- Split View
-
Views
-
Cite
Cite
Adam T. Greer, Robert K. Cowen, Cedric M. Guigand, Margaret A. McManus, Jeff C. Sevadjian, Amanda H.V. Timmerman, Relationships between phytoplankton thin layers and the fine-scale vertical distributions of two trophic levels of zooplankton, Journal of Plankton Research, Volume 35, Issue 5, September/October 2013, Pages 939–956, https://doi.org/10.1093/plankt/fbt056
- Share Icon Share
Abstract
Thin layers of phytoplankton are well documented, common features in coastal areas globally, but little is known about the relationships of these layers to higher trophic levels. We deployed the In Situ Ichthyoplankton Imaging System (ISIIS) to simultaneously quantify the three trophic levels of plankton, including phytoplankton, primary consumers (copepods and appendicularians) and secondary consumers (gelatinous zooplankton). Over a 2-week sampling period, phytoplankton thin layers, primarily composed of Pseudo-nitzschia spp., were common on two of the five sampling days. Imagery showed copepods aggregating in zones of lower chlorophyll-a fluorescence, while appendicularians were more common at greater depths and higher chlorophyll-a levels. All gelatinous zooplankton generally increased in abundance with depth. Bolinopsis spp. ctenophores underwent a ‘bloom,’ and they were the only species observed to aggregate within phytoplankton thin layers. The vertical separation between copepods, phytoplankton and gelatinous zooplankton suggests that copepods may use the surface waters as a predation refuge, only performing short migrations into favorable feeding zones where gelatinous predators are much more abundant. Thin layers containing dense diatom aggregates obstruct light reaching deeper waters (>10 m), which may allow gelatinous zooplankton to avoid visual predation as well as improve the effectiveness of contact predation with copepod prey.
INTRODUCTION
Thin layers are dense aggregations of phytoplankton or zooplankton spanning a few centimeters to meters in depth and sometimes several kilometers horizontally (Dekshenieks et al., 2001; McManus et al., 2003). The concentrations of organisms within thin layers can be orders of magnitude greater than above or below the layer (Donaghay et al., 1992), and persist on time scales of hours to days depending upon physical, chemical and biological conditions (Sullivan et al., 2010). Thin layers are of interest ecologically because they may serve as zones of enhanced biological interactions in the vertical dimension (Alldredge et al., 2002), much like fronts do in the horizontal, with the trophic impact of a thin layer increasing in relation to its temporal persistence (Cowles et al., 1998). As thin layers can occur in a variety of marine systems including estuaries (Donaghay et al., 1992), coastal shelves (Cowles and Desiderio, 1993) and fjords (Holliday et al., 1998; Dekshenieks et al., 2001; Alldredge et al., 2002), they may be important contributors to community structure in shallow water environments.
Trophic interactions in relation to phytoplankton thin layers are influenced by the degree of spatial overlap between thin layers, grazers and zooplankton predators. While it may seem that grazers would seek an aggregated food source within thin layers, several studies have produced counter-intuitive results where grazers were found to spend a majority of time just outside of a thin layer (Bochdansky and Bollens, 2004; Benoit-Bird et al., 2010; Talapatra et al., 2013). One possible explanation for these observations is that zooplankton predators (including gelatinous zooplankton) may influence the distribution of grazers through predation or modification of grazer behavior. Linking the distribution of phytoplankton, zooplankton and gelatinous zooplankton aggregations is challenging due to sampling limitations. For example, the relative positioning of different zooplankton taxa in relation to phytoplankton thin layers is poorly described (Holliday et al., 1998, 2003, 2010) because it is difficult or impossible to distinguish acoustic returns from organisms of similar acoustic impedance (typically similarly sized). Further, common gelatinous zooplankton are extremely difficult to sample with traditional net sampling systems [e.g. MOCNESS (Wiebe et al., 1976)], which destroy fragile gelatinous bodies. Gelatinous zooplankton are also thought to have low acoustic detectability, except for large (>10 cm) specimens (Monger et al., 1998; Brierley et al., 2005), thus their association with thin layers and grazers was not known prior to the present study.
Although gelatinous zooplankton are known to aggregate within temperature discontinuities (Arai, 1976; Graham et al., 2001), field studies relating gelatinous zooplankton distributions to the well-understood physical processes of thin layer formation are limited. The frontal zone at the edge of the upwelling shadow has been implicated as a retention mechanism for the large (30 cm bell diameter) scyphomedusa Chrysaora fuscescens, with highest concentrations located near the thermocline (Graham, 1993). Hydromedusae have been found to be abundant in Monterey Bay (Raskoff, 2002) and consistently aggregate in salinity discontinuities, regardless of whether or not prey is present indicating that they may use physical cues to aggregate there (Frost et al., 2010).
In Monterey Bay, CA, USA, thin layers of phytoplankton typically form during upwelling favorable (northwesterly) winds when the northern (sheltered) region of the bay tends to be thermally stratified (Graham and Largier, 1997; McManus et al., 2008). The upwelling season spans between the months of March–October (Pennington and Chavez, 2000). During upwelling events, cold, nutrient-rich filaments cross the mouth of the bay (Rosenfeld et al., 1994) and a cyclonic gyre, known as the ‘upwelling shadow,’ forms in the northeastern part of the bay. When present, the upwelling shadow increases the surface water residence time in the bay (Breaker and Broenkow, 1994; Graham and Largier, 1997). Water in the upwelling shadow is sheltered from the winds, which results in decreased mixing, and diurnal heating results in high thermal stratification (Graham, 1993). These characteristics provide optimal conditions for thin layer formation. Density discontinuities formed via thermal stratification may serve as a mechanism to slow the sinking rate of particles, creating thin phytoplankton layers of non-motile organisms such as diatoms (MacIntyre et al., 1995; Alldredge et al., 2002). When the upwelling winds subside or reverse direction, i.e. ‘relaxation events,’ surface waters in the upwelling shadow zone are advected from the bay to the northwest within 2–3 days if the event persists (Woodson et al., 2009). Under these conditions, California current waters, characterized by relatively warm temperatures, low salinity and low nutrient concentrations (Rosenfeld et al., 1994; Ramp et al., 2005; Ryan et al., 2008, 2009), are advected towards shore to replace exiting waters, disrupting the stratified conditions favorable to thin layer formation.
To elucidate the trophic effects of phytoplankton thin layers and directly sample the gelatinous community, we deployed a towed In Situ Ichthyoplankton Imaging System (ISIIS) (Cowen and Guigand, 2008) to synoptically sample zooplankton abundance and related environmental parameters (including chlorophyll-a fluorescence) in northern Monterey Bay. The goals of this study were (i) to describe the environmental conditions and their relationship to thin layers, (ii) to relate fine scale changes in abundance of different zooplankton taxa to environmental conditions and (iii) to compare and contrast the fine-scale vertical distributions of three distinct planktonic trophic levels (phytoplankton, primary consumers and gelatinous zooplankton), their relationship to physical discontinuities and the implications for predator/prey interactions.
METHOD
Study site and sampling period
Monterey Bay is an open embayment located on the central coast of California. The study area was in the northeastern part of the bay where the upwelling shadow tends to form during active upwelling (Fig. 1). All sampling was conducted over 5 days between 28 June and 7 July 2010 in conjunction with a large physical oceanographic study investigating lateral mixing on the inner shelf (Woodson et al., in revision).
Moored vertical profiler and wind data
To continuously monitor hydrographic and chlorophyll-a fluorescence variability in northern Monterey Bay waters, an autonomous vertical profiler (Brooke Ocean Technology SeaHorse) was deployed on the 20 m isobath in the center of the ISIIS sampling array at 36.9325°N 121.9244°W. The SeaHorse provided profiles every 30 min from near-surface to near-bottom with a Sea Bird 19 CTD, a Sea Bird 43 Oxygen sensor and a Wet Labs WetStar fluorometer with a sampling frequency of 4 Hz. For the period between 5 July and 7 July 2010, the SeaHorse collected data on 54% of its profiles. Some data loss occurred near the end of deployment due to issues with the gripper mechanism, creating gaps in the time series.
Regional scale wind forcing (and thus upwelling/relaxation conditions) was obtained using hourly averaged wind velocities from the National Data Buoy Center's buoy number 46042 located at 36.789°N122.404°W, ∼51 km WSW from the profiling sites. The direction and magnitude of these winds indicate whether or not active upwelling was occurring.
