Short-Term Impact of Decomposing Crown-of-Thorn Starfish Blooms on Reef-Building Corals and Benthic Algae: A Laboratory Study

Abstract

Outbreaks of crown-of-thorn starfish (COTS) have caused dramatic declines in reefs through predation on corals, but the post-bloom effects of COTS may still potentially threaten the environment and living organisms due to massive organic decomposition. This stimulation experiment showed that the decomposition of COTS debris triggered an extra mineralization process and resulted in acidifying, hypoxic, and eutrophic seawater. Consequently, the photosynthetic efficiency of coral symbionts decreased by 83%, and coral bleached after removing the stress within two days, then the coral skeleton dissolved at rates of 0.02–0.05 mg cm⁻² day⁻¹. Within two weeks, the photosynthesis and growth of benthic algae were suppressed by 27–86% and 1.5–16%, respectively. The mortality of turf algae and coralline algae indicated compromised primary productivity and limited coral recruitment, respectively. However, macroalgae, as coral competitors, became the only survivors, with increasing chlorophyll content. This study suggests a continuing decline of reefs during the collapse phase of COTS outbreaks and highlights the need for improving control strategies for the COTS population.

1. Introduction

Coral reef is an important marine ecosystem, but it has been experiencing multiple stressors. In addition to ocean warming [1], acidification [2], eutrophication [3], and hypoxia [4], an outbreak of crown-of-thorn starfish (COTS) [5] has become a major cause of coral reef declines. The increase in COTS larval food (e.g., phytoplankton) and reduction in predation pressure on COTS juveniles and adults are proposed as causes for COTS outbreaks [6]. Outbreaks of COTS have been occurring more frequently in response to the changing ocean [7]. Climate change has positive effects on the growth and feeding of juvenile COTS [8], as well as larval survivorship [9], potentially increasing the likelihood of COTS outbreaks. Many researches have focused on the predation of COTS on scleractinian corals [5,10], but few have addressed the ecological effect on coral reefs during the collapse of the COTS population.

The COTS population can grow rapidly during the outbreak, e.g., from 1 to 2 individuals to >200,000 within a year [6]. High densities of COTS (>1 individual m⁻²) have been observed on reefs of Hawaii and Fiji, etc. [11,12], more than ten thousand times the normal level [6]. It is noted that the COTS population may collapse for the following reasons: (i) with food scarcity, COTS individuals became aggregately stranded and died in shallow water [13], as they could consume the corals in a reef region within several years [7]; (ii) epidemic diseases or longevity, contributed to frequent and rapid collapses of the COTS population [12,14], infected individuals died after 8–9 days [15] and decomposed in the field for at least five days, depending on the location [16]; (iii) artificial control technologies, like single-shot injection with chemicals can induce mass mortality of COTS in short time (12–24 h) underwater [17,18]. The sudden collapse of a large population of COTS poses a risk to the coral reef community, as it may trigger an extra mineralization process of organic matter (OM) from COTS detritus.

Increasing amounts of OM have been found to be derived from river runoffs, and aerobic decomposition of OM induced spreading hypoxia ‘dead zones’ in the ocean [19]. Recent research also showed that the collapse of eruptive marine species, e.g., jellyfish or algal blooms, produced excessive OM and deteriorated seawater quality through organic decomposition [20,21]. From simulation experiments, the decomposition of jellyfish (800 g carrions in 30 L seawater) resulted in a depletion of oxygen level from 6.5 to 0.5 mg L⁻¹, a decrease in seawater pH from 8 to 7.1, and the massive release of carbon (C), nitrogen (N) and phosphorus (P) [20,22,23]. Moreover, the decomposition reduced the abundance of polychaeta, which increased the abundance of gastropods that scavenged on jellyfish carrions [24]. Similarly, the decomposition of eruptive algae Ulva (1–32 g algae in 4 L seawater) induced 30–100% mortality of abalone after 96 h [25] but promoted the proliferation of the raphidophyte Heterosigma akashiwo, a kind of red-tide species [26]. The findings suggest that the influence of massive OM decomposition on coral reefs is likely impactful since corals are susceptible to even slight changes in their environment. Ocean acidification (pH 7.76–7.91) has caused 10–80% lower net calcification rates of reef-building corals and crustose coralline algae [27]. The photosynthesis efficiency of symbionts in the coral Pocillopora is reduced by 50% under anoxic conditions [28]. Nutrient enrichment disfavors corals [3] but perhaps promotes their major competitors, e.g., macroalgae [29] and turf algae [30]. That has concerned managers due to the potential risks from high biomass of COTS to reef environments [31], but how the coral reef community responds to the stresses from COTS decomposition is still unclear. The effects may be complicated under ocean warming since warming perhaps enhances the biomass of microbes [32], the major decomposers [33].

In this study, incubation experiments were performed to explore the influences of COTS decomposition on seawater, including seawater acid/base balance, oxygen consumption, and nutrient release. Subsequently, we will reveal the responses of the benthic coral reef community involving four functional groups, i.e., live coral and coral skeleton, coralline algae, turf algae, and macroalgae.

2. Methods

2.1. Crown-of-Thorns Starfish Data Collection

Acanthaster sp. samples were collected from 1 to 15 m depth in the coral reefs of Xisha Islands (16°35′–16°26′ N, 111°29′–111°45′ E) in the South China Sea (SCS; Figure S1) in June 2020, and then stored in a freezer at −20 °C. Their characteristics (n = 10) were recorded in Table S1. The samples were thawed at room temperature (~26 °C) before the experiments. Subsamples (7.4–7.6 g wet weights of carrions) were used to measure their organic content, following the method proposed by Clarke et al. [34]. The wet weight, dry weight, and ash weight of the subsamples were measured in the frozen state after freeze-drying to a constant mass and after ignition at 500 °C for 24 h, respectively. The weight difference before and after ignition was recorded as the organic weight, and the difference between wet and dry weight was recorded as the water content (Table S2).

