2013 NORTHEASTERN NATURALIST 20(3):397–418 Rapid Assessment Tools for Conserving Woodland Vernal Pools in the Northern Blue Ridge Mountains Erik D. Lindquist1,*, David K. Foster1, Samuel P. Wilcock2, and Jeffrey S. Erikson1 Abstract - Woodland vernal pools, small-scale forested depressions that flood in the spring and dry as summer progresses, are an essential and rapidly disappearing component of breeding habitat for numerous amphibian and macroinvertebrate species. Here we propose five multiple linear regression models to assist in the rapid assessment of vernal pools to conserve those with the greatest amphibian, aquatic macroinvertebrate, and herbaceous vegetation features. These models are based on samples taken from 21 pools distributed throughout the South Mountain landscape, a largely forested landscape matrix in south-central Pennsylvania. A comparison of the vegetative community of vernal pools and upland sites using Morisita’s index of community similarity shows them to be quite distinct. Based on our analyses and models, we propose conserving woodland vernal pools that possess plant species indicative of seasonal inundation, a large water volume when inundated, high tree cover, presence of coarse woody debris, high phosphate level in the water, and low sphagnum presence along the pool perimeter. Vernal pools with high herbaceous, shrub, and tree diversity likewise predict high amphibian productivity (abundance). This rapid assessment is expected to be a set of invaluable tools for identifying and ranking woodland vernal pools for state and federal conservation agencies. Introduction Pennsylvania possesses roughly 87,000 ha (215,000 acres) of forested wetlands that boast a wide diversity of plant and animal species, including many of threatened and endangered status (PA DEP 2003). Despite the predominance of forest wetlands, the US Environmental Protection Agency (1996) reported that 56% of wetlands had been lost from Pennsylvania between the 1780s and the 1980s. Woodland vernal pools are small-scale ecosystems—natural depressions which are inundated with rainwater in the winter and spring months and desiccate throughout the summer months (Colburn 2004, Purer 1939). Studies on vernal pools have shown that unique species assemblages of plants and animals exist in such ecosystems, where the aquatic community is devoid of species requiring permanent water impoundments (De Meester et al. 2005). The distinctiveness of assemblages in small-sized, short-hydroperiod pools has been empirically confirmed in amphibians (Gibbs 1993, Semlitsch et al.1996, Snodgrass et al. 2000) and plants (Bauder 1989, Zhulidov et al. 1997). The role of such wetlands has been demonstrated to be important for recruitment in many amphibian meta- 1Department of Biological Sciences, Messiah College, Grantham, PA 17027. 2Department of Information and Mathematical Sciences, Messiah College, Grantham, PA 17027. *Corresponding author - quist@messiah.edu. E.D. Lindquist, D.K. Foster, S.P. Wilcock, and J.S. Erikson 2013 Northeastern Naturalist Vol. 20, No. 3 398 populations, as well as the persistence of unique macroinvertebrate communities (Boorse 1999, Dehoney and Lavigne 1984, King et al. 1996, Skelly et al. 2005, Werner et al. 2007). The degree of ephemerality, or hydroperiod, of such wetlands is thought to be an abiotic predictor of species presence (Bauder 2005 Hargeby 1990, Pechmann et al. 1989, Schneider and Frost 1996, Van Buskirk 2005, Werner et al. 2007). Plant species that inhabit vernal pools are those that are capable of enduring periods of soil saturation and desiccation (Bliss and Zedler 1998, Colburn 2004, Zedler 1981). The animals that occupy vernal pools and adjacent areas are not only adapted to these hydrological extremes, but rely upon temporal fluctuations to avoid predation and/or competitors (Ebert and Balko 1984; Grodhaus 1980; Schneider 1990, 1999; Skelly 1996; Wiggins et al. 1980; Wilbur 1972,1987). Woodlands adjacent to vernal pools often appear more as a combination of wetland forest (palustrine or riparian) and upland deciduous forest due to higher local ground saturation. These forests represent important terrestrial refugia for a variety of species requiring pools for some period of their life cycle (Colburn 2004). It is expected that localized densities would be higher for some species of reptiles, amphibians, aquatic macroinvertebrates, and herbaceous vegetation and ferns (combined as “herbaceous vegetation” hereafter) requiring moist to saturated soils as compared to those of upland forest (McDiarmid 1994). These biota are important to forests because of their ecological uniqueness and connectivity through food webs and nutrient cycling (Colburn 2004). Vernal pools serve as key stopover sites for migrating birds, as well as foraging sites for resident mammals and birds (Colburn 2004). They also serve as potentially important habitat islands for plants and invertebrates. Vernal pools are associated with Mediterranean climates or glaciated terrain in the United States, yet in south-central Pennsylvania they exist in unglaciated, moist deciduous forest. Therefore, a community-study approach can provide insight on a distinct category of vernal pools. The encroachment of housing developments and agriculture on woodland vernal pool complexes along this northwestern edge of the range is visible by air (E.D. Lindquist, pers. observ.) and poses a serious threat to the biological diversity of the landscape (Azous and Horner 2001). We expect that the progressive loss and destruction (conversion) of this critical habitat is detrimental to a number of ephemeral forest pool species. The chief objective of our study was to develop landscape-level models that predict amphibian species diversity, species richness, and abundance in vernal pools and immediately adjacent forests. The secondary objective was to develop the same type of models, but for aquatic macroinvertebrate species richness, herbaceous vegetation species diversity, and herbaceous vegetation species richness. The overarching goal of these objectives was to assist conservation and land managers to rapidly assess, identify, and adopt best-practice policies for conservation. 