec-tive for woody species and inhibits next year’s growth, and Arsenal (active ingredient imazapyr), which is nonselective and prevents amino acid synthesis.

First-year treatment effects were assessed in July 2009, after surviving scrub oak leafed out. Using GIS software (ArcMap v. 9.2, ESRI, Redlands CA) to stratify the treat-ment areas into cells, we established permanent 30 m transects, representing 2.5% randomly selected cells, to measure live scrub oak density and height structure, live versus dead zones of nontarget vegetation, and nectar plant diversity.

Using the point-intercept method, at every meter along the transect live scrub oak was measured in height classes (0–0.5, > 0.5–1.0, > 1.0–2.0, and > 2.0 m). Densities were arcsine–square root transformed and compared among sites and before and after treatment using confidence intervals, ANOVA, and protected Tukey-Kramer comparisons. To assess nontarget effects, observers noted whether each meter point fell within a live or dead zone of vegetation. These mutually exclusive zones were easily distinguished as a brown and green mosaic (Figure 1). Fisher’s Exact Test was used to analyze the total intercepts of live and dead zones for departure from unity.

Important bee and butterfly nectar species, such as northern dewberry (Rubus flagellaris) and dotted horsemint

(Monarda punctata), were documented in eight quad-rats (2 × 2 m) spaced two meters apart along each tran-sect. Nectar species composition and plot frequencies (% quadrats) were analyzed using transects as replicates. The multiresponse permutation procedure (MRPP) was used to compare sites, and the Mantel test (999 iterations) was used for correlations before and after treatment within sites; Bray-Curtis distances were calculated in both tests.

All transects in the two largest treated areas were previ-ously sampled for scrub oak density and nectar diversity in 2006, allowing a pre- and posttreatment comparison. The data reported here provide a baseline to assess rejuve-nation of scrub oak and nontarget vegetation in the next several years.

Scrub oak density was greatly reduced in the two sites with pretreatment data (Table 1); statistical comparison was unnecessary. These sites contained similar pretreat-ment scrub oak densities based on overlapping confidence intervals, but posttreatment scrub oak was significantly sparser in Karner Barrens West than in Kings Road Barrens (Table 1). Depending on the site, 68%–75% of scrub oak was intercepted below 0.5 m height, 12%–24% between 0.5 m and 1.0 m, 5%–11% between 1.0 m and 2.0 m, and 2%–8% above 2.0 m. Significantly more nontarget vegetation was alive than dead in Blueberry Hill, whereas the opposite was found in Karner Barrens West (Table 1). We conclude that scrub oak cover density was reduced overall to 5%–16% (outermost bounds of 2009 confidence ranges), well below the 30%–35% target.

Blueberries (Vaccinium angustifolium, V. pallidum), brambles (Rubus allegheniensis, R. flagellaris, R. idaeus, R. occidentalis), and whorled loosestrife (lysimachia quadri-folia) were clearly the dominant nectar taxa, accounting for 66% to 76% of the total species frequency among sites after treatment. Nectar species composition and abundance dif-fered among two or more sites (MRPP effect size = 0.072, p < 0.001) but was correlated pre- and posttreatment in both Karner Barrens West (Mantel’s r = 0.37, p < 0.001) and Kings Road Barrens (r = 0.40, p < 0.001), indicating no substantial community-level change. The total frequency of nectar species increased from 672 plots before treatment to 838 plots after treatment in Karner Barrens West, and from 662 to 690 in Kings Road Barrens. Although treat-ment appeared most lethal to scrub oak and nontarget vegetation in Karner Barrens West, this site showed greatest posttreatment amounts of whorled loosestrife, arrow-leaf violet (Viola sagittata), goldenrods (solidago spp.), New Jersey tea (ceanothus americanus), and other important bee/butterfly food plants. In addition to favored nectar plants, many annuals characteristic of recent wildland fire, such as American burnweed (Erechtites hieracifolia), were abundant in all treated areas.

The status quo of mow and burn meets many important ecological objectives, such as fuels reduction and exposure of mineral soil, but the thinning of scrub oak canopy is

Figure 1. View of Kings Road Barrens in the Albany Pine Bush Preserve on 13 July 2009, approximately ten months after the herbicide treat-ment that followed mowing to open up the dense scrub oak canopy. Photo by Lee Demick


very brief. Repeated mow and fire treatment becomes costly and logistically challenging, especially in an urban-ized landscape. Herbicide application offers a cost-effective alternative for restoring closed canopy scrub oak thicket to open barrens. If effective, APB managers plan to use herbicide only to initiate the early shrubland stage, not to maintain it. Frequent low-intensity prescribed fire will be the primary tool used to maintain scrub oak around 30%–35% density and less than 2.0 m height. The current overall estimate of 5%–16% density is less than ideal but certain to increase as surviving scrub oak matures over sub-sequent growing seasons. With less than one full growing season, most scrub oak (68%–76%) was less than 0.5 m in height; at maturity, scrub oak is 2–4 m tall. We anticipate that after three years scrub oak density will approximate the desired 30%–35%. If future monitoring reveals persistent low scrub oak density, then herbicide treatments to new treatment areas will target fewer scrub oak crowns.

Many rare butterflies and moths of the region special-ize on scrub oak (Wagner et al. 2003). The inland barrens buckmoth (Hemileuca maia), for example, may require dense scrub oak for growing larvae, oviposition and resting sites, climate mediation, and protection from predators and parasitoids (Haggerty 2006). However, recent find-ings in the APB indicate that overabundant scrub oak increases populations of non-native compsilura concinnata, a parasitic tachinid fly introduced to North America to control gypsy moth (lymantria dispar) and implicated in population declines of many saturniid moth species (Hoven 2009). Closed-canopy scrub oak thickets may also force buckmoth populations toward patch edges or into matrix openings where survival is limited (D. Parry, SUNY Col-lege of Environmental Science and Forestry, pers. comm.). Overgrown thickets may also limit the Karner blue and state-threatened frosted elfin (callophrys irus) (both present in APB), along with their larval host plant, wild blue lupine (e.g., Albanese et al. 2007, Pfitsch and Williams 2009). The newly open canopy appears to have stimulated violet populations, which host the regal fritillary (speyeria idalia), a species currently being considered for reintroduction to New York via the APB.

Many vertebrate taxa should also benefit from the reduced-thicket structure, including over two dozen locally or regionally rare amphibian, reptile, and bird species documented in the APB (Barnes 2003, Gifford et al. 2010). Rare species, such as the eastern hognose snake (Heterodon platirhinos) and prairie warbler (Dendroica discolor), were observed in the study sites before treatment and again in 2009, one year after treatment. Although too soon to tell, we think the inherent trade-offs of mow and herbicide restoration will mostly favor shrub-dependent wildlife.

AcknowledgmentsRestoration and monitoring work was supported by the New York State Department of Environmental Conservation. Thanks to Joel Hecht and Jesse Hoffman, who supervised the restoration treat-ments, and to vegetation monitors Keith Bergreen, Lee Demick, Jacob Humm, and Lindsey Kolar. Chris Reyes, Dylan Parry, and Lee Demick provided helpful comments on the article.

ReferencesAlbanese, G., P.D. Vickery and P.R. Sievert. 2007. Habitat

characteristics of adult frosted elfins (callophrys irus) in sandplain communities of southeastern Massachusetts, USA. Biological conservation 136:53–64.

Barnes, J.K. 2003. Natural history of the Albany Pine Bush: Albany and Schenectady Counties, New York. New York State Museum Bulletin No. 502.

Finton, A.D. 1998. Succession and plant community development in pitch pine–scrub oak barrens of the glaciated northeastern United States. MS thesis, University of Massachusetts.

Forrester, J.A., D.J. Leopold and S.D. Hafner. 2005. Maintaining critical habitat in a heavily managed landscape: Effects of power line corridor management on Karner blue butterfly (lycaeides melissa samuelis) habitat. Restoration Ecology 13:488–498.

Gifford, N.A., J.M. Deppen and J.T. Bried. 2010. Importance of an urban pine barrens for the conservation of early-successional shrubland birds. landscape and Urban Planning 94:54–62.

Haggerty, S.A. 2006. Land management implications for Hemileuca maia (Lepidoptera: Saturniidae) habitat at Manuel F. Correllus State Forest, Martha’s Vineyard, Massachusetts. MS thesis, University of Massachusetts.

Table 1. Baseline data for mow and herbicide treatment in major shrubland areas of the Albany Pine Bush Preserve, New York State, USA. Mean percentage of live to dead vegetation zones is out of 30 point-intercepts per transect (* p = 0.015, ** p = 0.006). Mean cover (95% CI) of live scrub oak (Quercus ilicifolia, Q. prinoides) was measured three growing seasons before (2006) and one growing season after (2009) treatment; values with different superscripts are significantly different (F = 3.52, df = 3,120, p = 0.017).

Area (ha) Live : Dead (%)Scrub oak density

Site 2006 2009Blueberry Hill 5.2 58 : 42** no survey 0.10 (0.07–0.14)AB

Karner Barrens West 18.0 45 : 55* 0.69 (0.62–0.74) 0.07 (0.05–0.10)B

Kings Road Barrens 16.5 52 : 48 0.62 (0.57–0.67) 0.13 (0.10–0.16)A

Madison Ave Pinelands 5.2 47 : 53 no survey 0.10 (0.07–0.14)AB


Hoven, B.M. 2009. The effect of restoration and maintenance treatments on host plant quality and natural enemies of the endemic barrens buckmoth, Hemileuca maia Drury (Lepidoptera: Saturniidae) in a scrub oak-pitch pine ecosystem. MS thesis, State University of New York at Syracuse.

Malcolm, G.M., D.S. Bush and S.K. Rice. 2008. Soil nitrogen conditions approach preinvasion levels following restoration of nitrogen-fixing black locust (Robinia pseudoacacia) stands in a pine-oak ecosystem. Restoration Ecology 16:70–78.

Pfitsch, W.A. and E.H. Williams. 2009. Habitat restoration for lupine and specialist butterflies. Restoration Ecology 17:226–233.

Wagner, D.L., M.W. Nelson and D.F. Schweitzer. 2003. Shrubland lepidoptera of southern New England and southeastern New York: Ecology, conservation, and management. Forest Ecology and Management 185:95–112.

