Increasing water clarity in Lake Ontario has led to a vertical redistribution of phytoplankton and an increased importance of the deep chlorophyll layer in overall primary productivity. We used in situ fluorometer profiles collected in lakewide surveys of Lake Ontario in 2008 to assess the spatial extent and intensity of the deep chlorophyll layer. In situ fluorometer data were corrected with extracted chlorophyll data using paired samples from Lake Ontario collected in August 2008. The deep chlorophyll layer was present offshore during the stratified conditions of late July 2008 with maximum values from 4–13 μg l−1 corrected chlorophyll a at 10 to 17 m depth within the metalimnion. Deep chlorophyll layer was closely associated with the base of the thermocline and a subsurface maximum of dissolved oxygen, indicating the feature's importance as a growth and productivity maximum. Crucial to the deep chlorophyll layer formation, the photic zone extended deeper than the surface mixed layer in mid-summer. The layer extended through most of the offshore in July 2008, but was not present in the easternmost transect that had a deeper surface mixed layer. By early September 2008, the lakewide deep chlorophyll layer had dissipated. A similar formation and dissipation was observed in the lakewide survey of Lake Ontario in 2003.
Deep chlorophyll layers (DCL) can be important during the summer in deep oligotrophic lakes (Brooks and Torke, 1977; Moll and Stoermer, 1982; Abbott et al., 1984; Fahnenstiel and Scavia, 1987a). In Lake Michigan, 30 to 60% of areal primary production has been attributed to the DCL (Moll et al., 1984; Fahnenstiel and Scavia, 1987b); however, recent observations (2007–2011) suggest it has disappeared (Pothoven and Fahnenstiel, 2013). There are several non-exclusive hypotheses for why DCLs are formed, generally differentiating active and passive processes (Cullen, 1982). Active processes include increased growth of phytoplankton stimulated by higher nutrient availability in the metalimnion. Phytoplankton growth is dependent on sufficient light levels at these depths, and therefore the photic zone must extend below the depth of the surface mixed layer (SML) for these nutrients to be utilized (Moll and Stoermer 1982). Passive processes include sedimentation of phytoplankton cells accumulating due to buoyancy. The role of photoadaptation in DCLs is recognized; phytoplankton adapted to a low light environment typically have a higher chlorophyll to carbon ratio than light adapted phytoplankton (Pilati and Wurtsbaugh, 2003; Reynolds, 2006). Photoadaptation led Barbiero and Tuchman (2001, 2004) to question whether the DCL represented a subsurface phytoplankton biomass and production maximum in Lake Ontario and the other Great Lakes.
Remote sensing archives track changes in water clarity through attenuation (kd490) and remote sensing reflectance (Rrs555) data products. These time series suggest that summer secchi depth of Lake Ontario had nearly doubled (from 2–4 m to 5–8 m) from the time period 1979–1985 (Coastal Zone Color Scanner) to 1998–2006 (SeaWifs) (Binding et al., 2007). This change in secchi depth translates to a deepening of the photic zone from 6–11 m to 14–22 m, potentially extending the photic zone below the SML. Therefore, this increase in water clarity may have had a major impact on the development and extent of the DCL in Lake Ontario. Higher light levels at depth can potentially make DCL phytoplankton more productive if not limited by nutrients. The main objective of this article is to delineate the spatial and temporal extent of the DCL in Lake Ontario in 2008 with comparisons with 2003. A hypothesis that is tested with the lakewide data is whether DCLs occur where photic zone depth exceeds SML depth. Our approach is to use the lakewide collection of water column profiles for in situ fluorescence (corrected with extracted chl a measurements) and other parameters collected during these intensive sampling years.
Correction of in situ fluorescence readings with extracted chl a measurements
Annual US EPA Great Lakes National Program Office (GLNPO) surveys monitor eight standard offshore sites in Lake Ontario in April and August of every year. The timing in summer corresponds to highest surface water temperatures and peak stratification that are ideal for DCL formation. The timing of the Lake Ontario cruise in 2008 was August 11–13. Samples for extracted chl a were collected from discrete depths using 8L Niskin bottles on a 12 bottle array. Water samples (250 ml) were filtered onto 47 mm diameter GF/F filters under low vacuum pressure (<5 psi). Filters were wrapped in foil and frozen until analysis following EPA Standard Operating Procedure (SOP) LG 404. Filters were extracted in acetone overnight and a calibrated fluorometer (Turner Designs 10-AU, Sunnydale, CA) measured chl a (μg l−1) following the Welschmeyer (1994) non-acidification method. Continuous profiles of in situ fluorescence were collected by a Seapoint chlorophyll fluorometer (Seapoint Sensors, Inc., Exeter, NH) on a SBE-911 CTD package (Seabird Electronics, Seattle, WA). Paired extracted chl a and in situ fluorometer data are reported for each depth within the GLENDA database (http://www.epa.gov/glnpo/monitoring/data_proj/glenda/). We evaluated several models relating in situ fluorometer readings to extracted chl a measurements that included the variables vertical sample depth, site depth, and time of day (day or night). Non-linear terms were included in the selection of the most parsimonious model.