399 E.D. Lindquist, D.K. Foster, S.P. Wilcock, and J.S. Erikson 2013 Northeastern Naturalist Vol. 20, No. 3 Methods Study sites In south-central Pennsylvania, the South Mountain landscape is physiographically unique in that it rises as the northern terminus of the Blue Ridge, ranging over 600 miles from the southern terminus in North Carolina. The landscape was identified by the Pennsylvania Bureau of Topologic and Geologic Survey of the Department of Conservation and Natural Resources as one of eleven distinct state ecological regions (Fike 1999). A forest matrix block of over 15,000 acres in the South Mountain landscape has emerged as a conservation priority for government agencies and non-governmental organizations alike, and has become a particular priority due to the pervasive wetland and vernal pool complexes on the landscape’s northwestern border. Through aerial orthophotoquad image analysis (of winter images), we counted and estimated that around 600 vernal pools were present on the South Mountain landscape before 1960. To our best knowledge, roughly 40% have been drained, filled, or otherwise removed as v iable habitat. Across the South Mountain landscape in Pennsylvania, we studied 21 vernal pools and adjacent terrestrial areas (vernal pool units) and 8 upland terrestrial areas (upland units) between 10 February 2005 and 10 February 2007 (29 study units total; Fig. 1). This effort followed a pilot study that identified sites from more than 75 vernal pools mapped by recent aerial photography, a leaf-off flyover in April 2004 ( E.D. Lindquist, pers. observ.), and field confirmation. Each Figure 1. The South Mountain landscape representing the terminus of the Blue Ridge. E.D. Lindquist, D.K. Foster, S.P. Wilcock, and J.S. Erikson 2013 Northeastern Naturalist Vol. 20, No. 3 400 woodland vernal pool study unit consisted of a vernal pool and its respective adjacent terrestrial area to ensure sufficient sampling evenness . Terrestrial transects within woodland vernal pool units were included to increase the probability of encountering vertebrate species that require vernal pools at some time throughout their life cycle as well as plants adapted to intermittently saturated soils. Within a geographic site where vernal pool complexes are extensive, study pools were randomly selected to minimize observer bias. Upland sites were located minimally 500 m away from vernal pool areas. Study sites (some with multiple pool areas) were located on both public and private lands in order to identify potential biodiversity issues with respect to land use. Permission of landowners and public lands authorities was secured before site inclusion in the study. Time, date, geographic position (lat./long.), and weather conditions were noted for each vernal pool and upland unit. Notes on slope, aspect, drainage, and prominent geological features were also recorded. Each pool and terrestrial transect was photographed from a south position looking north for reference and archival purposes. Potential and actual human features, threats, and impacts (e.g., houses, roads, mines, clearcuts, etc.) were also noted. Between 1 April and 30 September (high field season), each site was visited biweekly. Site visits for physical measurements also occurred during the remainder of the year (low field season). Woodland vernal pool sampling Vegetation, aquatic macroinvertebrate, and vertebrate species in each vernal pool and within 1 m of the shoreline were identified during sampling periods. Rooted plants within and around the pool (to 1 m outside shoreline) were identified. Canopy closure and DBH of canopy trees found within the vernal pool sampling were noted. This survey was repeated twice during the study period (May and July 2005) to capture peak cover for spring and summer vegetation. In order to minimize impact on each vernal pool, sampling with a 500-μmmesh D-frame dip net occurred at one end of a pool so as not to stir up sediment across the pool. We attempted to standardize a five-minute dip-net sampling protocol in the following manner: first, along the shore for 1 m in the water column, then a 1-m return sweep through the pool bottom (with vertical chopping motions), and then a final return sweep back through the water column. Dip net contents were spread over white plastic sorting trays to separate macroinvertebrates and vertebrates from the leafy detritus. This procedure was repeated in three different locations to get representation of macroinvertebrate activity. Extensive leaf and net inspections were conducted to thoroughly inventory the sample. Representative vertebrate individuals were keyed out to the species level and returned to the location they were taken from. During the egg-laying period for ambystomatid (mole) salamanders and Lithobates sylvaticus (Wood Frog), censuses of visible egg masses were conducted. Macroinvertebrates captured by net sampling were preserved in 70% ethanol for laboratory identification. Macroinvertebrates were identified to the lowest 401 E.D. Lindquist, D.K. Foster, S.P. Wilcock, and J.S. Erikson 2013 Northeastern Naturalist Vol. 20, No. 3 taxonomic level using Merritt and Cummins (1996) and Pennak (1989). Aquatic sampling was conducted up to drying events. Because a central goal of this study was to construct a rapid assessment tool and because we sought to minimize pool turbidity and maintain a food source for larval amphibians, aquatic macroinvertebrate species richness was determined without assessing quantitative abundance of species. Before biotic sampling commenced, water in each vernal pool was tested for nitrates, phosphates, sulfates, turbidity, and pH. Nitrates, sulfates, dissolved oxygen, phosphates, and turbidity were measured using a HACH DR-850 Colorimeter (HACH Company, Loveland, CO). Water pH was tested using a portable Oakton Instruments Basic pH Tester (Oakton Instruments, Vernon Hills, IL). Water temperature was recorded with a standard glass organic chemistry thermometer. At the end of each day of data collection, the daily high and low temperature, wind speed and direction, sky conditions, precipitation, dew point, relative humidity, and barometric pressure were gathered from the National Weather Service reporting stations for York and Harrisburg, PA and Hagerstown, MD. Lastly, a Trimble ProXR/TSCe (Trimble, Sunnyvale, CA) was used to obtain precision shoreline positions by point averaging, and GPS Pathfinder Office v. 4.00 (Trimble, Sunnyvale, CA) was used to calculate the area of each vernal pool during peak period of rainfall of the second year. GPS measurements were done once during the study at approximate maximum water levels. Average pool depth was calculated by taking measurements at meter intervals between shorelines. Volume was estimated by multiplying the pool area by the average pool depth (n = 17). Due to rapid pool drying in five ephemeral pools, we were unable to make volume measurements (n = 4), and in one pool, we were unable to determine dissolved oxygen. Volumetric and area relationships among the majority of vernal pool dimensions are given in Figure 3. Pool-related and pool-unrelated adjacent forest sampling Radiating from the pool shore, or from a central point for upland sites, a 50-m-diameter sampling area was established to conduct a modified point-quarter measure of associations of biological diversity between study units. Within the 50-m-diameter survey area, five randomly positioned 2-m x 2-m vegetation plots and one randomly positioned 