619 W. Dripps, G.P. Lewis, R. Baxter, and C.B. Andersen 22001133 SOUSToHuEthAeSaTstEeRrnN N NaAtuTrUaRlisAtLIST 1V2o(3l.) 1621,9 N–6o3. 73 Hydrogeochemical Characterization of Headwater Seepages Inhabited by the Endangered Bunched Arrowhead (Sagittaria fasciculata) in the Upper Piedmont of South Carolina Weston Dripps1,*, Gregory P. Lewis2, Rachel Baxter1,3, and C. Brannon Andersen1 Abstract - Sagittaria fasciculata (Bunched Arrowhead) is an endangered plant known to grow only in Greenville County, SC, and Henderson and Buncombe counties, NC. This study compared the hydrogeochemical characteristics (hydrologic setting, water chemistry, substrate grain size, and organic matter content) of fourteen Bunched Arrowhead sites across Greenville County. All Bunched Arrowhead were found in partially to fully shaded and saturated discharge areas in close proximity to and fed by continuous and consistent groundwater seeps. The plants grew in sandy substrata with highly variable organic matter contents. Surface waters at these sites were shallow, dilute, and acidic (pH 4.5–5.7). However, the degree to which water chemistry influences the plant’s growth and survival remains to be determined. Introduction Rare plant species may be associated with habitats that are themselves restricted in area (Kruckeberg and Rabinowitz 1985). Therefore, human alteration or destruction of those habitats significantly increases the species’ risk of extinction. A good example of such a rare species in the southeastern United States is Sagittaria fasciculata E.O. Beal (Bunched Arrowhead), a small, emergent perennial endemic to Greenville County, SC, and Henderson and Buncombe counties, NC (Porcher and Rayner 2001). In Greenville County, Bunched Arrowhead grows in a limited number of springhead seepage forests in the Upper Piedmont physiographic province (Porcher and Rayner 2001). The species was listed as federally endangered in 1979 due to its restricted range and increasing anthropogenic threats to known populations (USFWS 1979). Many of the remaining populations are threatened by land development and competition from exotic invasive plant species (Newberry 1991). Previous research on the Bunched Arrowhead and its habitat has been limited. In nearly all populations examined in South Carolina, the plant grows in hydrated sandy-mucky sediments fed by a constant but gradual flow of water from groundwater seeps in areas shielded from direct sunlight during the growing season by forest canopies (Newberry 1991, Snipes et al. [date unknown]). Increased impervious surface associated with land development can modify local hydrologic budgets by generating increased surface runoff, reducing infiltration, and diminishing groundwater recharge (Paul and Meyer 2001). Therefore, changes to the hydrologic budget in areas in which 1Department of Earth and Environmental Sciences, Furman University, Greenville, SC 29613. 2Department of Biology, Furman University, Greenville, SC 29613. 3Current address - Department of Earth and Environmental Sciences, Lehigh University, Bethlehem, PA 18015. *Corresponding author - weston.dripps@furman.edu. W. Dripps, G.P. Lewis, R. Baxter, and C.B. Andersen 2013 Southeastern Naturalist Vol. 12, No. 3 620 the Bunched Arrowhead grows have the potential to threaten the plant, given its specific habitat requirements. In particular, increased runoff may deliver too much water and possibly sediment to the plant’s habitats. Likewise, a decline in the water table may cause water-level reductions in the springs and seeps inhabited by the Bunched Arrowhead. Snipes et al. [date unknown] conducted a hydrogeologic characterization of Bunched Arrowhead habitats at two locations in South Carolina and one location in North Carolina. Their analyses included measurements of the organic matter content, mineralogical composition, and grain-size distribution of sediments, as well as a partial analysis of the surface-water chemistry. Their study was the first systematic attempt to characterize the Bunched Arrowhead habitat, but they focused primarily on making recommendations regarding boundaries for land acquisition for what has become the Bunched Arrowhead Heritage Preserve in northern Greenville County. Whether the three sites studied by Snipes et al. [date unknown] represent Bunched Arrowhead habitats more broadly was unknown. The purpose of our study was to characterize the hydrogeochemical characteristics of springhead seepages at all Bunched Arrowhead locations known in South Carolina as of 2007. We analyzed both surface-water chemistry and substrate characteristics at each location. This analysis is a critical first step towards documenting physical and chemical characteristics of individual sites as well as identifying habitat commonalities among the sites that may be important requisites for ensuring the plant’s survival. Further, our study provides what is, to our knowledge, the first examination of the biogeochemistry of headwater seepages in the South Carolina piedmont. Field Site Description We examined all 14 of the Bunched Arrowhead sites known to exist in South Carolina as of 2007 (Fig. 1). All of the sites occurred in northern Greenville County. Eight of the sites were found within preserves managed by the South Carolina Department of Natural Resources, one site was on the campus of Furman University, and the remaining sites resided on private properties. The sites all occurred within the headwaters of the Reedy and Enoree River drainages. In this region, the predominant underlying bedrock is composed of granites and gneisses (Overstreet and Bell 1965), which are relatively resistant to weathering and thus yield low levels of dissolved constituents. As a result, stream waters in the region typically have low conductivities (<100 μS/cm; Andersen et al. 2001, Lewis et al. 2007). Within the drainages containing the Bunched Arrowhead sites, the dominant upland soils are mapped as well-drained sandy loams or clay loams in the Cecil, Hiwassee, and Pacolet series (Camp 1975). The land in the immediate vicinity of the Bunched Arrowhead populations was either forested or shrub/grass dominated (Table 1). The land cover upgradient of the communities varied among the sites and comprised different proportions of forest, shrubs, grasses, and/ or residential areas (Table 1). All sample sites occurred in low-lying saturated depressions or in streams fed by groundwater seeps that were in close proximity (≤10 m) to each Bunched Arrowhead population. 