Kelp detritus provides high-quality food for sea urchin larvae
Abstract
Highly productive kelps release abundant particulate organic matter into the nearshore environment due to their constant fragmentation and erosion by ocean waves. The contribution of kelp detritus to coastal planktonic food webs has not previously been examined. Here, we demonstrate that detritus derived from a dominant kelp in the Northeast Pacific, Nereocystis luetkeana, provides high-quality food for planktonic sea urchin larvae. Our findings challenge the paradigm that phytoplankton are the main diet for zooplankton in nearshore regions, with implications for modeling of ocean production. Furthermore, at the benthic adult stage, sea urchins can destructively graze kelps causing the kelp ecosystem to collapse; hence, our results have implications for understanding feedback mechanisms that may determine the resilience of kelp ecosystems.
Kelps (order Laminariales) provide food and habitat for diverse organisms along temperate and polar coastlines (Dayton 1985; Steneck et al. 2002). With an estimated 82% of annual kelp production released into detrital pathways (Krumhansl and Scheibling 2012a), kelp detritus can be a main source of food for benthic detritivores (e.g., crabs and sea urchins) and suspension feeders (e.g., mussels and barnacles) in kelp beds or forests and in adjacent areas of lower productivity, such as deep basins where light is insufficient to support photosynthesis (Duggins et al. 1989; Vetter 1995; Britton-Simmons et al. 2012; Filbee-Dexter and Scheibling 2014a). As kelp detritus degrades and is colonized by microbes, its nutritional quality increases due to an increase in % nitrogen and associated decrease in C : N ratio (Mann 1988; Duggins and Eckman 1997; Norderhaug et al. 2003; Krumhansl and Scheibling 2012b; but see also Dethier et al. 2014). Degradation of kelp tissue also decreases the concentration of secondary metabolites, such as phlorotannins, which act as a grazing deterrent; thus, degradation improves palatability for diverse consumers (Duggins and Eckman 1997). Suspended particulate kelp detritus is transferred broadly into the nearshore environment (Kaehler et al. 2006; Lowe et al. 2014). It can be a dominant source of organic carbon at the ocean surface at a range of 10s of km offshore of kelp beds or forests (Kaehler et al. 2006; Ramshaw et al. 2017). However, despite the abundance and extent of suspended kelp detritus, the quality of this material as food for planktonic consumers, such as invertebrate zooplankton, has not been examined.
Invertebrate zooplankton form a critical link in marine food webs as prey for larger consumers such as fishes. Many species of zooplankton are assumed to depend almost entirely on phytoplankton for food, as indicated by patterns of increased zooplankton abundance and biomass in regions of enhanced phytoplankton productivity (Munk et al. 2003; Molinero et al. 2008). However, in areas of low phytoplankton productivity, zooplankton persist on other food sources. For example, planktotrophic echinoderm larvae living under Antarctic sea ice have been found to preferentially ingest marine bacteria over phytoplankton (Rivkin et al. 1986). Furthermore, in the euphotic zone, invertebrate zooplankton colonize and consume marine snow (organic aggregates often composed of phytodetritus) (Kiørboe 2000). Such diversity in the feeding modes of zooplankton (i.e., herbivory, bacterivory, and detritivory), in combination with evidence that kelp detritus is a high-quality food for benthic organisms, suggests that suspended kelp detritus is an important nutritional reservoir for zooplankton in temperate and polar coastal regions.
Here, we challenge the assumption that phytoplankton are the optimal diet for zooplankton in coastal waters of the Northeast Pacific by testing the quality of suspended particulate detritus derived from a dominant kelp, Nereocystis luetkeana, as food for larvae of the sea urchin Strongylocentrotus droebachiensis. In the Salish Sea, off the west coast of Washington State, U.S.A., Lowe et al. (2014) found that during periods of low phytoplankton abundance, detritus is the dominant source of organic carbon in surface waters. The essential fatty acid 20:4n-6 (a common macroalgal biomarker; Kelly and Scheibling 2012) was overrepresented in detritus-dominated waters, indicating a contribution of kelp to the detrital pool (Lowe et al. 2014). Moreover, Ramshaw et al. (2017) found through stable isotope analysis that kelp-derived detritus accounts for an average of 33% of particulate organic matter in the 20 to 63-μm size fraction during the summer off the west coast of Vancouver Island, Canada.