5-m x 5-m herpetological plot was surveyed in each of four quadrants opening toward each cardinal bearing (N, E, S, and W). Each plot was randomized within its corresponding cardinal quadrant (Fig. 2). We estimated percent cover within vegetation plots for each vegetation size category (herbaceous, ferns, shrubs, and trees). Herpetological sampling included rolling and replacing coarse woody debris and stones, and temporarily raking away superficial leaves within plots. Macroinvertebrates were only sampled with dip nets in the woodland vernal pools and not in the terrestrial transects. Community type (Fike 1999), coarse woody debris cover, basal area factor, and browse condition estimates were noted. E.D. Lindquist, D.K. Foster, S.P. Wilcock, and J.S. Erikson 2013 Northeastern Naturalist Vol. 20, No. 3 402 Statistical analysis Statistical analysis proceeded in a three-pronged approach. First, Morisita’s index of community similarity (IM) (Brower and Zar 1984) was used to compare the similarity of herb, fern, and shrub/vine species composition from within 50 m of woodland vernal pools and from adjacent upland sites. We used herbaceous vegetation and shrub species, but not tree data, for the IM values (Brower and Zar 1984), as tree data integrates a broader area and often reflects the conditions near the pool instead of in it. Secondly, principal components analysis (PCA) using Figure 2. Vegetation and Herpetological Survey. Numbered diamonds in a vegetative survey represent 2-m x 2-m sample plots for herbaceous, shrub, seedling, sapling, pole-tree, and large-tree cover observations. Lettered diamonds indicate 5-m x 5-m herpetological sample plots. The dashed line around the vernal pool perimeter demarcates location for presence/absence survey of plant species within 2 m of the shoreline. Northern and eastern cardinal quadrants are illustrated; surveys were completed twice in each of the four cardinal quadrants. 403 E.D. Lindquist, D.K. Foster, S.P. Wilcock, and J.S. Erikson 2013 Northeastern Naturalist Vol. 20, No. 3 SAS® was used to ordinate sites along axes of principal variation in potentially important ecological variables. Variables considered are as follows: for vernal pool physical and chemical parameters (pool volume [m3], pool area [m2], pool perimeter [m], turbidity [FAU], pH, phosphates [mg/L PO4], nitrates [mg/L NO3], sulfates [mg/L SO4] and dissolved oxygen [DO]); and for pool perimeter and adjacent terrestrial plot (sum tree cover [%], tree species richness, tree species diversity [Shannon-Wiener index; H'], shrub cover [%], shrub species richness, shrub species diversity [H'], herbaceous vegetation cover [%], herbaceous vegetation species richness, herbaceous vegetation species diversity [H'], perimeter sphagnum cover [%], coarse woody debris [>30 cm diameter] cover [%], aquatic macroinvertebrate species richness, and reptile species richness). Lastly, multiple linear regression (MLR) with forward addition of variables using SAS® was employed to identify variables significantly correlated to amphibian species richness, herbaceous vegetation species richness, and aquatic macroinvertebrate species richness. Environmental parameters used in developing MLR models were the same used for PCA. PCA axis scores were not used as variables in MLR modeling. MLR models were constructed for each of the following parameters of interest: species diversity, species richness, and abundance of amphibians; species richness of aquatic macroinvertebrates; and species diversity and species richness of herbaceous vegetation. These six models represent important ecological features for conservation of vernal pools (Colburn 2004). Due to the low encounter rate of reptiles, statistical modeling with reptile species richness was not considered and reptile species diversity was not calculated (see Results, Zoological section). MLR models were developed by stepwise addition of the variables beginning with the one most correlated to the parameter of interest. Variables needed to have a maximum α value of 0.10 to enter the model and then to remain in subsequent steps as additional variables were added. MLR modeling for amphibians and aquatic macroinvertebrates was calculated based on the 16 pools that had data sets with volume and DO values (see Woodland vernal pool sampling section). The procedures determining the predictors of herbaceous vegetation species diversity and species richness excluded data on volume and DO because analyses including volume and DO were not found to be significant using the 16 pools that had these data. Therefore, the herbaceous vegetation models were calculated using data from all 21 woodland vernal pools. Results Botanical and zoological parameters showed marked variation across the woodland vernal pool and upland study sites. Physical and chemical parameters also varied substantially across these sites, providing a basis for MLR model construction (Table 1, Fig. 3). E.D. Lindquist, D.K. Foster, S.P. Wilcock, and J.S. Erikson 2013 Northeastern Naturalist Vol. 20, No. 3 404 Botanical At vernal pool study sites combined, we encountered 76 herbaceous plant species, 7 fern species, 33 shrub species, and 27 tree species. For upland sites, we identified 34 herbaceous plant species, 7 fern species, 18 shrub species, and 20 tree species. The ten most abundant herbaceous vegetation and shrub/woody vine species for the vernal pool and adjacent upland sites are similar (Table 2); however, their order varies considerably. More informative is the comparison of all species. Despite the ten most common species being similar, the vernal pool sites were botanically very different from adjacent upland sites. Of the sites having both vernal pools and adjacent upland sites (n = 8), the vernal pool sites sampled from the pool margin out to 50 m had 105 species of herbaceous vegetation and shrub/ Table 1. Chemical and physical parameter data for vernal pools (n = 21) in the study. Mean Std. dev. Minimum Maximum Turbidity (FAU) 71.8 54.8 26.0 239.0 pH 5.7 0.6 4.5 6.8 D.O. (mg/L; n = 20) 9.4 3.7 4.7 16.5 NO3 (mg/L) 22.3 23.5 3.7 80.9 PO4 (mg/L) 2.2 0.9 0.2 2.8 SO4 (mg/L) 73.1 18.0 14.0 80.0 Perimeter (m) 96.9 60.5 32.7 223.8 Area (m2) 428.4 397.0 49.7 1408.7 Volume (m3; n = 17) 167.8 166.4 15.5 475.3 Figure 3. XY scatter plot of vernal pool area to volume ratio for 17 of the vernal pools under consideration; four of the 21 vernal pools lacked volume data due to rapid pool drying. 