Managing diverse and complex systems, such as coral reefs, is increasingly difficult due to unprecedented disturbance regimes associated with climate heating and other human pressures. Predictive models that capture the complexity of processes occurring on coral reefs across multiple spatial scales can help to unlock new information about reef ecology, generate projections of reef futures, and evaluate the ecological benefits, risks and uncertainties of different management strategies.

In our recent paper we present C~scape, a new modelling framework to project coral dynamics at scales ranging from sites within reefs to clusters of reefs across several 10s of kilometres. This modelling framework arose from efforts to identify the major mechanisms of change on coral reefs and the processes that influence coral distribution and resilience. We considered everything from reef geomorphology to water temperature and circulation, to acute disturbances, community composition, and genetics. By conceptually mapping these processes, we highlighted gaps in existing models of coral reefs; specifically, the need for a model structured around the coral life cycle using robust demographic data, that could account for the considerable variation in conditions within reefs (over scales of 100s of metres to kilometres) and the potential for coral communities to be linked over scales of tens of kilometres among reefs via larval dispersal and recruitment. 

Science and management of coral reefs in the 21st century faces novel questions and challenges. Working with The Reef Restoration and Adaptation Program we also considered what would be needed to quantitatively assess proposed intervention and restoration actions. We then set about bringing together a diverse array of data and expertise to build a new quantitative framework for exploring coral reef dynamics.

SPATIAL SCALE

Coral reef scientists often survey and model particular sites or habitats on a reef (due to time and cost restraints), and then extrapolate across the reef or reef system, yet there is significant variation within reefs. For example, differences in coral taxa, growth forms and abundance, driven by variation in environmental conditions, are clearly evident between different reef habitats, such as the reef crest, slope or lagoon, but similar variation also occurs among sites within these habitats. Some areas may be more suitable for certain types of corals, some may be more or less exposed to disturbances, and some may be particularly important for generating coral larvae. 

Emerging technology and improved satellite mapping products now allow us to spatially quantify some of the variation within and among reefs. We borrowed principles from landscape ecology as a way to partition reefs into a spatial mosaic of relatively homogenous ‘sites’ that became the base spatial units for our coral communities in the modelling efforts. Each site was characterised using additional spatial information and modelling that targeted specific processes. For example, we used benthic habitat maps to determine coral habitat suitability, bathymetry to assign median depth (given that depth influences the exposure to disturbances like marine heatwaves), and hydrodynamic modelling to determine larvae connectivity and recruitment among sites.

DEMOGRAPHY

With sites defined, we needed to simulate coral population changes over time. We first identified the key transitions in the coral life cycle, from the production of a microscopic egg, fertilisation into a larva, settlement onto the seafloor into a tiny coral that might grow to an adult over several years and then produce hundreds of thousands of eggs each year. We developed demographic population models (Integral Projection Models, IPMs) using empirical data on growth, survival, and fecundity from thousands of coral colonies. The IPMs are the engine of our C~scape modelling framework, describing the rate of coral population growth. 

OUTPUT

In our paper, we present the full methodology for the C~scape framework. We hindcast coral cover dynamics across a cluster of five reefs on the Great Barrier Reef and assessed how well the model reproduced AIMS Long-Term Monitoring data over the past decade, finding that the model performed well when parameterised with the satellite derived habitat maps. This is the first spatially-explicit, mechanistic model for coral reefs, to our knowledge, capable of reproducing observed reef dynamics at scales ranging from hundreds of metres to tens of kilometres, across decades. 

APPLICATIONS AND FUTURE PLANS

Developing meaningful models for a future that will be characterised by both unprecedented disturbance regimes and unprecedented management and intervention strategies is challenging, but is needed now more than ever. We need tools to quantitatively assess and communicate the likely future for coral reefs under different scenarios of greenhouse gas emissions reductions as well as across a wide range of traditional and novel management actions. The C~scape framework, synthesising processes spanning from individual organisms to populations, communities and metacommunities in a spatially explicit manner, marks a significant leap in our ability to explore coral reef dynamics at scales relevant to management and restoration in the 21st century.

There are a many questions for which the model will be applied, such as investigating the relative importance and uncertainty surrounding the other processes contributing to within-reef variability — larval connectivity, depth, wave exposure, spatial heterogeneity in acute stressors and coral demography.

In response to the overwhelming threat posed by climate change, innovative interventions focussing on reducing acute heat stress, increasing bleaching tolerance or aiding recovery through outplanting natural or genetically adapted corals are being investigated. Logistical and economic constraints limit the area that can be targeted with such interventions, so identifying deployment locations that might generate the highest system-level benefit is crucial. As part of The Reef Restoration and Adaptation Program we will investigate how the choice of restoration site location influences reef-wide benefits.  

C~scape models every coral colony on the reef, and the life stages that may be targeted by interventions (e.g. coral larvae or juveniles) so there is much to explore regarding coral abundance, density and size structure across different coral types and different reef habitats. Billions of coral larvae may be produced naturally on a healthy reef, how does this compare to what may be possible to produce as part of a restoration action? These are the types of questions we can investigate with the C~scape framework.

The model has already advanced significantly from the version published in this study: we have added more coral types, expanding from two to six (through collaboration with the EcoRRAP program), and we now include variation in coral growth and survival as a function of variables like wave exposure and depth. We have also been developing the capacity to model coral adaptation to thermal stress, and have started to predict future reef dynamics as well as hindcasts. 

~ ~ ~ ~ ~

This study was funded by the Reef Restoration and Adaptation Program, a partnership between the Australian Government’s Reef Trust and the Great Barrier Reef Foundation.

~ ~ ~ ~ ~

We acknowledge the Wulgurukaba, Bindal, Whadjuk Noongar, Turrbal and Jagera Peoples of the land on which the authors conducted the model development, analysis and writing for this study. We acknowledge the Gooreng Gooreng, Gurang, Bailai, Taribelang Bunda, Wunambal Gambera and Gunggandji peoples as the Traditional Owners of the Sea Country where the empirical data for this study were sourced.