621 W. Dripps, G.P. Lewis, R. Baxter, and C.B. Andersen 2013 Southeastern Naturalist Vol. 12, No. 3 Methods We visited each of the sites on two occasions, once during fall (October 24– November 2) 2006, and again during winter (February 15–March 14) 2007. During the fall visit at each site, we conducted a visual site assessment and collected sediment samples. During both fall and winter visits, we conducted in situ water quality measurements and collected water samples for chemical a nalyses. Site assessments The visual site assessment qualitatively characterized the hydrogeology (i.e., degree of saturation, water depth, water velocity, topographic setting), surrounding land cover (i.e., main vegetation type, degree of shading), and general substrate composition. Water-depth measurements were made during the fall sampling. A measurement was taken at the maximum and minimum depth in which the Bunched Arrowhead plants were found growing within each population. Flow rates at each site were categorized during the fall sampling as “stagnant”, “near stagnant”, “low flow”, or “flow” depending on the observed Figure 1. Location of the 14 Sagittaria fasciculata (Bunched Arrowhead) sites that were sampled in Greenville County, SC. Each site is designated by a black circle with its respective site number. W. Dripps, G.P. Lewis, R. Baxter, and C.B. Andersen 2013 Southeastern Naturalist Vol. 12, No. 3 622 surface water movement through the population. Land cover was assessed visually on site immediately surrounding each population, and aerial photographs within Google Earth (Version 6) were used to characterize the land cover for the upland drainage areas for each site. Sediment sample collection and processing A single 5.1-cm-diameter core sample of the upper 10 cm of substrate, the estimated rooting depth of the Bunched Arrowhead, was collected from the middle of each Bunched Arrowhead population for analyses of grain size and organic matter content. We chose not to collect more than one sediment sample per population in order to minimize damage to the plants and their habitats. Samples were dried at 60 °C for 2 days, and 50 g of each dried sample then was sieved using a 0.5-phi interval, with mean grain size and sorting calculated in phi units (φ = -log2d, where d = grain diameter) using the method of moments (Folk 1980). Also, two 1.5-g subsamples of each dried sediment sample were combusted at 400 °C for 16 hours to determine loss on ignition (LOI). A temperature of 400 °C (as opposed to a higher temperature) was used to avoid dehyradation of clays and ferrihydrite so that our estimate of percent organic matter would not be inflated by loss of water from the mineral fraction of the samples (Heiri et al. 2001, Schumacher 2002). Therefore, we assumed that LOI would be equivalent to percent organic matter in the sediments. Water sample collection and analyses During each visit to each location, we measured water temperature, dissolved oxygen (DO) concentration using a YSI 55 meter, conductivity using a YSI 30 meter, and pH using a Fisher Scientific AP62 Accumet pH meter. Also, a single surface-water sample was collected from each site at each visit using a pre-cleaned Table 1. Summary of the site characteristics from the visual assessment for the 14 Sagittaria fasciculata (Bunched Arrowhead) locations conducted during October–November 2006 in Greenville County, SC. “On-site land cover” indicates the dominant vegetation in the immediate vicinity of each Bunched Arrowhead population. “Upland land cover” indicates the dominant land cover type(s) in the drainage basin of each site. Site Water depth (cm) Water velocity On-site land cover Upland land cover 1 1–5 Near stagnant Shrub/grass Forest/residential 2 1–10 Near stagnant Shrub/grass Shrub/grass 3 7–12 Near stagnant Small trees (<1 m) Forest 4 1–3 Near stagnant Forest Forest/residential 5 1–2 Near stagnant Forest Forest/shrub 6 3 Low flow Forest Forest/shrub 7 7–10 Flow Forest Forest 8 1–7 Flow High grass (≈1 m) Shrub/residential 9 5 Flow Forest Forest/residential 10 1–5 Low flow Forest Forest 11 1–3 Low flow Forest Forest 12 1–5 Flow Forest Forest/residential 13 1–5 Flow Forest Forest/residential 14 1 Near stagnant Forest Grass 623 W. Dripps, G.P. Lewis, R. Baxter, and C.B. Andersen 2013 Southeastern Naturalist Vol. 12, No. 3 high-density polyethylene bottle which was rinsed three times with water from the site before filling. Water samples were transported to the laboratory on ice and then were filtered through 0.45-μm-membrane filters. An aliquot from each sample was preserved with trace-metal-grade nitric acid for cation analysis, a second aliquot was preserved with trace-metal-grade sulfuric acid for total dissolved nitrogen (TDN) analysis, and a third aliquot was left unpreserved for dissolved organic carbon (DOC), alkalinity, ammonium (NH4 +), and anion analyses. Solute analyses were conducted as in Lewis et al. (2007). Anion concentrations (F-, Cl-, Br -, H2PO4 -, NO2 -, NO3 -, and SO4 2-) were measured using a Dionex 120 ion chromatograph. Gran titration was used to measure alkalinity, from which bicarbonate (HCO3 -) concentrations were calculated. Concentrations of base cations (Na+, K+, Mg2+, Ca2+) and dissolved silicon (as Si4+) were measured using a Varian 2000 ICP-AES. Total dissolved nitrogen concentrations were measured with an O.I. Analytical Flow Solution IV autoanalyzer. Because of instrument error, only samples from fall 2006 were analyzed for TDN. Ammonium concentrations were measured using a Turner Designs 10-AU fluorometer, and DOC concentrations were measured using a Tekmar-Dohrmann Phoenix 8000 total organic carbon analyzer. Charge balances