Changes in a Restored Wetland during 18 Years of Management (Ohio)Denis conover (Dept of Biological sciences Ml 0006, Univer-sity of cincinnati, cincinnati, oH 45221-0006, 513/556-0716, [email protected]) and John Klein (Hamilton county Park District, 10245 Winton Rd, cincinnati, oH 45231, 513/521-7275, [email protected])

In 1991, the Hamilton County Park District began res-toration of a wetland area at Miami Whitewater Forest

west of New Haven, where the Whitewater Shaker Com-munity used to be. Land survey records indicate that the area in question once contained wetland and prairie habi-tat. After European settlement, the wetland was drained for agricultural use. In order to restore the area to wetland, the park district used a backhoe to locate and break drain tiles and a bulldozer to recontour the terrain, including construction of some dikes (Klein 1992). Thanks to water from natural springs and precipitation, the area became a 50-hectare ephemeral wetland. This area, which contains standing water during winter and spring, is known as the Shaker Trace Wetlands.

Around the same time as this restoration, the Hamilton County Park District established the Shaker Trace Seed Nursery from wild plants growing in relic prairie and wetland plant communities within 160 km of the nursery. Plants and seeds produced in the nursery were planted in parts of the Shaker Trace Wetlands and more than 120 hectares of planted prairie that surround the wetland area. Part of the wetlands was left unplanted to determine whether the wetland could become revegetated naturally without being planted. Here we share our observations of the plant community dynamics over two decades.

A vascular plant survey in the wetland/prairie complex was conducted from 1992 to 1996 via weekly visits during

Figure 1. Hydro-Axe operation to clear woody vegetation in Shaker Trace Wetlands in Ohio: a) resulting tire rut—such depressions hold water for longer periods into the dry season than do the surrounding areas; b) willows resprouting afterward, now being controlled by bush-hogging during the dry season. Photos by Denis Conover

the growing season (early March through October) and monthly visits from November to February. Most species were identified in the field, but specimens of some plants were collected for closer examination with a dissecting microscope. Plants were identified mainly using Glea-son and Cronquist (1991) and Holmgren (1998), and some specimens were deposited in the herbarium at the University of Cincinnati.

During that survey, 527 species in 84 plant families were identified in the wetlands/prairie complex (Conover 1996). Approximately 23% of these species were planted. Species that were not planted presumably came from propagules in the soil seed bank and from natural vec-tors of seed dispersal. The wetlands attract large numbers of waterfowl and other wildlife such as turtles that may introduce plant propagules to the area. Some of the species that quickly appeared in wet areas without being planted were sedges (carex spp.), spike rushes (Eleocharis spp.), rushes ( Juncus spp.), water pimpernel (samolus valerandi),


clammy hedge hyssop (gratiola neglecta), false pimpernel (lindernia dubia), cattails (typha spp.), arrowhead (sagit-taria latifolia), five-angled dodder (cuscuta pentagona), and bulrushes (schoenoplectus spp.), including the rare state-listed Pursh’s bulrush (s. purshianus), just one of several rare plant species that have spontaneously appeared in the wetlands. At the end of the 1992–1996 survey, we concluded that the wetland would become vegetated on its own, but that diversity could be increased by planting additional species (Conover and Klein 1994).

The prairie areas surrounding the wetlands have been maintained over the years by mowing and periodic pre-scribed burning to restrict tree growth. Mowing and burning of the wetlands was not always possible during extremely wet years. As a result, much of the wetland area had changed and become dominated by willows (salix spp.) and cattails. In October 2006, the park district removed much of the woody vegetation from the wetlands by using a Hydro-Axe machine to cut down willows, other small trees, and cattails and thus opened extensive areas to more light. The large tires of the Hydro-Axe machine also churned up soil in the wetlands and left new depres-sions that hold water for longer periods of time into the summer drought period (Figure 1a). Since the Hydro-Axe operation, resprouting willows (Figure 1b) have so far been kept under control by yearly mowing with a Bush Hog.

During a second survey in 2005–2008, 478 species in 84 plant families were identified in the wetland/prairie complex. Of these, 22% were introduced by park staff, and 73% were indigenous to the region (Conover 2008). Some changes took place in the wetlands/prairie complex between the two plant surveys. Certain species, such as four-angled spikerush (Eleocharis quadrangulata), were found during both of the plant surveys, but other species such as Pursh’s bulrush, that were found during the first plant survey were not found during the second plant survey.

As the wetlands mature, the disappearance of some species and the appearance of others whose propagules are brought in by natural vectors are to be expected.

Several new unplanted species were found during the second survey. Apparently it took longer than five years for these to reach the wetlands through natural dispersal or to become established from the seed bank. Such spe-cies include Mississippi arrowhead (sagittaria calycina), groundnut (apios americana), water parsnip (sium suave), and the state-endangered upright burhead (Echinodorus berteroi). A new non-native species that is spreading rapidly in the wetlands is bog bulrush (schoenoplectus mucronatus). Competition from the non-native invasive bog bulrush, as well as from cattails and willows, may have contributed to the disappearance of the rare Pursh’s bulrush. Some new non-native invasive species found in higher areas such as dikes include Chinese lespedeza (lespedeza cuneata), Cal-lery pear (Pyrus calleryana), and winter creeper (Euonymus fortunei).

By removing woody vegetation from the wetlands, the Hydro-Axe operation served to provide more habitat for herbaceous species. For instance, the establishment of upright burhead in the wetlands seems to have been asso-ciated with the Hydro-Axe operation. In July 2007, a population of this endangered species was discovered for the first time in the wetlands in tire ruts (Figure 2). It is possible that the seeds were present in the soil seed bank and were exposed by the Hydro-Axe operation’s churning of the soil, or they could have been brought in by waterfowl from a different wetland. Either way, upright burhead has benefited from the extra sunlight resulting from cutting down the willows.

Another change was that many species bloomed earlier during the 2005–2008 survey than during the 1992–1996 survey. Of 269 species that were observed in flower during both surveys, 39% bloomed earlier during the second time period. Between 1992 and 1996, the average annual temperature was 11.9°C, while between 2005 and 2008 the average was 12.9°C (Alan Black, NOAA Midwestern Regional Climate Center, pers. comm.). The earlier flow-ering of certain species might be due at least in part to the warmer temperatures associated with global warming (Root et al. 2003).

In conclusion, what in 1991 had been mostly corn and soybean fields has now become prime habitat for a wide variety of wild plants and animals. The Shaker Trace Wet-lands/Prairie Complex supports a large number of native as well as non-native vascular plant species. Most of the native species found in the planted prairies were introduced by Land Management staff, whereas most of the native species found in the restored wetland areas were not planted.

Future management of the area will include bush-hog-ging the wetlands periodically to cut off the willows before they become too large and periodically burning the planted prairie. Efforts will be made to control invasive non-native

Figure 2. State-endangered upright burhead (Echinodorus berteroi) appeared in tire rut depressions after the Hydro-Axe operation to control woody vegetation in Shaker Trace Wetlands in Ohio. Photo by Denis Conover


species. New depressions may be scooped out and some of the existing depressions in the wetlands may be deepened so that they will continue to hold water longer into the summer drought period. This would benefit not only the endangered burhead, but also amphibian species that need more time to develop into adults.

AcknowledgmentsWe wish to thank Marjie Becus, for identifying schoenoplectus mucro-natus and for reporting the locations of rare plants to ODNR; Allison Cusick and Rick Gardner of ODNR, for examining plant specimens sent to them; and Vic Soukup, for preparation of specimens for deposition in the herbarium at the University of Cincinnati.

ReferencesConover, D.G. 1996. Vascular plant survey of the Shaker Trace

wetlands/prairie complex at Miami Whitewater Forest. Unpublished Report for Hamilton County Park District, Cincinnati OH.

___. 2008. Vascular plant survey of the Shaker Trace wetlands/prairie complex at Miami Whitewater Forest. Unpublished Report for Hamilton County Park District, Cincinnati OH.

Conover, D.G. and J. Klein. 1994. Both planted and non-planted species populate restored wetland (Ohio). Restoration & Management notes 12:195–196.

Gleason, H.A. and A. Cronquist. 1991. Manual of Vascular Plants of northeastern United states and canada, 2nd ed. Bronx: New York Botanical Garden.

Holmgren, N.H. 1998. The illustrated companion to gleason and cronquist’s Manual. Bronx: New York Botanical Garden.

Klein, J. 1992. Drain-tile removal, recontouring, planting characterize wetland restoration project (Ohio). Restoration & Management notes 10:186–187.

Root, T.L., J.T. Price, K.R. Hall, S.H. Schneider, C. Rosenzweig and J.A. Pounds. 2003. Fingerprints of global warming on wild animals and plants. nature 421:57–60.

Potential Terrestrial Arthropod Indicators for Tallgrass Prairie Restoration in IowaJessica M. orlofske (Dept of Biology, University of new Brunswick, Fredericton, new Brunswick E3B 5a3, canada, [email protected]), Wayne J. ohnesorg (University of nebraska—lincoln, Extension, Pierce nE) and Diane M. Debinski (Dept of Ecology, Evolution and organismal Biology, iowa state University, ames ia)

Prairie management and restoration have traditionally emphasized plant communities but could benefit from

the development of ecological indicators that incorporate animal communities to provide tangible benchmarks of ecosystem function and process (Hodkinson and Jackson 2005). Arthropods are useful bioindicators in terrestrial and aquatic systems because, in addition to being diverse

and abundant, they represent key components in food webs, nutrient cycles, plant reproduction, and soil forma-tion (Hodkinson and Jackson 2005, Kremen et al. 1993).

A more comprehensive understanding of arthropod habitat specificity and resilience is needed before defining holistic ecological indicators for tallgrass prairies. Previous research has focused on either responses of prairie arthro-pods to management practices, specifically prescribed fire, or the status of particular taxa (Shepherd and Debinski 2005). Panzer and colleagues (1995) provide a preliminary list of remnant-dependent insects that qualify as poten-tial ecological indicators. This list could be improved by 1) eliminating a priori assumptions regarding potential indicators; 2) using consistent sampling across prairie sites; and 3) comparing indicator response at restored and reconstructed prairies. Our objective is to identify insect and spider families that are informative bioindica-tors of tallgrass prairie condition—a surrogate of biotic integrity—for use as measures of restoration progress. We accomplish this objective by 1) broadly examining insect and spider taxa without assuming their indicator potential; 2) collecting invertebrates using a standardized method; and 3) comparing the indicators at a suite of sites relevant to those interested in prairie conservation and restoration.

