Abundance modulates the ecosystem functional contributions of two sympatric Caribbean sea cucumbers

Study location and study species

Research was conducted under a Marine Scientific Research Permit issued by the Department of Marine Resources, Government of The Bahamas, to the Cape Eleuthera Institute, and in accordance with the Canadian Council of Animal Care (Protocol No. 1301B-19).

The study was conducted on 35 separate coral reef patches along the southwestern coast of Eleuthera Island, The Bahamas, from May to August 2019 (Fig. 1A). Reef patches were located in Rock Sound, a large, shallow (<5 m depth) sandy basin. They ranged in hard-bottom area from 2 to 209 m² (mean ± sd: 35 ± 43 m²) and depth from 2.6 to 4.5 m (mean ± sd: 3 ± 0.5 m) and were separated from the nearest patch by a minimum of 100 m. All patch reefs were immediately surrounded by a halo of seagrass, Thalassia testudinum, that extended up to 9.6 m away from the patch edge (Fig. 1B). Beyond this distance, seagrass was either sparse or absent. The two focal sea cucumber species, H. mexicana and A. agassizii (family Holothuriidae; Fig. 1C), are distributed widely across the Caribbean region (Hendler et al., 1995). In Rock Sound, we found both species co-occurring in seagrass beds and on or near coral patches. Holothuria mexicana feeds approximately 12 h per day, whereas A. agassizii feeds approximately 10 h each day (Hammond, 1982). There was no sea cucumber fishery in The Bahamas at the time of this study. A small experimental sea cucumber fishery in the Bahamas took place in 2010 on Andros Island but closed after 11 months due to stock depletion (Sherman et al., 2018).

Seagrass area, sea cucumber body sizes and density

We estimated seagrass area at each patch by measuring the circumference at the outer edge of the high-density seagrass halo (Fig. 1B), as well as that of the hard-bottom area of the coral patch reef itself. We converted perimeters to areas and subtracted the hard-bottom area from the total area to obtain seagrass area. Divers counted, identified to species, and measured the length and midbody girth of every sea cucumber encountered on reefs and within the dense seagrass halo. Beginning at a recognizable landmark and moving in a clockwise fashion, two divers swam side-by-side, and systematically searched in the seagrass for sea cucumbers, and then searched the reef, carefully looking in crevices and overhangs for sea cucumbers. Sea cucumber species density is expressed as individuals per m² of seagrass area.

Sediment processing and movement

We quantified hourly sediment processing by A. agassizii and H. mexicana following Lee et al. (2018, adapted from Uthicke, 1999), where we assumed that the quantity of sediment egested is equal to the quantity ingested. We selected A. agassizii (mean length ± sd = 20 ± 2 cm, range: 14–23 cm, n = 20) and H. mexicana (mean length ± sd = 24 ± 3 cm, range: 18–30 cm, n = 20) individuals that represented the commonest size classes across Rock Sound. We measured sediment processing and movement for 20 individuals of each species from 11:00 to 16:00 hrs in July 2019 on two patch reefs that were separate from our 35 survey patches. To do so, divers tracked sea cucumbers by planting a metal stake labelled with flagging tape in the sediment at a standardized distance (ca. 1 cm) from the posterior end of the focal sea cucumber. At the end of each hour (for three consecutive hours), the number of fecal pellets egested by each individual was counted, and the linear distance moved by each individual was recorded. The stake was re-placed near the posterior end of the focal individual to serve as starting point for the next hour-long observation. After the last observation period, up to 10 (average = 9 ± 0.4 pellets; median = 10 pellets; range = 0–10 pellets) of the most recently defecated fecal pellets for each individual (i.e., the pellets closest to the individual) were collected in Falcon^® tubes. More recently released pellets were chosen because they are easier to collect as they have not yet disintegrated. The length and girth of each sea cucumber were also measured. Fecal pellets were frozen and transported to Simon Fraser University.

After thawing and combining pellets for each individual, we placed them in a drying oven for 24 h at 60 °C. Pellets were then weighed to determine dry weight (DW) on an analytical balance to the nearest 0.001 g. Dried pellet samples were transferred into porcelain crucibles, placed in a muffle furnace for 2 h at 550 °C, then reweighed to obtain ash weight (AW). We calculated ash-free dry weight (AFDW = DW –AW) to determine the organic matter (OM) content in the fecal pellets.