Our study species, S. droebachiensis, has a broad geographic distribution that includes the Northeast Pacific and the eastern and western North Atlantic (Scheibling and Hatcher 2013). Its life history involves a benthic adult and a planktotrophic larva, which allows for broad dispersal in the marine environment. S. droebachiensis is of great interest to ecologists, as the adult sea urchin facilitates nutrient cycling and energy flow in kelp ecosystems by consuming live and detrital kelp tissue and releasing this material as feces that are highly nutritional food for microbes and detritivores (Mamelona and Pelletier 2005; Sauchyn and Scheibling 2009). Adult S. droebachiensis also can destructively graze kelp beds or forests resulting in the formation of sea urchin barrens, or areas denuded of kelps and other seaweeds. Sea urchin barrens can persist for decades, and are thus considered a collapsed state of the kelp ecosystem (Filbee-Dexter and Scheibling 2014b). Here, we demonstrate that sea urchin–kelp interactions extend beyond the adult life-stage (Fig. 1). We show that sea urchin larvae thrive on kelp detritus as a source of food, with potential implications for coastal planktonic food webs and the resilience of kelp ecosystems.

Conceptual diagram of a coastal marine ecosystem showing detrital production by kelp N. luetkeana, with erosion of the blade at the distal end and the release of particulate kelp detritus into the water column. Our study shows that kelp detritus provides high-quality food for sea urchin larvae. These larvae metamorphose into juveniles and recruit to the benthos where as adults they consume large detrital fragments or destructively graze attached kelp resulting in the formation of sea urchin barrens. Illustration by Andrea Dingeldein. [Color figure can be viewed at wileyonlinelibrary.com]
Materials and methods
Larval culture
Adult S. droebachiensis were collected on 21 November 2015 from a N. luetkeana kelp forest on the southwestern tip of Shaw Island, San Juan Archipelago, Washington with SCUBA. Sea urchins were maintained in seawater aquaria in a flow-through system at Friday Harbor Laboratories, and fed ad libitum on blades of N. luetkeana, a preferred food (Vadas 1977). On 06 February 2016, spawning was induced in one female and three males by injection with 0.53 M KCl (Strathmann 1987). Eggs were fertilized with a drop of sperm from each male diluted in 0.37-μm filtered seawater (FSW) to produce embryos from three parental lineages. Fertilization success was estimated at > 95%. Following formation of a fertilization membrane (< 1 h), embryos were rinsed in FSW and parental lineages were combined. Embryos were divided into 12 replicate 4 L glass jars containing 2 L of FSW at a concentration of 1 embryo mL−1. At age 22 d when larvae had reached the eight-arm stage, cultures were diluted to 0.2 larvae mL−1 to accommodate increased feeding rates of late-stage larvae. Jars were placed in an aquarium within a flow-through seawater system to maintain cultures near ambient seawater temperature. Seawater temperature in the aquarium during the experiment was 9.4 ± 0.2°C (mean, SD). Cultures were stirred constantly with paddles operated by a 6-RPM motor mounted to a PVC frame (Strathmann 2014).
Feeding treatments and experimental design
Beginning at 4 d post-fertilization, larvae were fed every other day on one of four diets to test the quality of detritus derived from kelp N. luetkeana as a food source. At each feeding, ∼ 75% of the culture water was replaced with new FSW. To simulate kelp detritus as particulate and dissolved organic material, we ground batches of 60 g of fresh N. luetkeana blades measured on an analytical scale (0.1 g precision) in 400 mL of FSW in an industrial-grade blender, filtered the resultant slurry through a 70-μm pore nylon mesh to remove particulates larger than an early-stage larva can consume, and aged the kelp in the dark at ambient seawater temperature with constant stirring for 1 week. Kelp was aged to mimic the decay process and allow for microbial colonization (Dethier et al. 2014). The filtered, aged kelp slurry was used in feeding for 1 week so that larvae were fed kelp-derived organic matter that was between 1 and 2 weeks old. In an “Aged Kelp 100%” treatment, larvae were fed the slurry at a particle density of 5000 kelp-derived particles mL−1, as determined with a Guava easyCyte flow cytometer (Millipore EMD); a putative optimal food concentration for larvae of S. droebachiensis and another echinoid based on previous experiments with phytoplankton diets (Burdett-Coutts and Metaxas 2004; Schiopu et al. 2006). To test whether 5000 kelp-derived particles mL−1 is a satiation diet for sea urchin larvae, a second kelp treatment, “Aged Kelp 75%,” was included wherein larvae were fed 75% of the 5000 kelp-derived particles mL−1 diet (3750 kelp-derived particles mL−1). If growth and development of larvae were comparable on the 75% and 100% concentration diets (3750 vs. 5000 kelp-derived particles mL−1), this would suggest that we had reached an upper limit of food concentration required for optimal growth (i.e., satiation). To compare growth and development of larvae fed an aged kelp diet to that under an optimal high-ration phytoplankton diet for echinoid larvae (Burdett-Coutts and Metaxas 2004; Schiopu et al. 2006), in a third feeding treatment, “Phytoplankton High,” larvae were fed a mixed ∼ 1 : 1 ratio (by cell volume) of Dunaliella tertiolecta and Isochrysis galbana at an approximate volume-equivalent of 5000 cell mL−1 of D. tertiolecta (the quantity of cells of each species was determined based on an estimated volume of I. galbana ∼ 1/8 that of D. tertiolecta). In our fourth and final treatment, “Phytoplankton Low,” larvae were fed the mixed phytoplankton diet of D. tertiolecta and I. galbana at the approximate volume-equivalent of 500 cell mL−1 of D. tertiolecta, which represents a low-ration, limiting diet (Burdett-Coutts and Metaxas 2004). Each of the four feeding treatments was randomly assigned to n = 3 replicate jars.