In June–August of 2006–2007, we sampled arthropods in central Iowa at prairie remnants (n = 9), isolated restora-tions/reconstructions (n = 10), and integrated reconstruc-tions (n = 11). Site types are defined according to Shepherd and Debinski (2005) (Figure 1). The surrounding landscape was primarily agricultural, but integrated reconstructions were nested within larger managed prairies. We selected sites to minimize geographical, soil type, and size variation; however, specific management practices varied across the sites. This approach allowed us to characterize taxa based on their association with a particular prairie type, provid-ing a way to evaluate consistency among managed prairies.

Sweep net transects were used to collect arthropods from foliage. Sweep net collections provide a rapid, inexpensive, and easily standardized protocol, although they do not sample the entire arthropod community, missing arthro-pods found on or below the ground surface. At each site, three 25 × 2 m belt transects were chosen with a random starting point and direction during each monthly visit. Transects were located at least 10 m from edges, trails, or water and were nonintersecting. Sampling occurred between 9:30 a.m. and 4:30 p.m. on days above 18°C with less than 60% cloud cover and calm winds (gusts less than 17 kph; Shepherd and Debinski 2005). Each transect was swept with a canvas net (diam. 30.5 cm) with 20 pendulum sweeps. Collected arthropods were transferred to labeled zip-top bags and stored in a −20°C freezer until identified.

Insect and spider specimens were identified to family according to Triplehorn and Johnson (2005) and Ubick and others (2005). We excluded taxa that were consid-ered impractical for practitioners to identify based on


Judson (JUD)

Sheeder (SH)

Moeckly (MOE) POLK

STORY

WEBSTER

HAMILTON

MARSHALL

JASPER

BOONE

GUTHRIE DALLAS

Liska-Stanek (LS) Briggs Woods (BW)

Richard’s Marsh (RM)

A.C. Morris (MOR)

R-Unit (RU)

Drobney Prairie (DP)

Marietta (MAR)

Grimes Farm (GF)

McFarland (MF)

Grant Ridge (GR)

Prairie Flower (PF)

Big Creek (BC)

Colo Bogs (CB)

Doolittle (DO)

Meetz (ME)

Stargrass (ST)

Chichaqua Bottoms

Neal Smith NWR

42° 30’

41° 30’

42° 00’

93° 00’ 94° 30’ 93° 30’ 94° 00’

N

Remnant

Isolated Restoration/Reconstruction

Integrated Restoration/Reconstruction

Site Types

0 20 km

Christopher D. Tyrrell 2010

Figure 1. Map of central Iowa prairie locations used in this study of potential arthropod indicators. Site symbols are proportional to the sizes of the sites (6–42 ha). Integrated sites, which are part of larger prairie complexes, are all shown the same size for clarity.

size (Collembola, Thysanoptera, Psocoptera, and micro-Hymenoptera) and taxa primarily associated with aquatic habitats (Ephemeroptera, Odonata, Plecoptera and primar-ily aquatic families of Diptera and Coleoptera). We also excluded Lepidoptera owing to the high proportion of indistinguishable larval specimens. To identify insect and spider families representative of each prairie type, we used indicator species analysis (ISA), which evaluates a taxon’s site specificity, the abundance of the taxon at a specific subset of sites, and site fidelity, the number of sites of a particular type where the taxon occurs, in order to calcu-late an indicator value for each taxon for each site type (see equation in Dufrêne and Legendre 1997). Arthropod abundances were log transformed to reduce skewness, and data from each sampling year were analyzed separately and combined to determine if there were effects attributable to sampling year. The ISA was performed using the duleg function in the labdsv package in the R statistical envi-ronment (version 2.6.0, R Foundation, Vienna). Using random permutation, ISA identified specific arthropod families that were significantly indicative (p < 0.05) and marginally indicative (p < 0.10) of a particular site type.

A total of 56,247 insects (28,179 in 2006, 28,068 in 2007) from 106 families (85 in 2006, 95 in 2007) and 3,246 spiders (1,913 in 2006, 1,333 in 2007) from 12 families (12 in 2006, 11 in 2007) were collected and analyzed. In 2006, five insect and one spider family were significantly indicative of remnant sites, one insect and one

spider family of isolated sites, and no significant taxa for integrated sites (Table 1). Insect family diversity and the number of potential indicators increased slightly in 2007. Significant indicators of remnant sites for 2007 included nine insect families and one spider family, two insect and one spider family for isolated sites, and only one insect family for integrated sites (Table 1).

Comparing between years, remnants shared two signifi-cantly indicative insect taxa, Formicidae (Hymenoptera) and Otitidae (Diptera), and the spider family Thomis-idae. Of the other taxa that appeared in both sampling years, Cixiidae (Hemiptera) was significantly indicative in 2007 and marginally indicative in 2006, and Tettigoniidae (Orthoptera) was marginally indicative in both years. For isolated sites, Aphididae (Hemiptera) was significantly indicative in 2007 and marginally indicative in 2006, and Tingidae (Hemiptera) was marginally indicative in both years. The combined data for both years yields a conserva-tive list of five significant indicator taxa for remnants: For-micidae (ants), Tettigoniidae (long-horned grasshoppers), Cixiidae (plant hoppers), Otitidae (picture-wing flies), and Thomisidae (crab spiders). Each of these groups was either a significant or marginally significant indicator of remnant sites in both the separate and combined analyses.

Our data corroborate Panzer and colleagues’ (1995) proposal of katydids as remnant indicators and suggest the further investigation of plant hoppers. Grasshoppers, specifically katydids, are large, easy to collect, ubiquitous


Table 1. Insect and spider families identified as indicators of prairie type. The indicator value (IV) ranges from zero to one and represents the frequency and abundance of each arthropod family for a particular prairie type. Using random permutation, the significance of each reported taxa was determined (** p = 0.05, * p = 0.10).

Remnant Isolated IntegratedFamily IV Family IV Family IV

Insects2006 Acanaloniidae * 0.33 Aphididae * 0.42

Alydidae ** 0.42 Coccinellidae ** 0.48Anthocoridae ** 0.46 Tingidae * 0.38Bombyliidae * 0.36Calliphoridae * 0.22Cixiidae * 0.38Elateridae ** 0.44Formicidae ** 0.42Otitidae ** 0.61Tettigoniidae * 0.42

2007 Anthomyzidae * 0.22 Aphididae ** 0.30 Gryllidae ** 0.46Anthribidae * 0.24 Sacarabaeidae * 0.26 Hemerobiidae * 0.27Cantharidae ** 0.56 Sepsidae ** 0.37 Melyridae * 0.28Chloropidae * 0.36 Tingidae * 0.35Cixiidae ** 0.44Dolichopodidae * 0.41Formicidae ** 0.39Lygaeidae * 0.44Megachilidae ** 0.39Membracidae * 0.42Miridae ** 0.40Muscidae ** 0.42Mutillidae ** 0.33Otitidae ** 0.55Sacrophagidae * 0.42Tabanidae * 0.28Tachinidae ** 0.38Tettigoniidae * 0.48Vespidae * 0.23

Total Anthocoridae * 0.43 Aphididae ** 0.40 Gryllidae ** 0.41Bombyliidae * 0.39 Coccinellidae ** 0.41 Hemerobiidae * 0.27Cantharidae ** 0.46 Phalacridae ** 0.42Cixiidae ** 0.45 Sepsidae * 0.43Dolichopodidae * 0.39Elateridae ** 0.46Formicidae ** 0.39Lygaeidae * 0.41Megachilidae * 0.34Membracidae * 0.42Miridae * 0.38Muscidae ** 0.39Otitidae ** 0.59Tachinidae ** 0.31Tettigoniidae ** 0.42Vespidae * 0.38

Spiders2006 Dictynidae ** 0.46 Salticidae ** 0.42 Lycosidae * 0.27

Thomisidae * 0.38

2007 Thomisidae * 0.44 Lycosidae * 0.25Philidromidae ** 0.55

Total Thomisidae ** 0.38 Philidromidae ** 0.44Dictynidae ** 0.47 Salticidae ** 0.38


insects in prairies. As potential indicators, katydids may provide information about site vegetation characteristics, since they feed on and oviposit into plant tissues. Several families of plant hoppers are abundant in tallgrass prairies (Hamilton 2005). Our analysis suggests at least one plant hopper family, Cixiidae, has potential indicator qualities. Unlike other plant hopper families (especially Cicadellidae) that include taxa endemic and often host-specific to prairie plant species and therefore often studied and identified as potential indicators (Hamilton 2005), Cixiidae are not well studied, especially in prairie ecosystems, and may merit further investigation.

In contrast to Panzer and colleagues (1995), our data support the inclusion of ants as remnant indicators. Ants are proposed indicator taxa in other habitats to monitor a variety of restoration and reclamation efforts (Andersen et al. 2004). This designation for tallgrass prairie underscores the important functional role of ants who contribute to soil quality by constructing subterranean tunnels, consuming plants, and scavenging, and who arguably constitute a majority of arthropod biomass in prairies (Trager 1998).

Our study expands the list of potential indicators to include crab spiders. Spiders are important insect predators in many ecosystems. Although spiders are typically general-ist predators, hunting strategy varies by taxa. In prairies, cryptic crab spiders (Thomisidae) are frequently located on stems and flowers of plants (Ubick et al. 2005). Crab spiders may reflect properties of the vegetation structure and abundance of floral resources within remnant sites.

In addition, we suggest that picture-wing flies, and true flies in general, deserve greater investigation in tallgrass prairie to determine what aspects of the prairie they may represent. Many families of true flies (Diptera), including the picture-wing flies (Otitidae) have not been studied in great detail in tallgrass prairies. The association of Otiti-dae with remnant sites is uncertain and should motivate additional study of this insect family, and perhaps Diptera in general because of the great diversity within the order (Triplehorn and Johnson 2005).

Our study provides an independent assessment of pre-viously proposed indicator or remnant-dependent taxa as well as an evaluation of a greater range of arthropod taxa, restricted by sampling design rather than a priori selection, across different prairie types. Family-level identification provides an efficient method to incorporate arthropod community information into terrestrial biomonitoring. Restorationists and managers do not need to become taxonomic specialists, and the scope is not limited to only a few well-known taxa such as butterflies. This analysis demonstrates that by examining a larger cross-section of arthropod taxa, we can not only empirically support cur-rent bioindicators, but also suggest additional taxa that may provide additional ecological information to resource professionals.