The Lake Ontario Lower foodweb Assessment (LOLA) program was a joint USA/Canada effort to monitor the status of the lower foodweb in Lake Ontario in 2003 and 2008 (Holeck et al., 2008; Holeck et al., 2015; Rudstam et al., 2015). Several components including nutrients, phytoplankton, zooplankton, and the microbial foodweb were sampled. As many as 35 sites arranged within several north-south transects provided lakewide coverage by the US EPA R/V Lake Guardian and the Canadian Coast Guard Ship Limnos in spring, summer, and fall. Here we focus on the summer (20–26 July) and fall (2–4 September) cruises of 2008 and also include the summer (10–22 August) and fall (21–25 September) cruises of 2003.
Water column profiles were taken on the Lake Guardian using the same SBE-911 CTD package described previously. From 20–26 July 2008 (two weeks prior to the previously described calibration) Seabird profiles were collected at 35 sites throughout the lake. Downcast descent rate was near 0.5 m sec−1 and data was collected at 24 Hz. Parameters measured include in situ fluorescence by the Seapoint fluorometer. Temperature and dissolved oxygen (SBE 43 sensor) were also measured. The bottom of the surface mixed layer (SML) was defined as the depth where the temperature gradient was >1°C m−1 following Lamont (2004). Oxygen data was advanced by 2 s to account for the time for the water parcel to be pumped to the sensor. This advance improved the overlay of down and upcast profiles. Percent oxygen saturation was then calculated using ambient temperature and the equation from Baca and Arnett (1976). A transmissometer (Wetlab C-Star) measured beam attenuation (660 nm wavelength) across a 10 cm pathlength. Photosynthetically Available Radiation (PAR) Irradiance was measured using a Biospherical/Licor sensor. An extinction coefficient (kd) was determined from the decreasing slope of natural logarithm transformed PAR with depth in the top 20 m of the water column. Photic zone depth was calculated as the 1% of surface light level using the calculated kd. Secchi depth was also measured at each site sampled during daylight hours.
Latitudinal sections (40 m depth) for corrected chl a were generated using the software Ocean Data View (Version 4.3.10, Reiner Schlitzer, 2011; http://odv.awi.de) that interpolates data between the six to seven vertical profiles of each transect. The VG method used constructs a variable resolution rectangular grid where grid spacing varies with data density. Interpolation uses simple weighted averaging between grid nodes. In situ fluorescence was converted to extracted chl a using the relationship determined from samples collected at discrete depths during the GLNPO survey (see above). Bottom topography was generated from concurrent hydroacoustic surveys.
Correction of in situ fluorometer readings with extracted measurements
This depth-variable relationship was applied to each LOLA July 2008 profile for converting in situ chl profiles to extracted chl. These corrected chl a values were used in our visualization of the DCL. Note that the highest extracted chl value for DCLs within the calibration data set was 6.72 μg l−1 (in situ sensor reading of 2.42 μg l−1).
Individual site profiles
Four site profiles illustrate gradients from western to eastern Lake Ontario (Figure 2). They represent four north-south transects including the Western (Site 12, 79.35 W Longitude), Central (Site 40, 78.01 W), Sodus (Site 64, 76.93 W) and Eastern (Site 72, 76.53 W) transects (Figure 3). In July 2008, the thermocline deepened toward the east reflecting the prevailing wind patterns (Figure 2a). At Site 12 on the Western transect, the thermocline was between 5 and 8 m depth, while Site 72 on the Eastern transect had a much deeper thermocline from 18 to 22 m.