Morphological measurements and photomicroscopy
Larval growth and development in feeding treatments were monitored with measurements over time of larval morphology, beginning 4 d post-fertilization (at first feeding), and then every other day from 10 to 22 d and every 4 d from 22 to 30 d post-fertilization. Four body dimensions were measured with a compound microscope with an ocular micrometer (5 μm resolution): length of the two postoral arm rods (larval skeleton), length of the stomach (food storage organ), length at the body midline (a measure of overall size), and length of the juvenile rudiment defined as beginning at first contact of the ectodermal invagination with the hydrocoel (Fig. 2). Formation of the juvenile rudiment within a larva is a measure of development toward competency for metamorphosis to the benthic juvenile stage. Postoral rods were visualized with cross polarization filters for measurement. Lengths of the two postoral rods per individual were averaged to give a mean postoral rod length. For each treatment, two larvae per jar (n = 3 jars per treatment) were removed and measured without replacement at each time point. Removal of larvae for sampling reduced concentrations in cultures by < 1% overall.

Sea urchin larval morphology in kelp and phytoplankton feeding treatments at three ages: (A–C) 10 d, at first observation of ectodermal invagination (In) in all treatments; (D–F) 16 d, at first observation of a juvenile rudiment (R) in all treatments; and (G–I) 30 d. Feeding treatments include: (A,D,G) Phytoplankton High; (B,E,H) Phytoplankton Low; and (C,F,I) Aged Kelp. Morphological measurements taken during the experiment are shown in (A,D) (B = body midline length; S = stomach length; P = postoral arm rod length; R = rudiment diameter). Primary podia of the developing rudiment (Po) are also shown. The farthest level of development achieved within a treatment at each age is shown. Scale bars = 150 μm.
At 10 d (first observation of ectodermal invagination in all treatments), 16 d (first observation of rudiment formation in all treatments), and 30 d post-fertilization, larvae were photographed with a smart phone camera mounted to the eyepiece of a compound microscope with a Snapzoom Universal Digiscoping Adapter.
Survival rate of larvae in each feeding treatment at 30 d was determined by removing replicate (n = 2) 20 mL aliquots of culture water from each jar and counting the number of larvae present under a dissecting microscope.
After 35 d, metamorphosis of larvae into benthic juveniles was observed in all Phytoplankton High treatment jars, and the proportion of larvae settled was determined by counting the total number of larvae and settlers (following treatment with MgCl2 to narcotize settlers to release them from the jar) in replicate (n = 2) 25 mL aliquots of culture water. Settlement was monitored thereafter in each treatment at 1–3 d intervals until 68 d when remaining cultures were terminated.
Statistical analysis
To examine the effect of the feeding treatments on larval growth and development, we used separate one-way ANOVAs to test the effect of feeding treatment (fixed factor, four levels: Aged Kelp 75%, Aged Kelp 100%, Phytoplankton High, Phytoplankton Low) on each measurement (postoral rod length, stomach length, body midline length, and rudiment diameter) at age 30 d post-fertilization (final day of measurements). We found no difference in larval growth and development between the Aged Kelp 75% and Aged Kelp 100% diets, indicating that the kelp diets were near satiation (see Fig. S1 in Supporting Information Material); therefore we reanalyzed the data using only Aged Kelp 100% (hereafter simply “Aged Kelp”). Tukey's HSD test (α = 0.05) was used to compare levels of factors that were significant in ANOVA. Statistical tests were run with Statistica 64 (StatSoft). Assumptions of homoscedasticity were tested with Levene's test (α = 0.05).
