Variation in purple sea urchin (Strongylocentrotus purpuratus) morphological traits in relation to resource availability

Survey design

This study was conducted along the southern end of Monterey Bay, California, USA. In 2014, overgrazing by purple sea urchins shifted a once expansive kelp forest landscape to a patchy mosaic of remnant forests interspersed with sea urchin barrens (Smith et al., 2021). While this study was conducted just 3 years after the initiation of the urchin outbreak, the spatial extent of barren patches continued to expand both during and for several years following this study.

Underwater surveys were conducted from June to September in 2017. A total of 83 randomly assigned subtidal survey sites within the study region were sampled to evaluate sea urchin size structure and density, habitat type (i.e., barrens, kelp forests), and algae cover (Fig. 1). Survey sites were located between the Breakwater (36.61237578 N, 121.8947703 W) and Point Pinos (36.63878914 N, 121.9246315 W) in Monterey, CA, in 5 to 20 m of water on rocky reef substratum. Each site was categorically assigned in situ as one of two habitat types: kelp forest (characterized by high macroalgae presence and low sea urchin density), or sea urchin barren (characterized by low macroalgae presence, dominated by coralline algae, and with high sea urchin density). These two categories formed the basis for all categorical comparisons and analyses.

Each site was surveyed using a radial sampling design following the methods described in Smith et al. (2021). Briefly, a total of eight 5-m long transects were surveyed at each site, radiating from a fixed central position. A single transect was assigned to each cardinal (N, S, E, W) and inter-cardinal (NE, NW, SE, SW) direction around the survey site. Two randomly placed 1m x 1m photographic quadrats were sampled per transect (16 total quadrats per site). The location of the quadrats along the transects were assigned using a randomly stratified design so that the quadrats were not intentionally biased towards either the center or outer edge of the site. Within each quadrat, the total number of sea urchins were quantified by searching in crevices and carefully examining dense algae. By utilizing the same sampling design at each site, each of our 83 survey sites represent a replicate sample.

Variation in sea urchin morphology between habitats

To test the hypothesis that lantern morphology and body shape varies as a function of habitat type, we conducted a series of laboratory dissections with sea urchins collected from spatially referenced field survey sites (study animal collection approved by the California Department of Fish and Wildlife, permit no. SC-389). Two sea urchins were randomly collected from fixed opposite corners of each quadrat for a maximum of 32 sea urchins per site. Only urchins greater than 3 cm were collected for dissections. These size classes of urchins are too large to have settled from the plankton later than 2014. Therefore, most of the dissected urchins were likely present during the 2014 habitat transformation event.

A total of 1,058 purple sea urchins were collected for this study. Upon collection, sea urchins were immediately ‘fixed’ following the methods of Harrold & Reed (1985) to preserve gonads for dissection. Sea urchins were injected with 3–15 mL of 10% neutral-buffered formalin (depending on animal size) through the peristomal membrane. After a 24-h fixation period, sea urchin tests were measured to the nearest 0.5 mm using Vernier calipers. Test diameter was recorded as the maximum distance across the test, excluding spines, and test height was measured as the length between the oral and aboral surfaces (to the nearest 0.05 mm). After these measurements were obtained, sea urchins were dissected for gonads and to measure lantern morphology. Our measure of lantern length was recorded as the distance from the oral tip to the aboral surface of the epiphysis. Lantern width was measured as the maximum diameter of the aboral epiphysis surface.

We evaluated both sea urchin lantern length and gonad weight as independent functions of test diameter between habitat types (barrens, forest) using separate analysis of covariance (ANCOVA) tests. In both models, we tested the homogeneity of slopes assumption by evaluating the effect of test diameter, habitat type (barren, forest), and their interaction. We found that the interaction was not significant, therefore we ran reduced two-term ANCOVAs with test diameter and habitat type as independent model terms.

Sea urchin gonad weight serves as a useful metric for assessing animal health and reproductive capacity (Walker, 1981; Lawrence & Lane, 1982). To account for the relationship between urchin size and gonad weight, we used a gonad index to comparatively evaluate sea urchin gonad condition and lantern morphology. Sea urchin gonads were removed from each individual, blotted dry, and wet-weighed to the nearest 0.001 g. Gonad index was calculated as the percent body mass comprised of gonads:

(1) $$GonadIndex\left(GI\right)={\displaystyle \frac{GonadWetWeight\left(g\right)}{\left(TotalWetWeight\left(g\right)\right)}\times 100}$$

Therefore, higher GI values indicate greater gonad mass relative to the animal’s total body mass.