for 21 of the total 28 samples were within ±5%, and charge balances for the remaining 7 samples were within ±10%. We estimated dissolved organic nitrogen (DON) concentration as TDN concentration minus dissolved inorganic nitrogen (DIN) concentration (NH4-N + NO3-N + NO2-N concentrations). Dissolved carbon dioxide (CO2(aq)) concentrations were estimated using the procedure described in Lewis et al. (2007). We calculated total dissolved carbon (TDC) concentrations as the sum of DOC and dissolved inorganic carbon (DIC) concentrations (the sum of C from HCO3 - and CO2(aq)). Partial pressure of CO2 dissolved in surface water at each site (PCO2) was calculated from water temperature, pH, and HCO3 - concentrations using methods modified from Andersen (2002). We also calculated PCO2 saturation, the ratio of PCO2 in surface water to PCO2 in the atmosphere (Andersen 2002). Statistical analyses All statistical analyses were conducted using JMP 9.0.0 (SAS Institute, Inc., Cary, NC). Data were tested for normality using Shapiro-Wilk tests. For two-group comparisons, equality of variances was checked using F-tests. For two-sample tests, if data were normal but variances were unequal, independent-samples t-tests assuming unequal variances were used. If data were not normal, Mann-Whitney tests were used. We used these independent-samples tests to compare sites with stagnant and flowing water within each of the two seasons. We also used independent- samples tests to compare sites with forest and non-forest vegetative cover within each season. We could not test for interactions between water flow and vegetation cover using factorial analysis of variance because only one site fell into the category of non-forest cover with flowing water. For comparing data from repeated samplings of all 14 sites in fall and winter, paired t-tests were used if the differences between the dates were normally distributed. Otherwise, Wilcoxon signed rank tests were used. Correlations between W. Dripps, G.P. Lewis, R. Baxter, and C.B. Andersen 2013 Southeastern Naturalist Vol. 12, No. 3 624 variables were conducted using Pearson’s correlations if both variables were normally distributed and the relationships were linear. Otherwise, Spearman’s rank correlations were used. For solute concentrations which were below detection limits, we assigned values of one-half of the detection limit for statistical purposes. This procedure was used for analyses involving three variables: NH4 + (<0.01 mg/L in 4 of 28 samples), NO3 - (<0.05 mg/L in 4 of 28 samples), and TDN (<0.1 mg/L in 4 of 14 samples). Results Vegetative cover and hydrologic conditions Vegetation cover at the Bunched Arrowhead populations varied among sites (Table 1). At ten of the populations, forest canopies shaded the plants. The remaining four populations grew in more open shrub- or grass-dominated settings. At those four sites, the shrubs or grasses provided some shade to the Bunched Arrowhead populations during the growing season. All sites were fully saturated with water present at the land surface during both the fall and winter visits. The plants themselves were partially submersed, growing in water depths that varied from 1 to 12 cm (Table 1). At six sites, the water flow was nearly stagnant (Table 1). At the other eight sites, water was flowing at very slow but observable rates. There was no significant relationship in frequency of occurrence between vegetation cover (mature forest vs. other covers) and water flow rate (stagnant vs. flowing) among the sites (Fisher’s exact test: P = 0.24). All of the sites appeared hydrologically stable with no observable field evidence (stressed or disturbed vegetation, extensive erosion or deposition, or flood debris) that they had been recently subjected to flooding or other hydrologic disturbance. Based on visual inspection, an organic-rich layer of varying thickness overlay sandy sediment at all sites. Sediment characteristics Sediments among the sites ranged from coarse to fine sands (mean phi values of 0.47 to 2.69) that were poorly to very poorly sorted (sorting of ± 1.11 to 3.03 phi). The organic matter content ranged widely among sites from 0.8% to 33.1% (mean = 10.7%, median = 7.2%). Percent organic matter correlated positively with both mean grain size and sorting (Fig. 2). Thus, sites with more organic matter also tended to be finer grained and more poorly sorted. Percent organic matter in the sediment samples did not differ significantly between flowing and stagnant sites (independent samples t-test: P = 0.13). However, the percent organic matter at forested sites (mean = 7.4%, SE = 2.6%) was significantly lower than that of non-forested sites (mean = 19.2%, SE = 4.1%) (Figs. 2, 3). Stagnant and flowing sites did not differ significantly in either mean grain size (t-test: P = 0.33) or sorting (t-test: P = 0.067). Mean grain size did not differ significantly between forested and non-forested sites (t-test: P = 0.11), but forested sites were significantly better sorted (Figs. 2, 4). 625 W. Dripps, G.P. Lewis, R. Baxter, and C.B. Andersen 2013 Southeastern Naturalist Vol. 12, No. 3 Figure 2. Relationships between sediment grain size, sorting, and percent organic matter at 14 seepages inhabited by Sagittaria fasciculata (Bunched Arrowhead) in Greenville County, SC, October–November 2006. Both relationships were statistically significant (grain size: Spearman’s rho = 0.82, P = 0.0004; sorting: Spearman’s rho = 0.95, P < 0.0001). Figure 3. Sediment organic matter content of seepages inhabited by Sagittaria fasciculata (Bunched Arrowhead) in Greenville County, SC, October–November 2006. Forested sites (n = 10) had significantly (independent- samples t-test, P = 0.032) lower organic matter content than did non-forested sites (n = 4). The solid horizontal line within each box represents the median, and the dashed horizontal line represents the mean. The lower and upper limits of the boxes represent the 25th and 75th percentiles, respectively. Whiskers represent the 10th and 90th percentiles. Filled circles represent values beyond the 10th and 