Imaging system sampling
The ISIIS contains a Piranha II line scan camera from Dalsa with 68 μm pixel resolution, imaging plankton in the size range of 680 μm to 13 cm. The ISIIS uses a shadowgraph imaging technique, in which a collimated light source is projected across a sampled parcel of water, and the silhouettes created by the plankton blocking the light source are then captured by the camera (Cowen and Guigand, 2008). The ISIIS line scan camera shoots a continuous image at 36 000 scan lines per second, but parses the image into frames that correspond to a 13 × 13 cm area of view with a depth of field of ca. 35 cm, giving an individual image volume of 6.4 L. At the usual tow speed (2.5 ms−1), 1 m3 of water is sampled every 7.7 s. In addition to the camera system, ISIIS was equipped with motor actuated fins for depth control, a Doppler velocity log (600 micro, Navquest) and environmental sensors including a CTD (SBE49, Seabird Electronics) and fluorometer [ECO FL (RT), Wetlabs chlorophyll-a fluorescence]. All sensors sampled water <1 m above the imaging area at a rate of 2 Hz, and a correction was applied to address this offset.
The system was deployed in a ‘tow-yo’ fashion behind an 18 m research vessel, the R/V Shana Rae, running at a constant speed of 2.5 ms−1 through the water, with ∼4–5 water column undulations per transect from the near surface to a maximum depth of ∼18 m to stay at least 5 m from the bottom. A total of 10 transects (3 km each) centered over the 20 m isobath were performed on each day of sampling (Fig. 1), with the exception of 5 July, when only 5 transects were completed due to temporary technical issues. All sampling was conducted during the day except for samples on 30 June, which were collected at night. Due to the optical technique utilized by ISIIS, ambient light has no effect on image quality (Cowen and Guigand, 2008).
Thin layer identification
In the ISIIS profiles, thin layers of fluorescence were defined as in Sullivan et al. (Sullivan et al., 2010), but the maximum layer thickness criterion was adjusted from <3 to <5 m (sensu Sevadjian et al., submitted). The method by Sullivan et al. (Sullivan et al., 2010) describes the intensity and thickness of the chlorophyll maximum in each fluorescence profile. ISIIS chlorophyll-a profiles were first smoothed using a low pass filter, and the first derivative was calculated to determine background fluorescence, layer thickness, and intensity. The depth of the fluorometer was calculated using the pressure sensor mounted on the upper pod of ISIIS and corrected for the minor physical offset and vehicle pitch.
Phytoplankton sampling
Water samples were taken to assess the phytoplankton community on the days of ISIIS sampling using a 5-L Aquatic Research Instruments discrete point water sampling bottle attached <0.2 m from the intake tubes on a high-resolution profiler on a separate small vessel. The bottle was triggered using real-time depth and chlorophyll-a fluorescence activity, with samples taken near surface (∼3 m), within the chlorophyll maximum and below the chlorophyll maximum (>15 m). The bottle samples were stored on ice and gently mixed before analysis in the laboratory within 5 h of returning from the field. Phytoplankton were counted and identified to genus using a PhycoTech 0.066 mL phytoplankton counting cell at 40× magnification and a Zeiss A1 Axoioscope. Triplicates were performed for each sample and averaged.
Image data analysis
ISIIS images were viewed using the VisionNow software (Boulder Imaging, Inc.), and a standard ‘flat-fielding’ transformation was applied to remove background noise from each image. For quantification of gelatinous zooplankton, three vertical profiles from each of 10 daily transects were examined (30 profiles on each sampling day, 15 profiles on 5 July). The profiles were approximately evenly spaced across the study area (∼1 km separation) and corresponded to different sections of the transect (offshore, middle and inshore). We refer to these units as ‘profiles’ even though they spanned a horizontal range of ∼300 m and an average depth of 15 m (∼5 m above bottom). Gelatinous zooplankton were identified to genus or species level. Identification of the ctenophores and siphonophores was verified by experts in the field.
Measurements of length and angle of swimming orientation were made for a subset of ctenophore specimens (n = 225 for Pleurobrachia spp., n = 274 for Bolinopsis spp., n = 200 for Bolinopsis spp. size frequencies). The angle of orientation of the mouth (an indicator of general swimming direction) was measured using ImageJ v1.44p (Rasband, 1997–2012) by bisecting the imaged specimen from aboral to oral end. The software recorded the orientation angle and length of each specimen. Differences in size distributions of Bolinopsis spp. among days were assessed using one-sided Kolmogorov–Smirnov tests. For sampling the highly abundant copepods and appendicularians (< 5 mm in size), 1/6th of a frame was subsampled on one profile from each transect (10 profiles per day) and 1 out of every 20 frames was examined, generating ∼2 samples per m of depth on each profile. This subsampling procedure was sufficient because there were typically several copepods and appendicularians in each image.
Because diatom aggregates (also known as ‘flocculations’ or ‘flocs’; Alldredge et al., 2002) were overwhelmingly the most common specimens in the images, we could use automated particle analysis available in ImageJ (Rasband, 1997–2012, ‘Particle Analyzer’) to calculate the particle size and shape features from one profile through a thin layer on 5 July. First, the images were thresholded (converted to black and white pixels) and particles >250 pixels in size were counted and sized, which effectively removed all of the copepods and appendicularians. ImageJ's ‘Particle Analyzer’ measured the major and minor axes of an ellipse fit to each particle, and the information from the maximum sized particle per frame was extracted. Since most of the diatom flocs were oblong in shape, the minor axis was used as a proxy for diameter. The ratio of the major to minor axis was used as a shape descriptor (lower values mean the particle is round and higher values more oblong). The values in pixels were converted to mm based on a known 13-cm field of view. Thin layer particles from 28 June were not measured because flocs were so dense that the individual particles could not be distinguished for analysis.
Data processing and generalized linear mixed models
Physical and biological data were merged to yield to precise environmental values (temperature, salinity, depth and chlorophyll-a fluorescence) for each gelatinous organisms found in the images. To obtain the average vertical distributions, all 30 profiles from each day were used, and the volume sampled in each 1 m depth bin was calculated based on the amount of time ISIIS spent in that depth bin. Counts of copepods and appendicularians were converted to concentrations by calculating the volume of water sampled in 1/6th of an ISIIS image and multiplying. Data analysis and visualization were performed in R (v2.15.2) (R Core Team, 2012) using the packages ‘plyr’ (Wickham, 2011) and ‘ggplot2’, respectively (Wickham, 2009).
Physical and biological data were processed to quantify the average environmental values and the total count of each gelatinous taxon for each m3 of water sampled with ISIIS. This was accomplished by binning gelatinous organism data by the time it took to sample 1 m3 of water (7.7 s) to yield ind. m−3 for each taxon; only the five most abundant taxa [Pleurobrachia spp., Bolinopsis spp. (small and large size classes), Eutonina indicans, Muggiaea spp. and Sphaeronectes spp.] were binned. Then, using the average timestamp within each gelatinous organism bin, each count per m3 was matched to the nearest timestamps of chlorophyll-a fluorescence, temperature, salinity and depth. Depths and the thin layer analysis (discussed previously) were used to create thin layer categorical variables: each m3 was characterized as being ‘above a thin layer’, ‘within a thin layer’, ‘below a thin layer’, or ‘thin layer absent from the profile.’
Generalized linear mixed models (GLMMs) with log link function were implemented to determine the influence of sampling date, depth, fine-scale chlorophyll-a fluorescence and thin layer categorical variables (no thin layer, above, below and within the thin layer) on the number of gelatinous organisms per m3. Although many ecological processes are expected to be nonlinear, preliminary plots showed some linear trends with respect to several explanatory variables. Also, the use of a linear model allowed for more intuitive interpretation of coefficients (similar interpretation as standard least squares modeling). Due to the collinearity of depth and temperature, temperature was dropped from the model and salinity was removed because there was little change along each profile. Models were fitted in R (v2.15.2) (R Core Team, 2012) with the package ‘lme4’ (Bates et al., 2012) using the Laplace approximation, and the significance of model coefficients was assessed using Wald z tests. Interaction terms were not used because they can obscure the effects of the individual predictor variables (Gotelli and Ellison, 2004), and preliminary analyses produced no indication of strong interactions.
GLMM is an approach to generalized linear modeling that allows for a correlation structure to be incorporated into a model by differentiating between fixed and random effects. For this study, the spacing of organism counts was on the scale of meters; therefore, there was the potential for adjacent observations to be correlated (violation of independence) within a sampling profile (Zuur et al., 2009). This correlation structure was accounted for by including the profile number as a random effect in the model. The model output produces coefficients that are proportional to the effect of a 1 unit increase of the variable on the expected concentrations of the organisms. Therefore, a positive coefficient indicates that the response variable (organism concentration) will increase in proportion to the value of that coefficient. GLMMs allow for modeling non-normal distributions (e.g. Poisson, binomial, etc.), and the use of random effects which essentially enables the user to parse out clusters of data that may contain autocorrelation (Bolker et al., 2009). Correlation of residuals that was present when performing ordinary generalized linear modeling was eliminated through the use of a random profile effect.