The subsurface chlorophyll maximum varied in intensity, depth of occurrence and thickness. The Central transect had the strongest maximum to 13 μg l−1 corrected chl a within the metalimnion at 16 to 20 m depth (Figure 2b). Note that this chl value is well above the range of our calibration. Site 12 in the Western transect also had a strong (7 μg l−1) DCL that was shallower (8–12 m). Site 64 on the Sodus transect had a comparable (7 μg l−1) DCL at 18 m depth. Site 72 on the Eastern transect had no DCL, with highest chl (4–8 μg l−1) within the surface layer to 12 m depth.
The three sites with prominent DCLs also had subsurface maxima in oxygen saturation and beam attenuation (particle maximum). Oxygen saturation maxima deepened from west to east from Site 12 (7 m), Site 40 (13 m) and Site 64 (16 m) (Figure 2c). Supersaturation levels were similar at these three sites (near 130%). The particle maxima were deeper than the oxygen saturation maxima (Figure 2d). Site 12 and 64 had the largest particle maxima (light attenuation near 6% m−1) at 11 and 19 m depth, while Site 40 had values half that within a broad peak from 14–20 m.
Site 40 with the strongest DCL also had the deepest photic zone of the four sites extending to 21 m (kd = 0.22), below the SML at this site. Sites 12 and 64, two other sites with DCLs had photic zone depths at 15 m (kd = 0.31) and 17 m (kd = 0.27) that were also below SML depth. Site 72, the only site of the four without a DCL, had a photic zone at 12 m (kd = 0.40), much shallower than the SML at this location.
Extent of deep chlorophyll layer for Lake Ontario
Latitudinal sections illustrate the spatial extent of the DCL in July 2008 (Figure 3). The DCL was present at offshore sites for much of the lake at a depth of 11.5 m in the west and 18 m toward the east. The feature was nearly 10 m thick with peak values of 4 to 13 μg l−1 corrected chl a. The DCL was most prominent within deep (>120 m) water columns on the southern side of the central Mississauga (Sites 39, 40, 41 of the Central transect) and Rochester (Site 65 of the Sodus transect) Basins, but also prominent in the Niagara Basin (Sites 12 and 19 of the Western transect). No site of the Eastern transect had a DCL (Figure 3 and 4). On average, 77.3% of chlorophyll within the upper 50 m of the water column was below the epilimnion for offshore sites (s.e. 4.2%, n = 14). Raw in situ fluorescence profiles at Sites 715 and 63 of the Sodus transect included a weak DCL and the correction using a higher ratio of extracted to in situ chl in the surface mixed layer decreased the contrast of surface and metalimnion chl values.
Averaging all offshore sites within the four cross-lake transects tracks the pattern of the mixed layer deepening from 6 m in the west to 14 m in the east while the photic zone varied from 16 to 19 m (Figure 5). For the three transects with prominent DCLs, the oxygen saturation maxima lay 1 to 7 m below the mixed layer, 0–5 m above the DCL, and 0–9 m above the base of the photic zone. For the Western and Central transects, light levels at the oxygen maxima were at approximately 14% of surface light levels while light levels of the oxygen maximum for the Central and Sodus transect were near the 1–3% light level. For the Eastern transect, the SML was near the base of the photic zone and no DCL or subsurface oxygen maxima were observed.
For comparison to the 2003 survey, fluorometer profiles were only collected along the Western transect during the summer survey of 2003 on August 10 (Figure 6). A strong DCL (6 μg l−1 corrected chl a) was present for offshore sites (12, 18 and 19) from 15 to 20 m depth, similar to the DCL seen at these sites in 2008. Thermocline depth was similar but more variable along this transect in 2003 than in 2008. For the offshore sites, the SML averaged 6 m offshore while the photic zone averaged 21 m. The thermal structure in both years suggested downwelling on the northern end of the transect.
Integrated epilimnion chl levels increased (averaging 1.5 to 3.1 μg l−1) from the summer to the fall of 2003 while decreasing (3.1 to 1.7 μg l−1) from the summer to the fall in 2008 (Holeck et al., 2015). Despite this difference in surface chl, DCLs were absent or weak during fall lakewide surveys in both years, and instead highest chl concentrations were within the top 5–10 m (Figure 7 for 2008). In 2003, passage of the remnants of Hurricane Isabel deepened the SML to 20 m depth at many sites. In 2008, SML depth also generally increased by 2 to 9 m from summer to fall. Photic zone depth remained generally consistent at 16 to 19 m from summer to fall for both years.