90th percentiles. W. Dripps, G.P. Lewis, R. Baxter, and C.B. Andersen 2013 Southeastern Naturalist Vol. 12, No. 3 626 Surface water properties Overall, surface waters at the 14 sites were poorly to very well oxygenated (DO = 1.6–10 mg/L), acidic (pH = 4.5–5.7), and dilute (conductivity < 53 μs/cm). In general, the surface waters exhibited a sodium-bicarbonate ion composition or a mixed cation-bicarbonate composition with a high proportion of dissolved silicon (Fig. 5). Typically, nutrient concentrations were low. For example, H2PO4 - concentrations were <0.10 mg/L for all samples, NH4 + concentrations were ≤0.06 Figure 4. Sediment sorting in seepages inhabited by Sagittaria fasciculata (Bunched Arrowhead) in Greenvi l le County, South Carolina, October–November 2006. The means of the two groups differed significantly (ttest: P = 0.031; sample sizes: forested n = 10, nonforested n = 4). The solid horizontal line within each box represents the median, and the dashed horizontal line represents the mean. The lower and upper limits of the boxes represent the 25th and 75th percentiles, respectively. Whiskers represent the 10th and 90th percentiles. Filled circles represent values beyond the 10th and 90th percentiles. Figure 5. Solute compositions of water samples collected at 14 seepages inhabited by Sagittaria fasciculata (Bunched Arrowhead) in Greenville County, SC, October–November 2006 (“Fall 2006”) and February–March 2007 (“Winter 2007”). 627 W. Dripps, G.P. Lewis, R. Baxter, and C.B. Andersen 2013 Southeastern Naturalist Vol. 12, No. 3 mg/L in all but one sample, NO2 - concentrations were all <0.08 mg/L, and NO3 - concentrations were <1 mg/L in 21 of 28 samples. However, at site 8, NO3 - concentrations were ≈6 mg/L in both seasons. Overall, oxidized N species were the dominant dissolved N species (>50% NO3 + NO2) at sites with actively flowing water, whereas DON was the dominant N species (>70% DON) at the stagnant sites (Fig. 6). Dissolved C was predominantly CO2 (aq) at all sites (>81% for all samples; Fig. 7), with CO2 partial pressures that were 48 to almost 500 times atmospheric pressure. Across all sites, pH and DO were significantly higher in winter than in fall (Table 2). Water temperature, conductivity, and concentrations of base cations, Si, HCO3 -, DOC, and CO2(aq) were significantly higher in fall than in winter. However, concentrations of other anions and of NH4 + did not differ significantly between seasons. We found significant differences in concentrations of four solutes between forest and non-forest sites (Table 3). Mean HCO3 - concentrations were ≈2 times higher at forest than non-forest sites in fall. However, means did not differ significantly in winter. Median SO4 2- concentrations were nearly 8 times higher at non-forest sites compared to forest sites in fall, but the medians did not differ significantly in winter. Median F- concentrations were about 3 times higher in non-forest sites during the winter, but not in fall. Only in the case of dissolved Si was the difference between forest and non-forest sites consistent in both seasons, with Si concentrations 1.8 to 2.0 times higher at forest than non-forest sites (Table 3, Fig. 8). Figure 6. Proportions of dissolved nitrogen species in surface water samples from 14 seepages (6 with stagnant or near-stagnant water velocity and 8 with visibly flowing water) inhabited by Sagittaria fasciculata (Bunched Arrowhead) in Greenville County, SC, October 24–November 2 2006. W. Dripps, G.P. Lewis, R. Baxter, and C.B. Andersen 2013 Southeastern Naturalist Vol. 12, No. 3 628 Figure 8. Dissolved silicon (Si) concentrations at seepages inhabited by Sagittaria fasciculata (Bunched Arrowhead) in Greenville County, SC, in fall 2006 (October 24– November 2) and winter 2007 (February 15–March 14). Of 14 sampling locations, 10 had forest canopy cover (“forested”) and the remaining 4 were dominated by shrubs and/or grasses (“non-forested”). A single water sample was collected at each site in each season. The solid horizontal line within each box represents the median, and the dashed horizontal line represents the mean. The lower and upper limits of the boxes represent the 25th and 75th percentiles, respectively. Whiskers represent the 10th and 90th percentiles. Filled circles represent values beyond the 10th and 90th percentiles. Means of the “forested” and “non-forested” groups differed significantly (P < 0.05) in both seasons. Flowing sites had significantly higher Na+, NO3 -, and Si concentrations than did stagnant sites in both seasons (Table 4). In other cases, significant differences Figure 7. Proportions of dissolved carbon species in surface water samples from 14 seepages inhabited by Sagittaria fasciculata (Bunched Arrowhead) in Greenville County, SC, in fall 2006 (October 24–November 2) and winter 2007 (February 15–March 14). 629 W. Dripps, G.P. Lewis, R. Baxter, and C.B. Andersen 2013 Southeastern Naturalist Vol. 12, No. 3 Table 2. Selected physical and chemical properties of surface waters at seepages inhabited by Sagittaria fasciculata (Bunched Arrowhead) in Greenville County, SC. Fall samples were October– November 2006; winter samples were collected February–March 2007 (n = 14 for each season). Values are either means ± SE (for normally distributed variables) or medians with interquartile ranges. Units of measurement are mg/L except for pH or as noted. Variable Season Values Temperature (°C) Fall 11.7 ± 1.86W, ** Winter 9.2 ± 0.6 pH Fall 4.96 ± 0.07T, ** Winter 5.23 ± 0.08 Dissolved O2 Fall 4.4 ± 0.6W, ** Winter 6.6 ± 0.5 Conductivity† Fall 41.7 (37.5, 49.9)T, *** Winter 31.1 (26.4, 37.3) Na+ Fall 2.57 ± 0.25T, * Winter 2.27 ± 0.21 K+ Fall 0.82 (0.68, 1.13)W, ** Winter 0.50 (0.29, 0.67) Ca2+ Fall 2.20 (1.57, 2.69)W, *** Winter 1.39 (0.78, 1.61) Mg2+ Fall 1.02 ± 0.08T, *** Winter 0.67 ± 0.05 Si4+ Fall 5.87 ± 0.52T, ** Winter 4.96 ± 0.55 NO3 - Fall 0.19 (0.09, 0.52)T, NS Winter 0.49 (0.08,1.44) Cl- Fall 3.01 (2.16, 3.47)W, NS Winter 2.42 (2.02, 2.86) SO4 2- Fall 0.69 (0.36, 1.37)W, NS Winter 0.54 (0.40, 0.92) F- Fall 0.04 (0.02, 0.09)W, NS Winter 