In May of 2009, the bloom-forming diatom Didymosphenia geminata was first identified in the Upper Esopus Creek, a key tributary to the New York City water-supply and a popular recreational stream. The Upper Esopus receives supplemental flows from the Shandaken Portal, an underground aqueduct delivering waters from a nearby basin. The presence of D. geminata is a concern for the local economy, water supply, and aquatic ecosystem because nuisance blooms have been linked to degraded stream condition in other regions. Here we ascertain the extent and severity of the D. geminata invasion, determine the impact of supplemental flows from the Portal on D. geminata, and identify potential factors that may limit D. geminata in the watershed. Stream temperature, discharge, and water quality were characterized at select sites and periphyton samples were collected five times at 6 to 20 study sites between 2009 and 2010 to assess standing crop, diatom community structure, and density of D. geminata and all diatoms. Density of D. geminata ranged from 0–12 cells cm-2 at tributary sites, 0–781 cells cm-2 at sites upstream of the Portal, and 0–2,574 cells cm-2 at sites downstream of the Portal. Survey period and Portal (upstream or downstream) each significantly affected D. geminata cell density. In general, D. geminata was most abundant during the November 2009 and June 2010 surveys and at sites immediately downstream of the Portal. We found that D. geminata did not reach nuisance levels or strongly affect the periphyton community. Similarly, companion studies showed that local macroinvertebrate and fish communities were generally unaffected. A number of abiotic factors including variable flows and moderate levels of phosphorous and suspended sediment may limit blooms of D. geminata in this watershed.
Citation: George SD, Baldigo BP (2015) Didymosphenia geminata in the Upper Esopus Creek: Current Status, Variability, and Controlling Factors. PLoS ONE 10(7): e0130558. https://doi.org/10.1371/journal.pone.0130558
Editor: Kay C. Vopel, Auckland University of Technology, NEW ZEALAND
Received: October 22, 2014; Accepted: May 22, 2015; Published: July 6, 2015
This is an open access article, free of all copyright, and may be freely reproduced, distributed, transmitted, modified, built upon, or otherwise used by anyone for any lawful purpose. The work is made available under the Creative Commons CC0 public domain dedication
Data Availability: All relevant data are within the paper.
Funding: This research was funded by the New York State Department of Environmental Conservation (http://www.dec.ny.gov/) and U.S. Geological Survey (http://www.usgs.gov/) under agreement number 06E4NY24570025. The funders had no role in study design, data collection and analysis, decision to publish, or preparation of the manuscript.
Competing interests: The authors have declared that no competing interests exist.
The bloom-forming diatom, Didymosphenia geminata (Lyngbye) Schmidt, has historically been considered a wide-spread but rare species found in moderately flowing cold-water streams of North America, Europe, and Asia , and has more recently been introduced to New Zealand and parts of South America [2–5]. It has been termed a native invader in parts of its historical range because it has begun producing problematic blooms in some areas where it once existed in equilibrium [6–8]. Distribution patterns of D. geminata have also recently changed, resulting in greater spatial coverage and temporal persistence in streams worldwide . Not only has D. geminata expanded its geographic range; evidence suggests it has also broadened its tolerance of environmental conditions. Once believed to exist only in cold, oligotrophic streams, D. geminata has now demonstrated tolerance to more nutrient-rich lotic environments . In New York State, blooms of D. geminata have been confirmed in the Batten Kill (2006), East and West Branches of the Delaware River (2007 and 2008), Upper Esopus Creek (2009), Little Delaware River (2010), Neversink River (2011), Rondout Creek (2011), and others (A. Smith, New York State Department of Environmental Conservation, personal communication). Blooms of D. geminata may cover as much as 100% of stream beds with mats of extracellular mucopolysaccaride stalks that are many centimeters thick [2, 9]. The production of stalk material traps algae, macroinvertebrates, and detritus, and extensive blooms of D. geminata can severely alter benthic habitat, river hydraulics, and the condition of lotic freshwater ecosystems [9–12]. Nuisance blooms can also negatively impact recreational opportunities and local economies .
Major progress has been made over the past ten years towards understanding the factors that induce D. geminata to produce nuisance blooms. Blooms are caused primarily by the extensive production of stalk material and may not be associated with high rates of cell division [14, 15]. It is now believed that D. geminata produces extensive stalk material when it is phosphorous-limited, which may be a strategy to expose cells to the water column for greater acquisition of phosphorous [16–18]. Specifically, phosphatase activity in the stalks and nutrient cycling within the resulting mats may provide D. geminata with a competitive advantage over other diatoms in low-nutrient environments [19–21]. When D. geminata is not phosphorous-limited, it exhibits faster rates of cell division and may exist at comparatively higher cell densities for short periods of time . Under these conditions, extensive stalk production is less common  and D. geminata may exist in a non-nuisance capacity. More generally, blooms of D. geminata often occur under conditions of low nutrients, high light, low temperature, and infrequent hydrologic disturbances [9, 18]. The frequency of high flow events, particularly those that mobilize the streambed, is considered the best hydrologic predictor of D. geminata biomass [2, 18]. Bed-mobilizing events scour away existing periphyton biomass and effectively reset the periphyton successional process. Because D. geminata may be a late successional species , frequent high-flow events can limit cell density and stalk biomass. Thus, it is clear why D. geminata thrives, and sometimes reaches nuisance levels, in streams below impoundments which moderate flows and water temperature [2, 23].