AcknowledgmentsThe authors thank the landowners, land mangers, and agencies for permission to use their prairies. We also acknowledge assistance from G. VanNostrand, D. Woolley, H. Gilliland, and MJ Hatfield. We appreciate statistical assistance from D. Cook and manuscript comments from C. Tyrrell and S. Orlofske. This project was funded by the Iowa Science Foundation, The Nature Conservancy: Nebraska Chapter’s J.E. Weaver Competitive Grants Program, Prairie Biotic Research, Inc., and the Iowa Department of Natural Resources: Iowa Wildlife Diversity Grant.

ReferencesAndersen, A.N., A. Fisher, B.D. Hoffmann, J.L. Read and

R. Richards. 2004. Use of terrestrial invertebrates for biodiversity monitoring in Australian rangelands, with particular reference to ants. austral Ecology 29:87–92.

Dufrêne, M. and P. Legendre. 1997. Species assemblages and indicator species: The need for a flexible asymmetrical approach. Ecological Monographs 67:345–366.

Hamilton, K.G.A. 2005. Bugs reveal an extensive, long-lost northern tallgrass prairie. Bioscience 55:49–59.

Hodkinson, I.D. and J.K. Jackson. 2005. Terrestrial and aquatic invertebrates as bioindicators for environmental monitoring, with particular reference to mountain ecosystems. Environmental Management 35:649–666.

Kremen, C., R.K. Colwell, T.L. Erwin, D.D. Murphy, R.F. Noss and M.A. Sanjayan. 1993. Terrestrial arthropod assemblages: Their use in conservation planning. conservation Biology 7:769–808.

Panzer, R., D. Stillwaugh, R. Gnaedinger and G. Derkovitz. 1995. Prevalence of remnant dependence among the prairie and savanna inhabiting insects of the Chicago region. natural areas Journal 15:101–116.

Shepherd, S. and D. Debinski. 2005. Evaluation of isolated and integrated prairie reconstructions as habitat for prairie butterflies. Biological conservation 126:51–61.

Trager, J.C. 1998. An introduction to ants (Formicidae) of the tallgrass prairie. Missouri Prairie Journal 18:4–8.

Triplehorn, C.A. and N.F. Johnson. 2005. Borror and Delong’s introduction to the study of insects, 7th ed. Belmont CA: Thomson Brooks/Cole.

Ubick, D., P. Paquin, P.E. Cushing and V. Roth, eds. 2005. spiders of north america: an identification Manual. American Arachnological Society.


Preliminary Habitat Assessment of Floating Oyster (Crassostrea virginica) Gardens (Delaware)Frank Marenghi (Delaware state University Department of agriculture & natural Resources, 1200 north DuPont High-way, Dover, DE 19901, 302/857-6476, [email protected]) and gulnihal ozbay (Delaware state Uni-versity Department of agriculture & natural Resources, [email protected])

Community volunteers, known as “oyster gardeners,” are growing the eastern oyster (crassostrea virginica)

off private docks in coastal lagoons of southeastern Dela-ware, known locally as “Inland Bays,” to aid restoration efforts. Today in Delaware, the oyster gardening program is a fraction of the size of those in neighboring states but has expanded tenfold from its inception in 2003. Oyster aquaculture can provide many of the same services as natu-ral oyster reefs, such as providing a hard substrate for colo-nization by marine epifauna, which accordingly attracts fishes and mobile invertebrates (e.g., O’Beirn et al. 2004). Many species of economic and ecological importance are considered habitat limited in the Inland Bays, particularly in juvenile refuge and forage areas. Floating oyster gar-dens can provide additional habitat at small scales while supplementing oyster spawning stocks without difficult and costly types of habitat modifications.

Oyster gardening activities are taking place along most of the Atlantic Coast into the Gulf of Mexico across an extremely wide gradient of conditions. The ecological impact of oyster gardening programs has received little attention and is absent from the literature with regard to Delaware’s Inland Bays. This study is the first to identify the macroepifauna that use floating oyster gardens as habitat. We also quantified the growth and survival of oysters and several water quality parameters. No natural reefs remain in the Inland Bays; descriptions of reef fauna in the region are for Delaware or Chesapeake Bay, and direct compari-sons would not be appropriate. Our data will be used as a baseline as the oyster gardening program expands within the Inland Bays.

The Delaware Inland Bays have a surface area of 83 km2, into which a 777 km2 watershed drains. The average depth is 1.2 m. The wild oysters present are few and widely scat-tered ( John Ewart, Delaware Sea Grant Marine Advisory Program, pers. comm.). Many seagrass (Zostera marina and Ruppia maritima) beds have also disappeared, and there has been a net loss of approximately 800 ha of tidal wetlands in the watershed over the past century (DIBEP 1993). These bays are experiencing eutrophication, high turbid-ity, periodically hypoxic conditions, poor tidal flushing, annual fish kills, harmful algae blooms, and low species diversity, especially in the 2 km2 of man-made dead-end canals (DIBEP 1993).

Oyster gardening sites are located throughout the three Inland Bays, almost exclusively on canals (Figure 1). By placing oysters in floating cages, oyster gardeners use these filter feeders to remove sediments and algae from the water column to improve local water quality and clarity while facilitating the removal of nitrogen and phosphorous. Once the oysters are 40–50 mm (1–2 years old), they are transplanted throughout the Inland Bays in rip-rap used for shoreline stabilization.

We randomly selected and sampled three replicate oyster gardens that were not cleaned or otherwise disturbed by the oyster gardeners between July and October 2007 in each of three embayments (Rehoboth, South Bethany, and Fenwick Island), for nine total study floats. Each study site consisted of a floating cage constructed of PVC pipe (ca. 10 cm diameter) and vinyl-coated wire mesh (14-gauge, 25 × 25 mm mesh), each with two 46 × 46 × 23 cm wire baskets. Floats were tied to docks located at the residences of volunteer oyster gardeners. Before deployment, all oyster gardens were power-washed to remove all fouling organ-isms, in part because the sessile invertebrates and algae that grow on hard marine surfaces can lead to oyster mortality when occurring at high densities. This initial cleaning also allowed epifauna subsequently associated with the oysters to be attributed to the spatial and temporal location of stocking and not to prior events in the culture process.

Three different floats in each embayment were sampled at monthly intervals for species richness and abundance such that each float was sampled only once. We selected 15% of the live and dead oysters and measured shell length to the nearest 0.1 mm. Epifaunal sampling was conducted with a net (3 mm mesh) supported by a square PVC frame measuring 61 × 61 cm. The net was quickly placed under each of the baskets as they were lifted out of the larger oyster float with a boat hook (Figure 1). With the net enveloping it, each basket was agitated while the oysters were washed with a freshwater hose, rinsing the contents into the net. Specimens were picked from the net by hand and placed in 70% isopropanol solution, and later identi-fied and counted. Statistical analyses, although carried out, are omitted from this note for the sake of brevity.

Oyster and macroepifauna results are summarized in Table 1. Although the percentage of live oysters ranged from 33% to 87%, a markedly higher percentage of Fenwick oysters were alive than of those in Rehoboth or Bethany. Fenwick oyster gardens also consistently had the greatest abundance of macroepifauna (3,339 indi-viduals or 50.5% of total collection). In Rehoboth and Bethany, 2,051 (31%) and 1,227 (18.5%) individuals were collected, respectively. The most abundant taxon, grass shrimp (Palaemonetes spp.), comprised 77.1% of all individuals collected, with no notable differences by embayment or month. Blue crab (callinectes sapidus) abundance was higher in August than in October and also differed by embayment, with Fenwick having fewer


blue crabs than Rehoboth or Bethany. Oyster toadfish (opsanus tau) abundance did not vary by month but was considerably different by embayment, with Fenwick having the greatest abundance. Striped blennies (chas-modes bosquianus) and naked gobies (gobiosoma bosc), which are obligate oyster reef residents, were observed in floats in Fenwick canals inside oyster shells with eggs stuck to the inner surface.

At the embayment scale, Fenwick had a much greater abundance of motile macroepifauna, a greater percentage of live oysters at the end of the study, more oyster toadfish, and fewer blue crabs than Rehoboth and Bethany. Fenwick, located in Little Assawoman Bay, experiences the most tidal flushing, estimated as 20 to 50 days (Taiping Wang, Virginia Institute of Marine Science, unpub. data), while estimates for Rehoboth and Indian River Bays are as high as 80 and 100 days, respectively (DIBEP 1993). Lower flushing times (more tidal action) may allow more fishes and invertebrates, including larvae, to enter the Fenwick

Figure 1. Collecting baseline data on oyster gardens within Delaware’s Inland Bays, USA. The map (left) shows the distribution of oyster gardens, both those included in the study (solid circles) and those not included (open circles). Epifaunal sampling (right) of floating oyster gardens involves one person retrieving the oyster baskets one at a time while the other person slides the net underneath, catching the specimens. Photo by J. Dersham

canals. The appreciably lower oyster mortality may also stem from these hydrologic properties.

We hoped to avoid changes in epifauna due to time of deployment by simultaneous stocking in July to focus on the integrity of the initial assemblages as time in the water increased (Osman and Whitlatch 1998). Species diversity and abundance in oyster floats did not change measurably over time and developed in less than one month. The number of taxa may seem low compared to that in commercial systems (O’Beirn et al. 2004); however, previous reports have described invertebrate communities in the Inland Bays severely impacted to the point that no animals were observed in some canals (DIBEP 1993). Clear abundance and diversity criteria will need to be established for the Inland Bays to effectively rank the oyster garden fauna described here.

Of the 14 species collected from floating oyster gardens, many may be habitat limited in these soft-bottom canal systems, including species considered obligate oyster reef


residents: oyster toadfish, naked goby, black-fingered mud crab (Panopeus herbstii), white-fingered mud crab (Rhithro-panopeus harisii), and striped blenny (Tolley and Volety 2005). These fishes use oyster shells for spawning substrate and may be useful as indicators, along with mud crabs, of quality reef habitat (Breitburg 1998). In an environment without natural reefs, it is not unreasonable to assume that the presence of obligate reef residents is due to enhanced biological production. Juvenile and adult stages of com-mercially important species such as American eel (anguilla rostrata), mummichog (Fundulus heteroclitus), and blue crab use floating oyster gardens. Blue crab populations in particular are sensitive to predation of juvenile stages and may benefit from additional nursery habitat (Messick and Casey 2004).