2.2. Coral and Algae Samples Collection

Live corals and three common benthic algae were collected from the coral reefs of the SCS. The coral colonies included branching coral Pocillopora sp., monotypic coral Fungia sp., and massive coral Favites sp. The algae included crustose coralline algae Hydrolithon sp., turf algae Helminthocladia sp., and macroalgae Caulerpa sp. that attached to living stones. They were transferred to the laboratory and cultured in tanks, supplied with a cycle flow of artificial seawater (artificial seawater was prepared by dissolving BLUE DIAMOND PRO coral salt in RO water) at 26–27 °C and a photoperiod of 12:12 h light: dark cycle. Small coral fragments were taken from a single coral colony for each morphology, and the benthic algae were removed from the live stones and then separated into small segments. After a month of recovery, epiphytes and fauna attached to the surface of the live organisms were cleaned carefully with a soft brush before the experiments.

Coral skeleton fragments were taken from recently dead coral colonies of branching Acropora, foliose Montipora, and massive Porites. The fragments were treated with 10% sodium hypochlorite for 24 h to remove the coral tissues and provide an abiotic framework. Then, they were soaked in deionized water for 24 h with triple-rinsed and dried at 60 °C for 72 h.

2.3. Experimental Design

The experiment to examine the responses in seawater and coral skeleton includes the following four treatments (Figure S2), (i) low-density treatment: 2.5 g L⁻¹ (12.5 g carrions + 5 L seawater), 26–27 °C; (ii) high-density treatment: 5 g L⁻¹ (25 g carrions + 5 L seawater), 26–27 °C; (iii) heating treatment: 2.5 g L⁻¹ (12.5 g carrions + 5 L seawater), 32 °C; and (iv) control: without COTS, 26–27 °C. The two values of COTS content in seawater were calculated on the basis of field surveys during outbreaks. The low and high-density COTS were set at 5–10 individual m⁻², referred to previous literature [12]. The weight of the COTS individual was taken as 1400 g, according to the maximum weight of the Acanthaster sp. samples we recorded (Table S1). The COTS carrions were assumed to be completely diluted by vertical mixing and uniformly distributed in a 3 m water column. The water depth was defined since COTS aggregations were generally observed on the reef flat and slope (3–10 m) [13,35]. Thus, the variables of COTS content in seawater were calculated as 2.5 g L⁻¹ (low-density treatment) and 5 g L⁻¹ (high-density treatment). The temperatures in the heating treatment were elevated from 26 to 32 °C, according to the monthly mean sea surface temperature of Xisha Islands of the SCS (26–27 °C), and increased by 3.3–5.7 °C by the end of this century, as predicted by Representative Concentration Pathway (RCP) 8.5 of IPCC [36].

In the seawater and coral skeleton experiment (14 days), three 5-L beakers for the heating treatment were soaked in a 32 °C buffer tank, and nine 5-L beakers for the other three treatments were placed in a 26 °C buffer tank. Each beaker contained one fragment of every form of coral skeleton, i.e., branching Acropora, foliose Montipora, and massive Porites. For each treatment, three beakers were applied for three replicates. This allowed for a comparative analysis of the reef habitat formed by coral skeletons of different morphologies. Freshwater (100 mL) was added daily into the beakers to balance evaporation.

The experiment to examine the responses in live corals and algae included only two treatments (Figure S2), considering the susceptibility of live organisms: (i) low-density treatment: 2.5 g L⁻¹ (12.5 g carrions + 5 L seawater), 26–27 °C; and (ii) control: without COTS, 26–27 °C. One fragment of every coral species (i.e., branching Pocillopora, monotypic Fungia, and massive Favites) was placed in a 5-L beaker. Each treatment has three replicate beakers. The experiment on live corals lasted for six days. Once the health of the corals was threatened, they were transferred to the culture tank so as to examine their potential recovery when removed from the influence of COTS decomposition. Meanwhile, another six beakers were applied for the algae experiment (14 days). Four replicates of algal segments of every species (i.e., Hydrolithon sp., Helminthocladia sp., and Caulerpa sp.) were placed in each 5-L beaker. All twelve beakers were soaked in 26 °C buffer tanks. The supplementation of freshwater was the same as in the seawater and coral skeleton experiment.

2.4. Seawater Quality Analysis

We quantified the parameters of seawater acidification, eutrophication, and hypoxia because they are related to the main decomposition products of OM (Figure 1) and are also the main causes of coral reef decline. Seawater pH, dissolved oxygen (DO), temperature, and salinity were monitored daily by using a YSI (PRO PLUSS, Yellow Springs, OH, USA) device. The pH (NBS scale) and conductivity probes were calibrated by NIST-certified buffers every third day to prevent drift. As particulate OM release changes seawater turbidity, we monitored light intensity under seawater using a light recorder (HOBO MX2202, Onset Computer, Bourne, MA, USA). Total alkalinity (TA) and dissolved inorganic carbon (DIC) were measured to determine the acid/base balance in seawater, and concentrations of NH₄⁺, NO_x, and PO₄³⁻ were measured to detect nutrient enrichment.