Indices of patchiness and spatial overlap
RESULTS
Water column properties and thin layers
The study period (28 June to 7 July 2010) began after ∼14 days of generally consistent upwelling favorable winds (Fig. 2a). However, a small relaxation event on 28 June and brief relaxation events shortly after likely led to a breakdown in thermal stratification in the study area by 30 June (Fig. 2a and b). Although upwelling (northwesterly) winds became consistent after 30 June, persistent, strong stratification did not occur again until 5 July. The winds on 5 July weakened and changed direction, but stratification did not decrease until the final day of sampling (7 July) when the winds had been calm for 2–3 days. 7 July was marked by lower overall chlorophyll fluorescence (a mean chlorophyll-a maximum of 1.004 V) and reduced thermal stratification. Salinity was fairly constant throughout the study (Δ0.06), but the autonomous profiler showed a distinct small drop in salinity just prior to the 2 July sampling date (Δ 0.03), potentially indicating the influx of California current water (J.P. Ryan, Monterey, pers. comm.). However, temperature/salinity diagrams from a concurrent study (Woodson et al., in revision) demonstrated that offshore influx of water was minimal.
Thin layers were present on all days of sampling with the exception of 30 June, which was the only sampling performed at night. Thin layers were most common on 28 June and 5 July (thin layers observed on 47 and 87% of profiles with thin layers, respectively), which were also the days with higher chlorophyll-a maxima and strong thermocline and oxycline (Table I, Fig. 2b). Pseudo-nitzschia spp. was the dominant phytoplankter on all sampling days with some temporal variation in concentration (see Timmerman, 2012).
Date . | 28 June . | 30 June . | 2 July . | 5 July . | 7 July . |
---|---|---|---|---|---|
Mean chlorophyll max (V) (SE) | 2.163 (0.086) | 1.306 (0.031) | 1.504 (0.027) | 1.610 (0.022) | 1.004 (0.023) |
Mean depth of chlorophyll max (m) (SE) | 8.620 (0.291) | 12.249 (0.354) | 6.189 (0.310) | 7.031 (0.162) | 9.626 (0.381) |
Profiles with thin layers | 14 | 0 | 7 | 13 | 6 |
Percent of profiles with thin layers | 47 | 0 | 23 | 87 | 20 |
Mean copepod concentration (ind. m−3) (SE) | 5301 (97.54) | 5067 (81.66) | 4282 (71.78) | 8453 (264.37) | 4644 (90.55) |
Maximum copepod concentration (ind. m−3) | 20 421 | 23 612 | 29 994 | 37 014 | 22 336 |
Copepods Lloyd's patchiness | 1.23 | 1.26 | 1.38 | 1.53 | 1.42 |
Mean appendicularian concentration (ind. m−3) (SE) | 5870 (79.81) | 2176 (38.45) | 5474 (104.01) | 4835 (126.16) | 10 610 (102.47) |
Maximum appendicularian concentration (ind. m−3) | 17 868 | 10 848 | 22 974 | 19 145 | 25 527 |
Appendicularian Lloyd's patchiness | 1.08 | 1.17 | 1.56 | 1.29 | 1.08 |
Date . | 28 June . | 30 June . | 2 July . | 5 July . | 7 July . |
---|---|---|---|---|---|
Mean chlorophyll max (V) (SE) | 2.163 (0.086) | 1.306 (0.031) | 1.504 (0.027) | 1.610 (0.022) | 1.004 (0.023) |
Mean depth of chlorophyll max (m) (SE) | 8.620 (0.291) | 12.249 (0.354) | 6.189 (0.310) | 7.031 (0.162) | 9.626 (0.381) |
Profiles with thin layers | 14 | 0 | 7 | 13 | 6 |
Percent of profiles with thin layers | 47 | 0 | 23 | 87 | 20 |
Mean copepod concentration (ind. m−3) (SE) | 5301 (97.54) | 5067 (81.66) | 4282 (71.78) | 8453 (264.37) | 4644 (90.55) |
Maximum copepod concentration (ind. m−3) | 20 421 | 23 612 | 29 994 | 37 014 | 22 336 |
Copepods Lloyd's patchiness | 1.23 | 1.26 | 1.38 | 1.53 | 1.42 |
Mean appendicularian concentration (ind. m−3) (SE) | 5870 (79.81) | 2176 (38.45) | 5474 (104.01) | 4835 (126.16) | 10 610 (102.47) |
Maximum appendicularian concentration (ind. m−3) | 17 868 | 10 848 | 22 974 | 19 145 | 25 527 |
Appendicularian Lloyd's patchiness | 1.08 | 1.17 | 1.56 | 1.29 | 1.08 |
Date . | 28 June . | 30 June . | 2 July . | 5 July . | 7 July . |
---|---|---|---|---|---|
Mean chlorophyll max (V) (SE) | 2.163 (0.086) | 1.306 (0.031) | 1.504 (0.027) | 1.610 (0.022) | 1.004 (0.023) |
Mean depth of chlorophyll max (m) (SE) | 8.620 (0.291) | 12.249 (0.354) | 6.189 (0.310) | 7.031 (0.162) | 9.626 (0.381) |
Profiles with thin layers | 14 | 0 | 7 | 13 | 6 |
Percent of profiles with thin layers | 47 | 0 | 23 | 87 | 20 |
Mean copepod concentration (ind. m−3) (SE) | 5301 (97.54) | 5067 (81.66) | 4282 (71.78) | 8453 (264.37) | 4644 (90.55) |
Maximum copepod concentration (ind. m−3) | 20 421 | 23 612 | 29 994 | 37 014 | 22 336 |
Copepods Lloyd's patchiness | 1.23 | 1.26 | 1.38 | 1.53 | 1.42 |
Mean appendicularian concentration (ind. m−3) (SE) | 5870 (79.81) | 2176 (38.45) | 5474 (104.01) | 4835 (126.16) | 10 610 (102.47) |
Maximum appendicularian concentration (ind. m−3) | 17 868 | 10 848 | 22 974 | 19 145 | 25 527 |
Appendicularian Lloyd's patchiness | 1.08 | 1.17 | 1.56 | 1.29 | 1.08 |
Date . | 28 June . | 30 June . | 2 July . | 5 July . | 7 July . |
---|---|---|---|---|---|
Mean chlorophyll max (V) (SE) | 2.163 (0.086) | 1.306 (0.031) | 1.504 (0.027) | 1.610 (0.022) | 1.004 (0.023) |
Mean depth of chlorophyll max (m) (SE) | 8.620 (0.291) | 12.249 (0.354) | 6.189 (0.310) | 7.031 (0.162) | 9.626 (0.381) |
Profiles with thin layers | 14 | 0 | 7 | 13 | 6 |
Percent of profiles with thin layers | 47 | 0 | 23 | 87 | 20 |
Mean copepod concentration (ind. m−3) (SE) | 5301 (97.54) | 5067 (81.66) | 4282 (71.78) | 8453 (264.37) | 4644 (90.55) |
Maximum copepod concentration (ind. m−3) | 20 421 | 23 612 | 29 994 | 37 014 | 22 336 |
Copepods Lloyd's patchiness | 1.23 | 1.26 | 1.38 | 1.53 | 1.42 |
Mean appendicularian concentration (ind. m−3) (SE) | 5870 (79.81) | 2176 (38.45) | 5474 (104.01) | 4835 (126.16) | 10 610 (102.47) |
Maximum appendicularian concentration (ind. m−3) | 17 868 | 10 848 | 22 974 | 19 145 | 25 527 |
Appendicularian Lloyd's patchiness | 1.08 | 1.17 | 1.56 | 1.29 | 1.08 |
Automated particle counting revealed changes in the fine-scale structure of diatom flocs above, within and below a thin layer. Above the layer, particles were few in number but increased in concentration and size with depth (Fig. 3a). The higher major-to-minor axis ratio indicates the particles above the layer were oblong in shape and likely sinking (Fig. 3b). Within the high chlorophyll-a thin layer, flocs were numerous, large in size and relatively round (Fig. 3a and b). Below the layer, flocs were few in number but highly variable in size and shape.
Copepod and appendicularian abundances and vertical distributions
Copepods and appendicularians displayed strong temporal variability in both their overall abundances and vertical distributions. The mean copepod concentration was highest on days when thin layers were present, but the vertical distributions indicate that they did not aggregate within zones of high chlorophyll-a fluorescence (Table I, Fig. 4). The mean appendicularian concentration was highest on the last day of sampling (7 July). Lloyd's index of patchiness was consistently >1 for both groups on all days, indicating that there was aggregation (Table I).
The GLMMs revealed differences between the vertical distributions of appendicularians and copepods. When a thin layer was present in a profile, appendicularians were more abundant in all zones of the water column, while the copepod concentrations were not influenced by the presence of thin layers (Table II). The two groups had opposing responses to chlorophyll-a fluorescence; copepods, although not influenced by thickness of the chlorophyll-a maximum, generally tended to aggregate outside of zones of high chlorophyll-a and appendicularians were slightly more common when chlorophyll-a was higher (Table II). The model results are supported by 1 m bin averaged concentrations (Fig. 4), where copepods displayed a bimodal distribution with peak concentrations outside the zones of high chlorophyll-a fluorescence, whereas appendicularians showed a weak increase with depth and near zones of higher chlorophyll-a. The spatial overlap index showed that appendicularians consistently had more fine-scale spatial overlap with gelatinous zooplankton than did copepods, with an exception being the one night sampling period when the overlap indices were approximately equal (30 June, Fig. 5).