To examine the effect of the feeding treatments on larval form, we constructed bivariate plots of rudiment diameter vs. postoral rod length. Because the length of the postoral rod is a proxy for length of the ciliary band required for larval feeding, this relationship can demonstrate food-limitation or satiation as indicated by allocation of resources toward capacity to catch food (ciliated band length) and metamorphic competence (rudiment diameter) (Bertram and Strathmann 1998).
Results
Trajectories of larval growth based on morphological measurements taken at 2–6 d intervals indicate that larvae in the Aged Kelp treatment outperformed larvae in the Phytoplankton Low treatment, and larvae in the Phytoplankton High treatment outperformed larvae in both Aged Kelp and Phytoplankton Low (Figs. 2, 3). At 30 d post-fertilization, larvae differed significantly among treatments in postoral rod lengths, stomach lengths, and rudiment diameters (one-way ANOVAs; Postoral rod length: F2,4 = 16.3, p = 0.004; Stomach length: F2,4 = 29.0, p = 0.001; Rudiment diameter: F2,4 = 25.2, p = 0.001) (Fig. 3). Postoral rods and stomachs were significantly longer in Phytoplankton High than Aged Kelp and Phytoplankton Low (Tukey's HSD tests, p < 0.05), which did not differ significantly from one another (Tukey's HSD tests, p > 0.05). Rudiments were significantly larger in Aged Kelp than Phytoplankton Low, and significantly larger in Phytoplankton High than Aged Kelp and Phytoplankton Low (Tukey's HSD tests, p < 0.05). At 30 d post-fertilization there was no significant difference among treatments in body midline length (F2,4 = 3.7, p = 0.09) (Fig. 3).

Growth and development of sea urchin larvae in kelp and phytoplankton feeding treatments. Treatments include: Aged Kelp, Phytoplankton High, and Phytoplankton Low. Four different morphological measurements as in Fig. 2, with age d = 0 at fertilization (A–D). Data are means for n = 3 replicate jars per treatment (error bars = SE with some error bars within the diameter of the symbols). [Color figure can be viewed at wileyonlinelibrary.com]
The timing of metamorphosis of larvae into benthic juveniles was similar among Aged Kelp and Phytoplankton High with ≥ 50% metamorphosis of larvae observed within replicate jars after 38.7 ± 2.9 d and 35 ± 0 d (mean ± SD for n = 3 jars) in the respective treatments. In contrast, metamorphosis was achieved in only one of three replicates of Phytoplankton Low after 68 d, after which the cultures were terminated.
Bivariate plots of rudiment diameter vs. postoral arm rod length revealed phenotypic plasticity in larval form in response to the feeding treatments, with similar growth forms of larvae in Phytoplankton High and Aged Kelp and a different series of forms in Phytoplankton Low (Fig. 4). Larvae in Phytoplankton High and Aged Kelp had greater rudiment diameter at a given postoral arm rod length as compared to larvae in Phytoplankton Low. This pattern is particularly apparent when comparing postoral arm rod lengths and rudiment diameters of larvae at age 30 d in Aged Kelp and Phytoplankton Low (Fig. 3C,D). While postoral arm rod lengths were similar among these treatments, rudiment diameter was substantially greater in Aged Kelp than Phytoplankton Low.

Phenotypic plasticity in sea urchin larval form in response to kelp and phytoplankton feeding treatments. Treatments are Aged Kelp, Phytoplankton High, and Phytoplankton Low. Rudiment diameter plotted against postoral rod length. Data are larvae measured within three replicate jars of each feeding treatment at 2–6 d intervals from age 4 to 30 d post-fertilization (n = 30). Lines indicate second order polynomial relationships (for equations see Table S1 in Supporting Information Material). [Color figure can be viewed at wileyonlinelibrary.com]
High survival of larvae occurred in all feeding treatments at termination of the experiment (96.0% ± 10.3%, mean ± SD, for three feeding treatments).
Discussion
We discovered that sea urchin larvae grow and develop to metamorphic competence on a kelp detritus diet. While larvae at age 30 d were overall larger when fed an optimal concentration of phytoplankton, bivariate plots of postoral arm rod length vs. rudiment diameter indicate that growth was allocated to postlarval structures in preference to capacity for capturing food when fed either aged kelp or a high concentration of phytoplankton. In contrast, larvae fed a low concentration of phytoplankton appeared food limited as compared to the aforementioned treatments. Food limitation of larvae was indicated by longer postoral arm rod lengths at a given rudiment diameter. The length of the postoral arm rod acts as a proxy for the length of the ciliary band used for larval feeding, and therefore, investment in larval feeding structures (Bertram and Strathmann 1998). Thus, our results illustrate a trade-off between allocation of resources toward development of metamorphic competence and capacity to catch food, driven by diet quality.