In May 2009, D. geminata was first identified in the Upper Esopus Creek, a key tributary to the New York City water-supply and a popular recreational stream. Although it is unknown if D. geminata is native to the Upper Esopus, subfossil records indicate that it was historically present on Long Island  and at the mouth of the Delaware River in New Jersey , both of which drain this region . Regardless, water quality and aquatic biota are extensively monitored in the Upper Esopus and it is unlikely that significant blooms occurred prior to 2009. The identification of D. geminata is concerning because nuisance blooms could threaten aquatic food webs, recreation (fishing and tubing), and therefore the regional economies that depend on the Upper Esopus and other Catskill Mountain Rivers. Although recent publications have helped better define the effects of watershed and water quality parameters on D. geminata, the basic ecological knowledge necessary to design management strategies that might control or mitigate nuisance blooms is still limited [2, 9, 10, 26]. The primary objectives of this study are to: 1) ascertain the current extent and severity of the D. geminata invasion, 2) determine the impact of supplemental flows from an inter-basin aqueduct on D. geminata, and 3) identify potential limiting factors for D. geminata in the Upper Esopus Creek watershed.
Ethics and data availability statement
Study sites were distributed across public and private property and landowner permission was obtained prior to sampling at privately owned properties. Permits were not required and no protected species were sampled during this project. All relevant data are included herein and thus are publically available.
Study scope and area
The Upper Esopus Creek is located in the south central Catskill Mountain Region of southeastern New York (Fig 1). The Creek follows a 41.8 km semi-circular course from its headwaters at Winnisook Lake, around Panther Mountain, to its impoundment downstream of Boiceville, where it forms the Ashokan Reservoir. The watershed area of the Upper Esopus Creek is 497.3 km2 and drains some of the most rugged and mountainous terrain in the Catskills. Forested land comprises over 95% of the watershed and its surficial geology features lacustrine clay deposits that contribute suspended sediment to the system . Turbidity and other potential water quality impairments are a major concern in this watershed because the Ashokan Reservoir provides close to 40% of New York City’s drinking water . Nine major tributaries (Table 1) deliver waters to the Upper Esopus in addition to the Shandaken Portal, the terminus of an inter-basin aqueduct which diverts water from Schoharie Reservoir to its confluence with the Upper Esopus in Shandaken. Discharge from the Portal can increase natural flows on the Upper Esopus by a factor of two or greater and the supplemental flow usually has a moderating effect on ambient stream temperature (cooler in the summer, warmer in the winter) .
Asterisks denote seasonal sites that were sampled during all five surveys.
Periphyton samples were collected from 20 study sites on two occasions and from six of these sites on three other occasions for a total of five surveys between 2009 and 2010. During August 2009 and August 2010, periphyton samples were collected from all 20 study sites across the watershed (Table 1). Ten sites were located on the Upper Esopus, including four upstream and six downstream of the Shandaken Portal. Nine other sites were located on tributaries near their confluences with the Upper Esopus, and the last site was approximately 3 km upstream in the largest tributary, Stony Clove Creek (Fig 1). Periphyton samples were also collected from six main stem sites (three upstream and three downstream of the Portal) during November 2009, April 2010, and June 2010 to assess seasonal variation (herein termed seasonal sites).
Although this study did not quantitatively define a bloom, most field studies have found that blooms are positively correlated with cell density [6, 19], and thus cell density of D. geminata and periphyton standing crop are used to estimate the spatiotemporal variability of D. geminata. Periphyton samples were collected using methods described in the U.S. Environmental Protection Agency periphyton protocol for single habitat sampling . Periphyton was sampled from riffles because assemblages within a single habitat type are more homogeneous than those from across multiple habitats, and therefore are more sensitive to subtle differences in water quality .
Quantitative periphyton samples were collected to determine standing crop of periphyton using chlorophyll a (chl a) and ash-free dry mass (AFDM) and to identify attached diatoms as follows. Three replicate samples were collected at each site: one near the left bank, one near the right bank, and one from the center of the channel. For each replicate, the scrapings were composited from a delineated area of the surface of three rocks classified as boulder or large cobble using the Wentworth Scale . The volume of the slurry was measured and subsamples were taken for determination of chl a, AFDM, and diatom identification. For chl a and AFDM, the sample was mixed thoroughly with an electric mixer and a 5-mL subsample was vacuumed through a glass fiber filter. Each filter was placed in a petri dish, covered with foil, and kept on ice until it could be frozen. For diatom identification, the sample was mixed with an electric mixer and a 20-mL subsample was placed in a glass vial and preserved with 5 mL of formalin. The scraped area of each rock was outlined in chalk, overlain with a wire screen of known mesh size, and photographed. Total area of each rock scrape was calculated from these photographs using digital image analysis software .
Chlorophyll a concentrations were determined using standard fluorometric methods with a correction for pheophytin a . Filters from each sample were extracted in acetone and centrifuged, and fluorescence was read before and after the addition of hydrochloric acid. Hold time for chl a samples ranged from roughly 4–13 months depending on sample date. The total chl a concentration was expressed as μg cm-2. Ash-free dry mass was calculated as the difference between the dried weight and ashed weight of filters and expressed as mg cm-2. Filters were oven dried at 100°C for 24 h, weighed, ashed at 500°C for 2 h, and reweighed to determine AFDM . The AFDM for three sites (USOP-03B, USOP-04A, and USOP-06) sampled in June 2010 could not be determined because cellulose filters were used. These filters could not be ashed and could have affected chl a measurements at these sites as well.
Samples for diatom identification were shipped to a contract laboratory (Rhithron Associates, Inc., Missoula, Montana) and identified to lowest possible taxon (generally species). Permanent diatom slides were prepared from acid-washed subsamples from each replicate at each site. A transect was scribed on each slide and the first 600 valves (300 cells) along the transect were identified. Unfortunately, upon reviewing the data it became apparent that this method did not sufficiently document D. geminata cell density. In some samples, D. geminata was not among the first 600 diatoms identified, yet a brief visual scan of the slide clearly showed it was abundant. It has been demonstrated that such fixed count methods are biased towards smaller diatoms [2, 35]. Because D. geminata cells are very large relative to other diatoms, it was hypothesized that this methodology was insufficient to meet the study objectives. Consequently, all slides were rescanned entirely (no transects used) and all D. geminata cells were counted. The density of the entire diatom community was calculated using the ratio of area needed to count 300 cells along a transect = area of the entire slide/total number of cells on the slide. The laboratory pipetted approximately 0.6 mL onto each slide, which enabled the density of both D. geminata and the entire diatom community to be expressed in terms of cells per square centimeter of rock surface area.