Model parameters . | Copepods . | Appendicularians . | ||
---|---|---|---|---|
Coefficient . | P . | Coefficient . | P . | |
Date 30 June | −0.0303 | ns | −0.7718 | ** |
Date 2 July | −0.4422 | * | −0.4557 | * |
Date 5 July | 0.2197 | ns | −0.3453 | ns |
Date 7 July | −0.3388 | ns | 0.6776 | ** |
Depth | −0.0020 | ns | 0.0073 | *** |
Fluorometry | −0.5006 | *** | 0.1858 | *** |
Above layer | 0.2708 | ns | 0.3737 | * |
Below layer | −0.2777 | ns | 0.4160 | * |
Within layer | 0.0519 | ns | 0.3708 | * |
Model parameters . | Copepods . | Appendicularians . | ||
---|---|---|---|---|
Coefficient . | P . | Coefficient . | P . | |
Date 30 June | −0.0303 | ns | −0.7718 | ** |
Date 2 July | −0.4422 | * | −0.4557 | * |
Date 5 July | 0.2197 | ns | −0.3453 | ns |
Date 7 July | −0.3388 | ns | 0.6776 | ** |
Depth | −0.0020 | ns | 0.0073 | *** |
Fluorometry | −0.5006 | *** | 0.1858 | *** |
Above layer | 0.2708 | ns | 0.3737 | * |
Below layer | −0.2777 | ns | 0.4160 | * |
Within layer | 0.0519 | ns | 0.3708 | * |
Profile number was treated as a random effect to account for spatial autocorrelation between nearby samples. Model coefficients and significance levels are shown. Significance was assessed using Wald z tests.
Model formula:
Count ∼ profile number + date + depth + fluorometry + thin layer category.
P value significance codes are as follows: P > 0.05 = ns, 0.01 < *P < 0.05, 0.001 < **P < 0.01 and ***P < 0.001.
Model parameters . | Copepods . | Appendicularians . | ||
---|---|---|---|---|
Coefficient . | P . | Coefficient . | P . | |
Date 30 June | −0.0303 | ns | −0.7718 | ** |
Date 2 July | −0.4422 | * | −0.4557 | * |
Date 5 July | 0.2197 | ns | −0.3453 | ns |
Date 7 July | −0.3388 | ns | 0.6776 | ** |
Depth | −0.0020 | ns | 0.0073 | *** |
Fluorometry | −0.5006 | *** | 0.1858 | *** |
Above layer | 0.2708 | ns | 0.3737 | * |
Below layer | −0.2777 | ns | 0.4160 | * |
Within layer | 0.0519 | ns | 0.3708 | * |
Model parameters . | Copepods . | Appendicularians . | ||
---|---|---|---|---|
Coefficient . | P . | Coefficient . | P . | |
Date 30 June | −0.0303 | ns | −0.7718 | ** |
Date 2 July | −0.4422 | * | −0.4557 | * |
Date 5 July | 0.2197 | ns | −0.3453 | ns |
Date 7 July | −0.3388 | ns | 0.6776 | ** |
Depth | −0.0020 | ns | 0.0073 | *** |
Fluorometry | −0.5006 | *** | 0.1858 | *** |
Above layer | 0.2708 | ns | 0.3737 | * |
Below layer | −0.2777 | ns | 0.4160 | * |
Within layer | 0.0519 | ns | 0.3708 | * |
Profile number was treated as a random effect to account for spatial autocorrelation between nearby samples. Model coefficients and significance levels are shown. Significance was assessed using Wald z tests.
Model formula:
Count ∼ profile number + date + depth + fluorometry + thin layer category.
P value significance codes are as follows: P > 0.05 = ns, 0.01 < *P < 0.05, 0.001 < **P < 0.01 and ***P < 0.001.
Gelatinous zooplankton abundance and vertical distribution
A total of 35 208 gelatinous animals were identified in the ISIIS profiles. Five species groups were most common, including the ctenophores Pleurobrachia spp. and Bolinopsis spp., E. indicans and the siphonophores Muggiaea spp. and Sphaeronectes spp. (Fig. 6). Several other species were also encountered in the images, including rare taxa such as Sarsia tubulosa, hydromedusae from the family Pandeidae, anthomedusae from the family Moerisiidae and the scyphomedusa C. fuscescens. Bolinopsis spp. was divided into two size classes because initially many small ctenophores without tentacles were extracted from the images, but it was later determined that almost all (>99%) of these ctenophores were juvenile Bolinopsis spp.
Overall abundances and the vertical distributions of the five most common gelatinous taxa (Fig. 6) changed dramatically throughout the study period. On 28 June abundances were low, but there were distinct vertical patterns with Pleurobrachia spp. most abundant near the surface, all taxa relatively common near the chlorophyll-a fluorescence maximum, and E. indicans was more common at depth (Fig. 7). By 2 July, E. indicans and small Bolinopsis spp. were the most abundant, with E. indicans highly patchy and aggregated at depth (Table III, Fig. 7). The fifth of July was marked by a significant increase in Bolinopsis spp. ctenophores (an instantaneous exponential growth rate of 0.54 between 2 July and 5 July—assuming no advection), as well as peak abundances for the other two ctenophore groups (Fig. 7). The two siphonophore species (Muggiaea spp. and Sphaeronectes spp.) also reached relatively high abundances on 5 July and remained common on 7 July (Fig. 7). The concentration of all six common gelatinous organism taxa combined tended to increase in relation to depth on all days of sampling (except for 28 June), regardless of the prevalence of thin layers (Table IV, Fig. 7).
Species . | Date . | Mean concentration (inds m−3) . | Standard error . | Maximum concentration (inds m−3) . | Lloyd's patchiness . | Proportion of m3 sampled with >1 gelatinous organisms per m3 . |
---|---|---|---|---|---|---|
Pleurobrachia spp. | 28 June | 0.8175 | 0.1002 | 9 | 2.7515 | 0.4079 |
30 June | 2.0591 | 0.2762 | 20 | 2.5039 | 0.6517 | |
2 July | 1.3516 | 0.1991 | 16 | 2.9293 | 0.4726 | |
5 July | 4.3396 | 0.4989 | 37 | 2.3829 | 0.7090 | |
7 July | 3.6206 | 0.7682 | 44 | 3.9983 | 0.5832 | |
Small Bolinopsis spp. | 28 June | 0.7299 | 0.1050 | 13 | 3.8088 | 0.3274 |
30 June | 3.1833 | 0.4730 | 54 | 3.5806 | 0.7088 | |
2 July | 5.3894 | 0.9034 | 68 | 3.7191 | 0.6144 | |
5 July | 6.0784 | 1.1129 | 72 | 3.6205 | 0.5896 | |
7 July | 1.1194 | 0.2392 | 28 | 5.7178 | 0.2955 | |
Large Bolinopsis spp. | 28 June | 0.5188 | 0.06850 | 8 | 3.3044 | 0.2880 |
30 June | 0.3014 | 0.09649 | 10 | 7.5519 | 0.1690 | |
2 July | 1.4234 | 0.3218 | 35 | 5.5256 | 0.3932 | |
5 July | 28.1157 | 2.4464 | 144 | 2.0584 | 0.8022 | |
7 July | 8.8258 | 1.4861 | 75 | 3.3481 | 0.6634 | |
E. indicans | 28 June | 0.5206 | 0.07849 | 10 | 3.2553 | 0.3077 |
30 June | 0.3747 | 0.05064 | 5 | 2.0379 | 0.2668 | |
2 July | 4.6711 | 1.4786 | 251 | 14.4520 | 0.6522 | |