Increased time as a larva can lead to increased risk of mortality, which results in a selective advantage of a shorter larval duration (Vaughn and Allen 2010). Similar time to metamorphic competency (indicated by settlement) of larvae fed a high concentration of phytoplankton or aged kelp indicates that these food sources were comparable in terms of development of larvae to the benthic juvenile stage. Greater age at competence on the low concentration phytoplankton diet indicates a lower quality diet that would lead to longer exposure to risk in the plankton.
Our results conclusively demonstrate that kelp detritus is a high-quality food for larval sea urchins. Kelp detritus also may be a high-quality food for myriad other zooplankton inhabiting coastal regions, including herbivorous copepods and diverse invertebrate larvae. While the main food source for many zooplankton is assumed to be phytoplankton or derived from phytoplankton, our results indicate that suspended kelp detritus can rival phytoplankton in terms of quality. Hence, we may be overlooking an abundant food for zooplankton in coastal regions, with implications for modeling of ocean production (e.g., fisheries production). In a study on the west coast of Vancouver Island, Canada, ∼ 200 km (linear distance) from our study location, Ramshaw et al. (2017) documented suspended kelp-derived detritus in the water column up to 30 km offshore in summer (July) of 2009, and winter and summer (January and July) of 2010. In summer, when feeding stages of larval Strongylocentrotus spp. are expected to be present in the water column (Strathmann 1987), kelp detritus contributed on average 33% of particulate organic matter in the 20–63 μm particle size-range (Ramshaw et al. 2017). Particles of this size should be easily consumed by late-stage larval Strongylocentrotus spp. (Strathmann 1971). Phytoplankton blooms also occur in this region in the spring and summer (Masson and Peña 2009; Ramshaw et al. 2017). We posit that kelp detritus acts to subsidize a phytoplankton diet during blooms, or may replace phytoplankton as a major food source in years of low phytoplankton biomass or when blooms occur out of synchrony with the location or timing of larval release. However, there is still much to learn about the role of kelp detritus as a food source for zooplankton. There is need for repetition of the present experiment with additional zooplankton species, including copepods, which often are the numerically dominant members of the zooplankton (Kiørboe 2011). Field sampling of zooplankton for analysis of biochemical markers (e.g., fatty acids or stable isotopes) to examine diet history will be informative in determining the contribution of kelp detritus to the diet of diverse zooplankton, and at varying distances from kelp beds or forests. Finally, the effect of the state of decay and colonization and growth of microorganisms on the quality of kelp detritus for zooplankton remains unknown.
Kelp ecosystems can undergo regime shifts—or abrupt changes in the configuration of the ecosystem—when sea urchins destructively graze kelps to produce sea urchin barrens (Filbee-Dexter and Scheibling 2014b; Ling et al. 2015). Destructive grazing events occur following migration of sea urchins into kelp beds or forests or recruitment pulses of planktonic larvae that yield local increases in sea urchin density. Feedback mechanisms that alter the stability of the kelp ecosystem can determine resilience to a shift to urchin barrens (Filbee-Dexter and Scheibling 2014b). For example, the three-dimensional structure of kelp beds or forests provides habitat for benthic predators that limit sea urchin recruitment (Scheibling 1996), thus stabilizing the kelp ecosystem (Filbee-Dexter and Scheibling 2014). In combination with observations that kelp-derived detritus is abundant in the water column (Ramshaw et al. 2017), our observation that kelp detritus is an optimal food for sea urchin larvae suggests that kelp detrital production and transfer into planktonic food webs could be a mechanism mediating sea urchin larval dynamics. It remains to be determined whether production of kelp detritus acts as a negative feedback, decreasing the resilience of the kelp ecosystem to a phase shift to barrens, by promoting sea urchin larval growth and survival, and thus recruitment into kelp beds/forests.
References
Acknowledgments
Thank you to the staff at the University of Washington's Friday Harbor Laboratories. We thank A. Dingeldein for providing the illustration. CJF was supported by a Friday Harbor Laboratories Postdoctoral Fellowship. J. Hodin, L. Francis, M. Brown, and three anonymous reviewers provided helpful comments on the manuscript.
Conflict of Interest
None declared.