High turbidity and sedimentation rates lead to poor flow through the floats, lower oxygen levels, and, collectively, inefficient filtering by oysters. In our study, oysters were often covered with silt at the time of sampling, despite being suspended in the water column, offering a possible explanation for slow growth and mortality. Paradoxically, oyster mortality was highest at a site that was sampled after one month in the water and was lowest at a site that was sampled after three months in the water, suggesting site differences can outweigh fouling time. Our results show that the fouling of oyster gardens for 1–3 months during the summer does not measurably affect the habitat of the mobile macroepifauna collected. Site selection, however, clearly affects oyster growth, survival, and community composition in oyster gardens in Delaware’s Inland Bays.

The goals of the Inland Bays oyster gardening program are to produce an annually spawning adult population, enhance the potential for natural recruitment, improve water clarity and nutrient cycling, and provide habitat lost owing to the decimation of the oyster population (DCIB 2009). Our study provides some baseline data on oyster survival and growth and community composition that will be used to evaluate progress as the oyster gardening program develops and we conduct annual quantitative assessments.

AcknowledgmentsFunding for this project was provided by USDA/CSREES 2004-38820-15154 and a USDA Evans-Allen grant. Thanks to John Ewart for facilitating the oyster gardening program and providing techni-cal input. Great thanks to the all of the volunteer oyster gardeners, without whom this study would not have been possible. Thanks to the Delaware Center for the Inland Bays for the use of their oysters and oyster aquaculture gear. Thanks to Kate Rossi-Snook for her assistance with field work, specimen processing, and lab work. Thanks to Amanda Treher, Brian Beal, and Christopher Heckscher for comments on an earlier draft of this paper.

ReferencesBreitburg, D.L. 1998. Are three-dimensional structure and

healthy oyster populations the keys to an ecologically interesting and important fish community? Pages 239–250 in M. Luckenbach, R. Mann and J.A. Wesson (eds), oyster Reef Habitat Restoration: a synopsis and synthesis of approaches. Gloucester Point VA: VIMS Press.

Table 1. Oyster and macroepifaunal results from 2007 field season for oyster gardens in three embayments of coastal Delaware’s Inland Bays. Mean (± SD) shell lengths for eastern oyster (Crassostrea virginica) and presence (x) or absence (−) for each species are given.

EmbaymentFenwick Rehoboth Bethany

Oyster VariableAlive (%) 66.3 44.2 47.2Final density (oysters/m2) 108 ± 13 54 ± 16 99 ± 89Final shell length (mm) 74.1 ± 7.2 71.8 ± 9.5 81.6 ± 0.1Change in live length (mm) 2.9 ± 1.0 −2.0 ± 6.3 3.9 ± 4.7Macroepifauna Taxon American eel (Anguilla rostrata) x x xAsian shore crab (Hemigrapsus sanguineus) x x −black-fingered mud crab (Panopeus herbstii) x x xblue crab (Callinectes sapidus) x x xclam worm (Neanthes succinea) x x xgrass shrimp (Palaemonetes spp.) x x xinland silverside (Menidia beryllina) − − xmummichog (Fundulus heteroclitus) x x xnaked goby (Gobiosoma bosc) x x xoyster toadfish (Opsanus tau) x x xrainwater killifish (Lucania parva) x − xscuds (Gammarus spp.) x x xstriped blenny (Chasmodes bosquianus) x x xwhite-fingered mud crab (Rhithropanopeus harrisii) x x x


Delaware Center for the Inland Bays (DCIB). 2009. Shellfish gardening program. www.inlandbays.org/cib_pm/comments .php?id=33_0_31_0_C

Delaware Inland Bays Estuary Program (DIBEP). 1993. Delaware Inland Bays Estuary Program characterization summary. Dover: DIBEP Science and Technical Advisory Committee.

Messick, G. and J. Casey. 2004. Chapter 8.6. Status of blue crab, callinectes sapidus, populations in the Maryland Coastal Bays. Pages 82–91 in Maryland’s Coastal Bays: Ecosystem health assessment. Maryland Department of Natural Resources Document No. DNR-12-1202-0009.

Newell, R.I.E. 2004. Ecosystem influences of natural and cultivated populations of suspension-feeding bivalve mollusks: A review. Journal of shellfish Research 23:51–61.

O’Beirn, F.X., P.G. Ross and M.W. Luckenbach. 2004. Organisms associated with oysters cultured in floating systems in Virginia, USA. Journal of shellfish Research 23:825–829.

Osman, R.W. and R.B. Whitlatch. 1998. Processes controlling local and regional patterns of invertebrate colonization: Applications to the design of artificial oyster habitat. Pages 179–197 in M. Luckenbach, R. Mann and J.A. Wesson (eds), oyster Reef Habitat Restoration: a synopsis and synthesis of approaches. Gloucester Point VA: VIMS Press.

Tolley, S.G. and A.K. Volety. 2005. The role of oysters in habitat use of oyster reefs by resident fishes and decapod crustaceans. Journal of shellfish Research 24:1007–1012.

SPECIAL THEME: ECOLOGICAL RESTORATION IN MExICO

Creating Refuges for the Axolotl (Ambystoma mexicanum)Elsa Valiente (Departamento de Zoología, instituto de Biología, Universidad nacional autónoma de México [UnaM]), armando tovar (Departamento de Zoología, UnaM), Homán gonzález (Departamento de Zoología, UnaM), Dionisio Eslava-sandoval (Umbral axochiatl, 2ª cerrada de Yucatán Barrio la concepción tlacuapa 16000 DF) and luis Zambrano (Departamento de Zoología, instituto de Biología, Universidad nacional autónoma de México, apdo Post 70-153. ciudad Universitaria, México DF, México, [email protected])

Cropping in the naturally fertilized Xochimilco wet-lands during seasonal floods helped Aztecs to build

one of the most advanced civilizations of the Americas. At the edge of Mexico City, Xochimilco hosts more than 140 migratory birds and an endemic crayfish (cambarel-lus montezumae) and fish (Menidia humboldtiana). It is the last remnant habitat of one of the most iconic animals from central Mexico: the axolotl (ambystoma mexicanum) (Figure 1). This salamander has been associated with Mexican culture as twin of the most important Aztec god Quetzalcoatl and inspiration for writers and philosophers.

Mexico City requirements for ecosystem services pro-vided by Xochimilco wetlands have intensively perturbed this aquatic system in the last 60 years. Its water quality has decreased, with high bacteria, nutrient, and heavy metal concentrations (Zambrano et al. 2009). Consequently, axolotl populations have decreased alarmingly in the last decade, with high probabilities of extinction in the wild by 2019 (Zambrano et al. 2007). The few remaining popula-tions are scattered and isolated, and therefore restoration measures are critical.

Top-down management and restoration attempts in Xochimilco have not worked, leading to abandonment

of traditional agriculture for greenhouse production, reducing even more the water quality and axolotl habitat. An aquaculture fish introduction program resulted in an overabundance of tilapia (oreochromis niloticus) and carp (cyprinus carpio), which affect axolotl populations by predation and modifying food webs. A program of axolotl reintroduction conducted by an academic institution and sponsored by local government, using salamanders grown in captivity, is not expected to increase population numbers (see Zambrano et al. 2007). Moreover, if siblings are used for the introduction program, the genetic diversity of wild populations may be reduced.

Long-term success of restoration measures requires involvement of local people in the management program (Gall and Staton 1992). Scientists together with local farmers (chinamperos) are working to create an alterna-tive axolotl restoration program. Traditionally, these local farmers produce in “chinampas,” which are land struc-tures forming a labyrinth of canals connecting lakes and wetlands (Contreras et al. 2009). Chinamperos use canal sediments for cropping, reducing chemical fertilizers. Also, canals surrounding the chinampas increase the wetlands’ spatial heterogeneity that seems to be necessary for axolotl survival. This indicates that it is necessary to conserve the traditional agriculture in order to restore the wetland.

By sharing information with chinamperos, we have enough ecological and empirical knowledge about the axo-lotl to create a restoration program focused on improving its habitat. We now understand that population dynamics depend mostly on the survival of eggs and larvae (Zam-brano et al. 2007), which is closely related to plants used by females to lay their eggs (Valiente 2006, Marín 2007). The axolotl is at the top of the food web, but shares most of its food sources with the non-native carp and tilapia (Zambrano et al. 2010). Axolotl distribution is limited to a few small canals, surviving away from urban and greenhouses areas (Contreras et al. 2009). City demand


Figure 1. The xochimilco wetlands (top) next to Mexico City provide habitat for the culturally iconic axolotl (Ambystoma mexicanum), an aquatic salamander (bottom) that is rapidly disappearing from the increasingly degraded wetlands and has been bred in captivity in early restoration efforts. Photos by Luis Zambrano (top) and Carmen Loyola (bottom)

for Xochimilco ecosystem services makes it impossible to restore the entire wetland, and isolating the conservation of the axolotl from human activities will break the already weak link between the salamander and citizens. Therefore chinamperos and scientists decided to generate refuges within chinampas border canals.

Canals used as refuges were isolated from the system with rustic filters to exclude non-native fish and improve the water quality (Figure 2). Many chinamperos water their lands manually by carrying water from the canals, and improved water quality will also yield farm products that test free from bacteria and heavy metals, which allows chinamperos to better market their products. This part-nership helps both the axolotls and the chinamperos by improving habitat and preserving traditional agriculture.

We are comparing the first experimental refuge canal, established in January 2009, with control canals and evalu-ating individual axolotl growth and health. The first evalu-ation suggests an improvement of water quality after four months. Refuge water was significantly lower in turbidity (15%), ammonium (77%), and nitrates (87%) than water in the control canal. Invertebrate abundance did not vary

significantly, except for Odonata larvae, which were denser in the refuge canal. In this first evaluation, only 30% of the 12 previously marked axolotls were recaptured. The rest of them could have been hidden in mud or depredated by snakes. However, recaptured axolotls gained, on average, 16% of their weight, which is more than the growth of individuals in experimental colonies within the same period of time (Zambrano et al. 2007).