Spearman correlations were used to assess relationships between D. geminata cell density, total diatom cell density, measures of standing crop, and basic hydrologic variables (mean discharge, discharge coefficient of variation (CV), and mean temperature of the 30 days preceding each survey) which were collected by or modeled from USGS stream gages (Table 2). A general linear mixed effects model was used to assess spatial and temporal differences in log(x+1)-transformed D. geminata cell density data with survey period, Portal (upstream or downstream), and the interaction term period*Portal as fixed factors and site (nested within Portal) as a random factor to account for repeated sampling of sites over time. Pairwise comparisons of significant effects were conducted with Tukey’s HSD test. The above analyses were performed on the mean values of the three rock scrape replicates, only considered the six seasonal sites (Table 1), and were conducted using Minitab v17.1 software.
Results of the diatom identifications were used to assess community structure using multivariate techniques with Primer-E v6 software with PERMANOVA+ [36–38]. The replicates for each sample were combined, fourth-root transformed, and used to form a resemblance matrix of Bray-Curtis similarities comparing all samples. Samples were plotted in “species-space” on a non-metric multidimensional scaling (MDS) ordination [39, 40] according to the non-parametric ranks of their Bray-Curtis similarities . A two-way crossed Analysis of Similarities (ANOSIM) test was applied to the resemblance matrix to test for significant effects of period and density class of D. geminata (no detection, 0–100 cells cm-2, >100 cells cm-2) on the diatom community. The homogeneity of multivariate dispersion within groups was assessed using PERMDISP . Although the ANOSIM test produces P-values, the value of the R statistic is considered more important for assessing differences between groups . An R value of less than 0.25 indicates barely separable groups, whereas an R value of greater than 0.5 indicates separate but overlapping groups, and values greater than 0.75 indicate well separated groups . It was hypothesized that the composition of the diatom community would be altered at sites where D. geminata was present or abundant and that these sites would group separately in the ordination.
Density of D. geminata and total diatoms, percentage D. geminata, chl a, and AFDM values are presented in Table 3 as means and standard deviations of the three replicates collected at each site. Density of D. geminata ranged from 0 cells cm-2 (observed frequently) to 2,574 cells cm-2 (observed at USOP-03B in November 2009). Total diatom density ranged from 19,085 cells cm-2 at USOP-04A during April 2010 to 419,956 cells cm-2 at USOP-03A during November 2009. The lowest chl a (0.82 μg cm-2) and AFDM (0.36 mg cm-2) concentrations were observed at USOP-00 during August 2009 and the highest chl a (20.77 μg cm-2) and AFDM (5.98 mg cm-2) concentrations occurred at USOP-03B during November 2009. Density of D. geminata was significantly correlated with total diatom density (r = 0.42, P = 0.022) and inversely correlated with CV of discharge (r = -0.52, P = 0.003) but was not significantly correlated with chl a (r = -0.05, P = 0.795), AFDM (r = 0.05, P = 0.810), mean discharge (r = -0.10, P = 0.590), or mean temperature (r = 0.24, P = 0.202). Total diatom density was significantly correlated with chl a (r = 0.43, P = 0.018) and AFDM (r = 0.63, P = 0.000) and standing crop measures (AFDM and chl a) were significantly correlated (r = 0.90, P = 0.000).
The density of D. geminata was high and variable at many sites during the November 2009 and June 2010 surveys while density was consistently low during the August 2009 and August 2010 surveys (Fig 2). No D. geminata cells were collected at any of the six sites during the April 2010 survey. The mixed model confirmed that period had a significant effect (P = 0.000) on the density of D. geminata (Table 4) and pairwise comparisons indicated the following grouping: June 2010: A, November 2009: AB, August 2010: BC, August 2009: C, April 2010: C (periods that do not share a letter are significantly different).Temporal changes in density of D. geminata were generally consistent with total diatom density and standing crop for most surveys. D. geminata cell density (Fig 3A), AFDM (Fig 3B), chl a (Fig 3C), and total diatom density (Fig 3D) were concurrently high during November 2009 and low during August 2009 and 2010. This relationship was not maintained in the April 2010 survey as no D. geminata cells were detected and total diatom density was low, yet standing crop measures were relatively high. Density of D. geminata and total diatoms was high in June 2010, but AFDM and chl a concentrations were relatively low during this survey.
Median value is indicated by the black center line, mean value is indicated by the black triangle, and the bottom and top of the box indicate the lower and upper quartiles, respectively. Whiskers represent the minimum and maximum values and hollow circles represent outliers.
The relative abundance of D. geminata ranged from 0–1.3% (Table 3) of the entire diatom community and generally increased during the same periods when total diatom density increased (Fig 3E). Increases in the percentage of D. geminata coincided with increases in AFDM and chl a during the November 2009 survey but not during the June 2010 survey. During the latter period, mean D. geminata density (621 cells cm-2) and relative abundance (0.4%) were among the highest observed during the study, yet mean chl a and the limited AFDM values were at moderate to low values (4.00 μg cm-2 and 0.80 mg cm-2, respectively).