405 E.D. Lindquist, D.K. Foster, S.P. Wilcock, and J.S. Erikson 2013 Northeastern Naturalist Vol. 20, No. 3 Table 2. The ten most common herbs and shrubs for vernal pool sites and their adjacent uplands. Although these appear similar, note the order is different and the average percent cover is quite different. Hay-scented Fern, a frequent indicator of heavy browsing by Odocoileus virginianus Zimmermann (White-tailed Deer), is the most abundant plant at both sites. Species Cover Vernal pool: 10 most common species Herbaceous vegetation Dennstaedtia punctilobula Michx. (Hay-scented Fern) 10.84 Osmunda cinnamomea L. (Cinnamon Fern) 5.32 Smilax rotundifolia L. (Greenbrier) 4.16 Symplocarpus foetidus L. (Skunk-cabbage) 1.08 Viola spp. (violets) 0.84 Maianthemum racemosum Link. (False Solomon’s-seal) 0.52 Polygonatum pubescens Willd. (Solomon's-seal) 0.40 Medeola virginiana L. (Indian Cucumber-root) 0.36 Smilax herbacea L. (Carrion-flower) 0.36 Veronica arvensis L. (Corn Speedwell) 0.36 Shrubs Vaccinium pallidum Aiton (Lowbush Blueberry) 7.40 Gaylussacia baccata Wangenh. (Black Huckleberry) 6.08 Vaccinium corymbosum L. (Highbush Blueberry) 3.00 Hamamelis viginiana L. (Witch-hazel) 2.88 Rubus hispidus L. (Swamp Dewberry) 2.68 Rubus pensilvanicus Poir. (Blackberry) 2.12 Rhododendron periclymenoides Michx. (Pinkster-flower) 1.04 Rhododendron spp. (rhododendrons) 1.04 Vaccinium stamineum L. (Deerberry) 0.68 Rubus enseleni Tratt. (Southern Dewberry) 0.68 Upland: 10 most common species Herbaceous vegetation Hay-scented Fern 33.84 Persicaria hydropiperoides Michx. (Water Smartweed) 7.36 Greenbrier 3.12 False Solomon’s-seal 2.56 Violet 2.44 Cinnamon Fern 2.08 Gaultheria procumbens L. (Teaberry) 1.88 Skunk-cabbage 1.52 Indian Cucumber 1.32 Amianthium muscaetoxicum Walt. (Fly-poison) 0.96 Shrubs Lowbush Blueberry 14.84 Witch-hazel 9.84 Black Huckleberry 9.12 Swamp Dewberry 5.16 Highbush Blueberry 4.64 Rhododendron 3.28 Ilex mucronata L. (Catberry) 1.92 Viburnum dentatum L. (Southern Arrow-wood) 0.96 Lindera benzoin L. (Spicebush) 0.92 Photinia melanocarpa Michx. (Black Chokeberry) 0.84 E.D. Lindquist, D.K. Foster, S.P. Wilcock, and J.S. Erikson 2013 Northeastern Naturalist Vol. 20, No. 3 406 woody vines, whereas the adjacent upland sites (500 m away) had 66 species, and these sites had 51 species in common. This result means that the vernal pool sites had 54 species not found in the adjacent uplands. The mean IM for any of the eight upland sites sampled compared with the composite species of all adjacent vernal pool sites was low, at 14.4%, with a maximum on one site of 32.4% and a minimum value of 5.5%. The herbaceous vegetation and shrub/woody vine species for upland sites were thus quite different from vernal pool sites. Species composition for vernal pool botanical communities was similar to the summary species lists published by Colburn (2004). Zoological At woodland vernal pool study sites, we encountered 15 amphibian species and 9 reptile species in our surveys. For upland sites, we identified 6 amphibian species and 1 reptile species in our surveys. Amphibians were seen in 25 of the 29 sampling units, with only one of the four absence sites being a vernal pool study unit. Reptiles were detected much less frequently, with species present in only 7 of the 29 sites. Of these sites, six were vernal pool units. Table 3 summarizes the most common amphibians and reptiles, and their frequency of encounter at vernal pools and adjacent upland sites. Thirty-five species of aquatic macroinvertebrates were observed within the vernal pools. Table 4 summarizes the ten most frequently encountered taxa. Of particular note, mosquito larvae in the family Culicidae were only the 19th most Table 3. The most common amphibians and reptiles for vernal pool sites. Names follow Collins and Taggart (2009). Species Frequency Amphibians Lithobates sylvaticus LeConte (Wood Frog) 0.76 Anaxyrus americanus Holbrook (American Toad) 0.43 Plethodon cinereus (Green) (Northern Redback Salamander) 0.38 Lithobates clamitans melanotus (Rafinesque) (Green Frog) 0.19 Ambystoma maculatum (Shaw) (Spotted Salamander) 0.14 Plethodon glutinosus (Green) (Northern Slimy Salamander) 0.14 Notophthalmus viridescens (Rafinesque) (Eastern Newt) 0.10 Eurycea bislineata (Green) (Northern Two-lined Salamander) 0.10 Eurycea longicauda (Green) (Long-tailed Salamander) 0.10 Hemidactylium scutatum (Temminck & Schlegel) (Four-toed Salamander) 0.10 Pseudacris crucifer (Wied-Neuwied) (Spring Peeper) 0.10 Lithobates catesbeianus Shaw (Bullfrog) 0.10 Ambystoma opacum Gravenhorst (Marbled Salamander) 0.05 Reptiles Diadophis punctatus edwardsii (Merrem) (Northern Ringneck Snake) 0.29 Terrapene carolina (L.) (Eastern Box Turtle) 0.19 Glyptemys insculpta (LeConte) (Wood Turtle) 0.10 Thamnophis sirtalis (L.) (Common Garter Snake) 0.10 407 E.D. Lindquist, D.K. Foster, S.P. Wilcock, and J.S. Erikson 2013 Northeastern Naturalist Vol. 20, No. 3 abundant macroinvertebrate, occurring in only 10% of vernal pools sampled. Two functional feeding groups, predators (Coleoptera, Diptera, Odonata, and Hemiptera) and collector-gatherers (Diptera, Ephemeroptera, and Naididae) were found in 20 of 21 pools. Predators were the most common (47.7%) followed by collector gatherers (30.5%). The least abundant functional feeding group was scrapers (Gastropoda) (2.27%), found in 4 of 21 pools. Most of the aquatic macroinvertebrates encountered are capable of living in low-oxygen waters. Forty-seven percent of organisms found had tolerance values of 5 or lower (Yuan 2004). Ephemeroptera and Trichoptera occurred in 10% or less of the samples and were amongst the least frequently encountered in the study. These two orders were found in four vernal pools that occasionally received streamwater spillover during periods of high precipitation. Species composition for vernal pool zoological communities was on par with summary lists published by Colburn (2004). Some deviation from this general trend occurred with the same four woodland vernal pools that occasionally received input from nearby streams during flooding events. Typically riparian amphibian species such as Eurycea bislineata (Northern Two-lined Salamander) and Eurycea longicauda (Long-tailed Salamander) were encountered at some of these sites. Principal components analysis PCA of sites by physical, chemical, vegetative, and faunal parameters indicated that there were three principal axes of variation ordinating the sites. In order, these were 1. variation in tree species diversity (H'), 2. variation in herbaceous vegetation species diversity (H'), and 3. variation in shrub species diversity (H'). Figure 4 shows XY scatterplots of the eigenvalues for sites along each axis. The scatterplots suggest that there is no strong correlation between these axes, a finding also confirmed by the fact that neither tree species diversity nor shrub species diversity axes emerge as predictive variables in MLR for herbaceous Table 4. Ten most frequently encountered macroinvertebrate taxa at vernal pool sites. Mosquitoes did occur but in 10% or less of sites, they were the 19th most abundant macroinvertebrate. Likewise, the majority of the most abundant organisms in this table are indicators of low oxygen content, especially Chironomid midges, here found in 86% of the pools sampled. Macroinvertebrate taxa Frequency Diptera: Chironomidae 0.86 Oligochaeta 0.67 Hemiptera: Notonectidae: Notonecta 0.57 Coleoptera: Dytiscidae: Acilius 0.52 Odonata: Aeshnidae: Aeshna 0.43 Anostraca: Branchiopoda: Eubranchipus 0.38 Diptera: Chaoboridae: Chaoborus 0.33 Diptera: Chaoboridae: Mochlonyx 0.33 Trichoptera: Phryganeidae: Ptilostomis 0.33 E.D. Lindquist, D.K. Foster, S.P. Wilcock, and J.S. Erikson 2013 Northeastern Naturalist Vol. 20, No. 3 408 Figure 4. Principal components analysis of sites by physical, chemical, vegetative, and faunal parameters revealed three strong axes of variation across sites. Sites are plotted herein using eigenvalues of site scores along these axes, which were tree species diversity, herbaceous species diversity, and shrub species diversity. 