We need more information to launch a multicanal refuge program, for example, the effectiveness of different plant types as shelter for axolotl eggs and larvae, or food web changes within the refuges. For this reason, a second phase with four additional canals in three chinampas is starting now. This second phase will yield results within the next year to inform a proposed larger refuge network to cover at least 10% of the chinampas canals. Refuges should function as bridges among areas where axolotls still survive (Contreras et al. 2009).

At every stage, scientists and chinamperos have to work together. Therefore we are holding participatory workshops to develop strategies so that chinamperos can manage refuges themselves in the long term.


The refuge program aims to conserve both this sala-mander and traditional agriculture. We are generating a market for farm products that will include the cost of saving local culture, Xochimilco, and axolotls. Restoration of the axolotl is a complex process that must be discussed among stakeholders (such as local government, fishermen, and tourists) in order to produce multiple actions and a monitoring program. Declining axolotl populations in the wild do not leave much time or room for mistakes. But a good restoration process will bring not only increased axo-lotl populations but also other benefits from a functioning wetland and better market opportunities for chinamperos.

ReferencesContreras, V., E. Martínez-Meyer, E. Valiente and L. Zambrano.

2009. Recent decline and potential distribution in the last remnant area of the microendemic Mexican axolotl (ambystoma mexicanum). Biological conservation 142:2881–2885.

Gall, G.A.E. and M. Staton. 1992. Conclusions. agriculture Ecosystems & Environment 42:217–230.

Marín, A.I. 2007. Preferencia de plantas para la oviposición del ajolote ambystoma mexicanum en condiciones del laboratorio. LicS thesis, Universidad Nacional Autónoma de México.

Valiente, E. 2006. Efecto de las especies introducidas en Xochimilco para la rehabilitación del hábitat del ajolote (ambystoma mexicanum). MS thesis, Universidad Nacional Autónoma de México.

Zambrano, L., V. Contreras, M. Mazari-Hiriart and A.E. Zarco-Arista. 2009. Spatial heterogeneity of water quality in a highly degraded tropical freshwater ecosystem. Environmental Management 43:249–263.

Zambrano, L., E. Valiente and M.J. Vander Zanden. 2010. Food web overlap among native axolotl (ambystoma mexicanum) and two exotic fishes: Carp (cyprinus carpio) and tilapia (oreochromis niloticus) in Xochimilco, Mexico City. Biological invasions doi: 10.1007/s10530-010-9697-8.

Zambrano, L., E. Vega, L.G. Herrera, E. Prado and V.H. Reynoso. 2007. A population matrix model and population viability analysis to predict the fate of endangered species in highly managed water systems. animal conservation 10:297–303.

Tropical Dry Forest Landscape Restoration in Central Veracruz, Mexicoguadalupe Williams-linera (instituto de Ecología ac, car-retera antigua a coatepec 351, Xalapa, Veracruz 91070, Mexico, [email protected]) and claudia alvarez-aquino (Universidad Veracruzana, instituto de investigaciones Forestales, Xalapa, Veracruz 91070, Mexico, [email protected])

Until recently, it was thought that native tropical dry forest (TDF) in central Veracruz was completely

gone. Fortunately, recent research using Landsat images and ground-truth verification showed the presence of TDF remnants. There is still 7% of the original forest in the region, although one-third is secondary vegeta-tion (F. López-Barrera, Instituto de Ecología A.C., pers. comm.).

Tropical dry forest landscape restoration has been car-ried out in only a few places. The most integral restoration experience started in Parque Nacional Santa Rosa, Costa Rica, which was expanded to become Area de Conservación Guanacaste, a region ten times larger ( Janzen 2008). Other restoration efforts are being carried out on the Pacific coast of Panama (Griscom et al. 2009), Paraná River Valley in Central Brazil (Sampaio et al. 2007), Hojancha and Cañas in Costa Rica (Fonseca-González and Morera 2008), the Ayuquila River watershed in Jalisco (Ortiz-Arrona et al. 2004), and the Tembembe river in Morelos, Mexico (Bonfil et al. 2004). Most TDF restoration experiences indicate that besides establishing native species plantations, resto-rationists should rely on the resprouting ability of the trees characteristic of these forests (Griscom et al. 2009) and different management techniques to facilitate this natural regeneration capacity (Sampaio et al. 2007). We recognize that natural tree regeneration is not only the least expensive method possible; it is also a tool for designing restoration efforts ( Janzen 2008, Viera and Scariot 2006).

This study is part of Project ReForLan, which focuses on the restoration of dryland forest landscapes for bio-diversity conservation and rural development in Latin America (Newton 2008). Our goal was to test restoration techniques and identify the main constraints to forest restoration. The two objectives were to define a reference system through the determination of forest structure and tree species composition, and to define a set of tree species to be used in restoration efforts.

Figure 2. Rustic filters in one of the axolotl refuges in xochimilco. The rustic filters improve water quality and keep out non-native fish. Note the chinampa in production on the left. Photo by Homan González


The study area is located in central Veracruz, Mexico (19°17'N, 96°26'W, 100–250 m asl). The climate is hot and dry. Mean minimum and maximum temperatures are 19.8°C and 30.7°C, respectively. Total annual precipita-tion is 966 mm (range: 502–1,466 mm). The dry season extends from October to May. Soils are mainly Cambisol and Vertisol. In this region, land is mainly used for cattle ranching, generally on a small scale by private landowners; but for common land tenants (ejidatarios), the main activ-ity is growing corn and other crops, such as papaya, bean, green chili, watermelon, sugar cane, and mango. This area is rich in history related to pre-Hispanic settlements (600 to 1500 A.D.) and the Mexican Independence (19th century). In addition, the region includes an important observation point, the River of Raptors, for one of the largest annual migrations of raptors in the world.

The TDF reference system was determined through the selection of ten forest remnants and five early successional sites, 1 to 72 months after the last abandonment, with different land use histories (Williams-Linera and Lorea 2009, Williams-Linera et al. 2010). Vegetation structure was characterized in terms of density, basal area, and height for canopy trees (≥ 5 cm dbh) and understory woody plants (< 5 cm dbh). Mean density was 1,014 ± 104 and 2,532 ± 2,272 individuals/ha, with basal area 30.2 ± 2.11 m2/ha and 1.96 ± 0.12 m2/ha, for canopy and understory vegetation, respectively. The forest mean height was 10 m and reached a maximum of 15 m. The early successional sites had a basal area and density ranging from 0.40 to 3.88 m2/ha and from 900 to 5,450 individuals/ha, respectively.

We recorded 122 woody plant species in the forest fragments (Williams-Linera and Lorea 2009). In the early successional sites, 45 woody species were recorded. Some species were represented only in forests or fallows, but 20 tree species were growing in both early secondary and mature forest sites (Williams-Linera et al. 2010), indicating that mature forest species were entering the successional process at very early stages.

Next, we assessed local knowledge and preferences through workshops, field visits, key informants, informal interviews, and specimen collection to determine tree species that were economically and ecologically valuable resources for restoration. This process identified 75 species as useful, rare, valuable, or important for wildlife (Suárez et al., forthcoming). Using the lists of tree species in forest or fallows as well as those often mentioned locally, we selected six native tree species for the restoration experiments. Two were timber species, one was used as forage, and the other three displayed potential to be used for such nontimber forest products as firewood, ornamentals, handcraft manu-facture, and the growth of an edible mushroom that is very popular locally (Table 1).

We established four restoration trials to evaluate the establishment and growth of these six species in fallows with different degrees of disturbance. The early succes-sional sites for the restoration experiments were fenced with barbed wire to exclude livestock. Seeds were collected in the study area and germinated in the local nursery where seedlings stayed four to six months prior to transplant. In September 2007, 960 seedlings were transplanted to four 12 × 20 m plots per site, with individuals 2 m apart. Plant survival and growth in basal diameter and height have been monitored every four months.

Seedling survival was statistically similar among species, except for cedro (cedrela odorata) (Figure 1). After the first dry season, survival of all species was greater than 55%, a result similar to that of restoration studies in pastures in central Brazil, where survival was greater than 60% (Sampaio et al. 2007) or 35%–77% (Vieira et al. 2007). After the second-year dry season, survival was the lowest for cedro (5.1%) and the highest for pochote (ceiba aescu-lifolia; 78.3%) and guácimo (guazuma ulmifolia; 64.3%) (Figure 1).

The relative growth rate in height (RGRh) was statisti-cally similar among species, whereas in diameter (RGRd) it was different among species (Table 1). Cedro had the highest RGRd. Cedro, despite its excellent growth perfor-mance, appeared unable to adapt to drought conditions, as has been previously reported (Ortiz-Arrona et al. 2005). Guácimo maintained a high survival percentage because of its resprouting ability; with the rainy season, some appar-ently “dead” individuals resprouted, as has been reported in other forests (e.g., Viera and Scariot 2006, Griscom et al. 2009). Pochote grew slowly because its stem top frequently breaks during the dry season, recovering when it rains.

Each tree species may contribute something different to the restoration effort. Results of transplant experiment suggest that all selected species can be used for restoration; however, each species requires different site conditions for transplanting. Timber species (cedro and roble, tabebuia rosea), which are the most important ones for local people, are severely drought intolerant. Therefore, they need to be planted near isolated trees that help to maintain humidity.

Figure 1. Survival percentages (October 2007–June 2009) for six tree species planted in restoration experiments in central Veracruz, Mexico. Survival of cedro (Cedrela odorata) was significantly lower than other species (one-way ANOVA, F = 3.7, df = 5,18, p = 0.02).


To avoid seedling mortality, sites with some tree or other canopy cover have been recommended because they offer a milder environment and moister soil than open sites (Viera and Scariot 2006). In contrast, pochote and guácimo are drought tolerant and may be used in disturbed areas with no woody vegetation; their resprouting ability allows establishment and survival in extremely dry conditions. Local people are not enthusiastic about nontimber spe-cies, but they are important for restoration by changing the microenvironment, providing suitable conditions for other species of economic importance to establish later.

The procedures used for TDF restoration must be adapted to this particular environment, instead of follow-ing techniques developed for temperate or moist forests (Viera and Scariot 2006). Seedling transplant represents one option; however, since primary species are present at early successional sites, it is worthwhile to consider the possibility of restoring TDF through passive restoration via natural colonization. Our next objective will be to determine whether secondary succession is efficient in the recovery of original vegetation. If not, we would enrich suc-cessional sites with nonresprouting species or some primary forest species that cannot be dispersed at early successional sites. Additionally, restoration should be implemented to strategically connect remnant forests, successional sites, and historical landmarks that may promote tourism and thus boost the local economy.

