Baseline Quantitative Vegetation Monitoring in Terrestrial and Wetland Communities
in the North Chicago Wetland Mitigation Site, Lake County, Illinois
John B. Taft, Zachary Kron, and Allen Plocher
2 February 2010
Illinois Natural History Survey Technical Report 2010(07)
Biological Surveys and Monitoring Associated with
Illinois Toll Highway Construction Activities
Illinois Tollway Authority
2700 Ogden Avenue
Downers Grove, Illinois 60515-1703
A request was received in 2009 from the Illinois Toll Highway Authority and the Illinois Department of Transportation to establish a vegetation monitoring program at the North Chicago Wetland Mitigation Site in Lake County, Illinois (Figure 1). The principal goal of the monitoring is to further assess habitat quality in terrestrial and wetland communities throughout the site and to provide a baseline reference for monitoring vegetation change with planned habitat management. Chief objectives of the 2009 baseline monitoring were to collect and analyze data in ground-layer vegetation on species composition and diversity and to quantify composition, stem density, and percent canopy cover in the shrub stratum. Proposed habitat management includes removal of invasive shrubs, control efforts with selected herbaceous wetland species, and localized seeding of native prairie and wetland species followed by habitat management involving prescribed fire. Previous work at this site includes extensive botanical surveys (e.g., Taft 1996 and 2006) and wetland mapping (Olson et al. 1991; Plocher et al. 1996, Plocher and Ketzner 2006a, and Plocher and Ketzner 2006b). Three state threatened plant species previously have been recorded from the site (Taft 2006): Elymus trachycaulus, Oenothera perennis, and Veronica scutellata.
Sample Design - A stratified vegetation sampling design was utilized with 10 parallel transects running west to east, each separated by 500 meter intervals. Five sample points were established on each transect separated by 250 m (the 10th transect, furtherest to the north, had four sample points). This array provided 49 sampling stations including 37 in non-wetland terrestrial vegetation and 12 in areas previously determined (Plocher et al. 1996) to be jurisdictional wetlands (Figure 2). In addition, eight plots were established in prairie remnants in the far southern portion of the study area and five plots were placed in selected wetland communities, mostly in the southern half of the study area (Figure 2), for a total of 62 vegetation sample plots. Specific plot location with the additional targeted sampling was determined randomly.
Vegetation Sampling - Vegetation in both wetland and terrestrial habitats was sampled with 25-m2 (5m x 5m) sampling plots with ground layer quadrats (1-m2 [1 m x 1m]) nested within. The baseline point for all sample plots was the southwest corner of the shrub/sapling plots, corresponding to the geographic coordinates associated with plot locations (Figure 2). Plot sides were oriented along cardinal directions (the southern boundary runs W-E at 90̊). Composition and stem density of shrubs and saplings (all woody stems > 1-m tall and < 10 cm dbh) were sampled within the 25-m2 plots. For terrestrial vegetation plots, percent shrub cover was determined using digital photography with a hemi-view lens oriented vertically in the plot center to photograph the canopy of the plot area (narrowed with a lens tube). Percent visible sky and leaf area index (LAI) were calculated from these images using HemiView Canopy Analysis Software, ver. 2.1. Percent canopy cover was calculated as 100 - % visible sky. A horizontal habitat image also was taken of each plot oriented from the southwest to the northeast corner. These will provide comparative images following vegetation management including shrub removal and prescribed fire. For wetland plots, shrub cover was visually estimated. Trees (woody stems > 10 cm dbh), scarce in the study area, were sampled in 200-m2 (14.14 m x 14.14 m) sample plots (n = 4) anchored at the SW corner of the shrub plot.
Ground layer vegetation was sampled with 3 quadrats (2 for wetlands) nested within each shrub plot, with quadrat placement in the southwest and northeast corners and, in terrestrial samples, one in the plot center. In two transects (T10 and T9), terrestrial plots were sampled with 5 quadrats. Data collected from each quadrat included species presence and percent cover for each species occurrence estimated with a modified Daubenmire cover-class scale (0-1 %, 1-5%, 5-25%, 25-50%, 50-75%, 75-95%, 95-100%). All species rooted within each quadrat frame were recorded to species including woody species < 1-m tall.