409 E.D. Lindquist, D.K. Foster, S.P. Wilcock, and J.S. Erikson 2013 Northeastern Naturalist Vol. 20, No. 3 species diversity (see below). We did not use the axis PCA scores for MLR modeling because our objective was to directly examine variation in amphibian species diversity, amphibian species richness, amphibian abundance, and aquatic macroinvertebrate species richness. Multiple linear regression models Six MLR models were constructed in this study, five of which are strongly predictive (with an r2 ≥ 0.6%, explaining greater than 60% the variability in the parameter of interest) and useful for faunal conservation and landscape management. Regression analysis does not assume that the raw variables are normally distributed, but does assume that the residuals of the model follow a normal distribution. For this reason, the residuals for each of the predictive models were checked for departures from normality. For each model, the residuals did not exhibit any indication of significant departure from normality. Therefore, there is no cause for concern about the normality assumption in any of the models presented in this paper. Each model is generally explained below and is in greater detail in Table 5. 1) Amphibian diversity. 72.75% of variability in amphibian diversity (AD) (H') is explained by the following linear model: AD = -0.857 + 0.055 herbaceous vegetation species richness + 0.001 percent tree cover + 0.397 mg/L PO4 - 0.236 percent perimeter sphagnum cover. 2) Amphibian richness. 89.42% of variability in amphibian diversity (AR) is explained by the following linear model: AR = -1.739 + 0.161 herbaceous vegetation species richness - 0.007 FAU turbidity + 1.222 mg/L PO4 + 0.005 m3 pool volume - 0.773 percent perimeter sphagnum cover. 3) Amphibian abundance. 88.23% of variability in amphibian diversity (AA) is explained by the following linear model: AA = -61.430 + 12.865 herbaceous vegetation species diversity + 7.814 shrub species diversity + 11.948 tree species diversity + 0.012 pool area + 2.694 coarse woody debris. 4) Macroinvertebrate species richness. 61.53% of variability in macroinvertebrate diversity (MR) is explained by the following linear model: MR = -17.945 + 0.255 tree species richness + 0.008 percent tree cover + 3.577 pH. 5) Herbaceous vegetation species diversity 16.09% of variability in herbaceous vegetation and fern species diversity (HFD) is explained by the following linear model: HFD = 1.576 + 0.094 percent perimeter sphagnum cover for the site. This model is not strongly predictive of herbaceous species diversity. 6) Herbaceous vegetation species richness. 67.02% of variability in herbaceous vegetation and fern species richness (HFR) is explained by the following linear model: HFR = -14.241 + 6.025 pH - 0.191 mg/L NO3 + 1.181 percent perimeter sphagnum cover for the site. E.D. Lindquist, D.K. Foster, S.P. Wilcock, and J.S. Erikson 2013 Northeastern Naturalist Vol. 20, No. 3 410 Table 5. Multiple linear regression models predicting value for six dependent variables important for assessing the health of vernal pools. For this study, conservation of vernal pools and immediately adjacent areas prioritized amphibian species diversity, amphibian species richness, amphibian abundance, macroinvertebrate species richness, herbaceous vegetation species diversity, and herbaceous vegetation species richness as the variables of interest. Variables added to the model intercept are listed by column in the order they were brought in through stepwise analysis (see Methods, Statistical analysis subsection). HerbVegSR = herbaceous vegetation species richness, HerbVegSD = herbaceous vegetation species diversity, ShrubSD = shrub species diversity, Tree SR = tree species richness, Tree SD = tree species diversity, CWD = coarse woody debris. Variable 1 Variable 2 Variable 3 Variable 4 Variable 5 Model dependent variable n Intercept (coefficient) (coefficient) (coefficient) (coefficient) (coefficient) r2 F P Amphibian species 16 -0.85665 HerbVegSR Tree cover PO4 Sphagnum - 0.7275 6.83 0.0061 diversity (0.05501) (0.00117) (0.39654) (-0.23628) Amphibian species 16 -1.73893 HerbVegSR Turbidity PO4 Pool volume Sphagnum 0.8294 9.72 0.0013 richness (0.16144) (-0.00655) (1.22235) (0.00511) (-0.77286) Amphibian abundance 16 -61.42953 HerbVegSD ShrubSD TreeSD Pool area CWD 0.8823 14.99 0.0002 (12.86532) (7.81390) (11.94848) (0.01172) (2.69367) Macroinvertebrate 16 -17.94493 Tree SR Tree cover pH - - 0.6153 6.4 0.0078 species richness (0.25457) (0.00841) (3.57680) Herbaceous vegetation 21 1.57635 Sphagnum - - - - 0.1609 3.64 0.0715 species diversity (0.09439) Herbaceous vegetation 21 -14.24138 pH NO3 Sphagnum - - 0.6702 11.52 0.0002 species richness (6.02525) (-0.19115) (1.18116) 411 E.D. Lindquist, D.K. Foster, S.P. Wilcock, and J.S. Erikson 2013 Northeastern Naturalist Vol. 20, No. 3 Table 6. Herbaceous vegetation species, shrub species, and tree species most correlated with amphibian species diversity, amphibian species richness, amphibian abundance, and macroinvertebrate species richness. Species that were included had a P value of 0.1 or less. The Pearson correlation coefficient (r), P-value, and Region 1 Indicator status (USACE 2012) is provided for each species and by each model dependent variable. Taxonomic designations follow Rhoads and Block (2007). Region 1 Dependent variable/Species r P indicator Herbaceous vegetation Amphibian species diversity Apocynum androsaemifolium L. (Spreading Dogbane) 0.44233 0.0447 Circaea alpina L. (Enchanter's Nightshade) -0.46583 0.0333 FACW Parthenocissus quinquefolia L. (Virginia Creeper) -0.37534 0.0936 FACU Pilea pumila (L.) A. Gray (Canadian Clearweed) 0.40761 0.0666 FACW False Solomon’s-seal -0.38732 0.0828 FACU Amphibian species richness Spreading Dogbane 0.59154 0.0047 Enchanter's Nightshade -0.43536 0.0485 FACW Galium boreale L. (Northern Bedstraw) 0.37345 0.0954 FACU Glechoma hederacea L. (Ground Ivy) 0.37345 0.0954 FACU Virginia Creeper -0.40296 0.0701 FACU Canadian Clearweed 0.45722 