AcknowledgmentsThis research was funded by the European Community under INCO Project ReForLan (CT2006-032132).

ReferencesBonfil, C., I. Trejo and R. García-Barrios. 2004. The

experimental station “Barrancas del Río Tembembe” for ecological restoration in NW Morelos, México. Paper presented at 16th Annual Conference of the Society for Ecological Restoration, Victoria BC, Canada, 23–27 August. Proceedings on CD-ROM.

Fonseca-González, W. and B.A. Morera. 2008. El bosque seco tropical en Costa Rica: Caracterización ecológica y acciones para la restauración. Pages 115–135 in

M. González-Espinosa, J.M. Rey-Benayas and N. Ramírez-Marcial (eds), Restauración de Bosques en américa latina. Mexico City: Mundi-Prensa.

Griscom, H.P., B.W. Griscom and M.S. Ashton. 2009. Forest regeneration from pasture in the dry tropics of Panama: Effects of cattle, exotic grass, and forested riparia. Restoration Ecology 17:117–126.

Janzen, D.H. 2008. Restauración del bosque seco tropical: Área de Conservación Guanacaste (ACG), noroeste de Costa Rica. Pages 181–210 in M. González-Espinosa, J.M. Rey-Benayas and N. Ramírez-Marcial (eds), Restauración de Bosques en américa latina. Mexico City: Mundi-Prensa.

Newton, A.C. 2008. Restoration of dryland forests in Latin America: The ReForLan Project. Ecological Restoration 26:10–13.

Ortiz-Arrona, C.I., P.R.W. Gerritsen, L.M. Martínez R. and M. Snoep. 2004. Restauración de bosques ribereños en paisajes antropogénicos, en el occidente de México. Presentation at the I Simposio Internacional sobre Restauración Ecológica, Santa Clara, Cuba, 17–21 November. www.secretariadeambiente.gov.co/sda/libreria/pdf/ecosistemas/restauracion/1_ar31.pdf

Sampaio, A.B., K.D. Holl and A. Scariot. 2007. Does restoration enhance regeneration of seasonal deciduous forests in pastures in central Brazil? Restoration Ecology 15:462–471.

Suárez, A., G. Williams-Linera, C. Trejo, J.I. Valdez-Hernández, V.M. Cetina-Alcalá and H. Vibrans. Forthcoming. Local knowledge helps select species for forest restoration in a tropical dry forest of central Veracruz, Mexico. agroforestry systems.

Viera, D.L.M. and A. Scariot. 2006. Principles of natural regeneration of tropical dry forests for restoration. Restoration Ecology 14:11–20.

Williams-Linera, G., C. Alvarez-Aquino, E. Hernández-Ascención and M. Toledo. 2010. Early secondary succession potential for tropical dry forest recovery in central Veracruz, Mexico. Unpublished manuscript in preparation.

Williams-Linera, G. and F. Lorea. 2009. Tree species diversity driven by environmental and anthropogenic factors in tropical dry forest fragments of central Veracruz, Mexico. Biodiversity and conservation 18:3269–3293.

Table 1. Tree species used in restoration assays in the TDF of central Veracruz, Mexico, and their mean (± SE) rela-tive growth rate in height (RGRh), and diameter (RGRd) during 19 months. The RGRh was similar across species (one-way ANOVA, F = 2.27, df = 5,67, p = 0.06), and RGRd was significantly different (F = 6.47, df = 5,67, p < 0.0001).

Relative Growth Rate

Use Common name Species FamilyHeight

(cm/cm/y)Diameter

(mm/mm/y)

Timber Cedro Cedrela odorata Meliaceae 0.54 ± 0.21 0.66 ± 0.10Roble Tabebuia rosea Bignoniaceae 0.41 ± 0.07 0.41 ± 0.09

Forage Guácimo Guazuma ulmifolia Sterculiaceae 0.29 ± 0.14 0.50 ± 0.06

Nontimber forest products

Pochote Ceiba aesculifolia Bombacaceae 0.09 ± 0.12 0.19 ± 0.05Patancán Ipomoea wolcottiana Convolvulaceae 0.35 ± 0.12 0.39 ± 0.07Algodoncillo Luehea candida Tiliaceae −0.01 ± 0.14 0.28 ± 0.02


Restoring the Vara de Perlilla in La Mesa, Mexicoconcepción Mendoza-Bautista (División de ciencias Fores-tales, Universidad autónoma chapingo, chapingo, Estado de México, cP 56230, km 38.5 carretera México-texcoco, México, [email protected]), Fortino garcía-Moreno (División de ciencias Forestales, Universidad autónoma chapingo, [email protected]) and Dante arturo Rodríguez-trejo (División de ciencias Forestales, Universidad autónoma chapingo, [email protected])

The vara de perlilla (symphoricarpos microphyllus), or rejagar, is a shrub found from New Mexico to Gua-

temala. In Central Mexico, it is found mostly in pine-oak forests and also in the true fir forests, where it blooms from July to September, produces fruits from October to December, and provides forage for wildlife such as deer. Its branches are widely used to make brooms and manu-facture Christmas crafts, making this shrub an important nontimber forest product in Mexican temperate areas. Despite its resprouting ability, vara de perlilla populations are starting to decline because of overuse and deforestation or forest degradation. In other cases, the individuals exist, but they are overutilized, too young, and small, so cannot be employed when required.

Propagation and reforestation of this shrub just began a few years ago, mostly for commercial purposes. Yet its seed propagation still is not successful, and the ideal conditions for establishing this species are not well known. Its light requirements have not been studied formally. We consid-ered light to be a key limiting factor because the species does not prosper under dense shade, but it does well under partial shade or full sun. The objective of this work was to determine appropriate light conditions for the establish-ment of restoration plantings with this species.

This study was carried out in La Mesa ejido, San José del Rincón municipality, in the state of Mexico (19°34'N and 100°10'W). The climate is temperate, with a mean annual precipitation of 904 mm and a mean annual temperature of 12°C, and soil is an Ocric Andosol. Cuttings were col-lected December 23–26, 2006 (winter, dry season, after fruit production), from local vara de perlilla populations. They were 15–20 cm long and were immediately dipped in Raizone Plus solution and planted in plastic bags filled with local pine-oak forest soil. The plants were grown in a small local forest nursery for six months. A 30% shade cloth protected the plants during five months but was removed one month before planting during the hardening phase. Irrigation was provided every other day, adjusting in case of rain, and was slightly reduced during the last month of hardening.

We planted a total of 900 six-month-old shrubs for this work: 450 small (20–35 cm length) and 450 large (45–60 cm length). Three planting sites (treatments) were chosen:

oak forest (encino laurelillo, Quercus laurina, with some individuals of aile, alnus jorullensis); six-year-old pine forest plantation (Pinus pseudostrobus, 3.5 m tall on average, 1,111 trees/ha density), on a former oak forest site; and an agricultural field cleared 20 years ago from oak forest. The treatments formed a gradient from less to more light: oak forest (1328.6 to 5224.1 MJ/m2/y), pine plantation (6388.4 to 9140.9 MJ/m2/y), and field (7050.2 to 9324.6 MJ/m2/y). The plots were approximately 150 m apart. In July 2007 during the summer rainy season, we planted 300 shrubs per treatment: 150 small and 150 large. The shrubs were planted one meter apart in six completely randomized blocks per treatment. Each block segregated 50 plants by size. To eliminate competition, grasses and forbs were removed in a circle of 1 m radius around each shrub. To measure light levels at planting time and one year later, we placed, leveled, and oriented to the north a digital camera with a hemispherical lens integrated into a frame on the three shrubs in the center of each plot to take 180° photographs toward zenith. Such hemispherical photographs were analyzed with the HemiView system (Dynamax, Houston TX), with date, latitude, longitude, and altitude to estimate direct, diffuse, and total solar radia-tion (MJ/m2/y) (averaged between the two observations taken in each site).

After one year of growth in the planting site ( July 2008), we measured several morphological variables, including shoot, root, and total biomasses. For this purpose, six plants from each block (3 per size category, 36 per planting site, for a total of 108) were carefully removed, cleaned, and oven-dried. All 900 planted shrubs were observed for survival. The probability of mortality was calculated with logistic regression, using solar radiation as the independent variable. Mortality and biomasses were compared among planting sites with the t-test; biomasses among size cat-egories were compared by analysis of variance and least significant difference.

Figure 1. Relationship between total solar radiation (X2) and probability of mortality (P) for the vara de perlilla (Symphoricarpos microphyllus) in an oak forest in Central Mexico.


The overall survival was 96.4%, with 91% for the oak forest, 98.7% in the pine forest plantation, and 99.7% for the agricultural field, without significant differences among treatments. The oak forest was the only treatment where the logistic model for the relationship between survival and radiation was significant (p < 0.00001), with mortal-ity increasing at higher shade levels (Figure 1). However, drought influences survival at different light levels. In dry conditions (for example, derived from late tree planting), shrub survival is higher under shade than in the opening (Hernández-García and Rodríguez-Trejo 2008).

One year after planting, in all cases large plants aver-aged more shoot biomass than small plants (p < 0.05): oak forest (10 g and 5.2 g, respectively), pine forest plantation (13.3 g and 8.7 g), and agricultural field (7.5 g and 4.9 g), with the higher biomass among treatments for the pine plantation (p < 0.05). Root biomass behaved similarly: oak forest (6.5 g and 3.2 g, respectively), pine forest plantation (9.0 g and 5.6 g), and agricultural field (6.5 g and 4.2 g), as well as total biomass (Table 1). The growth after one year, measured as percentage of original size, was higher for small plants and higher in the pine forest than in the other treatments (p < 0.05) (Table 1).

In our work, total solar radiation intermediate between the oak forest and the pine forest plantation seems to yield higher total biomass, although without differences in shoot/root dry weight ratio among conditions, indicating a semitolerance to shade by the vara de perlilla. In contrast, shade-tolerant species, like salal (gaultheria shallon) and evergreen huckleberry (Vaccinium ovatum), increase their shoot/root dry weight ratio under shade in comparison to direct solar radiation (Hawkins and Henry 2004).

In La Mesa, the vara de perlilla may be planted in partial shade or full sun, because it is semitolerant. Even at low light levels (for example, 2000 MJ/m2/y), as in dense oak forests, its one-year mortality probability is only 15%. Midlight levels will maximize shoot biomass and facili-tate high survival. In dry years, partial shade may help to prevent high mortalities.