Diatom community structure was strongly influenced by survey period but not by the cell density of D. geminata. Assemblages from each period clustered tightly in the MDS and April 2010 samples were most strongly isolated (Fig 4). A two-way ANOSIM test confirmed that differences between survey period were highly significant (Global R: 0.764, P = 0.001) and all pairwise comparisons between periods were significant (P<0.05). The density class of D. geminata (no detection, 0–100 cells cm-2, >100 cells cm-2) did not significantly affect the composition of the diatom community (Global R: 0.106, P = 0.053). PERMDISP indicated that multivariate dispersion (community homogeneity) differed significantly between periods (P = 0.002) and density class of D. geminata (P = 0.014). Pairwise comparisons indicated that diatom assemblages from the >100 cells cm-2 class were significantly more homogenous than the no detection class (P = 0.023) but did not differ from the 0–100 cells cm-2 class (P = 0.311). Sites where D. geminata was present in high densities, however, did not consistently separate in the ordination from nearby sites where it was not detected. For example, during the November survey, diatom assemblages at USOP-02 and USOP-04A had a high degree of similarity and were located close to one another in the ordination yet D. geminata was not detected at USOP-02 and was present at a density of 1,410 cells cm-2 at USOP-04A.
Spatial differences in the density of D. geminata were clearly evident during the study. Cell densities at seasonal sites were generally lowest at upstream sites, peaked abruptly at USOP-03B, and then declined gradually at sites further downstream. The highest mean density of D. geminata (637 cells cm-2) was observed at USOP-03B which is located immediately downstream from the confluence of the Shandaken Portal with the Upper Esopus. Portal was a significant (P = 0.029) factor in the mixed model and indicated that sites downstream of the Portal had greater densities of D. geminata than sites upstream of the Portal (Table 4). The interaction term period*Portal was also significant (P = 0.026) which suggests the effect of the Portal on D. geminata cell density differed by period.
Results of this investigation indicate that D. geminata may have expanded its range across parts of the watershed during the 12-month study. This seems likely because D. geminata was only identified in this highly monitored river system three months prior to our first (August 2009) survey. During this survey, D. geminata was only detected at main stem sites downstream of the Shandaken Portal (USOP-03B, USOP-04, USOP-04A, USOP-04B, USOP-05 and USOP-06) and at three tributary sites (STOC-00, STOC-01, and WODC-01) which all enter the Upper Esopus downstream of the Portal (Table 3). During the subsequent November 2009 survey, D. geminata was collected at the same three seasonal main stem sites, and at USOP-03A (Fig 5). This survey was the first to collect D. geminata upstream of the Shandaken Portal. D. geminata was not detected at any site during the April 2010 survey, but was found at all six of the seasonal main stem sites during June 2010, marking its first detection at USOP-02 and USOP-03. During the August 2010 survey, D. geminata was collected again at all six seasonal main stem sites (although the only detection at USOP-02 was a qualitative fourth replicate not included in this analysis), at additional downstream sites (USOP-04, USOP-04B, and USOP-05), and at three tributary sites (STOC-00, STOC-01, and BSNL-01). These findings suggest the diatom expanded its range on the main stem from a 17 km reach exclusively downstream of the Portal in August 2009 to an additional 2 km upstream of the Portal in November 2009, and to at least another 6 km upstream of the Portal by June 2010. D. geminata was not detected at the uppermost main stem site (USOP-00) or at six of the tributary sites during either of the comprehensive August surveys.
Though spatiotemporal variations of D. geminata in the Upper Esopus Creek watershed were significant and complex, blooms generally did not reach nuisance levels that would be expected to cause serious ecological effects. In streams from other regions with nuisance blooms, concentrations of AFDM and chl a often increased by a factor of 5–10 and exceeded guidelines for maximum desirable periphyton growth [2, 12]. For example, Kilroy  found that mean AFDM increased from 6.7 mg cm-2 to 33.2 mg cm-2 and mean chl a increased from 8.4 μg cm-2 to 45.3 μg cm-2 in stream reaches with D. geminata blooms compared to unaffected reaches of the Mararoa River, New Zealand. During our study in the Upper Esopus, densities of D. geminata and periphyton biomass were concurrently high during some periods (e.g., November 2009), but densities were also negligible during other surveys when periphyton biomass was high (e.g., April 2010). This suggests that the biomass of periphyton communities was not consistently dominated by D. geminata during the study. High densities of D. geminata only appeared to strongly affect standing crop during the November 2009 surveys at USOP-03B and USOP-04A. Additionally, D. geminata was only identified at four of the ten tributary sites: STOC-00 and STOC-01 during August 2009 and August 2010, WODC-01 during August 2009, and BNSL-01 during August 2010. Although it is possible that D. geminata has not been introduced to all of these tributaries, it is noteworthy that cell density never exceeded 13 cells cm-2 at any tributary site during August 2009 and was always below 1 cell cm-2 during August 2010 (Table 3). This observation may be similar to findings from a New Zealand study in which experimental introductions of D. geminata failed on spring-fed tributaries of larger rivers that supported D. geminata . Even the highest densities of D. geminata at main stem sites of the Upper Esopus (2,574 cells cm-2 at USOP-03B and 1,410 cells cm-2 at USOP-04A during November 2009, and 1,422 cells cm-2 at USOP-06 during June 2010) were similar to or below those values reported in other North American studies [3, 43]. The lack of nuisance blooms in the Upper Esopus is significant because excessive periphyton growth can impact water quality, biodiversity, and the aesthetic and recreational value of a stream  which may affect local and regional economies .
The overall effect of D. geminata on diatom communities in the Upper Esopus appears limited. The relative abundance of D. geminata was consistently low and peaked at 1.3% of the entire diatom community. These findings are consistent with research on streams in the western United States where abundance of D. geminata never exceeded 3% of the entire diatom community . Although density of D. geminata was positively correlated with total diatom density, results of the two-way ANOSIM indicate that the composition of diatom communities was not significantly affected by its presence or density. These results suggest that either D. geminata cell densities were not high enough to alter diatom communities or that cell density is a poor predictor of the impact of D. geminata on biota and benthic habitat. Since extracellular stalk material can comprise up to 90% of the D. geminata biomass , it is possible that a measure of D. geminata biomass or biovolume would have better identified changes to the diatom community.