Seawater samples were taken before and after the experiment and additional seawater samples were collected during the seawater response experiment. Each seawater sample (200 mL) was filtered through 0.45 μm filter membranes and divided into two subsamples. One (60 mL) for measuring TA and DIC was treated with 0.02 vol% saturated mercuric chloride (HgCl₂) to arrest biological activity. TA and DIC were analyzed by Apollo SciTech instruments (Apollo, AS-ALK2, and AS-C6, respectively, Newark, DE, USA) that were calibrated by certified reference materials (Batch 190, TA: 2218.4 ± 0.8 μmol kg⁻¹ and DIC: 2034.0 ± 0.7 μmol kg⁻¹) from the laboratory of Dr. Andrew Dickson (Scripps Institution of Oceanography, USA). The other subsample (140 mL) for measuring NH₄⁺, NO_x, and PO₄³⁻ was treated with 0.2 vol% CHCl₃ to inhibit microbial consumption. NH₄⁺, NO_x, and PO₄³⁻ were analyzed according to the seawater analysis standard of China (GB 17378.4-2007) [37], i.e., modified indophenol blue method, cadmium-copper reduction and Griess reaction, and phosphomolybdenum blue spectrophotometric method, respectively. All seawater samples were stored in the dark at −20 °C until they were used for analysis.

The bacteria abundance could reflect the respiration rates in seawater and the potential for remineralization [33]. The bacteria abundance was quantified using the plate counter method before and after the experiment in 10-mL seawater following the seawater analysis standard of China (GB 17378.7-2007) [38]. The seawater samples were diluted and plated on 2216E plates with marine agar, where bacteria were cultured at 25 °C for 7 days. After a week, colony-forming units (CFU) were counted to estimate the bacteria abundance.

2.5. Coral and Benthic Algal Data Collection

The dry weight of each coral skeleton was recorded before and after the experiment, as described above. The surface area of the coral skeleton was measured using aluminum foil technology [39]. The real-time photosynthetic efficiency (ΔF/Fm’) was used as a proxy for assessing photosynthesis and health status of coral-algal symbionts and algae. ΔF/Fm’ was measured by a pulse-amplitude modulated fluorometer (Monitoring-PAM, WALZ, Effeltrich, Germany) at 9:00 every day. We adjusted the distance between the probe and the sample and kept a stable Ft in the range of 200–500 to give optimal fluorescence signals. The chlorophyll content of the algae was used to indicate its photosynthetic capacity. Subsamples of coralline algae (each 0.5 cm²), turf algae (each 0.5 g wet weight), and macroalgae (each 0.5 g wet weight) were obtained at the end of the experiment and then stored at −20 °C. The chlorophyll of the subsamples was extracted by grinding 10 mL of 90% acetone in an ice bath in the dark. After storage at 4 °C and 24-h darkness, the extracts were centrifuged at 1664× g and 4 °C for 20 min, and the chlorophyll content in the supernatant was determined using a spectrophotometer (SHIMADZU UV-2700, Shimadzu, Kyoto, Japan) and the formulas of GB17378.7-2007 [38].

The coverage of live tissue, skeleton, and microalgae on the fragments of corals and coralline algae was extracted from the photos by measuring the area of the categories before and after the experiment using image analysis software (Adobe Photoshop CS6 version 13.0). To calculate the growth rate of the algae, we quantified the buoyant weight of coralline algae by using the buoyant weight technique [40]. The wet weight of turf algae and macroalgae was quantified after rinsing them with fresh seawater and drying the surface water with paper towels [41]. The buoyant weight and wet weight were measured with an analytical balance (± 0.1 mg precision).

2.6. Statistical Analysis

The relative consumption/release rate represents the flux of dissolved inorganic matters from COTS carrions to seawater and was calculated using the following Formula (1):

$$R=(C_f-{ C}_i) \times V/\left(W \times T\right)$$

(1)

where R is the relative consumption/release rate; C_f and C_i are the final and initial concentrations of dissolved inorganic matters (DO, NH₄⁺, and PO₄³⁻); V is the volume of the seawater (5 L); W is the wet weight of COTS carrions; and T is the incubation time (days). If R > 0 (e.g., NH₄⁺, NO_x, and PO₄³⁻), it indicates a net release of dissolved inorganic matters from carrions to seawater; if R < 0 (e.g., DO), it indicates a net consumption by decomposers.

The calcification rate (or dissolution rate) represents the accumulation (or loss) of CaCO₃ and is calculated using the following Formula (2):

$$C=(W_f-{ W}_i)/(A \times T)$$

(2)

where C is the calcification rate; W_i and W_f are the initial and final dry weights of coralline algae or coral skeleton; A is the surface area of coralline algal tissue or coral skeleton; and T is the incubation time (14 days). The dry weight of coralline algae was calculated from buoyant weight based on a calcite density of 2.71 g cm⁻³ [40]. If C > 0, it indicates positive calcification of the sample; if C < 0, it indicates net dissolution.

The relative growth rate was used to quantify changes in the biomass of benthic algae and is calculated using the following Formula (3):

$$\mathit{RGR}=\ln (W_f/{ W}_i)/T \times 100$$

(3)

where RGR is the relative growth rate; W_i and W_f are the initial and final wet weights of turf algae and macroalgae; and T is the incubation time. If RGR > 0, it indicates positive growth of benthic algae; if RGR < 0, it indicates negative growth.

Generalized linear models were performed to examine the effects of the treatments (COTS density or temperature), time, and their interaction on pH, DO, and ΔF/Fm’ of benthic algae with a Gaussian distribution and log link and on bacteria abundance with a quasi-Poisson distribution and log link. A two-way analysis of variance (two-way ANOVA) was used to analyze the effects of the treatments (COTS density or temperature), time, and their interaction on DIC, TA, nutrients, and ΔF/Fm’ of the coral-algal symbiont. Differences in dissolution rates of the coral skeletons at different temperatures and morphologies were analyzed using a one-way ANOVA. For the COTS density effect, we used another one-way ANOVA with the approximate method of Welch to adjust the degree of freedom due to unequal variances, as well as the Tamhane’T2 test (PMCMRplus R-package) for further pairwise comparisons. The Student’s or Welch’s t-test was performed to compare the weight and growth rate (or calcification rate) between the treatment and the control. Assumptions of normal distribution and equal variances were tested before analyses using the Shapiro test and Bartlett test, respectively. The above statistical analyses were performed in R 4.2.1. All data are presented as mean ± SE.