AcknowledgmentsWe thank the community of La Mesa, for allowing the establish-ment and conduction of this work, and the Universidad Autónoma Chapingo, for the financial support for developing this project (Project number 07100413).

ReferencesHawkins, B. and G. Henry. 2004. Effect of nitrogen supply

and seedling survival in two evergreen, ericaceous species. scandinavian Journal of Forest Research 19:415–423.

Hernández-García, J.D. and D.A. Rodríguez-Trejo. 2008. Radiación solar y supervivencia en una plantación de vara de perlilla (symphoricarpos microphyllus H.B.K.). Revista chapingo. serie ciencias Forestales y del ambiente 14:27–31.

Germination, Emergence, and Survival of Buddleja cordata in an Urban ForestPedro E. Mendoza-Hernández (Departamento de Ecología y Recursos naturales, Universidad nacional autónoma de México [UnaM], [email protected]), alma oro-zco-segovia (Departamento de Ecología Funcional, UnaM, [email protected]) and irene Pisanty (Departamento de Ecología y Recursos naturales, Facultad de ciencias, UnaM, ciudad Universitaria, México, DF 04510, +52 (55) 56 22 49 12, [email protected])

Mexico City, one of the world’s largest cities, has grown at the expense of different ecosystems. The

Parque Ecológico de la Ciudad de México (PECM) is located on Mt. Ajusco south of Mexico City (19°14'N and 99°15'W, elev. 2,600–2,900 m asl), one of the sur-rounding mountains that are important for conserving biodiversity and recharging aquifers. This park includes 730 ha, 200 of which were deeply disturbed in the early 1980s by an urban shantytown settlement (Cano-Santana et al. 2006).

Vegetation includes oak, pine, and pine-oak forests as well as xerophylous shrubland growing on basaltic sub-strate. Soils are well developed in the forests but scarce in the shrublands. The basaltic substrate has low water

Table 1. Total biomass (± SD) and growth increment of vara de perlilla (Symphoricarpos microphyllus) by treatment and plant size (L = large, S = small) in Central Mexico. Total biomass at one year was significantly different from initial biomass for all treatments and sizes (p < 0.05).

Biomass (g) IncrementTreatment Size Initial 1 y (g) (%)

Pine forest L 4.5 ± 1.3 22.3 ± 10.9 17.8 496S 2.1 ± 0.8 14.3 ± 9.9 12.2 681

Oak forest L 4.5 ± 1.3 16.5 ± 9.7 12.0 367S 2.1 ± 0.8 8.4 ± 4.2 6.3 400

Agricultural field L 4.5 ± 1.3 13.9 ± 4.0 9.4 309S 2.1 ± 0.8 9.2 ± 3.4 7.1 438


retention and wide temperature variation (0°C–50°C) (Olvera-Carrillo et al. 2009). Average annual temperature is 12°C–14°C, and annual precipitation, concentrated in the summer rainy season, is 700–1,000 mm. In the most disturbed area, volcanic rock was fragmented and removed. The most common pioneer trees in this area are known locally as tepozán (Buddleja cordata) and chapulixtle (Dodo-naea viscosa), while mala mujer (Wigandia urens) is the most common shrub. However, the harsh environmental conditions make plant recruitment difficult, and interven-tion to restore this area is necessary.

Tepozán is common in disturbed habitats from Mexico and Guatemala (Rzedowski and Rzedowski 2001). It flow-ers between July and December, and seed shedding occurs from October to March. The tiny seeds (0.0278 ± 0.0005 mg) are wind dispersed, need light to germinate, have high viability, and are partially dormant after collection (González-Zertuche et al. 2002). Tepozán has high litter production and growth rates, which make it a suitable species to restore ecosystem structure and functionality in the area (Vázquez-Yanes and Bátis 1996, Mendoza-Hernández 2003)

To make better use of tepozán in the restoration of disturbed shrubland, we tested the effects of seed age and collection time on the germination of this species and measured the effects of substrate and watering on the emergence and survival of seedlings. We seek a better understanding of these early life stages that are crucial to establishment of this important pioneer species.

We collected seeds from ten trees in October 1992 (autumn), November 1993 (autumn), January and Feb-ruary 1994 (winter), and March 1995 (winter). Seeds were cleaned, air dried, and stored at room temperature (23°C–25°C; 20%–50% RH), until use in June 1995. To compare germination among seeds produced during the same season, we used seeds collected in November 1993 and January and February 1994. Seeds from each collection date were divided into three replicates of 50 seeds each and germinated in Petri dishes on a 1% agar plate in a growth chamber (Model 125L, Conviron, Winnipeg, Canada) at 25°C and 12 h of light. We recorded the number of germinated seeds every three days. Final germination per-centages were arcsine transformed and compared using an ANOVA test (Statistica, vers. 5, Statsoft, Tulsa OK). Under constant temperature, seeds germinated simultaneously, beginning on the fifth day and ending on the tenth day. Seeds collected in the rainy autumn had higher germina-tion percentages (77.8 ± 2.4) than seeds collected in the dry winter (35.6 ± 1.1 and 28.5 ± 2.5, for the January and February seed lots, respectively) (F2,8 = 22.937, p = 0.0001), suggesting that environmental conditions had a significant effect on seed dormancy and vigor.

Next, we determined the effect of soil type, watering regime, and seed age on seedling emergence. We filled 24 pots with forest soil; half were sown with seeds collected in 1992 and the other 12 with seeds collected in 1995, using 50 seeds per pot. Pots were placed in a shadehouse in the PECM. To emulate precipitation variability, six pots from each cohort were watered every three days (frequent water-ing), and the other six only once a week (scarce watering). The same procedure was carried out with 24 pots filled with soil from the transition zone (between the forest and the secondary shrubland). Seedling emergence was recorded every three days. Emergence and survival percentages were arcsine transformed and compared using an ANOVA test (Figure 1). There were no significant differences in emergence between seed batches (F1,47 = 0.21, p = 0.64). Watering frequency and soil origin, however, significantly interacted to influence seedling emergence (F1,47 = 8.7, p = 0.002). Seeds sown in soil from the transition zone had higher emergence rates regardless of watering frequency. Seedling emergence began on the 14th day under frequent watering, while under scarce watering it started after 16 and 18 days on soil from the transition zone and forest, respectively.

All seedlings from the scarce-watering regime died one month after emergence (Figure 2). Seedlings from the 1992 collection growing under frequent watering and on forest soil showed the lowest mortality rate. The highest mortality rate occurred in seedlings from the 1995 collec-tion growing on the transition zone soil (F3,19 = 13.68, p = 0.0001). In the shadehouse, a high level of emergence was observed under frequent watering, probably because this species requires a high soil water potential to germinate

Figure 1. Emergence of tepozán (Buddleja cordata) seedlings in a shadehouse in 1995 in Parque Ecológico de la Ciudad de México from seeds collected in 1992 (●, ○) and 1995 (■, □) under different treatment conditions: A) transition soil and watering every three days (frequent); B) transition soil, watering every seven days (scarce); C) forest soil, frequent watering; and D) forest soil, scarce watering. Small letters in the graphs indicate significant differences.


(González-Zertuche et al. 2002). Soil in the transition zone is porous and contains small pebbles of volcanic rock, thus offering more root emergence opportunities for the seedlings than the more compact forest soil.

Under frequent watering, seedlings coming from seeds collected in autumn had higher survival rates than those coming from seeds produced in winter (F3,19 = 47.15, p = 0.00001). Seedling survival was low on soil from the transi-tion zone, where this species typically grows, and higher on forest soil, where it is rarely found, probably owing to low light conditions under the closed canopy. These results indicate that there is a fine balance between the requirements for seed germination and those for seedling establishment.

Since tepozán can be used for restoration in harsh envi-ronments with thin, undeveloped soils, such as the PECM

shrublands, we highly recommend nursery production using seeds collected in the autumn and forest soil mixed with coarse sand or small pieces of volcanic rock to increase germination and seedling emergence and survival. Consid-ering the small size of seeds and the fragility of seedlings, we recommend producing saplings grown from seeds in pots and transplanting them to the disturbed sites, along with the soil in which they grew. Additionally, seeds can be stored for at least three years without detrimental effects on their viability and vigor, allowing for large seed collections and nursery production over time.

AcknowledgmentsThis study was supported by the DGAPA IN 209292, Universidad Nacional Autónoma de México research program. We are grateful to María Esther Sánchez-Coronado for technical support, and to Chris Reyes and Roberto Lindig for their contribution to the improvement of the manuscript.

ReferencesCano-Santana, Z., I. Pisanty, S. Segura, P.E. Mendoza-

Hernández, R. León-Rico et al. 2006. Ecología, conservación, restauración y manejo de las áreas naturales y protegidas del Pedregal del Xitle. Pages 203–226 in K. Oyama and A. Castillo (eds), Manejo, conservación y Restauración de Recursos naturales en México. Mexico City: UNAM-Siglo XXI.

González-Zertuche, L., A. Orozco-Segovia, C.C. Baskin and J.M. Baskin. 2002. Effects of priming on germination of Buddleja cordata ssp. cordata (Loganiaceae) seeds and possible ecological significance. seed science and technology 30:535–548.

Mendoza-Hernández, P.E. 2003. El tepozán. ciencias 70:32–33.Olvera-Carrillo, Y., I. Méndez, M.E. Sánchez-Coronado,

J. Márquez-Guzmán, V.L. Barradas et al. 2009. Effect of environmental heterogeneity on field germination of opuntia tomentosa (Cactaceae, Opuntioideae) seeds. Journal of arid Environments 73:414–420.

Rzedowski, J. and G. Rzedowski. 2001. Flora Fanerogámica del Valle de México. Pátzcuaro MIC: Instituto de Ecología and CONABIO.

Vázquez-Yanes, C. and A.I. Bátis 1996. La restauración de la vegetación. ciencias 43:16–23.

Figure 2. Survival of tepozán (Buddleja cordata) seedlings in a shade house in 1995 in Parque Ecológico de la Ciudad de México from seeds collected in 1992 and 1995. Seedlings were grown in pots filled with forest soil (open symbols) and in transition soil (closed symbols). Small letters indicate significant differences.

riparian corridor-channel restoration and management in elm creek, minnesota

Documents