The ecological importance of habitat complexity to the Caribbean coral reef herbivore Diadema antillarum: three lines of evidence

Abstract

When Caribbean long-spined sea urchins, Diadema antillarum, are stable at high population densities, their grazing facilitates scleractinian coral dominance. Today, populations remain suppressed after a mass mortality in 1983-1984 caused a loss of their ecosystem functions, and led to widespread declines in ecosystem health. This study provides three lines of evidence to support the assertion that a lack of habitat complexity on Caribbean coral reefs contributes to their recovery failure. Firstly, we extracted fractal dimension (D) measurements, used as a proxy for habitat complexity, from 3D models to demonstrate that urchins preferentially inhabit areas of above average complexity at ecologically relevant spatial scales. Secondly, controlled behaviour experiments showed that an energetically expensive predator avoidance behaviour is reduced by 52% in complex habitats, potentially enabling increased resource allocation to reproduction. Thirdly, we deployed a network of simple and cost-effective artificial structures on a heavily degraded reef system in Honduras. Over a 24-month period the adult D. antillarum population around the artificial reefs increased by 320% from 0.05 ± 0.01 to 0.21 ± 0.04 m^-2 and the juvenile D. antillarum population increased by 750% from 0.08 ± 0.02 to 0.68 ± 0.07 m^-2. This study emphasises the important role of habitat structure in the ecology of D. antillarum and as a barrier to its widespread recovery.

Figure 1

Panel photograph showing tank setup for different complexity treatments. Top left natural-material-low-complexity, top right = artificial-material-low-complexity, bottom left = natural-material-high-complexity, bottom right = artificial-material-high-complexity.

Figure 2

Three-dimensional images showing complexity differences between: (A) a 2 m × 2 m area of reef without urchins on Utila, (B) a 2 m × 2 m area of reef with urchins on Utila, (C) a 2 m × 2 m area of control reef in La Ensenada, (D) an artificial reef deployed at La Ensenada. D_(5–15 cm) = fractal dimension of the model at the 5–15 cm spatial resolution. Mean ± 1SE is reported for (A)–(C), but all artificial reefs are identical therefore variation in fractal dimension for (D) is not reported.

Figure 3

Distribution of adult D. antillarum test diameters on the reefs of Utila (yellow; n = 139) and Tela (blue; n = 100). Dotted vertical line is the mean of all individuals from both locations.

Figure 4

Complexity signatures of reef areas devoid of D. antillarum (orange, n = 35), and areas inhabited by D. antillarum (blue, n = 5). Data shown in the main panel are mean ± 1SE fractal dimension (D), a measure of structural complexity within defined spatial resolutions, shown at: 1–5 cm, 5–15 cm, 15–30 cm, 30–60 cm and 60–120 cm, with 5–15 cm representing the size range of ecological significance to D. antillarum as predation refugia. D ranges from 2 to 3 and higher values are associated with greater structural complexity. Error bars represent ± 1SE from the mean.

Figure 5

Impacts of habitat complexity on the predator avoidance behaviour (PAB) of D. antillarum. Flat treatments were carried out in trial tanks without any structure, low complexity treatments with natural material occurred in tanks enriched with fragments of reef rubble, and low complexity treatments with artificial materials were conducted in tanks with fragmented breezeblocks. High complexity treatments with natural material used reef rubble to simulate a reef crevice, and during high complexity treatments with artificial material individuals were provided with a whole breezeblock. The study was fully factorial and n = 10 for each treatment. The bold horizontal line on each boxplot represents the mean, the box itself shows the interquartile range and the whiskers delimit the full range of the data. Letters above the plots show the result of a post hoc Tukey–Kramer analysis and show where significant differences between treatments occur.

Figure 6

Temporal changes in the density of adult and juvenile D. antillarum on small artificial reef structures and nearby control reefs. Data were collected at the same time each year, but points are jittered to aid visualisation. Data shown are mean ± SE.