3. Results

3.1. Acidifying Seawater

The pH value decreased sharply on day 1, from 7.83 ± 0.003 and 7.83 ± 0.003 to 7.12 ± 0.02 and 7.29 ± 0.01 in the high-density and the low-density treatments, respectively, and to 6.90 ± 0.04 and 7.11 ± 0.003 by day 3 (Figure 2a). The former had a lower mean pH value and a greater reduction rate of pH (7.32 ± 0.01; 0.31 ± 0.01 day⁻¹), compared with the latter (7.42 ± 0.02; 0.24 ± 0.002 day⁻¹). There were statistically significant effects of COTS density and time on changes in seawater pH (p < 0.001,

Table 1. Generalized linear models (GLM) and two-way analysis of variance (two-way ANOVA) for effects of density and time on variation in seawater quality.

Source of Variation	Density		Time		Density × Time
	df	p	df	p	df	p
pH (GLM)	1, 133	<0.001 ***	1, 132	<0.001 ***	1, 131	0.051
DO (GLM)	1, 133	<0.001 ***	1, 132	0.033 *	1, 131	<0.001 ***
Bacteria abundance (GLM)	1, 22	<0.001 ***	1, 21	<0.001 ***	1, 20	0.95
DIC (ANOVA)	1, 32	<0.001 ***	1	<0.001 ***	1	<0.001 ***
TA (ANOVA)	1, 32	0.011 *	1	<0.001 ***	1	0.012 **
NH4+ (ANOVA)	1, 23	0.0086 **	1	0.0013 **	1	0.64
NOx (ANOVA)	1, 23	0.014 **	1	0.88	1	0.72
PO43− (ANOVA)	1, 23	<0.001 ***	1	0.011 **	1	0.018 *

Note(s): * p < 0.05, ** p < 0.01, *** p < 0.001.

). The pH value in the control was stable, averaging at 7.99 ± 0.004. The mean pH values (7.47 ± 0.01) and reduction rates of pH (0.28 ± 0.01 day⁻¹) in the heating treatment were similar to that in the low-density treatment, and the time (p < 0.001) rather than the temperature (p > 0.05) was a statistically significant factor affecting pH dynamics (

Table 2. Generalized linear models (GLM) and two-way analysis of variance (two-way ANOVA) for effects of temperature and time on variation in seawater quality.

Source of Variation	Temperature		Time		Temperature × Time
	df	p	df	p	df	p
pH (GLM)	1, 88	0.21	1, 87	<0.001 ***	1, 86	0.96
DO (GLM)	1, 88	0.93	1, 87	<0.001 ***	1, 86	1
Bacteria abundance (GLM)	1, 22	0.71	1, 21	<0.001 ***	1, 21	0.99
DIC (ANOVA)	1, 32	0.69	1	<0.001 ***	1	0.44
TA (ANOVA)	1, 32	0.50	1	<0.001 ***	1	0.94
NH4+ (ANOVA)	1, 14	0.29	1	<0.001 ***	1	0.76
NOx (ANOVA)	1, 14	0.68	1	0.97	1	0.62
PO43− (ANOVA)	1, 14	0.027 *	1	<0.001 ***	1	0.89

Note(s): * p < 0.05, *** p < 0.001.

). The pH values were restored to 7.63–7.66 by the end but were still below the initial values. The results indicated acidifying seawater caused by COTS decomposition.

Seawater DIC increased from 1637 ± 4 μmol L⁻¹ and 1552 ± 83 μmol L⁻¹ to 4267 ± 461 μmol L⁻¹ and 3304 ± 85 μmol L⁻¹ (Figure 2c), and TA increased from 2139 ± 30 μmol L⁻¹ and 2020 ± 13 μmol L⁻¹ to 5666 ± 127 μmol L⁻¹ and 3893 ± 127 μmol L⁻¹ in the high-density treatment and the low-density treatment, respectively (Figure 2d). The COTS density, time, and their interaction exhibited significant effects on the dynamics of DIC and TA (p < 0.05, Table 1). In the control, DIC and TA ranged from 1803 μmol L⁻¹ to 2193 μmol L⁻¹ and from 2683 μmol L⁻¹ to 2728 μmol L⁻¹, respectively. Similar DIC and TA were observed between the heating treatment (DIC: 3406 ± 173 μmol L⁻¹; TA: 4078 ± 199 μmol L⁻¹) and the low-density treatment by the end. The time (p < 0.001) rather than the temperature (p > 0.05) was a statistically significant factor affecting the dynamics of DIC and TA (Table 2). The increasing DIC and alkalinity suggested elevated CO₂ and alkalinity matters (e.g., CO₃²⁻, HCO₃⁻, NH₃, and PO₄³⁻) from the process of COTS decomposition into seawater, indicating a disturbance of seawater acid/base.