The relatively low estimates of D. geminata cell density and standing crop during most surveys suggest that habitat or water quality in the Upper Esopus Creek watershed was not conducive to extensive bloom formation, at least during 2009 and 2010. It is well documented that bed-mobilizing high flows can scour algae from rock surfaces and effectively reset the successional process. Thus, the frequency of large floods can limit the growth and blooms of D. geminata [2, 18, 43, 45], in part because it is a late successional species . Additionally, it has been shown that moderate and stable base flows are correlated with occurrence and abundance of D. geminata [46, 47], and the significant negative correlation between density of D. geminata and the discharge coefficient of variation in the present study further supports this finding. Prolonged blooms of D. geminata may be unusual within most tributaries and main stem reaches of the Upper Esopus upstream of the Portal because channel forming flows are common and summer base flows can be extremely low [48, 49]. In contrast, the extended periods of relatively stable base flows downstream of the Portal appear to favor proliferation of D. geminata. The only two periods (November 2009 and June 2010) when D. geminata reached high densities were preceded by relatively stable and moderate flows. Accordingly, no D. geminata cells were detected at any site during the April 2010 survey which followed a period of hydrologic instability and a peak discharge that exceeded 600 m3 s-1 at USOP-06. Other factors, however, had to be responsible for the low densities of D. geminata observed at most sites during the August 2009 and 2010 surveys because stream flows immediately preceding these surveys were comparatively stable.
High water temperatures may limit the growth and blooms of D. geminata and promote die back during mid to late summer in parts of the Upper Esopus Creek. Although the upper thermal tolerance of D. geminata appears variable , peak biomass of D. geminata has been linked to water temperatures that do not exceed 18°C  and D. geminata is more frequently found in locations where average summer air temperatures remain below 20°C . Several laboratory studies using static tests confirmed that temperature was an important variable affecting the survival of D. geminata cells [51, 52]. Lagerstedt  found that cells were unable to survive more than 60 hours at 28°C and densities of viable cells gradually declined at 20°C. In the Upper Esopus, peak summer temperatures in the main stem consistently exceed 20°C. During 2010, water temperature at USOP-03A (and many other sites) exceeded 20°C for long periods of time and peaked at 28.1°C (Fig 6A). These temperatures far exceeded the preferred range of D. geminata and may be responsible for the low cell densities observed during the August 2010 survey.
The presence of D. geminata in the Upper Esopus is informative because suspended sediment concentrations are unusually high and blooms typically occur in clear oligotrophic streams with high light levels. Suspended sediment adversely affects lotic periphyton primarily through reduced light penetration . Because the stalk length of D. geminata cells is positively correlated with light level , it follows that turbid waters could cause decreased stalk length and therefore less problematic blooms. It has also been demonstrated that abrasion caused by suspended sediment during elevated flows can scour and remove benthic algae . However, the role of turbidity and suspended sediment in limiting the growth, density, and distribution of D. geminata is not well studied. In one of the few studies to identify sediments and turbidity as possible limiting factors, Kirkwood  found that moderate levels of turbidity and total suspended solids (means of 10.5 NTU and 9.35 mg L-1) may have restricted the growth of D. geminata in the Red Deer River (Alberta, Canada). In comparison, the median levels of turbidity and suspended sediment at 13 study sites in the Upper Esopus watershed from 10/1/2009 to 9/30/2010 ranged from 4.3 to 119.5 NTU and 3.0 to 136.0 mg L-1, respectively . Each measure peaked at STOC-00 (where D. geminata was observed at low densities) and values were an order of magnitude higher than those in the Red Deer River. In general, it is apparent that waters across the Upper Esopus Creek watershed were similarly or more turbid than those of the Red Deer River and could, accordingly, have adversely affected the growth and density of D. geminata.
The supplemental flows from the Portal appeared to promote proliferation of D. geminata at downstream reaches of the Upper Esopus Creek during 2009 and 2010. The mean density of D. geminata downstream of the Portal was significantly greater than that of sites upstream of the Portal. In addition, mean density of D. geminata was highest at the site immediately downstream of the Portal (USOP-03B), and gradually declined at sites further downstream. Favorable thermal and hydrologic conditions produced by the Portal are likely responsible for these observations. Mean temperature from July 1 –August 31 was 16.0°C at USOP-03A and 14.8°C at USOP-03B in 2009 compared to 19.5°C and 19.3°C at these sites in 2010 (Fig 6A). Additionally, daily mean discharge at USOP-03A was usually below 2 m3 s-1 during both summers (lowest flow 0.24 m3 s-1, August 21, 2010), while the daily mean discharge at USOP-03B never dropped below 3.5 m3 s-1 during either summer (Fig 6B). Accordingly, the discharge coefficient of variation, which was negatively correlated with density of D. geminata, was smallest at USOP-03B when averaged across the five study periods (Table 2). Conditions at downstream reaches apparently become progressively less favorable as any beneficial effects of the Portal dissipate. These findings are not surprising because blooms of D. geminata have frequently been observed in reaches downstream of impoundments with regulated flow and thermal regimes [2, 9, 22, 23]. Stable base flows, like those present on the Upper Esopus immediately downstream of the Portal, are also considered beneficial for D. geminata [46, 47]. Therefore, it is likely that the moderated thermal and hydrologic conditions produced by supplemental flows from the Portal favor growth and sustained blooms of D. geminata in the downstream reaches of the Upper Esopus.