Empirical estimates of ammonium excretion

To measure excretion rates of A. agassizii and H. mexicana, divers collected individuals of both species haphazardly from various reef patches that were separate from our sediment processing and movement observations. We made excretion estimates by following well-established methods by Layman et al. (2011) and Francis & Côté (2018) who modified slightly the methods of Schaus et al. (1997), Whiles et al. (2009), and Taylor et al. (2007). Individual H. mexicana and A. agassizii (n = 20 for each species) were brought to the Cape Eleuthera Institute (CEI, 24°49′54.46″N, 76°19′56.28″W) and allowed to recover in sea tables connected to a flow-through seawater system pumped directly from the ocean for 1–2 h before being placed gently but rapidly in individual 20-L acid-washed bags filled with a known volume of pre-filtered (0.7 µm Whatman GF-F filters) sea water. Bags containing sea cucumbers (n = 20 per species) and control bags of filtered sea water containing no sea cucumbers (n = three empty bags) were sealed and placed in sea tables to maintain ambient temperature (29–31 °C) for 60 min. Although handling might have increased excretion rate initially, the relatively long incubation period makes it likely that sea cucumbers were near resting rates for most of this period. At the end of the incubation period, we collected one 100 ml water sample from each bag using a sterile plastic syringe. Samples were filtered (0.45 µm Whatman GF-F filters), placed in dark bottles and refrigerated for immediate analysis of ammonium (NH₄⁺) content, a proxy for inorganic nitrogen, using fluorometric methods (Taylor et al., 2007). After incubation, we measured the wet weight (g), total length (cm), and midbody girth (cm) of each sea cucumber and allowed them to recover in sea tables for several hours before release onto their reef of capture. We randomized with respect to species the order in which we measured ammonium excretion of individuals.

Data analyses

We transformed sea cucumber counts at each patch into densities (i.e., numbers per m² of seagrass). We ran a Welch’s t-test to assess differences in mean density and mean proportion of total density between sea cucumber species across Rock Sound. To test for differences in size distributions between A. agassizii and H. mexicana, we ran a two-sample Kolmogorov–Smirnov test on sea cucumber length.

Body size and density

Across all patches surveyed, H. mexicana was significantly larger than A. agassizii (mean ± CI (range); Hm: 27 cm [25.9, 27.2] (17–47 cm), Aa: 22 cm [21.5, 22.3](13–45 cm), p < 0.001). The length distributions of the two species were also significantly different (p < 0.001; Fig. 2). There were, on average, 15 (± 2 SE) sea cucumbers per patch (range: 1 –62; Fig. 1). There were significantly more A. agassizii (11 ± 2) present, on average, than H. mexicana (4 ± 1) per patch (p = 0.003). Mean patch size was 35 (± 7 SE; range 1.3–209) m² and mean seagrass area was 54 (± 10 SE; range 5.3–296) m², yielding a mean overall density of 0.3 (± 0 SE; range 0.02–1.3) sea cucumbers per m² of seagrass area.

Sediment processing and movement

Actinopyga agassizii egested fecal pellets at approximately four times the rate of H. mexicana (p = 0.008; Fig. 3A). However, the fecal pellets egested by H. mexicana were seven times heavier than those egested by A. agassizii (p < 0.001; Fig. 3B). Combining these measures together, individual H. mexicana processed three times more reef sediment per hour, on average, than individual A. agassizii (p < 0.001; Fig. 3C). Fecal pellets egested by A. agassizii had a significantly higher OM by approximately 1.5% than those of H. mexicana (p < 0.001; Fig. 3D). Holothuria mexicana moved a maximum of 170 cm in a three-hour observation, while A. agassizii moved at most 125 cm. Speed did not differ significantly between A. agassizii (0.1 ± 0.03 m h⁻¹) and H. mexican a (0.2 ± 0.04 m h⁻¹) (p = 0.17; Fig. 3E).