Data Analysis - Species abundance is measured by Importance Value (IV 200), calculated as the sum of relative frequency and relative cover for ground-layer samples; for the shrub/sapling stratum, IV is calculated as the sum of relative frequency and density; and for trees, IV is calculated as the sum of relative density and basal area. Plot data are represented by both quadrat means (averaged among the 3  quadrats in each plot) and plot sum (combining data from all 3  quadrats). Cluster analysis (McCune and Mefford 1999) was utilized to produce a hierarchical classification of sites from the quantitative sample data based on the Sørensen similarity distance measure and flexible Beta linkage method (ß = -0.25). Indicator Species Analysis was used to determine non-random group affiliation for species with probability determined from 1,000 Monte Carlo permutations of the data. Indicator Values were calculated for each species with the following formula: Indicator Value = 100 (RA x RF), with RA = relative abundance and RF = relative frequency. A perfect indicator (IV = 100) would be a species that is both faithful (complete fidelity to a particular community type) and reliable (always present).
Vegetation data include parameters calculated at both quadrat and plot spatial scales. Parameters, defined below, include species richness (native and non-native), Shannon-Wiener Index of diversity (H́ [native species only]), Simpson’s Index of Dominance (including all species), and metrics for Floristic Quality Assessment (FQA) including calculations based on both native and all species. FQA metrics include Mean Coefficient of Conservatism and the Floristic Quality Index (Taft et al. 1997). Regarding parameters influenced by total richness (differing in total sample area), for plots sampled with 5 quadrats (most plots in transects 9 and 10) only 3 quadrats were used, as with all other terrestrial vegetation sampling.
Species-level metrics included native species richness, Shannon-Wiener diversity, Simpson’s dominance index, evenness, species richness index, native richness index, mean coefficient of conservatism, and floristic quality index (Whittaker 1975; Taft et al. 2006) were calculated as follows:
Native Species Richness: Total number of native species in a sample unit,
Shannon-Wiener Index of Diversity (H'): -∑ [pi ln(pi)], where pi is the relative abundance of each native species (based on importance values [IV200] calculated as the sum of relative cover and relative frequency),
Simpson’s Dominance Index: ∑ pi2, where pi is the relative importance value for each species in the sample area (transect),
Evenness: H'/ln (richness),
Native Richness Index (NRI): Mean Rn (√N), where Mean Rn = native species richness per quadrat and N is total native species richness,
Mean Coefficient of Conservatism (Mean C): ∑ CC/S, where CC = Coefficient of Conservatism and S = total species richness, and
Floristic Quality Index (FQI): Mean C (√N) where N = native species richness
All indices and parameters calculated from terrestrial vegetation samples were normally distributed. Two basic terrestrial vegetation types were perceived from field work and confirmed from Cluster Analysis (prairie and old field/shrubland). Comparisons of vegetation parameters among these two basic vegetation types were examined with means comparison tests (t-tests). Associations among parameters were examined with correlation analysis. These statistical tests were performed with Systat ver. 10.
The arrangement of sites, vegetation types, and species was examined with indirect gradient analysis of the species sample data using Detrended Correspondence Analysis (DCA) with scaling set for inter-species distances (ter Braak and Smilauer 2002). A species-weight inclusion rule of 2% or greater in the ordination, met by 51 taxa, was used for displayed species in the ordination biplot.
Direct gradient analysis using Canonical Correspondence Analysis (CCA) was applied to examine variance in species composition between sites attributable to selected “environmental” variables (ter Braak and Smilauer 2002) using vegetation parameters. Parameters used in the analysis include diversity, dominance, structure (e.g., % bare ground and vegetation cover, shrub density and % cover), and Floristic Quality Index (FQI) and Mean Coefficient of Conservatism (Mean C), indices used in Floristic Quality Assessment (Taft et al. 1997). Botanical nomenclature follows Taft et al. (1997), a modification from Mohlenbrock (1986). Non-native species in the report will be indicated with an asterisk (*).
RESULTS AND DISCUSSION
I. TERRESTRIAL VEGETATION
A total of 215 taxa of vascular plants were recorded from the combined stratified ground-layer samples (37 plots, 121 quadrats) and the targeted prairie samples (8 plots, 24 quadrats) including 171 native species, 33 non-native species, and remaining taxa undetermined to species (Appendix 1). The undetermined taxa were sterile seedlings that could not be determined with certainty to species; most likely are seedlings of recorded taxa. The top-10 dominant species overall were Rhamnus cathartica*, Cornus racemosa, Solidago juncea, Fragaria virginiana, Schizachyrium scoparium, Aster drummondii, Allium canadense, Potentilla simplex,
Lonicera X bella*, and Andropogon gerardii accounting for about 36% of the total importance value among all species. See Appendix 1 for a listing of common names associated with scientific names used for terrestrial species throughout the report.