5 July | 1.7201 | 0.2696 | 14 | 2.3148 | 0.5858 | |
7 July | 1.0391 | 0.2672 | 18 | 4.6522 | 0.3601 | |
Muggiaea spp. | 28 June | 0.1270 | 0.03056 | 3 | 3.3450 | 0.1020 |
30 June | 0.1568 | 0.02687 | 2 | 2.1685 | 0.1303 | |
2 July | 0.3081 | 0.05988 | 5 | 3.5151 | 0.1966 | |
5 July | 0.8209 | 0.2435 | 15 | 5.3587 | 0.3396 | |
7 July | 0.6732 | 0.06055 | 7 | 2.0605 | 0.3933 | |
Sphaeronectes spp. | 28 June | 0.1825 | 0.05863 | 7 | 10.6655 | 0.1002 |
30 June | 0.5601 | 0.05812 | 9 | 2.3047 | 0.3483 | |
2 July | 0.7713 | 0.1240 | 11 | 3.5151 | 0.2665 | |
5 July | 0.7351 | 0.2733 | 33 | 5.3587 | 0.2575 | |
7 July | 1.3659 | 0.1373 | 17 | 2.0605 | 0.4462 |
Species . | Date . | Mean concentration (inds m−3) . | Standard error . | Maximum concentration (inds m−3) . | Lloyd's patchiness . | Proportion of m3 sampled with >1 gelatinous organisms per m3 . |
---|---|---|---|---|---|---|
Pleurobrachia spp. | 28 June | 0.8175 | 0.1002 | 9 | 2.7515 | 0.4079 |
30 June | 2.0591 | 0.2762 | 20 | 2.5039 | 0.6517 | |
2 July | 1.3516 | 0.1991 | 16 | 2.9293 | 0.4726 | |
5 July | 4.3396 | 0.4989 | 37 | 2.3829 | 0.7090 | |
7 July | 3.6206 | 0.7682 | 44 | 3.9983 | 0.5832 | |
Small Bolinopsis spp. | 28 June | 0.7299 | 0.1050 | 13 | 3.8088 | 0.3274 |
30 June | 3.1833 | 0.4730 | 54 | 3.5806 | 0.7088 | |
2 July | 5.3894 | 0.9034 | 68 | 3.7191 | 0.6144 | |
5 July | 6.0784 | 1.1129 | 72 | 3.6205 | 0.5896 | |
7 July | 1.1194 | 0.2392 | 28 | 5.7178 | 0.2955 | |
Large Bolinopsis spp. | 28 June | 0.5188 | 0.06850 | 8 | 3.3044 | 0.2880 |
30 June | 0.3014 | 0.09649 | 10 | 7.5519 | 0.1690 | |
2 July | 1.4234 | 0.3218 | 35 | 5.5256 | 0.3932 | |
5 July | 28.1157 | 2.4464 | 144 | 2.0584 | 0.8022 | |
7 July | 8.8258 | 1.4861 | 75 | 3.3481 | 0.6634 | |
E. indicans | 28 June | 0.5206 | 0.07849 | 10 | 3.2553 | 0.3077 |
30 June | 0.3747 | 0.05064 | 5 | 2.0379 | 0.2668 | |
2 July | 4.6711 | 1.4786 | 251 | 14.4520 | 0.6522 | |
5 July | 1.7201 | 0.2696 | 14 | 2.3148 | 0.5858 | |
7 July | 1.0391 | 0.2672 | 18 | 4.6522 | 0.3601 | |
Muggiaea spp. | 28 June | 0.1270 | 0.03056 | 3 | 3.3450 | 0.1020 |
30 June | 0.1568 | 0.02687 | 2 | 2.1685 | 0.1303 | |
2 July | 0.3081 | 0.05988 | 5 | 3.5151 | 0.1966 | |
5 July | 0.8209 | 0.2435 | 15 | 5.3587 | 0.3396 | |
7 July | 0.6732 | 0.06055 | 7 | 2.0605 | 0.3933 | |
Sphaeronectes spp. | 28 June | 0.1825 | 0.05863 | 7 | 10.6655 | 0.1002 |
30 June | 0.5601 | 0.05812 | 9 | 2.3047 | 0.3483 | |
2 July | 0.7713 | 0.1240 | 11 | 3.5151 | 0.2665 | |
5 July | 0.7351 | 0.2733 | 33 | 5.3587 | 0.2575 | |
7 July | 1.3659 | 0.1373 | 17 | 2.0605 | 0.4462 |
Species . | Date . | Mean concentration (inds m−3) . | Standard error . | Maximum concentration (inds m−3) . | Lloyd's patchiness . | Proportion of m3 sampled with >1 gelatinous organisms per m3 . |
---|---|---|---|---|---|---|
Pleurobrachia spp. | 28 June | 0.8175 | 0.1002 | 9 | 2.7515 | 0.4079 |
30 June | 2.0591 | 0.2762 | 20 | 2.5039 | 0.6517 | |
2 July | 1.3516 | 0.1991 | 16 | 2.9293 | 0.4726 | |
5 July | 4.3396 | 0.4989 | 37 | 2.3829 | 0.7090 | |
7 July | 3.6206 | 0.7682 | 44 | 3.9983 | 0.5832 | |
Small Bolinopsis spp. | 28 June | 0.7299 | 0.1050 | 13 | 3.8088 | 0.3274 |
30 June | 3.1833 | 0.4730 | 54 | 3.5806 | 0.7088 | |
2 July | 5.3894 | 0.9034 | 68 | 3.7191 | 0.6144 | |
5 July | 6.0784 | 1.1129 | 72 | 3.6205 | 0.5896 | |
7 July | 1.1194 | 0.2392 | 28 | 5.7178 | 0.2955 | |
Large Bolinopsis spp. | 28 June | 0.5188 | 0.06850 | 8 | 3.3044 | 0.2880 |
30 June | 0.3014 | 0.09649 | 10 | 7.5519 | 0.1690 | |
2 July | 1.4234 | 0.3218 | 35 | 5.5256 | 0.3932 | |
5 July | 28.1157 | 2.4464 | 144 | 2.0584 | 0.8022 | |
7 July | 8.8258 | 1.4861 | 75 | 3.3481 | 0.6634 | |
E. indicans | 28 June | 0.5206 | 0.07849 | 10 | 3.2553 | 0.3077 |
30 June | 0.3747 | 0.05064 | 5 | 2.0379 | 0.2668 | |
2 July | 4.6711 | 1.4786 | 251 | 14.4520 | 0.6522 | |
5 July | 1.7201 | 0.2696 | 14 | 2.3148 | 0.5858 | |
7 July | 1.0391 | 0.2672 | 18 | 4.6522 | 0.3601 | |
Muggiaea spp. | 28 June | 0.1270 | 0.03056 | 3 | 3.3450 | 0.1020 |
30 June | 0.1568 | 0.02687 | 2 | 2.1685 | 0.1303 | |
2 July | 0.3081 | 0.05988 | 5 | 3.5151 | 0.1966 | |
5 July | 0.8209 | 0.2435 | 15 | 5.3587 | 0.3396 | |
7 July | 0.6732 | 0.06055 | 7 | 2.0605 | 0.3933 | |
Sphaeronectes spp. | 28 June | 0.1825 | 0.05863 | 7 | 10.6655 | 0.1002 |
30 June | 0.5601 | 0.05812 | 9 | 2.3047 | 0.3483 | |
2 July | 0.7713 | 0.1240 | 11 | 3.5151 | 0.2665 | |
5 July | 0.7351 | 0.2733 | 33 | 5.3587 | 0.2575 | |
7 July | 1.3659 | 0.1373 | 17 | 2.0605 | 0.4462 |
Species . | Date . | Mean concentration (inds m−3) . | Standard error . | Maximum concentration (inds m−3) . | Lloyd's patchiness . | Proportion of m3 sampled with >1 gelatinous organisms per m3 . |
---|---|---|---|---|---|---|
Pleurobrachia spp. | 28 June | 0.8175 | 0.1002 | 9 | 2.7515 | 0.4079 |
30 June | 2.0591 | 0.2762 | 20 | 2.5039 | 0.6517 | |
2 July | 1.3516 | 0.1991 | 16 | 2.9293 | 0.4726 | |
5 July | 4.3396 | 0.4989 | 37 | 2.3829 | 0.7090 | |
7 July | 3.6206 | 0.7682 | 44 | 3.9983 | 0.5832 | |
Small Bolinopsis spp. | 28 June | 0.7299 | 0.1050 | 13 | 3.8088 | 0.3274 |
30 June | 3.1833 | 0.4730 | 54 | 3.5806 | 0.7088 | |
2 July | 5.3894 | 0.9034 | 68 | 3.7191 | 0.6144 | |
5 July | 6.0784 | 1.1129 | 72 | 3.6205 | 0.5896 | |
7 July | 1.1194 | 0.2392 | 28 | 5.7178 | 0.2955 | |
Large Bolinopsis spp. | 28 June | 0.5188 | 0.06850 | 8 | 3.3044 | 0.2880 |
30 June | 0.3014 | 0.09649 | 10 | 7.5519 | 0.1690 | |
2 July | 1.4234 | 0.3218 | 35 | 5.5256 | 0.3932 | |
5 July | 28.1157 | 2.4464 | 144 | 2.0584 | 0.8022 | |
7 July | 8.8258 | 1.4861 | 75 | 3.3481 | 0.6634 | |
E. indicans | 28 June | 0.5206 | 0.07849 | 10 | 3.2553 | 0.3077 |
30 June | 0.3747 | 0.05064 | 5 | 2.0379 | 0.2668 | |
2 July | 4.6711 | 1.4786 | 251 | 14.4520 | 0.6522 | |
5 July | 1.7201 | 0.2696 | 14 | 2.3148 | 0.5858 | |
7 July | 1.0391 | 0.2672 | 18 | 4.6522 | 0.3601 | |
Muggiaea spp. | 28 June | 0.1270 | 0.03056 | 3 | 3.3450 | 0.1020 |
30 June | 0.1568 | 0.02687 | 2 | 2.1685 | 0.1303 | |
2 July | 0.3081 | 0.05988 | 5 | 3.5151 | 0.1966 | |
5 July | 0.8209 | 0.2435 | 15 | 5.3587 | 0.3396 | |
7 July | 0.6732 | 0.06055 | 7 | 2.0605 | 0.3933 | |