Through extrapolation of egestion rates and quantities, we found that individual A. agassizii and H. mexicana have the potential to process 5.9 (range: 4.3–7.5) and 12.5 (range: 9.4–16) kg of sediment y⁻¹, respectively. When we scaled up these individual egestion rates to population-level rates, A. agassizii populations turned over, on average, significantly more sediment (1.9 ± 0.2 SE kg m⁻² yr⁻¹) than H. mexicana (1.0 ± 0.4 SE kg m⁻² yr⁻¹; p = 0.019; Fig. 4). However, this difference appears due mainly to the absence of H. mexicana from several reefs. When we considered only reefs where both species were present (24 of 35 reef patches), A. agassizii populations turned over sediment at a similar rate (1.6 ± 0.2 SE kg m⁻² yr⁻¹) compared to H. mexicana populations (1.4 ± 0.5 SE kg m⁻² yr⁻¹; p = 0.77; Fig. 4). At reefs with both species, Actinopyga agassizii contributed more to sediment processing at 57% of reefs, H. mexicana contributed more at 40% of these reefs, and both contributed equally at a single reef (Fig. S1).

Ammonium excretion

The sea cucumbers used to assess ammonium excretion rates were slightly larger, on average, than those found on the study reefs, but they spanned the ranges of lengths observed on reefs (Fig. 2). We used slightly (∼11%) larger H. mexicana individuals (mean ± SE [range]; 787 ± 69 [361–1397] g) on average, than A. agassizii individuals (mean ± SE [range]; 706 ± 55 [110–1080] g) to obtain ammonium excretion rates. Species identity had a significant effect on sea cucumber ammonium excretion rate (p = 0.04), but there was no effect of wet weight (p = 0.07, r² = 0.17) (Fig. S2). On average, individual Holothuria mexicana excreted NH₄⁺ at a rate that was approximately 23% higher than individual A. agassizii (mean ± SE; Hm: 15.6 ± 1.1 µmol NH₄⁺ h⁻¹, Aa: 12.0 ± 1.0 µmol NH₄⁺ h⁻¹) (p = 0.023).

Reef-level estimates of excretion contributions showed that A. agassizii populations contributed 5.7 times more ammonium per unit area (3.1 ± 0.5 µmol NH₄⁺ m⁻² h⁻¹) than H. mexicana populations (0.54 ± 0.1 µmol NH₄⁺ m⁻² h⁻¹; p < 0.001; Fig. 4). When considering only reefs where both species occurred, A. agassizii populations contributed 3.3 times more ammonium per unit area (2.5 ± 0.6 µmol NH₄⁺ m⁻² h⁻¹) than H. mexicana populations (0.75 ±0.1 µmol NH₄⁺ m⁻² h⁻¹; p < 0.001; Fig. 4). When both species were present, A. agassizii contributed more ammonium at 83% of reefs, H. mexicana contributed more at 13% of reefs, and both species contributed equally at one reef (Fig. S3).

Sediment processing and movement

We highlight the ecological roles of H. mexicana and A. agassizii as motile sediment processors of patch reef sediments. At the individual level, A. agassizii egested pellets faster than H. mexicana, but their pellets were smaller. Assuming the total amount of sediment egested reflects the amount ingested (Uthicke, 1999; Lee et al., 2018), individual H. mexicana processed three times more sediment per hour than individual A. agassizii. The sediment processing rates of both species were within the range of Mediterranean species (Coulon & Jangoux, 1993) but considerably lower than those of Indo-Pacific species (Uthicke, 1999; Wolfe & Byrne, 2017; Lee et al., 2018).

At the reef scale, however, A. agassizii had a higher estimated sediment processing potential because of its higher abundance. Actinopyga agassizii populations processed 1.9 times more sediment per unit area than H. mexicana populations. Note that our estimates of sediment processing rates across patch reefs are likely conservative because we could only measure sediment processing rates of sea cucumbers between 09:00 and 16:00 hrs but both species feed and are active at night (Hammond, 1982). There is no information, to our knowledge, on the ecological consequences of sediment processing by our two target species. However, the ingestion and release of fecal casts and disturbance caused by locomotion by other deposit-feeding sea cucumbers play a role in redistributing surface sediments and influencing biotic interactions occurring at the sediment–water interface (Purcell et al., 2016). Moreover, halting bioturbation and feeding functions by experimental removal of sea cucumbers led to the development of cyanobacterial mats and a reduction in oxygen penetration depth into sediments (Moriarty et al., 1985; Uthicke, 1999; Michio et al., 2003; Lee et al., 2018).