0.0372 FACW Saururus cernuus L. (Lizard’s Tail) 0.37345 0.0954 OBL False Solomon’s-seal -0.39312 0.0779 FACU Stellaria holostea L. (Greater Stitchwort) 0.37345 0.0954 Thalictrum thalictroides L. (Rue Anemone) 0.37345 0.0954 Amphibian abundance Maianthemum canadense Desf. (Canada Mayflower) 0.72077 0.0002 FAC Macroinvertebrate species richness Chimaphila maculata L. (Striped Wintergreen) 0.37686 0.0922 Ageratina altissima L. (White-snakeroot) 0.42862 0.0525 FACU Northern Bedstraw 0.68214 0.0007 FACU Galium circaezans Michx. (Licorice Bedstraw) 0.4124 0.0632 UPL Ground Ivy 0.68214 0.0007 FACU Impatiens capensis Meerb. (Jewelweed) 0.66807 0.0009 FACW Oxalis acetosella L. (Northern Wood-sorrel) 0.46069 0.0356 FAC Canadian Clearweed 0.71348 0.0003 FACW Persicaria perfoliata L. (Mile-a-minute Weed) 0.39493 0.0764 FAC Lizard’s Tail 0.68214 0.0007 OBL False Solomon’s-seal -0.39533 0.0761 FACU Greater Stitchwort 0.68214 0.0007 Rue Anemone 0.68214 0.0007 Shrubs Amphibian species diversity Catberry -0.41161 0.0638 OBL Highbush Blueberry -0.43314 0.0498 FACWAmphibian species richness Catberry -0.37023 0.0985 OBL Highbush Blueberry -0.48082 0.0274 FACWAmphibian abundance - - - E.D. Lindquist, D.K. Foster, S.P. Wilcock, and J.S. Erikson 2013 Northeastern Naturalist Vol. 20, No. 3 412 Plant correlations Plant species correlating with amphibian species diversity. Statistically significant correlations between plants and AD are detailed in Table 6. Two features are clear from the plant species that are significantly correlated with AD. First, there are no obligate hydrophytic herbaceous species correlated to AD. Of the two species that are positively correlated with AD, Pilea pumila (Canadian Clearweed) is a facultative wetland species (FACW) and an annual plant, indicative of seasonal overabundance of soil moisture (Rhoads and Block 2007). Herbaceous species negatively correlated to AD tended to be those that do not normally occur in sodden ground. In the shrub layer, AD is negatively correlated to plants indicative of perennially wet soil, one of which (Ilex mucronata [Catberry]) is an obligate hydrophyte. In the tree layer, AD is negatively correlated only to two species, both of which are primarily grow on uplands, and positively correlated to Tsuga canadensis (Eastern Hemlock), which is a facultative upland species but occupies a wide range of soil moisture conditions, as its seedlings are highly sensitive to drought. Plant species correlating with amphibian species richness. AR repeats the same correlation patterns with herbaceous species as in AD, but with one exception: it is positively correlated with Saururus cernuus (Lizard’s Tail). This species, although not an annual, reaches peak abundance in seasonally flooded areas (Rhoads and Block 2007). Likewise, the same pattern is repeated between AR and shrub and tree species. Plant species correlating with amphibian abundance. AA does not correlate to most herbaceous vegetation species. However, it strongly correlates to one herbaceous species, Maianthemum canadense (Canada Mayflower), an herb that occupies similar site conditions to Eastern Hemlock (Rhoads and Block, 2007). No Table 6, continued. Region 1 Dependent variable/Species r P indicator Macroinvertebrate species richness Berberis thunbergii DC (Japanese Barberry) 0.44536 0.043 FACU Witch-hazel 0.4116 0.0638 FACTrees Amphibian species diversity Castanea dentata Marshall (American Chestnut) -0.39709 0.0747 Nyssa sylvatica Marshall (Blackgum) -0.4763 0.0291 FAC Tsuga canadensis L. (Eastern Hemlock) 0.39616 0.0754 FACU Amphibian species richness Blackgum -0.41172 0.0637 FAC Eastern Hemlock 0.40169 0.0711 FACU Amphibian abundance Betula lenta L. (Sweet Birch) 0.48519 0.0258 FACU Macroinvertebrate species richness Pinus strobus L. (Eastern White Pine) 0.54567 0.0105 FACU Quercus palustris Müenchh. (Pin Oak) -0.39287 0.0781 FACW Eastern Hemlock 0.52861 0.0138 FACU 413 E.D. Lindquist, D.K. Foster, S.P. Wilcock, and J.S. Erikson 2013 Northeastern Naturalist Vol. 20, No. 3 shrub species correlated with AA. The one tree species that correlated (positively) to AA was Betula lenta (Sweet Birch), a pattern which defies easy explanation. Plant species correlating with macroinvertebrate richness. The pattern of correlation of vegetative species to aquatic macroinvertebrates is less clear in that vegetative species occupy a wide variety of indicators in the US Army Corp of Engineers National Wetlands Plant List (2012). However, it is evident that MR, like AR, is positively correlated with Canadian Clearweed and Lizard’s Tail, both of which are indicators of seasonally wet or flooded ground. Discussion Our analysis suggests that species diversity, species richness, and abundance of amphibians; species richness of aquatic macroinvertebrates; and species richness of herbaceous vegetation in woodland vernal pool ecosystems may be predicted by a few relatively easily measured variables. Physical and chemical parameters such as area, volume, pH, and DO fell within the ranges summarized by Colburn (2004). Sulfate levels were low, while other parameters such as nitrate and phosphate revealed that the vernal pools in our study were eutrophic (Table 1; Swistock 2010). Principal components analysis PCA indicates that there are three primary axes differentiating sites from one another. These axes are all related to plant diversity features, namely tree species diversity, herbaceous species diversity, and shrub species diversity. While no regression analysis was performed of axis scores (eigenvalues) directly, all three of these vegetative features are significant predictors of amphibian abundance, as demonstrated from the MLR model (see Results). Diversity in herbaceous, shrub, and tree layers is an indicator of sites with conditions that promote higher abundance of amphibians. MLR modeling MLR modeling for amphibian species diversity indicates that the herbaceous species richness, tree cover, and phosphate levels were all positively correlated. Sites with high tree cover and high herbaceous richness indicate moderately high soil fertility, although soil fertility was not measured directly. High levels of nutrient influx presumably drive relatively high vernal pool phosphate levels via annual leaf fall (Colburn 2004). Conversely, amphibian species diversity is negatively correlated to percent perimeter sphagnum cover, an indicator of nutrient-poor, acidic soils. Amphibian species richness is predicted by similar coefficients, with the addition of positive correlation to vernal pool volume and negative correlation to turbidity. This model lacks percent tree cover as a variable. Amphibian species richness is thus greatest in larger pools with relatively little disturbance (turbidity) and relatively high nutrient conditions as indicated by high vernal pool phosphate level. Admittedly, the vernal