During the July 2008 lakewide cruise of Lake Ontario, a DCL (corrected chl a 4 to 13 μg l−1) was present through much of the offshore at a depth between 10–18 m. The DCL occurred within the metalimnion generally 5 m below the surface mixed layer and above the 1% light level (inferred from kd from PAR sensor) at 16–19 m depth. A DCL was also observed in August, 2003, but only the Western transect was sampled using a fluorometer. The DCLs in the two years were similar in intensity and depth despite large differences in offshore surface chl in 2003 (1.5 μg l−1) relative to 2008 (3.1 μg l−1). In both years the DCL dissipated by the time of the September cruises less than six weeks later.
In July 2008, on either end of each north-south transect there were nearshore sites <30 m depth that did not have a DCL. Therefore, the DCL was generally an offshore feature. Interaction between the water column and the shore zone causes a shear stress that likely dissipates by entrainment any established DCL.
Two deep sites (72 and 74) of the Eastern transect also did not have a DCL. It is likely that a deeper thermocline (SML near 14 m), higher surface chl (>4 μg l−1), and lower water clarity (secchi depth 5 m, kd = 0.3) contributed to the absence of the DCL at these sites. Phytoplankton likely had sufficient light and nutrients within the SML. Metalimnetic phytoplankton were potentially light limited and likely not actively growing.
As discussed previously, several mechanisms are thought to contribute to DCL formation and depth. Although we lack direct primary productivity measurements based on radiocarbon or other techniques, the dissolved oxygen maximum just above the DCL is an important indication of active photosynthesis below the SML. Therefore active growth of phytoplankton could have contributed to the DCL. High water clarity and a shallow SML in the west combined to produce light conditions sufficient for photosynthesis and phytoplankton growth within the metalimnion. Twiss et al. (2012) directly measured high phytoplankton growth rates within the Lake Ontario DCL in July 2008. Using a Fluoroprobe (bbe Moldaenke), these phytoplankton were identified as Heterokontophyta (includes diatoms) and Pyrrophyta (dinoflagellates) and also included cyanobacteria and Cryptophyta.
The correspondence of the DCL with a subsurface particle maximum suggests that the DCL in Lake Ontario also represents a biomass maxima. Transmissometers generally detect particles smaller than 20 μm that include phytoplankton that contribute to the chlorophyll signal but also non-photosynthetic organisms such as bacteria and inorganic particles. The high variation in the relationship of chlorophyll and beam attenuation in our profiles is similar to the broad variation of carbon:chl ratios (a factor of 10) previously observed in DCLs (Cullen, 1982). Alternatively, variable contributions of mixotrophic or heterotrophic organisms could contribute to this pattern (Flynn et al., 2013).
The depth of the DCL is often linked to the nutricline in marine systems, reflecting the balance of light and nutrient needs for growth of phytoplankton (Cullen, 1982). Low carbon:phosphorus ratios of seston within the DCL in the Great Lakes have been offered as evidence for this effect (Barbiero and Tuchman, 2001). In a model of Lake Superior, a nutricline was essential in the formation of a DCL and nutricline depth influenced DCL depth and magnitude (White and Matsumoto, 2012). However, nutriclines are often difficult to identify in the Great Lakes because phosphorus is at such low concentrations. In the summer cruises of 2003 and 2008 for Lake Ontario, soluble reactive phosphorus (SRP) was below detection limits within the offshore SML (Gouvêa et al., 2006; Holeck et al., 2015). In 2003, dissolved phosphorus increased with depth >100 m while total phosphorus (TP) decreased with depth at Site 64 (Gouvêa et al., 2006). In 2008, TP varied little with depth lakewide in Lake Ontario, but total filtered phosphorus (TFP) increased from 3 to 5.5 μg l−1 at a depth of 75 m much deeper than the DCL (EPA GLNPO GLENDA).
Although we cannot confirm that nutrients were more available below the SML, oxygen supersaturation at depth indicates that active photosynthesis and growth by deep phytoplankton likely contributes more to the DCL than passive processes such as settling from the epilimnion. This conclusion is consistent with studies of DCLs in Lake Michigan that found that sedimentation was only important at the time of early stratification and then at a rate much lower than in situ production (Fahnenstiel and Scavia, 1987a).