Actinopyga agassizii egested fecal pellets with higher OM content than H. mexicana. This might make pellets of A. agassizii more prone to bacterial and fungal growth, and result in a more rapid loosening of the mucous membrane that holds the fecal material and hence faster resuspension of organic matter (Conde, Diaz & Sambrano, 1991). The interspecific difference may stem from differences in the time each species spends on various substratum types. For example, in Jamaica, A. agassizii was observed mainly on algal turf and on macroalgae, whereas H. mexicana spent 90% of its time on sand (Hammond, 1982). Alternatively, interspecific differences may be attributed to the resident time of sediment in the digestive tract, where one species could be less ‘thorough’ in sediment processing than another, resulting in fecal pellets with higher organic matter. The fecal pellets of A. agassizii could also be higher in OM because they are smaller, meaning that there is more organic material surrounding the pellet relative to the amount of sediment inside the pellet, assuming that organic matter is concentrated on the surface of the pellet where most digestive reactions take place. Future studies should (1) use a stable isotope approach to identify the specific origins and diet sources of H. mexicana and A. agassizii (e.g., Slater, Carton & Jeffs, 2010), (2) determine pellet grain size distribution between species, which may have an effect on organic matter content and could reveal niche differentiation, and (3) quantify nocturnal sediment processing (and ammonium excretion) rates of sea cucumbers.

Ammonium excretion

Individual ammonium excretion rates by H. mexicana and A. agassizii were species specific but did not significantly vary with body size. The average rates estimated here for A. agassizii (12.0 µmol NH₄⁺h⁻¹) and H. mexicana (15.6 µmol NH₄⁺h⁻¹) are at the high end of the range reported for Western Pacific tropical species (1–18 µM; Mukai et al., 1989; Uthicke, 2001a; Wheeling, Verde & Nestler, 2007). Though H. mexicana had a higher average excretion rate, both species showed the same weak relationship with body size, indicating that individual H. mexicana excrete more nutrients than individual A. agassizii of the same size. Note that the relationship between ammonium excretion rate and sea cucumber body size was weaker than expected from physiology and mass–balance theory. Obtaining accurate but non-destructive mass and morphology measurements of holothuroids is notoriously difficult because they readily change shape and retain water in their body cavity (Wheeling, Verde & Nestler, 2007).

Despite having a lower per capita excretion rate, A. agassizii contributed more ammonium than H. mexicana at the reef scale owing to its higher abundance. Actinopyga agassizii contributed more to ammonium excretion at 83% of our study reefs and, on average, excreted 5.6 times more ammonium per unit area than H. mexicana. Ammonium excretion by tropical sea cucumbers has been shown to be an important source of limiting nutrients that promotes growth of microalgae (e.g., Uthicke, 2001b; MacTavish et al., 2012) and seagrass (e.g., Wolkenhauer et al., 2010).

Both sea cucumber species together contributed approximately 15% of the ammonium released by coral reef fishes on patch reefs in Rock Sound. We estimated that A. agassizii and H. mexicana excreted 3.1 ± 0.5 µmol NH₄⁺m⁻²h⁻¹ and 0.5 ± 0.1 µmol NH₄⁺m⁻²h⁻¹, respectively. In Rock Sound, all resident fishes together contribute, on average, ∼25 µmol m⁻²h⁻¹ during the daytime (Francis & Côté, 2018). These fishes included more than 45 species across 17 families. On a per-species basis, the role of sea cucumbers as nutrient providers is therefore substantial, and given their nocturnal behaviour, they may contribute more to the overall diel nutrient budget at night compared to daytime. In addition, migratory grunts (Haemulidae), which contribute more than twice as other reef-associated fishes, migrate seasonally and annually, resulting in an unpredictable nutrient supply (Francis & Côté, 2018). In contrast, some species of sea cucumbers are known to exhibit high site fidelity for years over time (Wolfe & Byrne, 2017), meaning that sea cucumbers may contribute more consistently to seagrass beds adjacent to reefs than reef fish do. In this way, sea cucumbers act as a ‘press’ of nutrient inputs, operating on time scales of days to months, or even years (Allgeier et al., 2014; Allgeier, Burkepile & Layman, 2017).