Although some recent studies indicate that blooms of D. geminata can alter invertebrate, algal, and fish communities, major ecosystem effects were not expected in the Upper Esopus Creek because of the low densities observed during most surveys. In other regions, the density of tolerant macroinvertebrates such as Chironomids and Oligochaetes increased, the number of Ephemeroptera, Plecoptera, and Trichoptera (EPT) taxa decreased, and overall community integrity declined in stream reaches impacted by blooms of D. geminata [10, 12, 56, 57]. Direct effects on fish assemblages have been more difficult to quantify and the deterioration of the brown trout (Salmo trutta L.) fishery in Rapid Creek, SD coincidental with the introduction of D. geminata in 2002 is one of the few studies to suggest fishery effects [9, 10, 57]. Although cell density across the Upper Esopus rarely exceeded 100 cells cm-2, densities >1,000 cells cm-2 were documented at sites downstream of the Portal during November 2009 and June 2010 and suggest that adverse impacts are possible. More recently (2012), Richardson et al  measured maximum D. geminata densities around 100,000 cells cm-2 in the Upper Esopus downstream of the Portal, and found that cell densities were negatively correlated with macroinvertebrate diversity, family richness, and EPT richness. Results from three companion studies on the Upper Esopus, however, did not identify any severe effects of D. geminata on resident fish and macroinvertebrate assemblages during 2009 through 2012. The mean New York State Biological Assessment Profile (BAP) score for the integrity of macroinvertebrate communities from eight samples downstream of the Portal from 2007 and 2008 (7.93) was nearly identical to the mean BAP score from 12 samples from 2009 and 2010 (7.94) after the appearance of D. geminata [58–60]. In addition, there were only minor differences in fish population or communities metrics at sites located upstream and downstream of Portal  (where densities of D. geminata were significantly different) and most were attributed to differences in habitat. Measures of physiological stress in brown trout were also lower or less evident at sites immediately downstream of the Portal (where D. geminata was most abundant) than at reaches upstream of the Portal .
Findings from several recent studies may further explain why dense blooms and elevated periphyton standing crop were not consistently observed in the Upper Esopus. The absence of dense sustained blooms and little indication of ecosystem impacts would intuitively suggest that habitat in this system is not optimal for D. geminata. Dense and problematic blooms, however, are apparently a response to phosphorous-limitation (soluble reactive phosphorus levels less than 2 mg m-3) in oligotrophic habitats [16, 17]. When phosphorous is not limiting, cell division rates increase for a short period of time, but stalk production decreases [15, 16] and D. geminata may be present in a non-nuisance capacity. Limited water samples collected from the Upper Esopus and tributaries during another companion study showed that orthophosphate ranged from roughly 6–11 mg m-3; which exceeds the 2 mg m-3 threshold for soluble reactive phosphorus identified in other studies. Additionally, macroinvertebrate samples from the Upper Esopus in 2007 and 2008 assessed using a Nutrient Biotic Index for phosphorous  suggest most main stem sites are mesotrophic . Therefore, the relatively high levels of available phosphorus likely contribute, at least partly, to our conclusion that D. geminata did not dominate aquatic ecosystems in the Upper Esopus to the extent observed in other regions. A more comprehensive study comparing the behavior of D. geminata in the Upper Esopus to that from more oligotrophic streams of the same region would be needed to further evaluate and test this hypothesis.
The authors extend appreciation to Anne G. Ernst and Justin Zimmerman of the U.S. Geological Survey; Alexander J. Smith, and Walter Keller (retired) of the New York State Department of Environmental Conservation; Tyler J. Ross of Cornell University; and Graham Markowitz of Cornell Cooperative Extension of Ulster County for field support. This research was funded by the New York State Department of Environmental Conservation and the U.S. Geological Survey. Any use of trade, firm, or product names is for descriptive purposes only and does not imply endorsement by the U.S. Government.
Conceived and designed the experiments: BB. Performed the experiments: BB. Analyzed the data: SG BB. Contributed reagents/materials/analysis tools: BB. Wrote the paper: SG BB.
- 1. Krammer K, Lange-Bertalot H. Bacillariophyceae: Naviculaceae. In: Ettl H, Gerloff J, Heynig H, Mollenhauer D, editors. Süßwasserflora von Mitteleuropa. 2/1. Heidelberg: Spektrum Akademischer Verlag; 2007.
- 2. Spaulding SA, Elwell L. Increase in Nuisance Blooms and Geographic Expansion of the Freshwater Diatom Didymosphenia geminata. Open-File Report. Reston, Virginia: US Geological Survey, 2007 OFR 2007–1425.
- 3. Kirkwood A, She T, Jackson L, McCauley E. Didymosphenia geminata in two Alberta headwater rivers: an emerging invasive species that challenges conventional views on algal bloom development. Can J Fish Aquat Sci. 2007;64:1703–9.
- 4. Reid BL, Hernández KL, Frangópulos M, Bauer G, Lorca M, Kilroy C, et al. The invasion of the freshwater diatom Didymosphenia geminata in Patagonia: prospects, strategies, and implications for biosecurity of invasive microorganisms in continental waters. Conserv Lett. 2012;5(6):432–40.
- 5. Kilroy C, Unwin M. The arrival and spread of the bloom-forming, freshwater diatom, Didymosphenia geminata, in New Zealand. Aquatic Invasions. 2011;6(3):249–62. pmid:CCC:000306277400002.
- 6. Carey MP, Sanderson BL, Barnas KA, Olden JD. Native invaders-challenges for science, management, policy, and society. Front Ecol Environ. 2012;10(7):373–81.
- 7. Simberloff D. Native Invaders. In: Simberloff D, Rejmánek M, editors. Encyclopedia of biological invasions. Los Angeles: Univ of California Press; 2011.
- 8. Richardson DC, Oleksy IA, Hoellein TJ, Arscott DB, Gibson CA, Root SM. Habitat characteristics, temporal variability, and macroinvertebrate communities associated with a mat-forming nuisance diatom (Didymosphenia geminata) in Catskill mountain streams, New York. Aquat Sci. 2014;76:553–64.
- 9. Whitton B, Ellwood N, Kawecka B. Biology of the freshwater diatom Didymosphenia: a review. Hydrobiologia. 2009;630(1):1–37.
- 10. Larson AM, Carreiro J. Relationship between nuisance blooms of Didymosphenia geminata and measures of aquatic community composition in Rapid Creek, South Dakota. Proceedings of the 2007 International Workshop on Didymosphenia geminata; 2008: Canadian Technical Report on Fisheries and Aquatic Sciences 2795.