0.02 (0.02, 0.04) HCO3 - Fall 9.36 ± 1.08T, ** Winter 5.59 ± 0.72 DOC Fall 3.33 (1.74, 4.66) W, *** Winter 0.88 (0.66, 1.67) CO2(aq) Fall 88.6 (54.2, 106.0)T, *** Winter 23.6 (16.4, 35.7) †Specific conductivity (μS/cm) at 25 °C. WWilcoxon signed rank test. TPaired t-test. NSP > 0.05. *P ≤ 0.05. **P ≤ 0.01. ***P ≤ 0.001. W. Dripps, G.P. Lewis, R. Baxter, and C.B. Andersen 2013 Southeastern Naturalist Vol. 12, No. 3 630 Table 3. Comparisons of solute concentrations (in mg/L) in Sagittaria fasciculata (Bunched Arrowhead) seepage habitats with either forest or non-forest vegetative cover in Greenville County, SC. Of 14 sampling locations, 10 had forest canopy cover (“Forested”) and the remaining 4 were dominated by shrubs and/or grasses (“Non-forested”). A single water sample was collected at each site in each season. Fall samples were collected between October and November 2006; winter samples were collected between February and March 2007. Values are either means ± SE (for variables which were normally distributed) or medians with interquartile ranges for variables which were not normally distributed. Variable Cover Fall Winter F- Forested 0.04 (0.03, 0.13)M, NS 0.02 (0.02, 0.02)M, ** Non-forested 0.04 (0.03, 0.10) 0.06 (0.04, 0.32) HCO3 - Forested 11.04 ± 1.05T, ** 5.63 ± 0.85T, NS Non-forested 5.18 ± 1.08 5.50 ± 1.56 SO4 2- Forested 0.54 (0.30, 0.87)M,* 0.47 (0.40, 0.92)M, NS Non-forested 4.03 (0.76, 14.17) 0.77 (0.23, 1.12) K+ Forested 0.82 (0.50, 0.97)M, NS 0.44 (0.23, 0.57)T, * Non-forested 1.05 (0.74, 1.91) 0.81 (0.49, 1.16) Si4+ Forested 6.74 ± 0.43 T, ** 5.78 ± 0.49T, * Non-forested 3.70 ± 0.70 2.92 ± 0.89 MMann-Whitney test. TIndependent samples t-test. NSP > 0.05. *P ≤ 0.05. **P ≤ 0.01. ***P ≤ 0.001. were found only in one of the two seasons (Table 4). For example, flowing sites had significantly higher water temperature and HCO3 - concentration in the fall, but not in winter. Dissolved organic N concentrations were significantly lower at flowing than stagnant sites in fall (no winter measurements were available). There was a tendency for NH4 + concentrations to be higher at stagnant than at flowing sites, although the differences were not significant (fall: P = 0.053, winter: P = 0.072) and in any case were low in magnitude (Table 4). In winter, flowing sites had significantly higher pH, DO, and conductivity, but lower DOC concentrations. For all other variables, there were no significant differences between flowing and stagnant sites in either season (T able 4). Discussion Physical setting Previous studies (Newberry 1991, Snipes et al. [date unknown]) have suggested that Bunched Arrowhead requires seepage habitats characterized by shallow, continuously flowing water. However, based on our field observations, the plants are at least intermittently able to tolerate borderline stagnant water. However, we do not have data on the growth rates of the populations visited, and it is possible that the stagnant conditions at some sites were detrimental to the 631 W. Dripps, G.P. Lewis, R. Baxter, and C.B. Andersen 2013 Southeastern Naturalist Vol. 12, No. 3 plants. Newberry (1991) has suggested that reduction of water flow in Bunched Arrowhead seepages by the invasive Murdannia keisak (Hassk.) Hand.-Mazz. (Asian Spiderwort) enhances siltation of the seepages and allows for establishment of other plant species that may hinder the survival of Bunched Arrowhead. Although most of the Bunched Arrowhead populations we studied grew under forest canopies, some populations grew in shrub or grass-dominated environments. Even in those settings, the Bunched Arrowhead received shade from taller Table 4. Physico-chemical properties of surface waters in Bunched Arrowhead (Sagittaria fasciculata) seepage habitats with either visible flow or stagnant waters in Greenville County, SC. Eight locations had visible water flow (“flowing”); the remaining 6 had stagnant water (“stagnant”). A single water sample was collected at each site in each season. Fall samples were collected October–November 2006; winter samples were collected February– March 2007. Values are either means ± SE (normally distributed variables) or medians with interquartile ranges (variables not normally distributed). Units of measurement are mg/L unless otherwise noted. ND = not determined. Variable Flow rate Fall Winter pH Flowing 5.04 ± 0.09T, NS 5.36 ± 0.08T, * Stagnant 4.86 ± 0.10 5.05 ± 0.13 Temp. (°C) Flowing 12.2 (11.7, 13.8)M, * 8.5 (7.6, 10.6)T, NS Stagnant 10.2 (9.2, 11.5) 10.5 (6.9, 11.5) Dissolved O2 Flowing 4.8 ± 0.8T, NS 7.5 ± 0.6T, * Stagnant 3.9 ± 1.0 5.5 ± 0.6 Conductivity† Flowing 45.1 (39.0, 51.0)M, NS 36.5 (30.7, 43.6) t** Stagnant 38.5 (35.6, 48.7) 25.1 (21.3, 30.6) HCO3 - Flowing 11.23 ± 1.46T, * 6.22 ± 0.92T, NS Stagnant 6.88 ± 0.97 4.76 ± 1.14 DOC Flowing 1.92 (0.90, 3.09)M, * 0.71 (0.57, 0.87) m* Stagnant 4.67 (3.39, 8.17) 1.61 (1.08, 2.27) Na+ Flowing 3.09 ± 0.29T, ** 2.76 ± 0.24T, ** Stagnant 1.87 ± 0.21 1.62 ± 0.14 Si4+ Flowing 6.93 ± 0.57T, * 6.03 ± 0.56T, * Stagnant 4.47 ± 0.58 3.53 ± 0.72 NO3 - Flowing 0.35 (0.16, 1.92)M, * 1.26 (0.62, 1.55) m** Stagnant 0.08 (0.03, 0.21) 0.18 (0.04, 0.36) NH4 + Flowing 0.02 (0.01, 0.03)T, NS 0.02 (0.01, 0.03)M, NS Stagnant 0.03 (0.02, 0.05) 0.03 (0.02, 0.07) DON Flowing 0.02 (BD, 0.03)M, ** ND Stagnant 0.16 (0.09, 0.58) ND †Specific conductivity (μS/cm) at 25 °C. MMann-Whitney test. TIndependent samples t-test. NSP > 0.05. *P ≤ 0.05. **P ≤ 0.01. ***P ≤ 0.001. W. Dripps, G.P. Lewis, R. Baxter, and C.B. Andersen 2013 Southeastern Naturalist Vol. 12, No. 3 632 vegetation, and water temperatures did not differ significantly between forest and non-forest sites in either fall or winter. It is possible that non-forest sites might have higher water temperatures than forest sites in summer months. On the other hand, proximity of the seeps inhabited by the Bunched Arrowhead to groundwater sources might minimize differences in water temperature between forest and non-forest sites. Additional research is needed to determine if the degree of canopy cover influences survival of Bunched Arrowhead populations