Throughout the study area there was an average species density per quadrat (1-m2) of 22 including 17.3 native and 4.66 non-native species. Average vegetative cover was 131.7%, accounting for overlapping vegetation, and average bare ground was 25.4%. Results from Floristic Quality Assessment based on mean quadrat values indicated a Mean C of 2.4 (Mean Cn = 3.05) and FQI of 10.2 (FQIn = 12.8).
Cluster Analysis - A cluster dendrogram based on between-plot rank similarity of ground-layer species abundance indicates two main terrestrial vegetation types in the study area (Figure 3). The Prairie cluster is comprised of 16 plots including all eight targeted prairie sample plots together with eight plots from the stratified transect samples. The other major grouping, the Old Field/Shrubland cluster, is comprised of the 29 remaining plots from the stratified samples (Figure 4). Subdivisions among these two groupings can be recognized. The eight targeted prairie sampling plots form two discrete groupings within the Prairie cluster of five and three plots each, while eight plots from the transect samples form another group within the Prairie cluster (Figure 3). The Old Field/Shrubland cluster forms two main groups in the cluster dendrogram.
Species Associations - The dominant species for the two major vegetation types are shown in descending rank order of abundance (based on IV200) in Table 1. The top-10 dominant species in the ground-layer samples of the Old Field/ Shrubland cluster are Rhamnus cathartica*, Cornus racemosa, Allium canadense, Lonicera X bella*, Aster drummondii, Fragaria virginiana, Vitis riparia, Potentilla simplex, Aster lateriflorus, and Viburnum lentago accounting for 45% of the total importance value. The 10 top-ranking dominant species in the Prairie cluster include Rhamnus cathartica*, Solidago juncea, Cornus racemosa, Schizachyrium scoparium, Andropogon gerardii, Sorghastrum nutans, Silphium terebinthinaceum, Parthenium integrifolium, Daucus carota*, and Carex lanuginosa. These species account for 31% of the total importance among all species.
A graphical depiction of the associations among plot clusters and characteristic species, using Detrended Correspondence Analysis (DCA), indicated the Old Field/ Shrubland cluster forms a fairly discrete group while the targeted reference prairie plots in the Prairie cluster form two groups (Figure 5). The remaining prairie plots, those from the stratified transects, are intermediate between the reference prairie plots and the old field/shrubland plots. Axis 1 in the DCA is explained by a strong inverse correlation with percent vegetation ground cover (Pearson Correlation Coefficient r = -0.89), FQIn (r = -0.84), and Mean Cn (r = -0.79), and a strongly positive correlation with percent shrub cover (r = 0.84). Consequently, Axis 1 is interpreted as a combination available light as well as habitat quality gradient. Plots to the right in the ordination biplot, members of the Old Field/Shrubland cluster (Figure 5), have lower percent vegetation cover, lower FQIn (and FQI), lower Mean Cn (and Mean C), and higher percent shrub cover compared to the plots of the Prairie cluster on the left side of the biplot.
Indicator Species Analysis identifies species that have statistically significant association patterns with particular clusters (i.e., greater than would be expected by random chance [p < 0.05], as determined from Monte Carlo permutation tests). Forty-eight species were found to have significant associations to the Prairie cluster (Table 2) including the forbs Achillea millefolium*, Antennaria neglecta, Aster ericoides, Daucus carota*, Leucanthemum vulgare*, Monarda fistulosa, Ratibida pinnata, Sisyrinchium albidum, Solidago juncea, Solidago nemoralis, Solidago rigida, Allium cernuum, Aster azureus, Helianthus rigidus, and Parthenium integrifolium and the prairie grasses Andropogon gerardii, Schizachyrium scoparium, and Sorghastrum nutans. Only Allium canadense, Circaea lutetiana, Sanicula canadensis, Carex umbellata, and Lonicera X bella* were found to be significant indicators for the Old Field/ Shrubland cluster.