- 11. Gillis C- A, Chalifour M. Changes in the macrobenthic community structure following the introduction of the invasive algae Didymosphenia geminata in the Matapedia River (Québec, Canada). Hydrobiologia. 2010;647(1):63–70.
- 12. Kilroy C, Larned ST, Biggs BJF. The non-indigenous diatom Didymosphenia geminata alters benthic communities in New Zealand rivers. Freshwat Biol. 2009;54(9):1990–2002.
- 13. Beville ST, Kerr GN, Hughey KF. Valuing impacts of the invasive alga Didymosphenia geminata on recreational angling. Ecol Econ. 2012;82:1–10.
- 14. Kilroy C, Bothwell M. Environmental control of stalk length in the bloom-forming, freshwater benthic diatom Didymosphenia geminata (Bacillariophyceae). J Phycol. 2011;47(5):981–9.
- 15. Bothwell ML, Kilroy C. Phosphorus limitation of the freshwater benthic diatom Didymosphenia geminata determined by the frequency of dividing cells. Freshwat Biol. 2011;56(3):565–78. pmid:CCC:000287092900013.
- 16. Bothwell ML, Taylor BW, Kilroy C. The Didymo story: the role of low dissolved phosphorus in the formation of Didymosphenia geminata blooms. Diatom Res. 2014;29(3):1–8.
- 17. Kilroy C, Bothwell ML. Didymosphenia geminata growth rates and bloom formation in relation to ambient dissolved phosphorus concentration. Freshwat Biol. 2012;57(4):641–53. pmid:CCC:000301227000001.
- 18. Cullis JDS, Gillis CA, Bothwell ML, Kilroy C, Packman A, Hassan M. A conceptual model for the blooming behavior and persistence of the benthic mat-forming diatom Didymosphenia geminata in oligotrophic streams. J Geophys Res (G Biogeosci). 2012;117(G2):G00N3. pmid:CCC:000305154100002.
- 19. Ellwood N, Whitton B. Importance of organic phosphate hydrolyzed in stalks of the lotic diatom Didymosphenia geminata and the possible impact of atmospheric and climatic changes. Hydrobiologia. 2007;592(1):121–33.
- 20. Aboal M, Marco S, Chaves E, Mulero I, García-Ayala A. Ultrastructure and function of stalks of the diatom Didymosphenia geminata. Hydrobiologia. 2012;695(1):17–24.
- 21. Sundareshwar P, Upadhayay S, Abessa M, Honomichl S, Berdanier B, Spaulding S, et al. Didymosphenia geminata: Algal blooms in oligotrophic streams and rivers. Geophys Res Lett. 2011;38(10).
- 22. Flöder S, Kilroy C. Didymosphenia geminata (Protista, Bacillariophyceae) invasion, resistance of native periphyton communities, and implications for dispersal and management. Biodivers Conserv. 2009;18(14):3809–24.
- 23. Kirkwood AE, Jackson LJ, Mccauley E. Are dams hotspots for Didymosphenia geminata blooms? Freshwat Biol. 2009;54(9):1856–63.
- 24. Lohman KE. Pleistocene Diatoms from Long Island, New York. U.S. Geological Survey, 1939 Professional Paper 189-H.
- 25. Woolman L. Annual Report of the State Geologist for the year 1894. Trenton, New Jersesy: Geological Survey of New Jersey, 1894.
- 26. Bothwell M, Spaulding S. Synopsis of the 2007 International Workshop on Didymosphenia geminata. Proceedings of the 2007 International Workshop on Didymosphenia geminata; 2008: Canadian Technical Report on Fisheries and Aquatic Sciences 2795.
- 27. CCE. Upper Esopus Creek Management Plan: Volume I Summary of Findings and Recommendations. Kingston, NY: Cornell Cooperative Extension of Ulster County; 2007.
- 28. Palmer PM, Wilson LR, O'Keefe P, Sheridan R, King T, Chen C-Y. Sources of pharmaceutical pollution in the New York City Watershed. Sci Total Environ. 2008;394(1):90–102. pmid:18280543
- 29. CCE. Upper Esopus Creek Management Plan: Volume III Watershed and Stream Characterization. Kingston, NY: Cornell Cooperative Extension of Ulster County; 2007.
- 30. Barbour MT, Gerritsen J, Snyder BD, Stribling JB. Rapid bioassessment protocols for use in streams and wadeable rivers: Periphyton, benthic macroinvertebrates, and fish (second edition). Washington, D.C.: U.S. Environmental Protection Agency, 1999 EPA 841–B–99–002.
- 31. Cummins KW. An evaluation of some techniques for the collection and analysis of benthic samples with a special emphasis on lotic waters. Am Midl Nat. 1962;67:477–504.
- 32. Pickle J. AnalyzingDigitalImages. Version 11 ed: Lawrence Hall of Science, University of California Berkeley; 2008.
- 33. APHA, AWWA, WEF. Standard Methods for the Examination of Water and Wastewater. 22 ed. Washington, D.C.: American Public Health Association, American Water Works Association, Water Environment Federation; 2012.
- 34. Steinman AD, Lamberti GA, Leavit PR. Biomass and pigments of benthic algae. In: Hauer FR, Lamberti GA, editors. Methods in stream ecology. Amsterdam: Academic Press; 1996. p. 357–80.
- 35. Spaulding S, Hermann K, Johnson T. Confirmed distribution of Didymosophenia geminata (Lyngbe) Schmidt in North America. Proceedings of the 2007 International Workshop on Didymosphenia geminata; 2008: Canadian Technical Report on Fisheries and Aquatic Sciences 2795.
- 36. Clarke KR, Gorley RN. PRIMER v6: User Manual/Tutorial. Plymouth, UK: PRIMER-E; 2006.
- 37. Anderson MJ, Gorley RN, Clarke KR. PERMANOVA+ for PRIMER: Guide to Software and Statistical Methods. Plymouth, UK: PRIMER-E; 2008.