either directly (in terms of light intensity) or indirectly (through effects on temperature). Lack of canopy shading may not necessarily be detrimental to the plant, however. Baugh and Schlosser (2012) report that removal of the overhead canopy at Bat Fork Bog in North Carolina initiated flowering and faster growth in a population of Bunched Arrowhead that had previously been shaded by exotic woody vegetation. The organic matter content of the sediments at our sample sites (0.8% to 33.1%) overlaps with the range (8.4% to 19.1%, median = 12.6%) reported by Snipes et al. [date unknown] for Bunched Arrowhead sites in the Enoree and Reedy River drainages that were included in our study. However, Snipes et al. [date unknown] reported that the organic matter content of the sediment at their site in Henderson County, NC (n = 3 samples, range = 20–57%, median = 29%), was higher than at the South Carolina sites. The reason for this difference was unclear. In our study, the greater organic matter content of sediments at nonforest sites might reflect higher productivity and detritus production of grasses and other herbaceous plants in high-light environments lacking forest canopies. Given that we collected only one sediment sample from each site, we do not know how much spatial variability in sediment characteristics occurred within each site. However, each Bunched Arrowhead population occupies a small area (less than ≈2 m2), and the populations as a whole reside in similar geomorphic settings. Surface-water characteristics Generally, the proportions of ions and other solutes in the seepages were similar to proportions in streams in the Upper (Inner) Piedmont (Andersen et al. 2001, Lewis et al. 2007). However, the seepages had relatively low conductivity (13–52 μs/cm) and lower pH (less than 5.8) compared to levels typically found in those streams (conductivity = 20–80 μs/cm, pH = 5.8–7.2). Given the small drainage areas of the seeps, we speculate that the groundwater entering the seeps follows relatively short flowpaths, which would provide limited time for dilute infiltrating rainwater to accumulate ions from minerals in the soil and bedrock before discharging at the surface and would account for the observed low ion concentrations. The seepages also occurred in drainages with limited developed land cover, which tends to increase conductivity and solute concentrations in Piedmont streams (Andersen et al. 2001, Lewis et al. 2007, Muthukrishnan et al. 2007). It is unclear whether residential land cover within the drainages of six of the sites influenced water chemistry of the seeps. On one hand, the sites (8 and 9) with the highest NO3 - concentrations and conductivity occurred within drainages in which residential land cover was present. On the other hand, the sites with the second highest NO3 - 633 W. Dripps, G.P. Lewis, R. Baxter, and C.B. Andersen 2013 Southeastern Naturalist Vol. 12, No. 3 concentrations, second highest conductivity, and highest SO4 2- concentrations (2, 3, and 6, respectively) occurred in drainages without residential land cover. CO2 generated by microbial and root respiration both in soils and in the seeps themselves may have contributed to the limited acid neutralizing capacity of the bedrock as well as the low pH of the surface waters. Dissolution of CO2 in water produces carbonic acid, which can increase the acidity of both groundwater and surface waters. In fact, PCO2 saturation in the seeps during fall (150–490 times atmospheric partial pressure) was considerably higher than summer PCO2 saturation in stream waters within the nearby Big Brushy Creek watershed (values up to ≈150), also located in the Upper Piedmont (Lewis et al. 2007). Because the seeps are located at or near points of groundwater discharge, CO2 in the groundwater would have had little time to out-gas, as compared to downstream reaches of streams (Dawson et al. 1995, Johnson et al. 2008). Thus, such headwater seeps could be important sites for out-gassing of CO2 from groundwater (e.g., Johnson et al. 2008). The higher CO2 concentrations and lower pH in the seeps in fall compared to winter is consistent with higher rates of respiration that would occur during warmer months. Further, DOC concentrations in the seeps were higher in fall than in winter, perhaps due to higher rates of DOC release from decomposing organic matter during the warmer months. Organic acids (a component of total DOC) released during decomposition could also contribute acidity to the surface waters (David et al. 1992, Driscoll et al. 1989). For solutes derived primarily from or associated with mineral dissolution (base cations, Si, and HCO3 -), concentrations in fall were higher than in winter. Ratios of fall to winter mean or median solute concentrations ranged from about 1.1 to 1.7 (Table 2). Similarly, median specific conductance in the fall (41.7 μS/ cm) was 1.3 times higher than in winter (31.1 μS /cm). We hypothesize that seasonal variation in evapotranspiration (Et) within the seepages’ catchments would account for much of this variation in water chemistry. With lower Et in winter (with dormancy of deciduous trees), groundwater would become more dilute chemically as more precipitation water recharges the water table and hydrologic residence times in the groundwater system are shorter. Although we did not make quantitative measurements of seep discharges, the observed flow appeared to be modestly higher during winter, consistent with lower rates of Et in the drainage basins. In contrast to base cations, Si, and HCO3 -, the fall/winter ratios of median DOC and CO2(aq) concentrations were 3.78 and 3.75, respectively, likely a result of biological activity (e.g., release of DOC from decomposing organic matter) that is higher under the warmer fall conditions. Why concentrations of anions other than HCO3 - did not differ significantly between seasons is uncertain, though the lack of difference may reflect processes other than mineral dissolution (e.g., biological influence or anion exchange with soils). Although not statistically significant, median