3.2. Hypoxic and Eutrophic Seawater

DO in the seawater was depleted from 6.30 mg L⁻¹ to 0.03 mg L⁻¹ in the high-density treatment and the low-density treatment (day 1–2; Figure 2b), and the consumption rates of DO reached 0.63–1.25 mg g⁻¹ day⁻¹. The seawater was anoxic (<0.5 mg L⁻¹) throughout the 14-day experiment in the high-density treatment, and the anoxic episode lasted for 10 days, with hypoxia (<2 mg L⁻¹) for 12 days in the low-density treatment. The high-density treatment had a lower average DO of 0.52 ± 0.01 mg L⁻¹ compared with the low-density treatment (1.01 ± 0.05 mg L⁻¹). COTS density, time, and their interaction exhibited statistically significant effects on DO dynamics (p < 0.05, Table 1). The DO was replete (6.22 ± 0.03 mg L⁻¹) in the control. The difference in DO was not statistically significant (p > 0.05, Table 2) between the heating treatment and the low-density treatment. The aerobic decomposition of COTS led to hypoxia, even anoxia in the seawater, spelling disaster for reef organisms.

In the COTS treatments, NH₄⁺ increased from <0.2 μmol L⁻¹ to 209–267 μmol L⁻¹, at rates of 3–8 μmol g⁻¹ day⁻¹, and PO₄³⁻ increased from <0.8 μmol L⁻¹ to 20–31 μmol L⁻¹, at rates of 0.4–0.6 μmol g⁻¹ day⁻¹ (Figure 2e,f). But NO_x decreased from 2.3–2.8 μmol L⁻¹ to 0–1.3 μmol L⁻¹ (day 1–4), then maintained at low levels (0.1–2.6 μmol L⁻¹; Figure 2g). The control had low values of NH₄⁺ and PO₄³⁻ (0.1–2 μmol L⁻¹) and increased NO_x (4–7 μmol L⁻¹). COTS decomposition triggered an extra mineralization process, inducing eutrophic seawater, primarily in NH₄⁺ and PO₄³⁻.

Similar trends in seawater were observed in the live coral and benthic algae experiments, which are detailed in the Supporting Information (Figures S3–S5). The light intensity was decreased by 40% in the turbid seawater (Figure S3).

3.3. Bacteria Abundance

The bacteria abundances increased by 4–5 orders of magnitude in the high-density treatment (1.81 × 10⁷ ± 6.86 × 10⁶ CFU mL⁻¹) and the low-density treatment (6.03 × 10⁶ ± 1.58 × 10⁶ CFU mL⁻¹), respectively. They are 1–2 orders of magnitude greater than that of the control (1.21×10⁵ ± 5.57×10⁴ CFU mL⁻¹; Figure 2h). The bacteria abundance was affected significantly by COTS density and time (p < 0.001, Table 1).

Before adding COTS carrions (day 0), the bacteria abundance in the heating treatment (1083 ± 19 CFU mL⁻¹) was two times more than that of the low-density treatment (517 CFU mL⁻¹). In the end, similar bacteria abundances were observed in the heating treatment (6.43 × 10⁶ ± 4.41 × 10⁶ CFU mL⁻¹) and the low-density treatment. The bacteria abundances were influenced significantly by the time (p < 0.001) rather than the temperature (p > 0.05; Table 2). It seemed that the OM decomposition in a greater COTS density would, in turn, promote the proliferation of the decomposer bacteria, while the promotion effect of heating was limited.

3.4. Responses of Live Coral and Coral Skeleton

In the COTS treatment, massive corals had decreasing ΔF/Fm’ from 0.60 ± 0.05 to 0.08 ± 0.003 by day 2 (Figure 3a), and that of branching and monotypic corals decreased from 0.68 ± 0.03 and 0.60 ± 0.05 to 0.10 ± 0.04 and 0.37 ± 0.14 on day 1, respectively (Figure 3b,c). The effect of the COTS treatment on ΔF/Fm’ dynamics was statistically significant (p < 0.001,

Table 3. Generalized linear models (GLM) and two-way analysis of variance (two-way ANOVA) for effects of treatment and time on the ΔF/Fm’ of benthic organisms.

Source of Variation	Treatment		Time		Treatment × Time
	df	p	df	p	df	p
ΔF/Fm’ of coralline algae (GLM)	1, 117	<0.001 ***	1, 118	0.91	1, 116	0.022 *
ΔF/Fm’ of turf algae (GLM)	1, 53	<0.001 ***	1, 54	0.56	1, 52	0.23
ΔF/Fm’ of macroalgae (GLM)	1, 117	0.0052 **	1, 118	0.032 *	1, 116	<0.001 ***
ΔF/Fm’ of massive coral (ANOVA)	1, 50	<0.001 ***	1	<0.001 ***	1	0.81
ΔF/Fm’ of monotypic coral (ANOVA)	1, 50	<0.001 ***	1	0.81	1	0.68
ΔF/Fm’ of branching coral (ANOVA)	1, 50	<0.001 ***	1	0.55	1	0.48

Note(s): * p <0.05, ** p <0.01, *** p <0.001.

). By contrast, the values of ΔF/Fm’ averaged 0.57 ± 0.02 (massive coral), 0.54 ± 0.01 (monotypic coral), and 0.60 ± 0.02 (branching coral) in the control. The photosynthesis of coral symbionts was inhibited in the eutrophic, acidifying, and hypoxic seawater induced by COTS decomposition (Figure S3).

All the coral fragments were transferred to the culture tank on day 2, and they lost tissues from day 3 until death. The relative cover of live tissue, skeleton, and microalgae on the coral fragments was 84.2%, 9.5%, and 6.3% on day 0, respectively, but the cover changed to 6.0%, 45.2%, and 48.8% on day 8 (Figure 3d–j). It seemed that the microalgae occupied the denuded coral skeleton, accounting for the increase in ΔF/Fm’ of the fragments to 0.37–0.42 by the end, but no signs of coral recovery were observed.

The coral skeleton dissolved at a faster rate in the acidifying water. The dissolution rate averaged 0.02 ± 0.004 mg cm⁻² day⁻¹ in the low-density treatment, which significantly doubled in the high-density treatment (0.05 ± 0.01 mg cm⁻² day⁻¹; p < 0.001,

Table 4. One-way analysis of variance (one-way ANOVA) for effects of density, temperature, and morphology on skeletal dissolution.

