Fig 1.
Diagram of our model structure that incorporates three herbivore functional groups (grazer, generalist, browser) and two algal resources (turf, macroalgae). Green arrow (γ) between “Turf Algae” and “Macroalgae” is a transition from turf to macroalgae. Our model incorporates the interactions and feedbacks (labeled 1–3 corresponding to their descriptions in S1 Appendix) from the [36] model.
Fig 2.
Shifts from Coral Dominance to Macroalgal Dominance.
A) Photos of a reef with high coral cover (left) versus with high macroalgal cover (right). The shift from high coral to macroalgal cover can result from increased fishing pressure and changes in the dominant herbivores present, which we explored with our model. Photos taken by Shayna Sura. B) Output of our model under 2 levels of fishing pressure (0.1, 0.9) for an even herbivore community and initial conditions and parameter values given in Table B in S1 Appendix. With increased fishing pressure, coral cover declines and macroalgal cover increases and persists over time. C) Output of our model under 2 herbivore community composition scenarios (generalist-dominated, grazer-dominated), for fishing pressure = 0.5. For a given fishing pressure, the benthic state of the reef can vary greatly depending on the herbivore community composition. Coral cover remains high for an herbivore community dominated by generalists, but macroalgal cover increases and persists over time for an herbivore community dominated by grazers, while a browser-dominated community leads to moderately high coral cover (Fig Dii in S1 Appendix). For both panels B and C, the Herbivores and Algal state variables are sums of their corresponding state variables (i.e., Herbivores = Grazers + Browsers + Generalists, and Alage = Turf + Macroalgae).
Fig 3.
Impacts of herbivore community composition and fishing pressure on coral reef recovery from disturbance.
Heatmaps showing how the initial abundance of herbivorous fishes that are grazers versus browsers affect final coral cover for fishing pressures ranging from 0 to 0.4 (rows) and for initial generalist abundance set to 0 (A-C), 0.2 (D-F), or 0.4 (G-I). Total herbivore abundance is restricted to 0.9; therefore, as initial generalist abundance increases, the maximum abundance of browsers+grazers has to decrease accordingly, resulting in smaller response spaces in the panels. Other initial conditions include: Ct=0 = 0.15, Tt=0 = 0.7, and Mt=0 = 0. Other parameter values: iC = 0.05, iT = 0.05, iM = 0, bC = 0.3, bT = 0.8, bM = 0.5, dC = 0.1, gT = 2, gM = 1, r = 1, ηT = 0, ηM = 1, αT = 0.25, αM = 0.5, σ = 0.6, and γ = 0.1. Note: final coral cover values show very little variation for scenarios resulting in coral-dominated states, with 91% of final coral cover values >0.5 (coral dominated) within the range of 0.65 – 0.72, hence the similarity in colors.
Fig 4.
Impacts of herbivore community composition and fishing pressure on algal cover after disturbance.
Heatmaps showing final cover of turf or macroalgae for the same scenarios as Fig 3. Specifically, these panels examine the initial abundance of grazers and browsers for fishing pressures ranging from 0 to 0.4 (rows) and for initial generalist abundance set to 0 (A-C), 0.2 (D-F), or 0.4 (G-I). Other initial conditions include: Ct=0 = 0.15, Tt=0 = 0.7, and Mt=0 = 0. See Fig 3 caption or Table B in S1 Appendix for other parameter values. For each scenario, the heatmap indicates the final cover of turf or macroalgae, depending on which had the higher final cover. Note that these results are complementary to the level of coral cover shown in Fig 3; specifically, turf and macroalgal cover values near zero (beige color) correspond to high coral cover values shown in Fig 3.
Fig 5.
Interaction of turf and macroalgal dominance and herbivore community composition on coral reef recovery from disturbance.
Heatmaps showing how initial cover of turf and macroalgae affect final coral cover for fishing pressures ranging from 0 to 0.6 (rows) and for initial herbivore proportions set to 1 of 3 scenarios: 1) browser-dominated (A-C): browsers = 0.6, generalists = 0.15, grazers = 0.15; 2) grazer-dominated (D-F): grazers = 0.6, generalists = 0.15, browsers = 0.15, or 3) generalist-dominated (G-I): generalists = 0.6, grazers = 0.15, browsers = 0.15;. Total initial benthic cover is restricted to 0.85, and initial coral cover (Ct=0) = 0.15; therefore, the amount of unoccupied space changes for each combination of initial turf and macroalgae abundance. See Fig 3 caption or Table B in S1 Appendix for other parameter values.
Fig 6.
Impacts of herbivore community composition on coral reef bistability and hysteresis.
Bistability plots (A-D) showing final coral cover for each combination of fishing pressure and initial coral cover conditions. Bistability is evident when, for a given fishing pressure, both low (<0.2) and high (>0.6) final coral covers occur, depending only upon the initial coral cover. Hysteresis plots (E-H) showing the final coral cover as fishing pressure is increased (forward) or decreased (reverse). Hysteresis is evident when there is a range of fishing pressures where the forward (red) and reverse (black) lines do not overlap. We examined four scenarios using our expanded model parameter values from Table B in S1 Appendix with different initial herbivore community compositions: herbivore groups even (A,E): generalists = grazers = browsers = 0.3; generalist-dominated (B,F): generalists = 0.6, grazers = 0.15, browsers = 0.15; browser-dominated (C,G): browsers = 0.6, generalists = 0.15, grazers = 0.15; or grazer-dominated (D,H): grazers = 0.6, generalists = 0.15, browsers = 0.15. See Fig 3 caption or Table B in S1 Appendix for other parameter values.