Relationship between sea urchin morphology and algae cover

Photographic quadrats were used to estimate algae cover in order to test for a relationship between algae cover, sea urchin morphology, and condition. Within each quadrat, still imagery was collected using a downward facing GoPro Hero4 camera and two LED Sola video lights. Still photos of each quadrat were analyzed in ImageJ using 16 uniform-points per quadrat. Photos were analyzed in the lab where an observer recorded the organism below each point to the lowest taxonomic level possible. However, because many species of algae are difficult to taxonomically identify from photo imagery, algae were organized into the following morphological groups: brown algae (predominately Macrocystis pyrifera), foliose red algae (predominately Chondracanthus spp.), and encrusting algae (includes encrusting red and encrusting coralline algae). These morphological groups were selected based on their known status as indicators for kelp forests (brown algae, foliose red) or sea urchin barren (encrusting algae) habitats (Filbee-Dexter & Scheibling, 2014). First, we used a series of linear regressions to test for associations between the cover of algae and two metrics of sea urchin morphology (lantern index, gonad index). Because lantern length was directly proportional to test diameter in both habitats, we used a lantern index to account for length-size relationships (e.g., larger animals have larger lanterns), defined as:

(2) $$LanternIndex\left(LI\right)={\displaystyle \frac{LanternLength\left(mm\right)}{TestDiameter\left(mm\right)}}$$

Therefore, higher LI values indicate a longer jaw length relative to body size. We then used a beta regression on the pooled quadrat data for each site to account for the proportional (0–1) response variable (Ferrari & Cribari-Neto, 2004), with mean sea urchin density as a continuous predictor. We examined sea urchin body size as the ratio of test diameter to test height between habitat types and found no significant difference. Additionally, lantern width was relatively proportional to body size.

Variation in sea urchin morphometrics as a function of habitat

Sea urchin density was greater in barrens (M = 13.79 ± 1.42 S.E. individuals/m²) than in the kelp forest habitat (M = 5.33 ± 0.69 S.E. individuals/m², P < 0.0001). Although there were differences in sea urchin density between habitat types, the mean test diameter for sea urchins from barrens (M = 36.58 ± 0.51S.E. mm) and kelp forests (M = 37.07 ± 0.62 S.E. mm) was not significantly different. Similarly, the relationships between lantern length and lantern width, and between test diameter and test height, did not vary with habitat type (Fig. 2).

An analysis of covariance (ANCOVA) revealed that relative lantern length was significantly greater in sea urchin barrens than in kelp forest habitats (F = 72.63, DF = 2 and 80, P < 0.0001). This difference in lantern length was most pronounced in individuals less than 49 mm (Fig. 3A). Although lantern length appeared to broadly converge between habitat types in individuals greater than 50 mm, the interaction between habitat type and test diameter was not significant. Conversely, gonad weight was significantly higher in the kelp forest habitat than in sea urchin barrens (F = 76.54, DF = 2 and 80, P < 0.0001; Fig. 3B).

The percent cover of algae was correlated with both lantern index and gonad index (Fig. 4). There was a positive and linear relationship between encrusting algae and lantern index (R² = 0.33, P < 0.001; Fig. 4A), while brown (R² = 0.14, P < 0.001; Fig. 4C) and foliose red (R² = 0.14, P < 0.001; Fig. 4E) algae were negatively correlated with lantern index. In general, these lantern index relationships were inversely correlated with gonad index. Cover of encrusting algae was negatively correlated with gonad index (R² = 0.4, P < 0.001; Fig. 4B), while brown (R² = 0.075, P < 0.02; Fig. 4D) and foliose red (R² = 0.35, P < 0.001; Fig. 4F) algae were positively correlated.

Habitat attributes of kelp forest and sea urchin barrens

We explored differences in the cover of key habitat-indicating algae groups (brown algae, foliose red, and encrusting algae) and sea urchin density between habitat types to determine factors that may influence variation in sea urchin morphological traits. A reduced two-term model with brown algae and red foliose algae explained the most variation in habitat type (AIC = 61.48, DF = 3, P < 0.0001). Based on the model coefficients, high cover of brown and foliose algae explained a greater likelihood of a forest habitat, while the cover of encrusting algae was negatively correlated with the forest habitat. Because of the linear relationship between lantern index and sea urchin density, identical relationships were found between percent cover of algae and sea urchin density. Finally, there was a positive relationship between encrusting algae and sea urchin density (F = 36.54, P < 0.0001), while both the cover of brown (F = 12.47, P = 0.0007) and foliose red (F = 14.02, P < 0.0003) were negatively correlated with sea urchin density (Fig. 5).