Source of Variation	Parameter	df	F	p
Dissolution rate of coral skeleton	density	2, 12	48.91	<0.001 ***
	temperature	1, 16	0.0053	0.94
	morphology	2, 24	0.88	0.43

Note(s): *** p < 0.001.

; Figure 4a). The rates were 20–50 times greater than that of the control. The difference in the dissolution rates was not statistically significant between the heating treatment (0.02 ± 0.003 mg cm⁻² day⁻¹) and the low-density treatment (p > 0.05, Table 4).

The branching and foliose corals had similar dissolution rates (0.03 ± 0.01 mg cm⁻² day⁻¹), which were greater than those of the massive coral (0.02 ± 0.004 mg cm⁻² day⁻¹; Figure 4b), while the difference was not statistically significant (p > 0.05, Table 4). The exacerbated dissolution of coral skeletons during COTS decomposition may contribute to the collapse of the reef habitat.

3.5. Responses of Turf Algae, Coralline Algae and Macroalgae

In the COTS treatment, the ΔF/Fm’ of the three benthic algae was significantly decreased (p < 0.01, Table 3), indicating a reduction in photosynthetic efficiency. The ΔF/Fm’ of the turf algae decreased from 0.51 ± 0.01 to 0.07 ± 0.002, and that of the coralline algae from 0.58 ± 0.03 to 0.13 ± 0.02 during days 1–3 (Figure 5a,b). They stayed at the low level for 2–3 days, then were restored to 0.5 by the end. The ΔF/Fm’ of the macroalgae decreased from 0.66 ± 0.01 to 0.48 ± 0.05, then increased to 0.75 ± 0.01 (Figure 5c). In the control, the ΔF/Fm’ was in the range of 0.60–0.73 for the turf algae, 0.51–0.69 for the coralline algae, and 0.72–0.84 for the macroalgae.

Compared to the control, the total chlorophyll content of the coralline algae decreased by 66% to 11.4 μg cm⁻² (Figure 5d, p < 0.01,

Table 5. Results of the t-tests showing significant differences in chlorophyll content, weight, and growth rate (or calcification rate) of benthic algae between treatments.

Source of Variation	Parameter	Coralline Algae		Turf Algae		Macroalgae
		df	p	df	p	df	p
Chlorophyll content	Treatment	6	0.0065 **	6	0.40	6	0.0018 **
Weight	Treatment	3	0.027 *	6	<0.001 ***	6	0.038 *
Growth rate (or Calcification rate)	Treatment	3	0.076	6	<0.001 ***	6	0.038 *

Note(s): * p < 0.05, ** p < 0.01, *** p < 0.001.

), and that of the turf algae decreased by 17% to 58.8 μg g⁻² (Figure 5e, p > 0.05, Table 5). Reversely, that of the macroalgae increased by 57% to 442.2 μg g⁻² (Figure 5f, p < 0.01, Table 5).

In the COTS treatment, the buoyant weight of the coralline algae decreased by 3.8 ± 1.8% (p < 0.05), showing net dissolution at a rate of 2.57 ± 2.12 mg cm⁻² day⁻¹ (Figure 5g; Table 5). The wet weights of the turf algae significantly decreased by 67.3 ± 2.7% (p < 0.05), and the macroalgae decreased by 18.3 ± 5.8% (p < 0.05); they had negative growth rates of −16.1 ± 1.3% and −1.5 ± 0.5%, respectively (Figure 5h,i; Table 5). In the control, the buoyant weight of the coralline algae increased by 3.1 ± 0.4%, showing net calcification at a rate of 0.93 ± 0.25 mg cm⁻² day⁻¹. The increases in wet weight of the turf algae (24.7 ± 0.7%) and the macroalgae (1.1 ± 4.6%) allowed them to have positive growth at rates of 1.6 ± 0.04% and 0.06 ± 0.32%, respectively.

The relative cover of live tissue, denuded skeleton, and microalgae on the coralline algae fragments before adding COTS (day 0) was 73.8%, 16.1%, and 10.1%, respectively, but the cover changed to 11.6%, 19.8%, and 68.6% by the end (Figure 5p). The coralline algae lost its pink, live tissues and then was taken by microalgae (Figure 5i,m), and the turf algae decayed (Figure 5k,n). Of the three benthic algae, the macroalgae were the only survivors (Figure 5l,o).

4. Discussion

COTS decomposition induced acidifying, eutrophic, and hypoxic seawater (Figure 1). The changes were much more dramatic compared with those caused by other eruptive marine species, e.g., jellyfish [23,42]. COTS decomposition had higher rates of DO consumption (0.63–1.25 mg g⁻¹ day⁻¹), longer duration of hypoxia (12–14 days), and higher release rates of nutrients (NH₄⁺: 6–23 mg g⁻¹ day⁻¹; PO₄³⁻: 0.5–2.5 mg g⁻¹ day⁻¹), compared with those of jellyfish (0.2 mg g⁻¹ day⁻¹; 7 days; NH₄⁺: <3 mg g⁻¹ day⁻¹; PO₄³⁻: 0.63 mg g⁻¹ day⁻¹), despite having less carrion content in the COTS experiment (2.5–5 g L⁻¹) than that in the jellyfish experiment (27 g L⁻¹) [23]. This is likely because COTS individuals have greater organic content (10%; Table S2) than jellyfish (3%) [43]. OM can be decomposed through bacterial metabolism in both aerobic and anaerobic ways [44] (Figure 1); in turn, OM supports bacterial proliferation [33]. In this study, the bacteria abundances increased by 3–5 orders of magnitude, and they consumed 97% DO on day 1. Meanwhile, bacterial respiration released CO₂, thereby lowering seawater pH from day 1 to 3 and increasing DIC and TA by 59–105% and 45–111%, respectively. Although the CO₂ escaping from seawater likely increased pH since day 3, the continual CO₂ releases during the denitrification process maintained seawater acidifying. OM can be degraded into DIM, including NO₃⁻ and PO₄³⁻, but NO₃⁻ further transforms to NH₄⁺ and N₂ in the process of denitrification. These processes explain NH₄⁺ and PO₄³⁻ enrichment, O₂ and NO_x depletion, and seawater acid/base disturbance during COTS decomposition.