Cl-, F-, and SO4 2- concentrations tended to be higher in fall than winter (Table 2), consistent with the trends in HCO3 -, base cations, and Si. Seasonal changes in NO3 - concentrations were W. Dripps, G.P. Lewis, R. Baxter, and C.B. Andersen 2013 Southeastern Naturalist Vol. 12, No. 3 634 highly variable among sites, with concentrations increasing from fall to winter at 8 sites, but decreasing from fall to winter at 5 sites. Although NO3 - concentrations did not change consistently by season among the sites, the influence of water flow rates on the nitrogen biogeochemistry of the seep waters was evident. For example, compared to sites with visibly flowing water, stagnant sites had lower concentrations of DO and NO3 - and higher DON concentrations. With longer residence time of water in the seeps, the potential for biological oxygen consumption would increase. Although none of our measurements indicated the surface waters were anoxic at those times, anoxic conditions in the sediments (in combination with available organic matter) would have allowed denitrification to remove NO3 - from surface waters (Harrison et al. 2011, Herrman and White 2008, Peterson et al. 2001), especially in the stagnant sites. In the stagnant sites, DON (rather than NO3 -) was the predominant form of dissolved N, accounting for 68–95% of TDN (in contrast to accounting for 0–43% of TDN at flowing sites). Ammonium concentrations were low at nearly all sites, possibly because of efficient uptake of NH4 + by plants and microbes and/or oxidation of NH4 + to NO3 - by nitrifying bacteria in the oxygenated surface waters. Effective NH4 + retention is typical of the biological systems of small headwater streams in many regions (Peterson et al. 2001). Other differences in water chemistry between stagnant and flowing sites were evident as well. The lower pH of the stagnant sites may have been due to higher concentrations of organic acids (as suggested by higher DOC concentrations in the stagnant sites). Because concentrations of CO2(aq) did not differ significantly between stagnant and flowing sites, differences in carbonic acid concentrations would not account for the pH differences. Given that conductivity and concentrations of HCO3 - and Na+ were lower at stagnant sites, it is also possible that groundwater flowpaths differed between stagnant and flowing sites. If groundwater entering sites with stagnant surface water had had less contact time with soil and bedrock minerals, there may have been less opportunity for neutralization of groundwater acidity and ion accumulation. The lower Si concentrations in stagnant sites compared to flowing sites may reflect greater opportunity for Si uptake by diatoms and graminoid plants (especially grasses and sedges) under low-flow conditions. Previous studies have provided evidence that diatoms and grasses can take up appreciable quantities of Si from surface waters in wetlands (e.g., Borrelli et al. 2012, Ding et al. 2004, Struyf and Conley 2009). Silicon uptake by diatoms and grasses may also explain the lower Si concentrations in the non-forested sites. For example, in streams in Shenandoah National Park, VA, Si concentrations declined after Porthetria dispar (L.) (Gypsy Moth) defoliated trees along the streams, allowing more light to reach the streams and presumably increasing diatom Si uptake (Grady et al. 2007). Conclusions This study represents the first comprehensive hydrogeochemical assessment of the habitat of the endangered Bunched Arrowhead. Overall, surface waters in the seep habitats were more acidic and dilute than stream waters in nearby areas 635 W. Dripps, G.P. Lewis, R. Baxter, and C.B. Andersen 2013 Southeastern Naturalist Vol. 12, No. 3 of the Upper Piedmont of South Carolina, probably reflecting the location of the seeps in the upper headwaters of streams. However, additional research is needed to determine if the Bunched Arrowhead’s survival is constrained to the range of physical and chemical parameters measured at our sample sites. Among the sites we studied, there was considerable variation in some physical and chemical conditions (widely ranging dissolved oxygen and sediment organic matter concentrations, as well as relatively high nitrate and sulfate concentrations at some sites). Further, Snipes et al. [date unknown] reported studying one population of Bunched Arrowhead in North Carolina that grew in sediments that were more organic-rich than the sediments we sampled in South Carolina. In spite of the variation in sediment characteristics and surface-water chemistry among habitats, all Bunched Arrowhead populations we examined occurred in shallow waters (≈70% of which had maximum depths ≤5 cm, and all were <13 cm deep). We hypothesize that seepage hydrology exerts more influence on Bunched Arrowhead growth and survival than does water chemistry, at least under current conditions. Still, additional research is needed to examine the hydrologic variability within the Bunched Arrowhead habitats. In contrast to our study, Baugh and Schlosser (2012) report a population of Bunched Arrowhead growing at depths of at least 1 m in a spring-fed pond along the Mills River, NC. Additional research will be critical to determining the extent to which land development (especially surface-water runoff and reduced groundwater recharge associated with increasing impervious surface cover) threatens the long-term survival of Bunched Arrowhead populations. Also, to our knowledge, no previous studies have made detailed analyses of the biogeochemistry of headwater springs and seeps in the South Carolina Piedmont. Previous studies in other regions have identified the importance of headwater streams to downstream water quality and nutrient dynamics (e.g., Alexander et al. 2007, Peterson et al. 2001). Thus, understanding the physical and biogeochemical processes (e.g., CO2 out-gasing and nitrogen retention and transformations) that occur within springs and seeps such as those inhabited by the Bunched Arrowhead will be important to understanding the function of stream systems in the southeastern Piedmont. Acknowledgments We thank the South Carolina Department of Natural Resources (SC DNR) for access to the field locations. We especially thank Mary Bunch (SC DNR), Amy Williams, and Guinn Garrett for field assistance and Lori Nelsen for solute analyses. This research was funded by Furman University and is a contribution of the River Basins Research Initiative at Furman University. Literature Cited Alexander, R.B., E.W. Boyer, R.A. Smith, G.E. Schwarz, and R.B. Moore. 