pool volume parameter makes a rapid E.D. Lindquist, D.K. Foster, S.P. Wilcock, and J.S. Erikson 2013 Northeastern Naturalist Vol. 20, No. 3 414 assessment difficult to perform in the field. Were conservation and land managers to exclude this parameter from the assessment, the loss of predictability would be approximately 19%. As stated by PCA results, diversity in herbaceous, shrub, and tree layers is highly positively correlated with amphibian abundance. When vernal pool area and coarse woody debris are added to the model, this increased prediction of amphibian abundance further. Site conditions provide a variety of hydrological microenvironments that produce high plant diversity through a seasonally wet and a perennially dry soil mosaic. Aquatic macroinvertebrate species richness is positively correlated with tree species richness, percent tree cover, and pH. The percent tree cover and circumneutral pH indicate high productivity for plants in general and thus a higher potential input of biomass into the macroinvertebrate food chain in the vernal pool. Our secondary objective in this study was to look at assessing herbaceous species diversity and species richness. The MLR model generated for herbaceous species diversity did not satisfy the level of statistical significance needed for inclusion into the overall rapid assessment protocol. On the other hand, herbaceous species richness can be predicted across these sites by a relatively simple MLR model in which it is positively correlated to vernal pool pH and percent perimeter sphagnum cover, but negatively correlated to pool nitrate concentration. Higher pH of the pool and simultaneous presence of sphagnum around its perimeter may indicate a diversity of microenvironments for plant growth. It is, however, unclear why herbaceous species richness is negatively correlated with nitrate concentration in the vernal pool. Linkages between vernal pool chemical concentration and its impact on adjacent site vegetation are beyond the scope of this study, and may merit further study. Plant correlations Our analysis of the community similarity between woodland vernal pools and upland sites demonstrates that on average 14.4% of the herbaceous vegetation and shrub/woody vine species were very dissimilar. That said, amphibian species diversity and species richness are both positively correlated with herbaceous and tree species that are found in upland sites but can facultatively tolerate wet conditions and plant species that are found in wetland sites that are seasonally but not perennially inundated. On the other hand, Amphibian species diversity and species richness are negatively correlated to plant species that have their primary distributions on upland sites. Additionally, amphibian species diversity and species richness are negatively correlated with shrub species associated with perennial wetlands (Table 6). To summarize, woodland vernal pools that have vegetation indicating seasonal, but no permanent inundation should be the most species rich and diverse for amphibians. Interestingly, macroinvertebrate richness is most strongly correlated with a variety of herbaceous vegetation that together indicate a diversity of soil 415 E.D. Lindquist, D.K. Foster, S.P. Wilcock, and J.S. Erikson 2013 Northeastern Naturalist Vol. 20, No. 3 moisture conditions adjacent to the vernal pools. This finding is also reflected with correlations with shrub and tree species (Table 6). Conservation As stated in the methods, we estimated that roughly 40% of vernal pools have been destroyed. In light of this finding, perhaps the greatest benefit of this study is that it has provided five simple modeling tools for the rapid assessment of vernal pools to conserve those with the greatest amphibian, aquatic macroinvertebrate, and herbaceous vegetation features.Our findings show similar trends to those of Werner et al. (2007), Egan and Paton (2004), and Rowe and Dunson (1995); yet, our more simplistic rapid assessment models presented here can be used and implemented in one field season, committing fewer human and financial resources. Likewise, our models focus on a greater number of amphibian species, several measures of amphibian community health (diversity, richness, and abundance), aquatic macroinvertebrate species richness, and herbaceous species richness—all in order to broadly assess vernal pool community health. We therefore recommend to state and private land managers that these tools be considered for implementation for rapid identification of potentially ecologically important vernal pools, before they are threatened by impending development. In conclusion, conservation and land managers should be interested in preserving woodland vernal pools that possess the following features: plant species indicative of only seasonal inundation, a large water volume when inundated, high tree cover, presence of coarse woody debris greater than 30 cm in diameter, high phosphate concentration in the water, and low sphagnum presence along the pool perimeter. Vernal pools with high herbaceous, shrub, and tree diversity predict high amphibian productivity. Acknowledgments We would like to thank the Pennsylvania Chapter of The Nature Conservancy and the Pennsylvania Department of Conservation and Natural Resources for funding this research. We are indebted to Ashley Rosenberger, Michael Currie, Charles DeCurtis, Miranda Demos, Jarrod Derr, Scott Forbes, Rebecca Kern, Lindsay Johnson, Alyssa Poplaski, Brook Reeve, and Megan Starace for their invaluable assistance in planning and carrying out data collection. We also thank Anne Barrett, Todd Sampsell, Tracy Coleman, and Nels Johnson for their encouragement and advocacy for our conservation work in Pennsylvania. Literature Cited Azous, A.L., and R.R. Horner. 2001. Wetlands and Urbanization: Implications for the Future. Lewis Publishers, Boca Raton, FL. 338 pp. Bauder, E.T. 1989. Drought stress and competition effects on the local distribution of Pogogyne abramsii. Ecology 70:1083–1089. Bauder, E.T. 2005. The effects of an unpredictable precipitation regime on vernal pool hydrology. Freshwater Biology 50:2129–2135. Bliss, S., and P. Zedler. 1998. The germination process in vernal pools: Sensitivity to environmental conditions and effects on community structure. Oecologia 113:67-73. E.D. Lindquist, D.K. Foster, S.P. Wilcock, and J.S. Erikson 2013 Northeastern Naturalist Vol. 20, No. 3 416 Boorse D.F. 1999. Faunal community ecology in prairie pothole wetlands. Ph.D. Dissertation. University of Wisconsin-Madison, WI. 256 pp. Brower, J.E., and J.H. Zar. 1984. Field and Laboratory Methods for General Ecology, 2nd Edition. W.C. Brown Publishers, Dubuque, IA. 226 pp. Colburn, E.A. 2004. Vernal Pools: Natural History and Conservation. MacDonald and Woodward Publishing Co., Blacksburg, VA and Granville, OH. 426 pp. Collins, J.T., and T.W. Taggart. 