Heating promoted bacterial growth [32]. Indeed, the abundance of bacteria doubled in the heating treatment, compared with that in ambient temperature (26–27 °C vs. 32 °C; Figure 2h). However, the OM content (indicated by COTS density), rather than seawater temperature, became the primary factor affecting decomposer (bacteria) and seawater quality during COTS decomposition (Table 1 and Table 2). The result is similar to the jellyfish decomposition, where global warming (e.g., heating +3 °C) produced little effect on sediment oxygen demand, NH₄⁺, and PO₄³⁻ in seawater [22]. However, the heating may mitigate the acidifying seawater to some extent; specifically, pH recovered to 7.5 in the heating treatment, three days earlier than that under ambient temperature (Figure 2a). This may be the effect of negative feedback, as ocean warming mitigates ocean acidification through decreasing CO₂ solubility in the seawater surface [45].

COTS decomposition posed negative impacts on live corals (Figure 3a–c), including both stress-tolerant species (massive coral) and stress-sensitive ones (branching and encrusting corals) [46]. The ΔF/Fm’ of coral symbionts sharply decreased from 0.6 to 0.1 on day 2. The low value of ΔF/Fm’ (<0.5) indicates an unhealthy status of corals [47]. The suppression of photosynthesis was likely related to the deteriorated seawater (Figure S3; pH = 7.37; NH₄⁺ = 55 μmol L⁻¹; PO₄³⁻ = 21 μmol L⁻¹; DO = 0.2 mg L⁻¹; a 40% decrease of light intensity). Light reduction inhibited the photosynthesis of coral symbionts, causing net respiration of OM in the bottom seawater during daylight hours [48]. When seawater pH decreased by 0.4–0.8 units to 7.6, net O₂ production of corals decreased by 50–100% [49]. Under high available CO₂ and pH < 7.5, coral, as a HCO_3⁻-user, showed decreasing photosynthesis [50]. External seawater pH declines simultaneously induced cell pH declines, which probably yields intracellular acidosis in corals [51]. Moreover, the nutrient tolerance thresholds for many corals are only 9 μmol L⁻¹ DIN and 0.3 μmol L⁻¹ DIP [3]. After adding 10 μmol L⁻¹ of NH₄⁺, the Fv/Fm of coral Siderastrea siderea decreased by 8–15% [52]. Nutrient enrichment induced rapid tissue loss in corals [52], as observed in this study. It is noted that hypoxic water (DO < 0.5 mg L⁻¹) caused a 12% decrease in the Fv/Fm of coral symbionts, or even 76% coral bleaching and 100% mortality [4,19,28,53]. Particularly, the seawater rapidly changed during the decomposition, e.g., pH declined by 0.9 units at a rate of 0.3 day⁻¹. This rate is much greater than the global ocean acidification rate, where pH has declined by 0.1 unit since pre-industrial times [36]. Benthic organisms are generally more susceptible to acute stresses than chronic stresses. For instance, chronic thermal stress yields a negative effect on coral growth rate, while acute thermal stress directly induces coral mortality [1]. This is because organisms do not have enough time to adapt to acute stresses [54]. All these dramatic changes in the bottom seawater during COTS decomposition become a formidable risk to corals, as well as other organisms of coral reefs.

COTS decomposition is fatal for live corals with short-term (12–24 h) exposure to deteriorated seawater (Figure 3d–i). Corals lost 93% of their tissue area, and then algae colonized 50% of the substrate area by the end (day 6). Corals have the potential to recover from damage within their tolerance. For instance, after removing seawater acidification stress (pH = 7.4 for 12 months), corals calcified again and reformed colonies [55]. After removing nutrient stress (10–20 μmol L⁻¹ NH₄⁺ and 2 μmol L⁻¹ PO₄³⁻ for 9 weeks), the coral growth rates partly recovered by 30–35% [56]. However, hypoxic stresses often cause severe, irreversible damage to corals. After a 12-h hypoxic stress, Acropora selago corals failed to recover from bleaching in subsequent reoxygenation phases and even had a further ~20% reduction in Symbiodiniaceae densities due to ineffective and delayed gene expression to support their hypoxia response systems [53]. This may explain the coral mortality after short-term exposure to hypoxia during the aerobic decomposition of COTS.