The three principal groupings within the prairie cluster are distinguished by specific species association patterns (Figure 5). Significant indicator species for the targeted reference prairie samples from the five-plot cluster (total n = 20 species), include Aster azureus, Bromus kalmii, Lespedeza capitata, Parthenium integrifolium, Silphium terebinthinaceum, Sorghastrum nutans, Schizachyrium scoparium, and Comandra umbellata. Indicators for the three-plot cluster (n = 17 species), plots placed adjacent to sedge meadow habitats, include Carex lanuginosa, Galium obtusum, Pycnanthemum virginianum, Solidago canadensis, Cacalia plantaginifolia, Euthamia graminifolia, Spartina pectinata, Glyceria striata, and Liatris spicata. Many of these latter species tend to be associated with wet-mesic prairie habitat. The second ordination axis of the DCA, then, is interpreted to be a moisture gradient with wetter plots positively associated with Axis 2 site scores (upper left side of Figure 5).
Most of these species from the reference prairie samples often are associated with high-quality prairie habitats in Illinois. The remaining prairie plots, those from the transect samples, are intermediate in composition between the high-quality prairie plots and the old field/shrubland plots (Figure 5). Significant indicators are Daucus carota*, Solidago juncea, Melilotus alba*, Ratibida pinnata, and Solidago nemoralis. Among the old field/shrubland plots, only Rhamnus cathartica* and Toxicodendron radicans, somewhat concentrated in one of the two main subclusters, were significant indicators differentiating the subgroupings.
Species Density and Diversity - The two main community types identified in Cluster Analysis, old field/shrubland and prairie, differ significantly in many measures of species richness and diversity. Mean values for native species density, native plot species richness, and percent vegetation cover were much greater in the prairie cluster plots compared to the old field/ shrubland plots, and the differences were significant (Table 3). For example, mean native species density for the old field/shrubland plots was 13.7 compared to 23.9 for the prairie plots. In contrast, Simpson’s Index of Dominance and percent bare ground were greater in the old field/shrubland samples compared to the reference prairie samples and the differences were significant. Mean percent vegetation cover in the old field/shrubland vegetation averaged 78% compared to 207% in the prairie vegetation while mean percent bare ground was 34% in the old field/shrubland plots and only 9% in the prairie plots (Table 3). Only the differences in number non-native species, at both the mean quadrat and plot scales, and the Shannon-Wiener Diversity index were not significantly different in paired comparisons after accounting for the multiple comparison error rates (Bonferroni adjustment [p = 0.05/n]).
Floristic Quality Assessment - The results of Floristic Quality Assessment (FQA) for the total species pools in the Old Field/Shrubland and Prairie communities (Appendix 1) indicate Mean C of 2.94 for the Old Field/Shrubland and 3.35 for the Prairie community (Mean Cn of 3.67 and 4.08, respectively). The Floristic Quality Indices were 34.75 for the Old Field/Shrubland and 38.1 for the Prairie community (FQIn = 43.4 and 46.3, respectively). However, FQI scores are influenced by species-area relationships, particularly total species richness. Comparisons among mean quadrat scores of Mean C (and Mean Cn) and FQI (and FQIn), values relatively independent of total sample area, allow for statistical comparisons and are described below.
All FQA indices were greater among prairie samples compared with the old field/shrubland data and the differences were significant (Table 3). For example, quadrat-level Mean C was 2.09 for the old field/shrubland sample compared to 2.94 in the prairie sample. Differences for Native Mean C were 2.74 and 3.62 for old field/shrubland and prairie samples, respectively (Table 3). Mean FQI was 7.86 for the old field/shrubland sample compared to 14.48 for the prairie sample data. The distribution of mean quadrat FQI scores across the study site indicate the highest quality areas are concentrated in the far southern portion of the study area, particularly the targeted reference prairie samples in the southeastern corner (Figure 6). For example, FQA applied selectively to the eight reference prairie plots indicates much higher scores compared to remaining terrestrial plots with Native Mean C of 4.4 compared to 2.8 and FQI of 19.1 compared to 8.3. The values for the high-quality prairie vegetation are similar to those calculated for mesic tallgrass prairie sites recognized as natural areas by the Illinois Natural Areas Inventory including sites dedicated as Illinois Nature Preserves (Taft, unpublished data). There were no significant differences in FQA scores in comparisons of the two major old field/shrubland plot groupings identified in cluster analysis.