2009. Standard Common and Scientific Names for North American Amphibians, Turtles, Reptiles, and Crocodilians. 6th Edition. The Center for North American Herpetology, Lawrence, KS. 44 pp. Dehoney, B. and D.M. Lavigne. 1984. Macroinvertebrate distribution among some southern California vernal pools. Pp. 154–160, In S. Jain and P. Moyle (Eds.). Vernal Pools and Intermittent Streams: A Symposium by the Institute of Ecology. University of California, Davis, CA. Publication No. 28. 280 pp. De Meester, L., S. Declerck, R. Stoks, G. Louette, F. Van De Meutter, T. De Bie, E. Michels, and L. Brendonck. 2005. Ponds and pools as model systems in conservation biology, ecology, and evolutionary biology. Aquatic Conservation: Marine and Freshwater Ecosystems 15:715–725. Department of Environmental Protection, Commonwealth of Pennsylvania (PA DEP). 2003. An introduction to wetlands. DEP Fact Sheet. Harrisburg, PA. 2 pp. Ebert, T.A., and M.L. Balko. 1984. Vernal pools as islands in space and time. Pp. 90–101, In S. Jain and P. Moyle (Eds.). Vernal Pools and Intermittent Streams: A Symposium by the Institute of Ecology. University of California, Davis, CA. Publication No. 28. 280 pp. Egan, R.S., and P.W.C. Paton. 2004. Within-pond parameters affecting oviposition by Wood Frogs and Spotted Salamanders. Wetlands 24(1):1–13. Fike, J. 1999. Terrestrial and palustrine plant communities of Pennsylvania. Pennsylvania Natural Diversity Inventory, Harrisburg, PA. 86 pp. Gibbs, J.P. 1993. Importance of small wetlands for the persistence of local populations of wetland-associated animals. Wetlands 13:25-31. Grodhaus, G. 1980. Aestivating chironomid larvae associated with vernal pools. Pp. 315–322, In D.A. Murray (Ed.). Chironomidae: Ecology, Systematics, Cytology, and Physiology. Pergamon Press, Oxford, UK. 374 pp. Hargeby, A. 1990. Macrophyte associated invertebrates and the effect of habitat permanence. Oikos 57:338–346. King, J.L., M.A. Simovich, and R.C. Brusca. 1996. Species richness, endemism, and ecology of crustacean assemblages in northern California vernal pools. Hydrobiologia 328:85–116. McDiarmid, R.W. 1994. Amphibian diversity and natural history: An overview. Pp. 5–15, In W.R. Heyer, M.A. Donnelly, R.W. McDiarmid, L.C. Hayek, and M.S. Foster (Eds.). Measuring and Monitoring Biological Diversity: Standard Methods for Amphibians. Smithsonian Institution Press, Washington, DC. 364 pp. Merritt, R.W., and K.W. Cummins (Eds.). 1996. An Introduction to the Aquatic Insects of North America, 3rd Edition. Kendall/Hunt Publishers, Dubuque, IA. 862 pp. Pechmann, J.H.K., D.E. Scott, J.W. Gibbons, and R.D. Semlitsch. 1989. Influence of wetland hydroperiod on diversity and abundance of metamorphosing juvenile salamanders. Wetlands Ecology Management 1:3–11. Pennak, R.W. 1989. Freshwater Invertebrates of the United States, 3rd Edition. J. Wiley and Sons, Inc., New York, NY. 628 pp. Purer, E.A. 1939. Ecological study of vernal pools, San Diego County. Ecology 20:217–229. 417 E.D. Lindquist, D.K. Foster, S.P. Wilcock, and J.S. Erikson 2013 Northeastern Naturalist Vol. 20, No. 3 Rhoads, A.F., and T.A. Block. 2007. The Plants of Pennsylvania: An Illustrated Manual, 2nd Edition. University of Pennsylvania Press, Philadelphia, PA. 1042 pp. Rowe, C.L., and W.A. Dunson. 1995. Impacts of hydroperiod on growth and survival of larval amphibians in temporary ponds of Central Pennsylvania, USA. Oecologia 102:397–403. Schneider, D.W. 1990. Habitat duration and the community ecology of temporary ponds. Ph.D. Dissertation. University of Wisconsin-Madison, WI. 191 pp. Schneider, D.W. 1999. Snowmelt ponds in Wisconsin: Influence of hydroperiod on invertebrate community structure. Pp. 299–317, In D.P. Batzer, R.B. Rader, and S.A. Wissinger, (Eds.). Invertebrates in Freshwater Wetlands of North America: Ecology and Management. John Wiley and Sons, New York, NY. 1100 pp. Schneider, D.W., and T.M. Frost. 1996. Habitat duration and community structure in temporary pools. Journal of the North American Benthological Society 15:64–86. Semlitsch, R.D., D.E. Scot, H.K. Pechmann, and J.W. Gibbons. 1996. Structure and dynamics of an amphibian community: Evidence from a 16-year study of a natural pond. Pp. 217–248, In M.L. Cody and J.A. Smallwood (Eds.). Long-term Studies of Vertebrate Communities. Academic Press, San Diego, CA. 597 pp. Skelly, D.K. 1996. Pond drying, predators, and the distribution of Pseuacris tadpoles. Copeia 1996:599–605. Skelly, D.K., M.A. Halverson, L.K. Freidenburg, and M.C. Urban. 2005. Canopy closure and amphibian diversity in forested wetlands. Wetlands Ecology Management 13:261–268. Snodgrass, J.W., M.J. Komoroski, A.L. Bryan, Jr., and J. Burger. 2000. Relationships among isolated wetland size, hydroperiod, and amphibian species richness. Conservation Biology 14(2):414–419. Swistock, B.R. 2010. Interpreting water tests for ponds and lakes. College of Agricultural Sciences, Cooperative Extension School of Forest Resources, The Pennsylvania State University, PA. 2 pp. United States Army Corps of Engineers (USACE). 2012. National wetlands plant list. Available online at http://rsgisias.crrel.usace.army.mil/NWPL/. Accessed 21 December 2012. United States Environmental Protection Agency: Office of Wetlands, Oceans, and Watersheds (USEPA). 1996. America’s wetlands: Our vital link between land and water. Washington, DC. 16 pp. Van Buskirk, J. 2005. Local and landscape influence on amphibian occurrence and abundance. Ecology 86(7):1936–1947. Werner, E.E., D.K. Skelly, R.A. Relyea, and K.L. Yurewicz. 2007. Amphibian species richness across environmental gradients. Oikos 116:1697–1712. Wiggins, G.B., R.J. Mackay, and I.M. Smith. 1980. Evolutionary and ecological strategies of animals in annual temporary pools. Archiv für Hydrobiologie Supplement 58:97–206. Wilbur, H.M. 1972. Competition, predation, and the structure of the Ambystoma-Rana sylvatica community. Ecology 53:3–21. Wilbur, H.M. 1987. Regulation of structure in complex systems: Experimental temporary pond communities. Ecology 68:1437-1452. Yuan, L.L. 2004. Assigning macroinvertebrate tolerance classifications using generalized additive models. Freshwater Biology 49:622–677. E.D. Lindquist, D.K. Foster, S.P. Wilcock, and J.S. Erikson 2013 Northeastern Naturalist Vol. 20, No. 3 418 Zedler, P.H. 1981. Micro-distribution of vernal pool plants of Kearny Mesa, San Diego County. Pp. 185–197, In S. Jainand P. Moyle (Eds.). Vernal Pools and Intermittent Streams: A Symposium by the Institute of Ecology. University of California, Davis, CA. Publication No. 28. 280 pp. Zhulidov, A.V., J.V. Headley, R.D. Roberts, A.M. Nikanorov, and A.A .Ischenko. 1997. Atlas of Russian Wetlands. M.J. Branned, M.J. (Ed.). Y.V. Flingeffman and O.V. Zhulidov (trans.). Environment Canada, National Hydrology Research Institute, Saskatoon, SK. 309 pp.