Sphaeronectes spp. | 28 June | 0.1825 | 0.05863 | 7 | 10.6655 | 0.1002 |
30 June | 0.5601 | 0.05812 | 9 | 2.3047 | 0.3483 | |
2 July | 0.7713 | 0.1240 | 11 | 3.5151 | 0.2665 | |
5 July | 0.7351 | 0.2733 | 33 | 5.3587 | 0.2575 | |
7 July | 1.3659 | 0.1373 | 17 | 2.0605 | 0.4462 |
Model parameters . | Pleurobrachia spp. . | Bolinopsis spp. (small) . | Bolinopsis spp. (large) . | E. indicans . | Muggiaea spp. . | Sphaeronectes spp. . | ||||||
---|---|---|---|---|---|---|---|---|---|---|---|---|
Coefficient . | P . | Coefficient . | P . | Coefficient . | P . | Coefficient . | P . | Coefficient . | P . | Coefficient . | P . | |
30 June | 1.0040 | *** | 1.6893 | *** | −0.3890 | ns | −0.1937 | ns | 0.4787 | ns | 1.7076 | *** |
2 July | 0.5112 | * | 1.9158 | *** | 1.0617 | *** | 1.6779 | *** | 0.9152 | ** | 1.8341 | *** |
5 July | 1.8751 | *** | 2.5120 | *** | 4.5574 | *** | 1.2858 | *** | 1.9323 | *** | 1.8660 | *** |
7 July | 1.3433 | *** | 0.0226 | ns | 3.6138 | *** | 0.4729 | ns | 2.2906 | *** | 2.9988 | *** |
Depth | 0.0816 | *** | 0.0256 | *** | 0.0902 | *** | 0.1165 | *** | 0.0855 | *** | 0.1926 | *** |
Chl-a | 0.2125 | *** | 0.8031 | *** | 0.8783 | *** | −0.0062 | ns | 0.3243 | ** | 0.1995 | ns |
Above layer | −0.0031 | ns | −0.6358 | * | 0.2510 | ns | −0.1618 | ns | 0.0306 | ns | −0.9434 | ** |
Below layer | 0.2975 | ns | 0.1850 | ns | 0.8286 | *** | 0.6243 | * | 0.5855 | * | 0.2359 | ns |
Within layer | 0.3037 | ns | 0.4358 | * | 0.7570 | ** | 0.1522 | ns | 0.3661 | ns | 0.2080 | ns |
Model parameters . | Pleurobrachia spp. . | Bolinopsis spp. (small) . | Bolinopsis spp. (large) . | E. indicans . | Muggiaea spp. . | Sphaeronectes spp. . | ||||||
---|---|---|---|---|---|---|---|---|---|---|---|---|
Coefficient . | P . | Coefficient . | P . | Coefficient . | P . | Coefficient . | P . | Coefficient . | P . | Coefficient . | P . | |
30 June | 1.0040 | *** | 1.6893 | *** | −0.3890 | ns | −0.1937 | ns | 0.4787 | ns | 1.7076 | *** |
2 July | 0.5112 | * | 1.9158 | *** | 1.0617 | *** | 1.6779 | *** | 0.9152 | ** | 1.8341 | *** |
5 July | 1.8751 | *** | 2.5120 | *** | 4.5574 | *** | 1.2858 | *** | 1.9323 | *** | 1.8660 | *** |
7 July | 1.3433 | *** | 0.0226 | ns | 3.6138 | *** | 0.4729 | ns | 2.2906 | *** | 2.9988 | *** |
Depth | 0.0816 | *** | 0.0256 | *** | 0.0902 | *** | 0.1165 | *** | 0.0855 | *** | 0.1926 | *** |
Chl-a | 0.2125 | *** | 0.8031 | *** | 0.8783 | *** | −0.0062 | ns | 0.3243 | ** | 0.1995 | ns |
Above layer | −0.0031 | ns | −0.6358 | * | 0.2510 | ns | −0.1618 | ns | 0.0306 | ns | −0.9434 | ** |
Below layer | 0.2975 | ns | 0.1850 | ns | 0.8286 | *** | 0.6243 | * | 0.5855 | * | 0.2359 | ns |
Within layer | 0.3037 | ns | 0.4358 | * | 0.7570 | ** | 0.1522 | ns | 0.3661 | ns | 0.2080 | ns |
Models use organism counts per m3 as a response variable to sampling date, depth, chlorophyll-a fluorescence (chl-a) and thin layer categorical variables. Sampling date coefficients are fit relative to the first date of sampling (28 June) and thin layer variables are fit relative to the ‘no thin layer present’ category. The profile number was used as a random effect to account for the correlation structure within a profile.
Model parameters . | Pleurobrachia spp. . | Bolinopsis spp. (small) . | Bolinopsis spp. (large) . | E. indicans . | Muggiaea spp. . | Sphaeronectes spp. . | ||||||
---|---|---|---|---|---|---|---|---|---|---|---|---|
Coefficient . | P . | Coefficient . | P . | Coefficient . | P . | Coefficient . | P . | Coefficient . | P . | Coefficient . | P . | |
30 June | 1.0040 | *** | 1.6893 | *** | −0.3890 | ns | −0.1937 | ns | 0.4787 | ns | 1.7076 | *** |
2 July | 0.5112 | * | 1.9158 | *** | 1.0617 | *** | 1.6779 | *** | 0.9152 | ** | 1.8341 | *** |
5 July | 1.8751 | *** | 2.5120 | *** | 4.5574 | *** | 1.2858 | *** | 1.9323 | *** | 1.8660 | *** |
7 July | 1.3433 | *** | 0.0226 | ns | 3.6138 | *** | 0.4729 | ns | 2.2906 | *** | 2.9988 | *** |
Depth | 0.0816 | *** | 0.0256 | *** | 0.0902 | *** | 0.1165 | *** | 0.0855 | *** | 0.1926 | *** |
Chl-a | 0.2125 | *** | 0.8031 | *** | 0.8783 | *** | −0.0062 | ns | 0.3243 | ** | 0.1995 | ns |
Above layer | −0.0031 | ns | −0.6358 | * | 0.2510 | ns | −0.1618 | ns | 0.0306 | ns | −0.9434 | ** |
Below layer | 0.2975 | ns | 0.1850 | ns | 0.8286 | *** | 0.6243 | * | 0.5855 | * | 0.2359 | ns |
Within layer | 0.3037 | ns | 0.4358 | * | 0.7570 | ** | 0.1522 | ns | 0.3661 | ns | 0.2080 | ns |
Model parameters . | Pleurobrachia spp. . | Bolinopsis spp. (small) . | Bolinopsis spp. (large) . | E. indicans . | Muggiaea spp. . | Sphaeronectes spp. . | ||||||
---|---|---|---|---|---|---|---|---|---|---|---|---|
Coefficient . | P . | Coefficient . | P . | Coefficient . | P . | Coefficient . | P . | Coefficient . | P . | Coefficient . | P . | |
30 June | 1.0040 | *** | 1.6893 | *** | −0.3890 | ns | −0.1937 | ns | 0.4787 | ns | 1.7076 | *** |
2 July | 0.5112 | * | 1.9158 | *** | 1.0617 | *** | 1.6779 | *** | 0.9152 | ** | 1.8341 | *** |
5 July | 1.8751 | *** | 2.5120 | *** | 4.5574 | *** | 1.2858 | *** | 1.9323 | *** | 1.8660 | *** |
7 July | 1.3433 | *** | 0.0226 | ns | 3.6138 | *** | 0.4729 | ns | 2.2906 | *** | 2.9988 | *** |
Depth | 0.0816 | *** | 0.0256 | *** | 0.0902 | *** | 0.1165 | *** | 0.0855 | *** | 0.1926 | *** |
Chl-a | 0.2125 | *** | 0.8031 | *** | 0.8783 | *** | −0.0062 | ns | 0.3243 | ** | 0.1995 | ns |
Above layer | −0.0031 | ns | −0.6358 | * | 0.2510 | ns | −0.1618 | ns | 0.0306 | ns | −0.9434 | ** |
Below layer | 0.2975 | ns | 0.1850 | ns | 0.8286 | *** | 0.6243 | * | 0.5855 | * | 0.2359 | ns |
Within layer | 0.3037 | ns | 0.4358 | * | 0.7570 | ** | 0.1522 | ns | 0.3661 | ns | 0.2080 | ns |
Models use organism counts per m3 as a response variable to sampling date, depth, chlorophyll-a fluorescence (chl-a) and thin layer categorical variables. Sampling date coefficients are fit relative to the first date of sampling (28 June) and thin layer variables are fit relative to the ‘no thin layer present’ category. The profile number was used as a random effect to account for the correlation structure within a profile.