Physiognomic Characteristics - Proportion of species among physiognomic groups in the two vegetation types, old field/shrubland and prairie, are very similar (Figure 7). Perennial forbs are dominant in both species number and proportion of the importance value (%IV) in both vegetation types. Old field/shrubland is distinguished by having a much greater proportion of its IV among shrubs captured in the ground-layer sample compared to prairie. Prairie is distinguished by having a much greater proportion of perennial warm-season grasses (P-Grass C4) compared to the old field/shrubland. Shrubs and trees in the ground layer sample (woody stems < 1 m tall) mostly were seedlings and resprouting plants.
Site Summary - Overall, including combined data from transect plots and targeted prairie sampling, total density of shrubs was 22,898 stems/ha. Average value per plot (25-m2) for shrub density was 57.2, mean Leaf Area Index (LAI) was 1.95, and mean percent shrub cover was 76%. Dominant species in the combined old field/shrubland and prairie samples were Rhamnus cathartica*, Cornus racemosa, Lonicera X bella, and Viburnum lentago accounting for nearly 80% of the sum importance values for all shrubs (Table 4). Shrub density calculated solely from the stratified transect samples totaled 25,200 stems/ha and samples solely from the targeted reference prairie samples totaled 12,250 stems/ha.
Plant Community Differences - The old field/shrubland community was characterized by having about 22,221 stems/ha, shrub cover of 87%, and average LAI of 2.51. As with the overall sample, the old field/shrubland samples were dominated by Rhamnus cathartica*, Cornus racemosa, Lonicera X bella*, and Viburnum lentago accounting for about 78% of the total importance for all species (Table 4). The prairie sample was characterized by having 24,125 stems/ha, percent shrub cover of 56%, and average LAI of 0.94. Dominant species were the same as with the overall and old field/shrubland samples accounting for about 84% of the total importance for all species (Table 4). Percent shrub cover and LAI were much greater in the old field/shrubland samples compared to prairie samples and the differences were significant, while shrub density was greater in the prairie plots but the difference was not significant (Table 3). Plots with high percent shrub cover occur throughout the study area (Figure 8); however, shrub cover is notably less in the reference prairie sample plots in the southeastern corner of the study area and in the wetland plots. These data indicate that the shrub stratum in the prairie plots was comprised generally of shorter, presumably younger, plants that have yet to shade the prairie vegetation as extensively as in the old field/shrubland plots. This suggests these plots are more recently invaded by shrubs and without management likely will transition towards the old field/shrubland vegetation type, as suggested in the DCA ordination biplot (Figure 5).
A 1995 survey of shrub density at this site along similar transects as in the current study indicated there were 7,396 stems/ha in the combined wetland and terrestrial plots (Plocher et al. 1996). During the 14-year interval to 2009, there was a three-fold increase in shrub density, much of this increase occurring in the formerly relatively open prairie habitat.
Four plots in the stratified transect samples included trees (woody stems > 10 cm dbh). In these plots, trees tended to be sparse with an estimated 250 stems/ha and basal area of 12.2 m2/ha. Only six species were recorded in the tree sample plots and Populus deltoides was the dominant with about 41% of the IV for all species (Table 5). Rhamnus cathartica*, the dominant shrub species and a species typically classified as a shrub (e.g., Taft et al. 1997), ranked second in importance among trees indicating some large individuals are present in the study area.
Threatened and Endangered Species
Amelanchier sanguinea, an endangered species in Illinois (Illinois Endangered Species Protection Board 2005) was discovered in the study area during 2009. Three individuals were found in a single shrub plot (plot 7D; Figure 9; Appendix 2). This is the first report of this species from Lake County, Illinois. In addition, seedlings of an Amelanchier species were found in 11 other sample plots; however, sterile individuals can not be determined with certainty to species.
Three of the four species listed by the Illinois Endangered Species Protection Board as threatened or endangered that are known from the site were recorded in quantitative sample plots. Oenothera perennis was found in one plot (7B) in the stratified transects and an additional 198 plants were found nearby about 170 m west of the Amelanchier sanguinea occurrence (Figure 9). Elymus trachycaulus was recorded from one of the reference prairie plots (Prairie Plot 2).
