Corallivory and algal dynamics on some coral reefs in the Persian Gulf

Macroalgae are a sign of degradation of coral reefs. Distribution of macroalgae on reefs is moderated by grazers including ﬁ sh and sea urchins. However, several ﬁ sh species including certain parrot ﬁ shes graze on live coral tissues, at times causing profound damage. In this paper, the potential role of macroalgae in suppressing parrot ﬁ sh predation on Porites corals, the dominant coral genus in Qeshm Island, is investigated at three research sites at Qeshm Island in the Persian Gulf between April and July 2014 and 2015. Macroalgae, which were abundant in April, decreased signi ﬁ cantly in frequency in July, while at the same time, the percentage of Porites colonies, the frequency of ﬁ sh bite marks on Porites colonies, and the overall area of live coral tissue, which was grazed by parrot ﬁ shes increased dramatically, all of which were only negligible in April (with certain exceptions). Nevertheless, no changes were observed in parrot ﬁ sh abundance. Despite partially supportive statistical data, because of the observed exceptions, this phenomenon is more likely to be due to other factors, in particular the increased nutritional values of the corals in July in comparison to April. However, to understand the cause(s) and mechanisms involved in this annual phenomenon, more investigations seem necessary.

The most important macroalgae-moderators, however, are herbivores, including sea urchin and fi sh [8,9]. In spite of this, corallivorous fi shes are a threat to coral reefs themselves [10,11].
These fi shes reduce coral abundance, diversity and colony size [11], coral expansion and spread, resistance to other stressors and resilience [12] and also increase coral mortality [13].
Although the role of herbivorous fi shes in moderating macroalgal assemblies as well as the impact of corallivorous fi shes on coral reefs is well documented, there is limited information on the effects of macroalgae on fi sh corallivory.
It was it discerned that under experimental conditions, macroalgae can provide protection for juvenile corals against parrotfi shes [14], which are among the most important coral predators preferring massive Porites corals on many reefs [12] and references therein). Moreover, fi eld studies mainly focus on interactions among algae, fi shes and juvenile corals (e.g. [15].
Hoey & Bellwood [16] showed that highly dense macroalgae could repel given fi sh species including parrotfi shes from some reefs. This study implies that if there are any coral colonies among those macroalgal assemblages, they could be sheltered from fi sh corallivory, this hypothesis still needs to be tested directly.
In this study, fi eld data is gathered to test if macroalgae are involved in moderating parrotfi sh predations of massive Porites colonies at Qeshm Island, the Persian Gulf. Several parameters, Citation: Kavousi

Material and methods
This study was done at the southeast of Qeshm Island (26°45'N, 55°49'E), situated in the Persian Gulf. Coral reefs in this place are located near-shore, and are mainly less than 6 m deep dominated by massive Porites spp. for 89% ± 12.2% (of live coral cover ± standard deviation) [17]. Three sites, namely 1, 2 and 3 were chosen at the two ends as well as the middle of the reef site ( Figure 1). Surveys of fi sh and benthic assemblages were performed in the middle of April, and then again in early fi sh bites on the colonies belonged to parrotfi shes, we only show the data for this fi sh. Furthermore, 400cm -2 -photo-quadrates were taken at 1m -1 intervals along each 30m -1transect to assess the frequency and extent of damage from parrotfi sh bite marks on Porites colonies between April and July. To this end, fi sh bites of parrotfi shes were categorized into spot bites (i.e. individual, superfi cial bites) and focused bites (i.e. repeated overlapping deep bites resulting in extensive live tissue removal) [18]. The number of spot fi sh bites and the area of both spot and focused bites on Porites surfaces were counted and estimated using the software CPCe [19].
From April 2015, we calculated the number of bites of S. persicus on both Porites corals and algae min -1 [20], in three month intervals.
As the data sets did not meet the assumptions of parametric tests, and transformations did not work, the non-parametric Kruskal-Wallis test was applied to compare each factor among the sites in each month. Also, the Mann-Whitney U test was utilized for pairwise comparisons. We used the Holm-Bonferroni method (Holm 1979) to correct the p values to test the respective signifi cant differences in each site between the months. The Signifi cance level, was set at 0.05. Either the Pearson correlation test (when data sets were normally distributed) or the Spearman rho test was used to investigate the relationships between each type of algal groups (MA, FA or AL cover) and other factors. Statistical analyses were performed using SPSS (ver. 22) [21].