GLMMs with a random profile effect accounting for small-scale autocorrelation revealed different influences of biological and physical parameters among the gelatinous taxa. Model coefficients showed taxa were more likely to be present and in higher concentrations (counts) at greater depths, and they were highly aggregated in their distributions (Tables III and IV). Eutonina indicans was most abundant at greater depths and had the most aggregated distribution, with extremely high concentrations on 2 July (a maximum concentration of 251 ind. m−3). The concentrations of Bolinopsis spp. were positively influenced by higher chlorophyll-a (Table IV), while the concentrations of other taxa were either slightly elevated (Pleurobrachia spp.) or unaffected. Only Bolinopsis spp. was more abundant within thin layers, but the concentrations of several taxa were higher below thin layers (Table IV).
Measurements of in situ orientation and daily size structure were made as indicators of behavioral characteristics and growth, respectively, of ctenophores within the bay. The in situ orientation of Bolinopsis spp. indicated that they were typically cruising the water column vertically, with oral end (and lobes) pointed down, perpendicular to the thin layers and/or chlorophyll maximum (Fig. 8a), while Pleurobrachia spp. had less consistent orientation (Fig. 8b), though with a predominately vertical (oral end up) orientation. The length/frequency histograms of a subsample Bolinopsis spp. for each day show a significant increase in size on subsequent sampling days (one-sided Kolmogorov–Smirnov tests, Fig. 9). A linear growth rate extracted from the modal size (rounded to nearest 0.2 mm) was 0.45 mm day−1, assuming ctenophores are being sampled from the same population.
DISCUSSION
Fine-scale distribution data obtained by ISIIS showed strong vertical heterogeneity in three separate trophic levels including phytoplankton, primary consumers and gelatinous zooplankton, with >2 orders of magnitude difference between surface and 15 m depth concentrations of the gelatinous zooplankton. Water column stratification likely caused passive accumulation of non-motile diatoms Pseudo-nitzschia spp., the most abundant phytoplankter in the study area. Thin layers were most common on days when the water column was thermally stratified (28 June and 5 July), and GLMMs fit to organism counts per m3 showed differing patterns between zooplankton taxa in relation to thin layers, chlorophyll-a fluorescence and depth.
Diatom dominated thin layers and primary consumers
The temporal persistence was the primary difference between the thin layers on 28 June and 5 July, and the lack of motility of the dominant phytoplankter had a large influence on the assumed mechanism of formation. The layer on 28 June could have been present for up to 10 days due to consistent, strong upwelling winds before the study began. Because of these winds, the upwelling shadow had likely retained water within the bay for up to 14 days, with Pseudo-nitzschia spp. located at the base of the pycnocline. These non-motile diatoms have been previously shown to accumulate at density interfaces (MacIntyre et al., 1995; Cheriton et al., 2009), and high thermal stratification likely kept Pseudo-nitzschia spp. in a thin layer near the pycnocline on 28 June, but the diatoms were possibly actively growing above the layer where oxygen was relatively high (ΔO2 2.1 mL L−1). The thin layer on 5 July was also driven by thermal stratification but was newly formed and had higher abundances of smaller phytoplankton (Timmerman, 2012). Size/shape descriptors obtained from thin layer particles on 5 July is consistent with the concept of diatoms flocculating and settling at a thermocline because within the thin layer, particles were larger, rounder and more abundant. In other cases, thin layers are composed of motile species of phytoplankton, such as the dinoflagellate Akashiwo sanguinea (Rines et al., 2010). In these cases, behavioral cues, rather than sinking and settlement in diatoms, may play a more important role.
Although there was a non-uniform distribution of fluorescence on all days, the primary diatom consumers, copepods, were not found in high concentrations within thin layers. A GLMM showed a negative influence of chlorophyll-a on copepod concentrations. Copepods were typically most abundant near the surface, where chlorophyll-a fluorescence was low, and also sometimes aggregated below thin layers (5 July). Copepod aggregations outside of chlorophyll-a maxima have been documented previously using water samples and acoustic methods (Herman, 1983; Nielsen et al., 1990; Jaffe et al., 1998; Alldredge et al., 2002; Holliday et al., 2003, 2010; McManus et al., 2005). The thin layers on 28 July were deeper, more intense in their chlorophyll-a fluorescence and thicker than the layers on 5 July (Table I), which may have affected the vertical distributions of copepods. Pseudo-nitzschia spp. in our study was producing the toxin domoic acid, which has been demonstrated to inhibit grazing in krill (Bargu et al., 2006; Timmerman et al., in press), and it may have a similar effect on other grazers, such as copepods, thereby increasing the persistence of a bloom. Although domoic acid has been shown experimentally to have no effect on feeding and growth in copepods (Lincoln et al., 2001), copepods could still preferentially occupy waters that contain fewer toxins. There were indications of bimodal vertical distributions in copepods on all sampling days, with the exception of 30 June, which was the only night sampling period and had a deeper and diffuse chlorophyll-a profile (no thin layers). Even on the night of 30 June, the copepods were not observed to aggregate inside the zone of high chlorophyll-a fluorescence. Diatoms within thin layers tended to form dense flocs; it is possible that on the edges of thin layers, copepods may find diatoms in a more ‘palatable’ form, or they utilize short excursions into the layers to feed (Bochdansky and Bollens, 2004).
Vertical distributions of zooplankton predators and prey
Although depth was not a significant predictor variable in the GLMM, copepod peak abundances tended to occur within the shallowest 5 m of the water column, where chlorophyll-a levels and gelatinous predator abundances were low. For every profile, the lowest measured values of chlorophyll-a fluorescence were found near the surface, which tended to be dominated by copepods. In contrast, Bolinopsis spp. ctenophores were particularly abundant within and below thin layers (and at greater depths in general). The spatial overlap index between copepods and gelatinous zooplankton was <1 indicating spatial separation, which is consistent with the concept of predation avoidance. Thus, it is possible that the copepods in Monterey Bay use the surface as a predation refuge and make short bouts into the deeper predator abundant waters to feed. Such a strategy would allow the copepods to minimize their predation risk while still being able to feed and is consistent with the ‘predation avoidance’ hypothesis of vertical migration (Zaret and Suffern, 1976) supported by several copepod behavior studies (e.g. Neill, 1990; Dawidowicz et al., 1990). Field evidence has also shown that copepods vertically migrate depending on their body condition, only risking vertical migration into predator dense zones if their oil sac is depleted (Hays et al., 2001). In addition, copepods grow faster and reproduce more in water with higher temperatures (Bonnet et al., 2009), so occupying the surface waters could have multiple positive effects on the population.
Appendicularians had a positive relationship with depth and chlorophyll-a fluorescence, potentially due to less predation pressure (when compared with copepods). Of the species in this study, appendicularians are only known to be preyed upon by E. indicans (Wrobel and Mills, 1998) and so when compared with copepods, the predation risk of occupying thin layers is likely substantially less. Food availability, therefore, may be the primary influence on their distributions. Appendicularians are capable of feeding on a wide range of phytoplankton sizes, including pico and nanoplankton, which are abundant in nutrient-poor offshore waters of Monterey Bay (Ryan et al., 2009). The positive association of appendicularians within and around thin layers may indicate that properties associated with thin layer formation (i.e., water column stability and concentrated sources of phytoplankton) are favorable to appendicularian population growth.
Gelatinous zooplankton (predators of copepods and appendicularians) were more abundant at greater depths on all sampling days, regardless of the stratification regime present, and this pattern may exist due to a combination of three factors: negative effects of copepods on gelatinous zooplankton reproduction, contact predator advantages and avoidance of visual predators. Copepods demonstrated less spatial overlap with gelatinous zooplankton compared with appendicularians. Although the ctenophores in this study tend to consume copepods (Reeve et al., 1978; Greene et al., 1986), they have a complex relationship with these prey. Studies of ctenophores using enclosures have shown that high concentrations of copepods can severely reduce the survival of larval ctenophores (Stanlaw et al., 1981). Therefore, above certain prey concentrations, ctenophores may become less effective predators due to inhibited reproductive success. Another potential benefit to gelatinous zooplankton occupying deeper waters is that ambient light levels influence the competitive advantage of contact versus visual predation. An experiment by Sørnes and Aksnes (2004) demonstrated that Bolinopsis spp. feeding reaches an asymptote at high prey concentrations, and when light levels are lower, tactile predation becomes more advantageous than visual predation. While irradiance was not measured in this study, a previous study in the same area found that several wavelengths of light experience ∼100-fold reductions below a thin layer of Pseudo-nitzschia spp. in Monterey Bay (Sullivan et al., 2010). Attenuation of light was almost certain to increase in the region of the pycnocline during our study, with diatom flocs larger and more abundant within thin layers. The thin layer of Pseudo-nitzschia spp., therefore, could set up microhabitats in which visual predation on copepods dominates above the thin layer (perhaps by planktivorous fishes), while tactile predation is more advantageous below. Jellies also have many visual predators (Oviatt and Kremer, 1977; Link and Ford, 2006) that can likely be avoided by occupying zones of lower light levels. Therefore, surface waters may be zones of the water column where jellies are less fecund, inferior predators (outcompeted by visual predators) and exposed to higher visual predation by their own suite of predators.