Multivariate Direct Gradient Analysis
Results from Canonical Correspondence Analysis comparing plant species abundance with environmental variables including species richness, diversity, floristic quality variables, and shrub abundance variables indicate that 70.7% of the total species-environmental relations were explained in 4 ordination axes. Marginal effects from forward selection showing the amount of variance each variable explains singly in the model indicate that Mean C, % ground cover, FQI, and % shrub cover explain the greatest amount of variation among the variables tested (Table 6 [covariables such as Mean Cn and FQIn were removed from the analysis]). Conditional effects, the remaining variance explained in the order variables are included in the model (once added in rank order of Marginal Effects, each variable is removed before examining proportion of remaining variance), indicate that Mean C, FQI, % Ground Cover, and Non-Native Species Richness explained significant proportions of the total variance in the model (Table 6). These results, where Mean C and FQI explain among the greatest amount of variance in the species abundance data among grassland habitats of varying quality, are very similar to findings in another prairie complex in northern Illinois (Taft et al. 2006). Mean C and FQI appear to be robust indicators of prairie habitat quality.
An ordination triplot indicates that old field/shrubland plots are associated with increasing shrub Leaf Area Index (LAI), % Bare Ground, % Shrub Cover, and Dominance from the ground layer samples (Figure 10). These plots occur in association with species identified in Indicator Species Analysis and include Allium canadense, Rhamnus cathartica*, Lonicera X bella*, and Aster drummondii. In contrast, the high-quality reference prairie plots occur associated with increasing Mean C, FQI, % Ground Cover, and Native Species Richness. These plots occur in association with Sorghastrum nutans, Schizachyrium scoparium, Silphium terebinthinaceum, Parthenium integrifolium, and Andropogon gerardii. Prairie/Old Field plots occupy an intermediate position in the ordination between high-quality prairie and old field/shrubland (Figure 10). Future monitoring will determine how these three basic vegetation types respond to ecological restoration. It is expected that the prairie/old field plots move in ordination (and ecological) space towards the reference high-quality prairie plots.
II. WETLAND VEGETATION
A total of 54 taxa of vascular plants were recorded from the combined stratified ground-layer samples (12 plots, 24 quadrats) and the targeted wetland samples (5 plots, 10 quadrats) including 45 native species, 7 non-native species, and 2 undetermined taxa (Appendix 3). The undetermined taxa were sterile plants that could not be determined with certainty to species (i.e., Carex sp. and Viola sp.). The top-10 dominant species overall were Typha angustifolia*, Carex stricta, Lythrum salicaria*, Phalaris arundinacea*, Calamagrostis canadensis, Polygonum punctatum, Leersia oryzoides, Impatiens capensis, Pilea pumila, and Lycopus americanus. See Appendix 3 for a listing of common names corresponding to scientific names.
There was an average species density per quadrat (1-m2) throughout the study area of 5.3 including 3.8 native and 1.5 non-native species. Average vegetative cover was 95%. Results from Floristic Quality Assessment based on the averaged quadrat values among all wetland plots indicated a Mean C of 1.88 (Mean Cn = 2.52) and FQI of 4.24 (FQIn = 5.28), values associated with relatively degraded vegetation.
Plant Communities, Species Richness, and Floristic Quality Assessment
Two wetland vegetation types, marsh and sedge meadow, were determined empirically at
the time of sampling. Twelve plots were categorized as marsh and five as sedge meadow. There was a total of 37 species in the marsh samples including 31 native, 6 non-native, and the 2 undetermined species (Table 7). Total percent vegetative cover in the ground-layer sample was 93.2%. The 10 top-ranking dominant species were Typha angustifolia*, Phalaris arundinacea*, Lythrum salicaria*, Polygonum punctatum, Impatiens capensis, Pilea pumila, Leersia oryzoides, Calamagrostis canadensis, Bidens frondosa, and Alisma plantago var. aquatica accounting for about 77% of the total importance among species. Average native species density in marsh habitat was 2.8 and average non-native species density was 1.88; non-native species comprised about 40% of the total (Table 8).
There was a total of 32 species in the sedge meadow plots including 30 native and 2 non-native species (Table 7). Total percent vegetative cover in the ground-layer sample was 100.3%. The 10 top-ranking dominant species were Carex stricta, Calamagrostis canadensis, Carex sartwellii, Spartina pectinata, Leersia oryzoides, Eupatorium maculatum, Helianthus grosseserratus, Lycopus americanus, Solidago gigantea, and Lythrum salicaria* accounting for 68% of the total importance among species. Average native species density in sedge meadow habitat was 6.1 and average non-native species density was 0.6; non-native species comprised about 9% of the total (Table 8).