Phytoplankton loss rates from zooplankton grazing can significantly balance phytoplankton growth rates (Fahnenstiel and Scavia, 1987a). Therefore, higher grazing pressure within the epilimnion could contribute to DCL formation. However, recent evidence suggests that epilimnetic zooplankton densities in Lake Ontario have decreased sharply and zooplankton are most abundant in the metalimnion (Rudstam et al., 2015). Microzooplankton grazing can play a major role in balancing phytoplankton growth as has been measured using dilution experiments, and both phytoplankton growth and grazing rates are often higher within the metalimnion than the epilimnion (Gouvêa et al., 2006; Twiss et al., 2012). Excretion by zooplankton and nightly migrating mysids could provide nutrients to the metalimnion that could enhance growth to balance grazing losses.
Phytoplankton at depth have more chlorophyll per cell than phytoplankton at the surface, a process known as photoadaptation. Barbiero and Tuchman (2001 and 2004) attributed most DCLs in Great Lakes systems to this effect that exaggerates the biomass differences of epilimnetic and DCL phytoplankton. They also noted that phytoplankton taxa were not significantly different for the two strata. For many historical DCL observations, measured primary production is highest well within the SML at depths much shallower than the DCL (Sterner, 2010 for Lake Superior). A key distinction of our study is that the subsurface oxygen maxima emphasize the importance of production by deep phytoplankton relative to phytoplankton within the SML in Lake Ontario. A shallow SML and high water clarity provide light levels up to 15–18% of surface levels within the metalimnion. Photoadaptation may contribute to the stronger chlorophyll signal of phytoplankton at the DCL relative to actively growing phytoplankton slightly shallower at the depth of the oxygen maximum.
In situ fluorometers can further exaggerate DCLs because of a depth- dependent increase in fluorescence as surface phytoplankton acclimatized to highlight levels often fluoresce less than phytoplankton occurring at deeper depths (Cullen, 1982). However, correction of measurements with extracted measurements accounts for this bias.
Dissipation of Great Lake DCLs are commonly observed in late August and have been attributed to a deeper mixed layer and a loss of stability within the metalimnion (Brooks and Torke, 1977). Lower availability of light and nutrients at this time combine to decrease phytoplankton growth within the DCL below loss rates. A fall increase in surface layer chlorophyll as we observed in 2003 could reflect mixing of DCL phytoplankton into the SML or renewed phytoplankton growth due to nutrient influx (Brooks and Torke, 1977). The decrease of chlorophyll within the SML observed in 2008 could be attributed to zooplankton grazing or may reflect loss due to particle settling after a whiting event that occurred from 17 August to 19 September (Watkins et al., 2013). In both cases, a well-established DCL had dissipated within 6 weeks, shifting phytoplankton production to the surface. In both years, the depth of large portions of the SML in Lake Ontario had increased from summer to fall, suggesting that the balance of light and nutrient needs of phytoplankton had been disrupted. However, this explanation falls short in explaining the loss of the DCL at several sites where the base of the SML continued to be within sufficient light levels for photosynthesis.
The spatial pattern of the DCL across Lake Ontario during July 2008 confirms that SML depth plays an important role in DCL formation. Phytoplankton are likely nutrient limited within the SML. Thus, the general deepening of the SML from the west to east increasingly pushes the light needs of deeper phytoplankton populations in utilizing available nutrients at depth. Recent changes in water clarity have likely relaxed light limitation of deep phytoplankton and enhanced DCL formation and production along this east-west gradient. As mentioned previously, in the early 1980s photic zone depth during the summer in Lake Ontario was only 6–11 m deep (Binding et al., 2007), restricting the potential of DCL formation to the western portion of the lake with a shallow SML. An important distinction of the current status of the DCL in Lake Ontario is that it has higher maximum chlorophyll levels (>5 μg l−1) than DCLs of Lakes Michigan, Huron and Superior (near 1.0 to 1.5 μg l−1, EPA GLNPO data). The intensity of the DCL in Lake Ontario is more similar to that of Lake Michigan prior to 2007 than today (Pothoven and Fahnenstiel, 2013). Oxygen maxima below the SML in late summer indicate that deep primary production is significant in Lake Ontario.
We thank the vessel staff of the R/V Lake Guardian and the EPA's Great Lake National Program Office (GLNPO) for their assistance throughout this project.
This study was supported by a grant from the Environmental Protection Agency within the Great Lakes Restoration Initiative (Improving Lake Ontario Environmental Management Decisions, Grant ID # 97220700-0). Additional support was obtained from the International Joint Commission and collaborating agencies. Reference to trade names does not imply endorsement by the U.S. Government. This article is Contribution #296 of the Cornell Biological Field Station and #1857 of the USGS Great Lakes Science Center.