2007. The role of headwater streams in downstream water quality. Journal of the American Water Resources Association 43(1):41–59. W. Dripps, G.P. Lewis, R. Baxter, and C.B. Andersen 2013 Southeastern Naturalist Vol. 12, No. 3 636 Andersen, C.B. 2002. Understanding carbonate equilibria by measuring alkalinity in experimental and natural systems. Journal of Geoscience Education 50:389–403. Andersen, C.B., K.A. Sargent, J. Wheeler, and S. Wheeler. 2001. Fluvial geochemistry of selected tributary watersheds in the Enoree River Basin, SC. South Carolina Geology 43:57–71. Baugh, T., and K.K. Schlosser. 2012. Management considerations for the restoration of Bunched Arrowhead, Sagittaria fasciculata. Natural Areas Journal 33:105–108. Borrelli, N., M. Osterrieth, A. Romanelli, M.F. Alvarez, J.L. Cionchi and H. Massone. 2012. Biogenic silica in wetlands and their relationship with soil and groundwater biogeochemistry in the Southeastern of Buenos Aires Province, Argentina. Environmental Earth Sciences 65(2):469–480. Camp, W. 1975. Soil Survey of Greenville County, SC. US Department of Agriculture, Washington, DC. 71 pp. + maps. David, M.B., G.F. Vance, and J.S. Kahl. 1992. Chemistry of dissolved organic carbon and organic acids in two streams draining forested watersheds. Water Resources Research 28(2):389–396. Dawson, J.J.C., D. Hope, M.S. Cresser, and M.F. Billett. 1995. Downstream changes in free carbon dioxide in an upland catchment from northeastern Scotland. Journal of Environmental Quality 24(4):699–706. Ding, T., D. Wan, C. Wang, and F. Zhang. 2004. Silicon isotope compositions of dissolved silicon and suspended matter in the Yangtze River, China. Geochimica et Cosmochimica Acta 68(2):205–216. Driscoll, C.T., R.D. Fuller, and W.D. Schecher. 1989. The role of organic acids in the acidification of surface waters in the eastern United States. Water, Air, and Soil Pollution 43(1–2):21–40. Folk, R. 1980. The Petrology of Sedimentary Rocks. Hemphill Publishing Company, Austin TX.184 pp. Grady, A.E., T.M. Scanlon, and J.N. Galloway. 2007. Declines in dissolved silica concentrations in western Virginia streams (1988–2003): Gypsy Moth defoliation stimulates diatoms? Journal of Geophysical Research 112:G01009. Available online at doi: 10.1029/2006JG000251. Harrison, M.D., P.M. Groffman, P.M. Mayer, S.S. Kaushal, and T.A. Newcomer. 2011. Denitrification in alluvial wetlands in an urban landscape. Journal of Environmental Quality 40:636–646. Heiri, O., A.F. Lotter, and G. Lemke. 2001. Loss on ignition as a method for estimating organic and carbonate content in sediments: Reproducibility and comparability of results. Journal of Paleolimnology 25:101–110. Herrman, K.S., and J.R. White. 2008. Denitrification in intact sediment cores from a constructed wetland: Examining the isotope pairing technique. Applied Geochemistry 23:2105–2112. Johnson, M.S., J. Lehmann, S.J. Riha, A.V. Krusche, J.E. Richey, J.P.H.B. Ometto, and E.G. Couto. 2008. CO2 efflux from Amazonian headwater streams represents a significant fate for deep soil respiration. Geophysical Research Letters 35:L17401. Available online at doi: 10.1029/2008GL034619. Kruckeberg, A.R., and D. Rabinowitz. 1985. Biological aspects of endemism in higher plants. Annual Review of Ecology and Systematics 16:447–479. Lewis, G.P., J. Mitchell, C.B. Andersen, D. Haney, M.-K. Liao, and K.A. Sargent. 2007. Urban influences on stream chemistry and biology in the Big Brushy Creek watershed, South Carolina. Water, Air, and Soil Pollution 182:303–323. 637 W. Dripps, G.P. Lewis, R. Baxter, and C.B. Andersen 2013 Southeastern Naturalist Vol. 12, No. 3 Muthukrishnan, S., G.P. Lewis, and C.B. Andersen. 2007. Chapter 24: Relationships between land cover, vegetation density, and nitrate concentrations in streams of the Enoree River basin, piedmont region of South Carolina, USA. Pp. 517–542, In D. Sarkar, R. Datta, and R. Hannigan (Eds.). Concepts and Applications in Environmental Geochemistry. Elsevier Press, New York, NY. 778 pp. Newberry, G. 1991. Factors affecting the survival of the rare plant, Sagittaria fasciculata E.O. Beal (Alismataceae). Castanea 56(1):59–64. Overstreet, W.C., and H. Bell III. 1965. Geologic map of the crystalline rocks of South Carolina. US Geological Survey Miscellaneous Geologic Investigations Map I-413, I sheet, scale 1: 250,000. USGS, Washington, DC. Paul, M.J., and J.L. Meyer, 2001. Streams in the urban landscape. Annual Review of Ecology and Systematics 32:333–365. Peterson, B.J., W.M. Wollheim, P.J. Mulholland, J.R. Webster, J.L. Meyer, J.L. Tank, E. Marti, W.B. Bowden, H.M. Valett, A.E. Hershey, W.H. McDowell, W.K. Dodds, S.K. Hamilton, S. Gregory, and D.D. Morrall. 2001. Control of nitrogen export from watersheds by headwater streams. Science 292(5514):86–90. Porcher, R.D., and D.A. Rayner. 2001. A Guide to the Wildflowers of South Carolina. University of South Carolina Press, Columbia, SC. 551 pp. Schumacher, B. 2002. Methods for the Determination of Total Organic Carbon (TOC) in Soils and Sediments. United States Environmental Protection Agency, Las Vegas, NV. Snipes, D.S., L.A. Sacks, J.A. Wylie, B.A. Israel, and S.E. Dawson. [Date unknown]. Hydrogeology of the Bunched Arrowhead. Technical Completion Report to Plant Conservation Program of the North Carolina Department of Agriculture, Raleigh, NC. 82 pp. Struyf, E., and D.J. Conley. 2009. Silica: An essential nutrient in wetland biogeochemistry. Frontiers in Ecology and the Environment 7(2):88–94. US Fish and Wildlife Service (USFWS). 1979. Determination that Sagittaria fasciculata is an endangered species. Federal Register 44:43700–43701.