The results of Floristic Quality Assessment (FQA) for the total species pools in the marsh and sedge meadow communities (Appendix 3) indicate Mean C of 2.52 for the marsh and 3.23 for the sedge meadow communities (Mean Cn of 2.73 and 3.33, respectively). The Floristic Quality Indices were 15.4 for the marsh and 18.6 for the sedge meadow community (FQIn = 16.6 and 19.1, respectively). However, FQI scores are influenced by species-area relationships, particularly total species richness, and there were a greater number of marsh plots (n = 12) compared to sedge meadow (n = 5). Comparisons among mean quadrat scores in the marsh community indicated a Mean C of 1.2 (Mean Cn = 1.96) and FQI of 2.53 (FQIn = 3.7) while for the sedge meadow Mean C was 3.5 (Mean Cn = 3.85) and the FQI was 8.3 (FQIn = 9.1). Sedge meadows had greater mean richness of native species, greater diversity, and higher scores for FQA variables compared to marsh. Marsh habitat had greater non-native species density and dominance compared to sedge meadow habitat. However, only differences for non-native species and floristic integrity scores (Mean C and FQI) were statistically significant (Table 8).
Site Summary - Only three wetland plots, all in marsh habitat, had shrubs in the wetland sample. Based on these data, there was an estimated 635 stems/ha among five species dominated by Rhamnus frangula* and Salix exigua, these species accounting for about 74% of all stems (Table 9).
III. COMPARISON OF TERRESTRIAL AND WETLAND VEGETATION
Ground layer species richness and diversity is greater in the terrestrial compared to the wetland vegetation and the differences for all but percent cover were significant (Table 10). Native species richness per quadrat (species density) was 3.79 compared to 17.3 in the wetland and terrestrial samples, respectively. Non-native species density also was higher in terrestrial plots compared to wetland plots (4.66 compared to 1.5); however, although there were only 1.5 non-native species in wetland plots, point-scale diversity was so low in the wetlands (some plots did not have any native species) there was a much greater proportion of non-native species in wetland (39.5%) compared to terrestrial plots (about 27%). In contrast to the diversity measures, Simpson’s Index of Dominance was greater in the wetland plots compared to terrestrial samples indicating that wetland samples were more strongly dominated by a few species (e.g., Typha angustifolia) compared to terrestrial plots.
Mean Coefficient of Conservatism (Mean C and Mean Cn) was greater in the terrestrial samples compared to wetland samples; however, the differences were not significant. In contrast, a comparison of prairie to wetland samples yielded significant differences for Mean C and Mean Cn (p = 0.01 and 0.007, respectively). Floristic Quality Index (FQI and FQIn) was significantly greater in the terrestrial compared to wetland samples (Table 10).
Shrub density and percent cover were much greater in the terrestrial samples compared to wetland samples (Table 10) and the differences were significant (p < 0.000001). Shrub density in the wetlands averaged about 1.6 per plot with average cover of 3.8%; shrub density in terrestrial samples average about 57 stems per plot with average cover of 76%.
SUMMARY AND CONCLUSIONS
Vegetation data from sample plots stratified throughout the North Chicago Wetland Mitigation Site provide insights to habitat characteristics in terrestrial and wetland communities. In terrestrial vegetation samples, additional data from local areas of high-quality prairie in the southern portion of the site provide a reference for comparison with trends following restoration activities in prairie habitats throughout the site. Hierarchical cluster analysis based on species abundance data identified two basic terrestrial (non-wetland) vegetation types in the sample data: prairie and old field/shrubland. The prairie community includes all plots from the targeted prairie reference samples and an additional eight plots from the stratified transect samples. These prairie plots have greater native species richness, diversity, and floristic integrity compared to the old field/shrubland samples while the latter are distinguished by greater species dominance and percent bare ground. An even greater difference in these parameters is found in comparisons between the reference prairie plots and all other samples. Two state-listed threatened species previously documented from the site, Oenothera perennis and Elymus trachycaulus, were recorded from prairie sample plots and the state endangered Amelanchier sanguinea was newly discovered for the site (and county) in a prairie plot during vegetation sampling.