Gelatinous zooplankton and diatom thin layers
The idea of gelationous organisms aggregating at density discontinuities is not new (Arai, 1976; Mills, 1984; Graham et al., 2001; Jacobsen and Norrbin, 2009), but the results of this study demonstrate that only one species (Bolinopsis spp.) tends to aggregate within thin layers and vertical density discontinuities. This may indicate that the vertical swimming of Bolinopsis spp. results in increased probability of thin layer encounter, as species with a horizontal or random swimming pattern would not be expected to encounter vertical density discontinuities as frequently as a vertical swimmer. Indeed, Bolinopsis spp. behavior in this study as well as others (Reeve et al., 1978; Toyokawa et al., 2003) suggests precise vertical orientation in this species is common. In contrast, Pleurobrachia spp. displayed a more variable swimming pattern, perhaps related its behavior of swimming straight or in an arc to spread its tentacles for feeding (Reeve et al., 1978). When feeding, Pleurobrachia spp. drifts passively (Purcell, 1991) and so it could be subjected to microturbulence causing variable orientation.
Of all the gelatinous zooplankton examined, only Bolinopsis spp. was found in significantly higher abundance within thin layers. This raises the question why a zooplankton predator would occupy a thin layer of phytoplankton, especially when its prey items are abundant above and below the layer. The ISIIS images displayed precise vertical orientation behavior of Bolinopsis spp., which is known to cruise the water column vertically in pursuit of prey (Reeve et al., 1978). Moving through a density interface (pycnocline) would cause the vertical cruising speed of the Bolinopsis spp. to be reduced (by physical mechanisms), thereby causing an accumulation of ctenophores where passively sinking Pseudo-nitzschia spp. tended to aggregate. In addition, while traces of diatoms have been found in the guts of Bolinopsis spp., they cannot maintain their size on these prey items alone, suggesting that diatom consumption may be a method to ward off starvation (Reeve, 1980). Indeed, carbon and nitrogen stable isotope analysis demonstrates Bolinopsis spp. occupies a relatively low trophic position in the food web (Toyokawa et al., 2003). Alternatively, ctenophores may actively seek temperature discontinuities and slow their swimming speed within them. Based on mesocosm experiments, it is likely that these ctenophores are actively responding to the density discontinuities (Frost et al., 2010), but further controlled studies on Bolinopsis spp. would be needed to distinguish between these passive and active aggregation mechanisms.
Gelatinous organism blooms
The gelatinous organism ‘bloom’, defined as a rapid increase in the population abundance (Graham et al., 2001), of Bolinopsis spp. between 30 June and 5 July may have been due to retention of ctenophores within the bay coupled with extremely fast growth rates. In captivity, Bolinopsis spp. has been shown to be capable of growing over 10 mm per day, and initiates reproduction when it becomes fully lobate at a size of ∼10 mm (Greve, 1970). According to temperature/salinity diagrams, the bay waters were retained during our study (i.e. not replaced by offshore California current water; Woodson et al., in revision) and so a bloom could have formed within the bay. Tidal ellipses occurred below the thermocline (Woodson et al., in revision) that may have acted as a retention mechanism, keeping ctenophores within the bay. Thin layers are also known as zones of reduced flow (McManus et al., 2005) and so Bolinopsis spp. occupying these zones within the thermocline could have increased retention, improving the chances of a bloom. When food and flow conditions are favorable, remarkable growth rates, combined with simultaneous hermaphroditism (capable of self-fertilization) and high fecundities (Baker and Reeve, 1974; Reeve et al., 1978; Costello et al., 2006) likely allow this species to form blooms. The reduction in Bolinopsis spp. abundance on 7 July may have been due to cannibalism, which is known to occur in a closely related species (Mnemiopsis leidyi) (Javidpour et al., 2009); however, trends in abundance with depth on 7 July indicate that ctenophores were possibly aggregating within 5 m of the bottom, which we were unable to sample. The increase in the length of Bolinopsis spp. over time, low Lloyd's patchiness and the circulation patterns suggest the phenomenon observed on 5 July was a ‘true bloom’ and not just a congregation of gelatinous animals due to physical convergence (Graham et al., 2001).
On the other hand, the high abundance of E. indicans on 2 July was likely an ‘apparent bloom’ because dense aggregations were present on only about three profiles in waters deeper than the 20 m isobath, causing Lloyd's patchiness to be extremely high (14.452, Table III). In addition, laboratory obtained size at age of E. indicans suggests the organisms we sampled were >50 days old (Rees, 1978), and therefore were likely advected into the area well before the study commenced.
Importance of behavior in thin layer trophic interactions
Zooplankton studies using high resolution instruments sometimes appear to show avoidance thin layers of diatoms. In a recent study by Talapatra et al. (Talapatra et al., 2013), two vertical profiles with an imaging system showed thin layers of Chaetoceros socialis and high particle counts within the vicinity of the pycnocline. In the first ascent, the zooplankton were located outside several, but not all, of the particle concentration peaks. In the second ascent, the zooplankton appeared to avoid the layers with elevated particle concentration (Talapatra et al., 2013). Other field studies demonstrate that trophic interactions can occur within thin layers on sub-hour time scales, and the grazers spend more time within thin layers when a higher fraction of water column phytoplankton is contained within the layers (Benoit-Bird et al., 2010; Holliday et al., 2010). The temporal resolution of sampling in our study may not have been adequate to capture these brief events often enough to yield a statistically significant overlap.
The primary fluorescent organisms in this study were Pseudo-nitzschia spp., which lack swimming ability, so if passive accumulation was occurring across all taxa, then all observed plankton would strongly overlap with the chlorophyll-a fluorescence activity. This was not the case. In fact, similarly sized copepods and appendicularians had very different vertical distributions, indicating that behavior is an important driver of vertical distributions of zooplankton in this system. The density discontinuities present near the chlorophyll maximum combined with consistent vertical swimming of Bolinopsis spp. may have led to accumulation and spatial overlap with higher chlorophyll-a levels and thin layers. Surface waters containing low abundances of gelatinous predators may serve as a predation refuge for copepods, which tend to aggregate at the surface despite low amounts of chlorophyll-a fluorescence activity in this zone. Further studies on copepod condition in different portions of the water column would elucidate some of the drivers of the strong spatial offset we observed between copepods and chlorophyll-a fluorescence. Copepods must strike a balance between predation avoidance and feeding, and the increases in gelatinous organism abundance with depth show that there are trade-offs to venturing into zones of higher phytoplankton concentrations.
In situ imaging technology provided a unique glimpse at an often overlooked component of marine food webs, gelatinous zooplankton, and placed them into the context of thin layers, which are common features in coastal environments. Acoustic surveys have detected thin layers but, unlike optical systems, are biased towards plankton with either an exoskeleton or gas bladder. The results of this study demonstrated that zooplankton of similar size, whose distributions are thought to be driven by physical forcing, can have strong differences in their vertical distributions. Optical systems serve as an ideal platform to better understand behavioral tendencies of zooplankton through both high resolution sampling and in situ orientation information. These characteristics should be considered (in conjunction with the physical environment) for studies on the causes of vertical heterogeneity in coastal ecosystems.
FUNDING
Funding was provided by an NSF RAPID grant to R.K. Cowen No. 1035047.
ACKNOWLEDGEMENTS
We thank Dorothy Tang for helping greatly with the identification and enumeration of zooplankton in the ISIIS images. Jessica Luo provided assistance in the field. Thanks to John Walter for help with statistical modeling. Su Sponaugle and Peter Ortner's comments and edits greatly improved early versions of the manuscript. Claudia E. Mills and Philip R. Pugh verified some of the ctenophore and siphonophore identifications. We would like to give a special thanks to Jim Christmann, captain of the R/V Shana Rae for making the field work a success. Brock Woodson and Ross Timmerman assisted in setting up the SeaHorse profiler. Raphael Kudela provided laboratory space for analysis of phytoplankton samples. We would also like to thank the other members of the Lateral Mixing Project including Steven Monismith, Derek Fong, Ryan Moniz, Mark Stacey, Susan Willis, John Ryan and Olivia Cheriton for productive conversations and data sharing.
References
Author notes
Corresponding editor: Roger Harris