Fish dynamics
There were only two species of parrotfi shes present at Simultaneously, severe fi sh biting on more than 80% of the Porites colonies could be observed [17]. Previously, t was found that experimental removal of algae from around the juvenile corals could increase parrotfi sh predations of those corals [14]. It was suggested that coral camoufl aging and creating microhabitats with lower graze rates were two mechanisms by which algae could protect juvenile corals against parrotfi shes [14]. However, none of these appeals to the present study because the massive Porites colonies in this work, which in some cases were 2-3 m in size, could not be sheltered from fi sh predation by either camoufl age or microhabitats. Thus, we have to consider other mechanisms before entertaining the idea whether macroalgae prevented fi sh predation of Porites corals at Qeshm Island.

Increase in adult parrotfi sh number in July rather than in April
It is shown that increase in the abundance of adult corallivore fi shes usually leads to higher grazing on corals [12,22,23]. In the present study, the unusual increases in fi sh bite marks on coral colonies and the number of grazed colonies at site A in April 2015 compared to July could be due to the signifi cantly higher abundance of adult fi shes of both species. On the other hand, in neither July 2014 nor July 2015, did fi sh abundance show changes compared to April. Therefore, we cannot relate the observed phenomenon to parrotfi sh abundance.

Declining food resources increased corallivory
Parrotfi shes are essential to some coral reefs because they feed on macroalgae [10,11]. Bellwood & Choat [24] showed Citation: Kavousi  that the majority of Scarus species in the Great Barrier Reefs consume epilithic algae. Carpenter, et al. [25] also reported that Scarus spp. in the Persian Gulf were mainly grazers on shallow benthic algae. However, neither the study by Carpenter, et al. [25] nor any other study presents any information on the diet of S. persicus. In spite of this, fi eld observations confi rm that this species seems to be a scraper (i.e. it takes a piece of substratum and its accompanying algae in each bite [26] and eats fi lamentous algae (P. Tavakoli-Kolour, pers. obsv). In September 2009, although macroalgae completely disappeared, short fi lamentous algae prevailed. Simultaneously, severe fi sh biting on more than 80% of the Porites colonies could be observed [17]. Filamentous algae are favorable food sources for herbivorous fi sh [28], and are more desirable, palatable and digestible than fl eshy algae [27][28][29]. Therefore, high corallivory rates in September, when fi lamentous algae were abundant at the research locations, cannot be attributed to scarcity of food.

Macroalgae prevented parrotfi sh grazing on Porites corals
It has been proven that herbivorous coral reef fi shes avoid dense macroalgae patches and prefer places with lower macroalgae [16]. Thus, if there are coral colonies among such dense macroalgae, they are most likely naturally protected from fi sh corallivory. In this study, there was no change in parrotfi sh distribution at any site between April and July, which suggests that macroalgae did not repel fi sh from the reefs. The increased fi sh abundance at site A in April 2015 backs this hypothesis.
Thus, the low fi sh bite rates in April must be attributed to other factors. For example, macroalgal cover might prevent fi sh from grazing on corals through other mechanisms, most likely producing chemicals, which are reported to exist in macroalgae [30]. Recently, it has been shown that both coral larvae and juvenile fi shes of different species avoid waters with water soluble chemicals produced by macroalgae [31]. However, the mixed statistical results and low macroalgal cover at two sites in April in each year (~5%) bring up two other hypotheses that may be involved in the fi ndings of this study.

Increased water temperature increased corallivory
Previous studies show that changes in water temperature can change the food intake rate [32][33][34]. Some studies show that increases in water temperature increase metabolism and nutrient requirement rate among fi shes [34][35][36][37]. Water temperature obviously increases in the Persian Gulf in summer, which could explain the potential increase in graze rate by the fi shes, which, however, did not happen. Indeed, there was a signifi cant increase in the number of fi sh bite min -1 (only on algae) in Febrauray 2016, when it was cold. Despite this, even if there was an increase in fi sh grazing rates due to elevated temperatures, the reason(s) why there was a shift from the main food source of S. persicus (whatever kind it is) to corals may not be explained by increased seawater temperature.  This study brings up the hypothesis of the naturally protective role of macroalgae of adult corals against parrotfi shes. This ecological process is an annual phenomenon happening among the southeastern reefs of Qeshm Island sometime in June, July or September of every year. Although there is some supporting data, this phenomenon is more likely to be related to enhanced corals' nutrition (or another factor), in particular, to the fact that the macroalgal cover was not dense enough to hide or somehow impact large coral colonies (~5% at sites A and B in April 2014, and 5% at sites A and C in April 2015). Therefore, the increased impacts of parrotfi shes on Porites colonies while macroalgal cover decreased at Qeshm Island may be coincidental.
The coral colonies grazed by parrotfi shes in September 2009 were found to have fully healed by November of the same year [17]; however, in our most recent study in August 2014, partial coral mortality due to fi sh biting was also observed, and in 2015, the grazed colonies were observed even in October. Therefore, healing is not always an option. To better understand this phenomenon, further investigation in particular at other reef sites is deemed necessary.