After corals die, the reef structure could provide a habitat for organisms for 4–7 years before the collapse [6]. However, COTS decomposition may hasten the collapse of the reef habitat since the acidifying seawater induced by the decomposition significantly enhanced the coral skeleton’s dissolution (Figure 4a). The dissolution rate was 0.05 mg cm⁻² day⁻¹ in the high-density COTS treatment (mean pH = 7.32). This rate is greater than that caused by global ocean acidification; for example, the net dissolution rates of the Great Barrier Reef ranged from 0.006 to 0.04 mg cm⁻² day⁻¹¹ [57]. The dissolution of the calcareous (CaCO₃) skeleton of corals mainly depends on pH. In response to the decreasing pH from 8.2 to 7.6, the dissolution rate of the CaCO₃ skeleton doubled in the treatment [58]. At present, global ocean acidification may not be fatal to live corals, as they still calcify and offset skeletal dissolution. But when coral tissues are entirely or partly ingested by predators (e.g., Acanthaster spp.), the denuded coral skeleton is exposed to acidifying seawater, exhibiting greater dissolution rates. Branching coral has 1.1–2.6 times greater dissolution rates than massive coral during COTS decomposition (Figure 4b). This is consistent with previous research, showing 2–3 times greater susceptibility of branching coral skeletons to erosion than massive coral skeletons [59]. The difference in dissolution rates may be related to the skeletal structure strength, density, pore size, surface area, and thickness [60]. The structural complexity of coral reef habitat is primarily formed by branching corals [61]. However, the skeletal dissolution, particularly in branching corals, tends to make the three-dimensional structure of the reef vulnerable to waves and storms [62]. The findings suggest an increased possibility of structural collapse of the habitat for reef organisms under the influence of COTS decomposition.

As for the three benthic algae, their photosynthesis was significantly inhibited during the decomposition (Figure 5a–c), and turf algae lost 67% biomass (Figure 5h). Benthic algae are fundamental producers in the reef ecosystem, contributing to >90% of primary productivity [63]. A meta-analysis showed that turf algae are widely thought to be the major contributors to primary productivity [64]. Turf algae covered 70% of space as the abundant component of algal assemblages, contributing to a primary productivity of 1.3–2.9 g C m⁻² day⁻¹ on South Australian coasts [65]. Therefore, the inhibited photosynthesis and biomass loss of the algae imply that the primary productivity of the coral reef ecosystem tends to be compromised during COTS decomposition.

Importantly, coralline algae serve as the preferred substrate for coral larval settlement, providing specific microbial and metabolomic cues on the algal surface [66]. However, the coralline algae were severely damaged, varying from net calcification to net dissolution (Figure 5g). The suppression of coralline algae may be mainly related to the acidifying seawater. Coralline algae consist of CaCO₃ in the form of high magnesium, which is more soluble than aragonite [67]. Therefore, they exhibited more sensitivity to the acidifying seawater from the decomposition than the coral skeleton. Further, the coralline algae lost 62% of its tissue area, and then microalgae occupied 68% of the substrate surface by the end (Figure 5p). This is because the dead substrate could be colonized by green algae within only seven days [19]. Green algae Ostreobium and cyanobacteria are considered primary micro-bioeroders [62], contributing to erosion rates of 0.09 g CaCO₃ cm⁻² year⁻¹ [60]. The bioerosion may further enhance the substrate dissolution of coralline algae. Hence, the destroyed coralline algae may be more fragile in waves, restricting coral recruitment and reef recovery.

However, macroalgae are the only survivors, although their growth was suppressed, showing more tolerance to the deteriorated seawater than live coral, turf algae, and coralline algae (Figure 5j–o). Nutrients (~80 μmol L⁻¹ NO₃⁻; ~80 μmol L⁻¹ NH₄⁺; ~4 μmol L⁻¹ PO₄³⁻) and CO₂ enrichment (pH = 7.62) promoted macroalgae growth, and phosphorus (30 μmol L⁻¹) stimulated photosynthetic inorganic carbon utilization and nitrogen uptake by macroalgae [29,68]. Hence, the tolerance and resilience of macroalgae allow them to compete for space and light with corals through direct contact and indirect chemical release [69]. More likely, the reef communities experience a coral-algal phase shift after multiple stressors, such as heating, disease, pollution, and overfishing, due to the greater tolerance and dominance of macroalgae [70]. This evidence supports the view that the stresses from COTS decomposition may drive continuing declines in the reef community.

Pollution discharges into the ocean might generally be expected to be diluted by water exchange. In fact, hypoxia ‘dead zones’ and ocean acidification linked with eutrophication are spreading worldwide [19,44]. In addition to permanent hypoxia, oxygen deficiency also occurs temporarily due to seasonal changes or specific events, such as excessive nutrients carried by extreme precipitation or released from decomposing jellyfish, etc. [20,71]. The hypoxia events have induced massive mortality of corals and other reef-associated organisms [19]. This suggests a limited capacity of water exchange to dilute excessive nutrients or pollutants within a short time. Especially in coral reef regions that often have complex topography, underwater reefs may create resistance, constraining the free movement of ocean currents. Therefore, the decomposition of high biomass of COTS could be a potential risk to reef environments. Moreover, the effects of the decomposition may be more complicated on long-term scales since enrichment may improve the food availability (e.g., phytoplankton) and increase the survival rate of COTS larvae [72], which have the potential to support the periodic COTS outbreaks

5. Conclusions

This experimental study explored the influence on coral reefs during the collapse phase of the COTS population outbreak. COTS decomposition triggered an extra mineralization process, disturbing the normal biochemical cycles of the coral reef ecosystem. As a consequence, the dramatic changes in seawater were fatal to the species that are sensitive to environmental changes, including the key functional groups in the ecosystem, e.g., reef-building coral and coralline algae, and major primary producers–turf algae. The unexpected promotion of tolerant species, e.g., macroalgae, however, may drive further declines in the coral reef ecosystem. Therefore, it is necessary to put emphasis on the post-outbreak phase of COTS on coral reefs, and strategies should be taken to reduce the risk of abrupt collapses of the massive COTS population under seawater. Instead of aiming for the complete elimination of all individuals, managers should focus on controlling the COTS population to levels below the outbreak thresholds. Before the collapse of the COTS population, proactive measures should be taken to control and remove them. It is essential to promptly remove deceased COTS individuals from the reefs. Those measures are crucial for mitigating the adverse impacts caused by the excessive release of organic matter on reef ecosystems.