In terrestrial communities, based on the stratified transect samples, there are about 25,000 shrubs and saplings per hectare throughout the study area, mostly Rhamnus cathartica*, Cornus racemosa, Lonicera X bella*, and Viburnum lentago. Shrub density throughout the site has increased three-fold during the last 14 years. Shrub density currently is slightly greater in the plots classified as prairie compared to the old field/shrubland plots, particularly those in the stratified transects. However, percent shrub canopy cover and leaf area index (LAI) are both significantly greater in old field/shrubland plots and, compared to patterns of shrub density, these data explain much more of the variation in the ground layer vegetation. These results suggest that compared to old field/shrubland plots, the prairie plots are characterized by a dense under-growth of relatively young shrubs that have yet to eliminate by shading characteristic prairie species. As evidence of ensuing change, prairie plots heavily infested with shrubs are intermediate in species composition between reference prairie vegetation and old field/shrubland vegetation and likely have been converting to the old field/ shrubland type with ongoing shrub invasion. Prairie areas probably were more widespread prior to extensive shrub infestation.
Two wetland communities were identified from vegetation sample data, marsh and sedge meadow. Sedge meadow samples, including plots from the stratified transects and targeted reference samples, had greater floristic integrity compared to marsh plots. Marsh plots were characterized by strong dominance by Typha angustifolia*. In a comparison of wetland and terrestrial plots, the old field/shrubland and prairie communities have greater species diversity and floristic integrity compared to wetland plots. Wetland plots have much lower shrub density and cover compared to terrestrial plots. Future vegetation monitoring will determine whether restoration involving shrub removal and prescribed fire lead to improvements in vegetation diversity and integrity throughout the North Chicago Wetland Mitigation Site.
The site maps (Figures 1, 2, 4, 6, 8, and 9 were prepared by Janet Jarvis (INHS) and her contributions and GIS skills are greatly appreciated.
Illinois Endangered Species Protection Board. 2005. Checklist of endangered and threatened animals and plants of Illinois. Illinois Endangered Species Protection Board, Springfield.
McCune B., and M. J. Mefford. 1999. PC-ORD.Multivariate Analysis of Ecological Data.
Version 4.34, MjM Software, Gleneden Beach, Oregon, U.S.A
Mohlenbrock, R. H. 1986. Guide to the vascular flora of Illinois. Revised and enlarged edition. Southern Illinois University Press, Carbondale. viii + 507 pp.
Olson, J., A. Nugteren, M. Morris, and T. Brooks. 1991. Wetland report and description of notable ecological features with Addendum for FAP 120 (US 41), Lake County, near Skokie, Illinois. 52 pages with appendix.
Plocher, A., P. Tessene, and J. Olson. 1996. Wetland mitigation site assessment of the US 41/IL 137 Illinois Department of Transportation project area. Technical report submitted to the Illinois Department of Transportation, Bureau of Design and Environment, Springfield.
Plocher, A., and D. Ketzner. 2006a. North Chicago Site - mitigation site assessment. Report to the IDOT. 11 pp.
Plocher, A., and D. Ketzner. 2006b. North Chicago Site - mitigation site assessment, Add. 1. Report to the IDOT. 60 pp.
ter Braak, C. J. F., P. Smilauer. 2002. CANOCO reference manual and CanoDraw for Windows user’s guide. Software for canonical community ordination (version 4.5). Microcomputer Power, Ithaca, NY
Taft,. J. 1996. Floristic survey and assessment of the Illinois Department of Transportation North Chicago (US Route 41/IL Route 137) Wetland Mitigation Site in Lake County. Technical Report 1996(29) submitted to the Illinois Department of Transportation, Springfield.
Taft, J. B. 2006. Floristic Survey and Assessment of the North Chicago Wetland Mitigation Site, Lake County, Illinois. Center for Wildlife and Plant Ecology. Illinois Natural History Survey Technical Report. Submitted to the Illinois Department of Transportation, Bureau of Location and Environment, 28 September 2006.
Taft, J. B., G. S. Wilhelm, D. M. Ladd, and L. A. Masters. 1997. Floristic Quality Assessment for vegetation in Illinois, a method for assessing vegetation integrity. Erigenia 15: 3-95.
Taft, J. B., C. Hauser, and K. R. Robertson. 2006. Estimating floristic integrity in tallgrass prairie. Biological Conservation 131:42-51.
Whittaker, R.H. 1975. Communities and Ecosystems, second edition Macmillan Publishing Co., Inc., New York, New York.