Introduction ============ Tropical coral reefs are an integral part of the earth's bio-geography, as they provide rich habitats for a vast variety of organisms that ultimately define these marine ecosystems (Wachenfeld [@b73]). However, recent observations reveal that coral reefs have been declining as a result of global climate change, human activities and other ecological stressors (Riegl *et al.* [@b60]; Hoegh-Guldberg *et al.* [@b34]). Many of the reef systems worldwide are either already overfished or under stress from intense human exploitation (Pandolfi *et al.* [@b51]; Riegl *et al.* [@b60]). These reef systems have lost the ecological resiliency needed for sustainability and have become increasingly vulnerable to climate change, ocean acidification and other anthropogenic stressors (Hoegh-Guldberg *et al.* [@b34]; Hoegh-Guldberg *et al.* [@b33]). In the Indo-Pacific Ocean, coral reefs are highly diverse and support \>20% of the world's fish biomass (Spalding *et al.* [@b67]; Kappel [@b39]). These reefs provide habitats for a diverse array of ecologically important fish species (Welsh *et al.* [@b74]), some of which are commercially exploited by the aquarium trade or are important dietary components of subsistence fishermen in developing countries (McCook [@b45]; Ruckelshaus *et al.* [@b63]; Bölling *et al.* [@b4]). In countries like Papua New Guinea, reefs are critical for livelihoods because of fishing and marine resource provision (Laurance & White [@b41]). Thus, the reef decline has broad-reaching socio-economic implications for island nations (Laurance *et al.* [@b42]; Le Souef *et al.* [@b40]). Furthermore, reef systems in the Indo-Pacific Ocean contain large reserves of undiscovered and potentially undiscovered species with the potential to reveal new taxa and evolutionary patterns of adaptation in the ecosystem (Rocha *et al.* [@b59]; Miller *et al.* [@b46]). They are also among the most important sources of genetic and protein resources globally, supporting the majority of commercial aquaculture fish (FAO-IVU [@b16]; Hedgecock & Wilson [@b31]; Veron & Boudreaux [@b72]). However, with the rapid degradation of Indo-Pacific reefs, the number of endemic fish species has decreased by 30--40% in recent years (Veron *et al.* [@b70]). Despite an on-going global decline in coral cover and reef systems over the past 20 years, little is known about the potential recovery of Indo-Pacific reefs in the future (Loya *et al.* [@b44]; Cinner *et al.* [@b9]). The main question that arises is, will Indo-Pacific reef systems still support the diversity and abundance of reef fishes in 50 years (Hoegh-Guldberg *et al.* [@b33]), or will the future be a reef ecosystem dominated by one or several dominant species? As climate change continues to modify the ocean, many of the world's coral reef systems could experience either increased coral cover or decreased coral cover, along with either decreased or increased fish species richness. This is because of the well-established relationships between coral cover and fish species richness (Nugues-Brazier *et al.* [@b50]) and the strong association between fish abundance and oceanographic conditions such as temperature (Hughes *et al.* [@b35]) and sea-level variations (Cinner *et al.* [@b9]). Given the ecological and economic importance of coral reefs worldwide, there is an urgent need to develop ecologically informed strategies for managing these systems sustainably. However, to design effective management strategies, the potential impact of environmental change on the dynamics of reefs must first be estimated. In the context of the coral reef meta-ecosystem, these dynamics could be determined using ecological models that seek to estimate the response of reef structure to changes in human exploitation and reef condition (Bellwood *et al.* [@b2]; Cinner *et al.* [@b9]). There are several factors that determine reef condition that influence the ability of reef fish species to co-exist. These include the amount of suitable habitat, environmental factors (e.g. ocean acidification) and trophic interactions (Nugues-Bravo *et al.* [@b49]; Priede *et al.* [@b54]). When the relative abundance of species and their niche partitioning change because of any of these factors, it may have significant effects on the assemblage structure and thereby alter the functioning of reef systems. The response of a meta-ecosystem to climate change can be modelled using the principles of population dynamics (Bellwood *et al.* [@b2]). There have been several recent efforts to use models of coral reef community dynamics to evaluate the likely outcomes of the current and future environmental changes, particularly in the context of climate change (Cinner *et al.* [@b9]; Cebrian & Goreau [@b5]; Leathwick *et al.* [@b40a]; Hughes *et al.* [@b35]). However, these previous studies often did not consider a full range of ecological processes that can influence the distribution of fish species across ecological networks. Despite a lack of explicit theoretical models for ecosystem dynamics on Indo-Pacific coral reefs, several ecologically important factors were identified in a review paper by Cebrian & Goreau ([@b5]). For example, it was shown that most fish species are limited by available prey and/or habitat (Cebrian & Goreau [@b5]). In their model, Cebrian & Goreau ([@b5]) used these ecological processes to project fish distributions under various scenarios. However, these models did not account for either the effect of other fish species or the effects of environmental changes on habitat quantity, quality and connectivity. While there has been an increase in the number of studies on the ecology and population dynamics of reef fish species (e.g. Meffe & Bellwood [@b44a]; Hughes *et al.* [@b35]), and the response of these fishes to changes in environmental parameters (Bellwood *et al.* [@b3]; Stobutzki & Bellwood [@b69]), there has been little attempt to integrate these factors into a model of the dynamics of whole reef-associated fish assemblages. In this paper, we present a meta-ecosystem model of the reef-associated fish community of Indo-Pacific coral reefs, where the spatial extent of the study area and data available are such that the effects of all of the important ecological processes on fish distributions can be identified. With this meta-ecosystem model, we show how the structure and dynamics of the whole assemblage can be controlled by key ecological processes. By developing models of community dynamics that include the dynamics of key species interactions, it is now possible to improve the design and use of ecological models of reef ecosystems. While this is a long-term research goal, a practical application of this model is that it can be used to simulate potential management strategies for reef systems (Amarasekare *et al.* [@b1]), which will assist in the formulation of effective sustainable management plans to mitigate adverse changes caused by environmental change (Hughes *et al.* [@b35]). In addition, this meta-ecosystem model can help us to understand how marine ecosystems will be impacted by environmental change. In this study, we developed a meta-ecosystem model for Indo-Pacific coral reefs that allows the distribution and abundance of fish species to be simulated under different scenarios, including changes in the quantity and quality of habitat, changes in the connectivity of habitats, changes in adult and juvenile survivorship, changes in population abundance and shifts in the fish life history traits. We show how each of these changes to environmental factors will control the dynamics of the reef fish species assemblages. We used this model to identify the potential key controls of reef fish community dynamics under scenarios of changing environmental parameters. In doing this, we found that the meta-ecosystem model showed a potential 'shift to dominance' when the abundance of a dominant fish species was high, which was driven by changes in the adult survivorship of coral-associated fish species. In addition, the meta-ecosystem model predicted a range of potential 'rescue' effects when there were changes in the quantity of the juvenile habitat that favoured a more herbivorous fish community. To test our model, we used species richness data from the present-day Indo-Pacific, which were combined with various scenarios of environmental change, in order to predict the potential effects of environmental change on the species composition and assemblage structure of Indo-Pacific reefs. Materials and methods ===================== Meta-ecosystem model -------------------- To model the structure of Indo-Pacific coral reef fish assemblages, we simulated a meta-ecosystem of interacting coral and fish populations. This system is dynamic and is described by a network of processes that include both population dynamics and species interactions (Fig. [1](#fig01){ref-type="fig"}). These processes were informed by experimental studies on the ecology and population dynamics of reef fishes (Bellwood *et al.* [@b2]), especially those related to the responses of populations to changes in environmental parameters (Fig. [1](#fig01){ref-type