The relative roles of recruitment and post-recruitment processes in the regulation of a coral reef Damselfish population

Assessing the processes by which populations are regulated is one of the most important goals in ecology. The local abundance of "open" marine populations may be regulated by recruitment and/or post-settlement mortality. These demographic rates could be influenced by varying levels of predation, competition and habitat structure, which may interact in complex ways. The relative importance of these factors in determining population size and structure has rarely been estimated. This thesis addresses this goal by focussing on the local dynamics of a coral reef fish Pomacentrus moluccensis at Lizard Island, northern Great Barrier Reef (GBR). Fundamental to this, was testing the impact of implant microtags on the growth and survivorship of P. moluccensis. These tags would enable the individual recognition required for mortality, migration and growth studies. Subsequently, the relative importance of recruitment, predation, habitat structure and their interactions in determining these patterns was investigated. These results were evaluated against generalisations emerging from studies on P. moluccensis at other locations on the GBR and other coral reef fishes. \ud \ud The impact of two types of microtag on P. moluccensis were tested: Firstly, visible implant fluorescent (VIF) tags which provided an external marker and secondly, coded wire tags (CWT) which were not visible externally but provided individual codes. Retention rates for both microtags in new recruits of this species (10-20mm SL) were high (100%). In addition, there was no significant difference in either growth or survivorship between fish with or without microtags. \ud \ud The basic life history traits and dynamics of P. moluccensis were assessed. Ageing P. moluccensis at Lizard Island by otolith increments was validated by inducing both tetracycline and stress markers on the otolith at a known time. Age and size structures varied among locations but growth rates were relatively uniform. Longevities at three sites at Lizard Island were found to be fairly similar with very few fish living longer than 6 years. This represents an extreme contrast to southern reefs where approximately 10% of individuals reached 18 years. This greatly reduced longevity at Lizard Island suggested that these populations turn over rapidly and are more likely to be recruitment limited. \ud \ud Patterns of adult and new recruit abundance were variable between sites over time at both local (transects meters to tens of meters apart) and broad (sites hundreds of meters to a few kilometers apart) scales. Recruitment to Lizard Island was up to an order of magnitude higher than recruitment recorded to the southern GBR. In contrast to the implications of reduced longevities at Lizard Island, higher recruitment is more likely to result in populations which are limited by post-settlement processes. Which would prevail? \ud \ud There was evidence of variable density-dependent mortality at two spatial scales. There was no relationship between adults and new recruits in transects at four local scale sites which suggests that at this level there may be factors other than density which were also important. However, there was a significant curvilinear relationship between adult densities and new recruit densities at nine broad scale sites implying density dependence at this scale. In addition, densities of several cohorts which were widely differing in year 1 showed convergence in the following year. These results suggested that post-settlement mortality was capable of altering initial patterns laid down at recruitment. \ud \ud The patterns of distribution of new recruits of P. moluccensis were used to assess the microhabitats with which they were most commonly associated. This species was highly associated with caespitose and corymbose acroporids, and also pocilloporid corals, but all other microhabitats were used less than would be predicted. The coral species with which P. moluccensis was highly associated, all show a highly complex internal structure. Could this attribute account for the observed patterns? A manipulative experiment was performed to test the effects of the peripheral branching structure of one of these coral species on the growth and survivorship of tagged P. moluccensis. Mortality rates were very high at all six sites around Lizard Island (67 - 93%). However, there was no significant difference in either growth or survivorship of fish between manipulated and unmanipulated coral heads. \ud \ud There was significant spatial variation in predator abundance among nine sites at Lizard Island. These predator densities correlated strongly with densities of new recruits suggesting a predator aggregative effect. Predation pressure was assayed by capturing newly settled individuals and then releasing them at the same nine sites. The majority of consumed individuals were taken by either Pseudochromis fuscus (Pseudochromidae) or Thalassoma lunare (Labridae). Small cephalopholids also took a relatively high proportion which was surprising because these species often hide or are subdued in the presence of a diver. Released new recruits selected the same coral species with which the unmanipulated individuals mentioned previously were highly associated. Mortality rates varied considerably according to microhabitat with lowest mortalities from the highly selected microhabitats. There was also significant variation in the percentage cover of these microhabitats between the nine sites, suggesting that microhabitat availability could influence local survivorship patterns. Other fishes behaving aggressively towards the released new recruits of P. moluccensis, did not appear to be important in determining their mortality. \ud \ud Patterns of mortality appeared to be a complex interaction between recruitment, predator densities and the availability of preferred coral. Two further experiments were performed to examine these interactions. The first tested the effects of microhabitat on the behaviour of three predator species and assessed their ability to catch P. moluccensis new recruits in aquaria. The second tested the effects of resident predators and microhabitat on the survivorship of P. moluccensis on patch reefs. In both cases microhabitat modified the effects of predation. The predatory abilities of Pseudochromis fuscus, Cephalopholis boenak and Thalassoma lunare varied according to microhabitat. In addition, there was no difference in the survivorship of P. moluccensis between patch reefs with no resident predator and those consisting of high complexity corals with a resident predator. This suggested that high complexity corals may offer almost complete refuge from predation. \ud \ud The factors regulating local populations of P. moluccensis at Lizard Island contrast with results from southern reefs where individuals were long lived and populations were found to be recruitment limited. Post-settlement processes at Lizard Island resulted in high mortality rates which varied with density and location. Both predator abundance and habitat structure were important in determining abundances of P. moluccensis. Hence, initial patterns of recruitment were modified by patterns of mortality to determine the size of the adult population. The difference between the northern and southern studies could be due to differences in either the scale or the locations. For example, the southern study considered processes occurring between reefs while this study looked at variation between and within sites on the same reef. Alternatively or in addition, traits specific to each location, such as the magnitude of recruitment, could be causing the differences observed. In conclusion, these results would suggest that single process theories of population regulation should be treated with caution, and instead attention should focus on the conditions which may lead to the dominance of any one process. This information would lead to the formation of universally applicable theories.\u