- 38. Clarke KR, Warwick RM. Change in Marine Communities: an Approach to Statistical Analysis and Interpretation. 2nd ed. Plymouth, UK: PRIMER-E; 2001.
- 39. Kruskal JB. Multidimensional scaling by optimizing goodness of fit to a nonmetric hypothesis. Psychometrika. 1964;29:1–27.
- 40. Shepard RN. The analysis of proximities: multidimensional scaling with an unknown distance function. Psychometrika 1962;27:125–40.
- 41. Ramette A. Multivariate analyses in microbial ecology. FEMS Microbiol Ecol. 2007;62(2):142–60. pmid:17892477
- 42. Sutherland S, Rodway M, Kilroy C, Jarvie B, Hughes G. The survival of Didymosphenia geminata in three rivers and associated spring-fed tributaries in the South Island of New Zealand. Southland Fish & Game report to MAF Biosecurity New Zealand. 2007.
- 43. Miller M, McKnight D, Cullis J, Greene A, Vietti K, Liptzin D. Factors controlling streambed coverage of Didymosphenia geminata in two regulated streams in the Colorado Front Range. Hydrobiologia. 2009;630(1):207–18.
- 44. Biggs BJF. New Zealand Periphyton Guideline: Detecting, Monitoring and Managing Enrichment of Streams. Wellington: Ministry for the Environment, 2000.
- 45. Cullis JDS. Removal of benthic algae in swift-flowing streams: The significance of spatial and temporal variation in shear stress and bed disturbance [PhD Thesis]: University of Colorado; 2011.
- 46. Schweiger W, Ashton IW, Muhlfeld CC, Jones LA, Bahls LL. The distribution and abundance of a nuisance native alga, Didymosphenia geminata, in streams of Glacier National Park: climate drivers and management implications. Park Sci. 2011;28(88–91).
- 47. Kumar S, Spaulding SA, Stohlgren TJ, Hermann KA, Schmidt TS, Bahls LL. Potential habitat distribution for the freshwater diatom Didymosphenia geminata in the continental US. Front Ecol Environ. 2009;7(8):415–20.
- 48. USGS. Water-resources data for the United States, Water Year 2010: Water-Data Report WDR-US-2010, site 01362200 Troy, NY: US Geological Survey; 2010. Available from: http://wdr.water.usgs.gov/wy2010/pdfs/01362200.2010.pdf.
- 49. USGS. Water-resources data for the United States, Water Year 2009: Water-Data Report WDR-US-2009, site 01362200 Troy, NY: US Geological Survey; 2009. Available from: http://wdr.water.usgs.gov/wy2009/pdfs/01362200.2009.pdf.
- 50. Lindstrøm EA, Skulberg O. Didymosphenia geminata—a native diatom species of Norwegian rivers coexisting with the Atlantic salmon. Proceedings of the 2007 International Workshop on Didymosphenia geminata; 2008: Canadian Technical Report on Fisheries and Aquatic Sciences 2795.
- 51. Lagerstedt MA. Didymosphenia geminata; an example of a biosecurity leak in New Zealand [M.S. Thesis]: University of Canterbury, New Zealand; 2007.
- 52. Kilroy C, Biggs B, Vieglais C. Didymosphenia geminata in New Zealand: a science response to help manage an unwanted, invasive freshwater diatom. Abstract book, ASLO Aquatic Sciences meeting; 2007; Santa Fe, New Mexico.
- 53. Newcombe CP, Macdonald DD. Effects of Suspended Sediments on Aquatic Ecosystems. N Am J Fish Manage. 1991;11(1):72–82.
- 54. Francoeur SN, Biggs BJ. Short-term effects of elevated velocity and sediment abrasion on benthic algal communities. Hydrobiologia. 2006;561:59–69.
- 55. McHale MR, Siemion J. Turbidity and suspended sediment in the upper Esopus Creek watershed, Ulster County, New York. U.S. Geological Survey, 2014 Scientific Investigations Report 2014–5200.
- 56. Larned S, Biggs B, Blair N, Burns C, Jarvie B, Jellyman D, et al. Ecology of Didymosphenia geminata in New Zealand: habitat and ecosystem effects–Phase 2. NIWA Client Report CHC2006-086 For Biosecurity New Zealand. 2006.
- 57. James D, Chipps S. The influence of Didymosphenia geminata on fisheries resources in Rapid Creek, South Dakota–an eight year history. Proceedings of the Wild Trout X Symposium–Conserving Wild Trout; 2010.
- 58. Smith AJ. Upper Esopus Creek Biological Assessment Albany, NY: New York State Department of Environmental Conservation; 2013. Available from: http://www.dec.ny.gov/docs/water_pdf/barupperesopuscreek09.pdf.
- 59. Duffy BT, Smith AJ, Heitzman DL, Abele LE. Upper Esopus Creek Biological Assessment Albany, NY: New York State Department of Environmental Conservation; 2011. Available from: http://www.dec.ny.gov/docs/water_pdf/sbuupesopscr08.pdf.
- 60. Smith AJ, Bode RW, Novak MA, Abele LE, Heitzman DL, Duffy BT. Upper Esopus Creek Biological Assessment Albany, NY: New York State Department of Environmental Conservation; 2008. Available from: http://www.dec.ny.gov/docs/water_pdf/sbuupesopscr07.pdf.
- 61. Baldigo BP, George SG, Keller WT. Fish assemblages in the Upper Esopus Creek: Current status, variability, and controlling factors. Northeast Nat. in press.
- 62. Ross T. Effects Of Anthropogenic Stream Alteration On Brown Trout Habitat, Movement And Physiology [M.S. Thesis]. Ithaca, NY: Cornell University; 2012.
- 63. Smith AJ, Bode RW, Kleppel GS. A nutrient biotic index (NBI) for use with benthic macroinvertebrate communities. Ecol Indicators. 2007;7(2):371–86.