Mapping the susceptibility of reefs to rubble accumulation across the Great Barrier Reef

doi:10.21203/rs.3.rs-3329362/v1

Download PDF

Research Article

Mapping the susceptibility of reefs to rubble accumulation across the Great Barrier Reef

https://doi.org/10.21203/rs.3.rs-3329362/v1

This work is licensed under a CC BY 4.0 License

Version 1

posted

You are reading this latest preprint version

Disturbance-induced rubble accumulations are described as “killing fields” on coral reefs as coral recruits suffer high post-settlement mortality, creating a bottleneck for reef recovery. The increasing frequency of coral bleaching events, that can generate rubble once coral dies, has heightened concerns that rubble beds will become more widespread and persistent. But we currently lack the tools to predict where rubble is most likely to accumulate. Here, we developed a modelling framework to identify areas that are likely to accumulate rubble across the Great Barrier Reef. The algorithm uses new high-resolution bathymetric and geomorphic datasets from satellite remote sensing. We found that 47 km of reef slope (3% of the entire reef), primarily in the southern region, could potentially reach 50% rubble cover. Despite being statistically significant (p < 0.001), the effects of depth and aspect on rubble cover were minimal, with a 0.2% difference in rubble cover between deeper and shallower regions, as well as a maximum difference of 0.8% among slopes facing various directions. Therefore, we conclude that the effects of depth and aspect were insufficient to influence ecological processes such as larval recruitment and recovery in different coral communities. Maps of potential rubble accumulation can be used to prioritise surveys and potential restoration, particularly after major disturbances have occurred.

Coral rubble

Coral restoration

Rubble accumulation

Reef recovery

Coral reefs worldwide are subject to increasingly frequent and severe damage from disturbances, including coral bleaching, cyclones, and anthropogenic activities, often creating large amounts of rubble (Done 1992; Fabricius et al. 2008; Fox et al. 2003; Harmelin-Vivien 1994; Riegl 2001). The term “rubble” refers to fragments of coral skeleton or reef rock derived from mechanical or chemical abrasion that have distinct morphology (branching, plate, and boulder) and size ranges (from 2 mm to 1 m) depending on the assemblage composition of the reef and the characteristics of disturbances (Rasser and Riegl 2002; Woodley et al. 1981) (Fig. 1). In addition to direct damage, indirect effects such as weakened coral skeleton and increased fragmentation after mortality can also contribute to rubble generation (Ceccarelli et al. 2020). Rubble formation is considered a part of the natural destruction and regeneration cycles that shape the diversity of the reef (Rasser and Riegl 2002). However, as climate change intensifies disturbances (McWhorter et al 2022), rubble formation is expected to accelerate and increase the prevalence of rubble on reefs (Cheal et al. 2017; Hoegh-Guldberg et al. 2007).

Persistent rubble fields can reduce reef structural complexity and inhibit reef recovery, raising concerns that future disturbance rates may surpass recovery rates (Ceccarelli et al. 2020; Fox et al. 2003; Knutson et al. 2020; Ortiz et al. 2018; Viehman et al. 2018). The “flattening” of reef structure has been documented in different regions as reef ecosystems degrade (Alvarez-Filip et al. 2009; Elliott et al. 2018). The loss of structural complexity can lead to significant ecological and socio-economic impacts, including a reduction in species diversity, fisheries productivity, and coastal protection offered by the reef, affecting the livelihoods of millions of people (Graham et al. 2006; Harris et al. 2018; Rogers et al. 2014). In addition to flattening the reef, rubble fields can also impede recovery processes (Kenyon et al. 2020; Rasser and Riegl 2002; Woodley et al. 1981). Coral recovery in rubble fields is largely determined by the stability of the substrate, which is essential for the survival of coral larvae (Yadav et al. 2015). Stabilised rubble fields are favourable for coral settlement owing to the abundance of cryptic microhabitats, which help reduce predation and competition for space with macroalgae (Biggs 2013; Edmunds et al. 2004; Doropoulos et al 2016). Coral recolonisation can take place in as little as ten months when the rubble field is consolidated through binding processes mediated by sponges and encrusting taxa (Wulff 1984). However, the stabilisation process can be interrupted in high-energy environments that promote rubble movement, leading to mechanical abrasion and smothering of newly settled coral recruits, as well as causing damage to adjacent coral colonies (Brown and Dunne, 2009; Clark and Edwards 2013; Viehman et al. 2018). A case study on the upper Florida Keys shelf observed a 52% loss of juveniles in a rubble field, as opposed to 10–40% at a nearby consolidated reef site, despite having similar densities of coral recruits and species composition (Cameron et al. 2016). Viehman et al. (2018) found that continued mobilisation of rubble in high-energy environments can result in a positive feedback loop in which rubble becomes smaller due to frequent erosion and mobilises at lower levels of hydrodynamic forcing. Thus, coral recovery on persistent rubble is therefore unlikely and may take decades even in the absence of further disturbances (Dollar and Tribble 1993; Riegl 2001). Indeed, several studies have documented little to no recovery in rubble-dominated sites years after previous disturbances (Chong-Seng et al. 2014; Fox and Caldwell 2006; Victor 2008). As a result, there is growing concern that future disturbance rates may exceed the capacity of coral reefs to recover, particularly when extensive areas are dominated by rubble fields, where windows for recovery need to be great (Cheal et al. 2017).

The mechanisms of rubble accumulation on reef slopes, where rubble is likely to become a persistent bottleneck for reef recovery, are heavily influenced by the area’s bathymetric properties including its slope and profile (Dollar and Tribble 1993; Harmelin-Vivien and Laboute 1986; Highsmith et al. 1980; Rasser and Riegl 2002; Shannon et al. 2012; Thornborough 2012). Although the majority of disturbance-generated rubble accumulates as talus at the foot of forereef slopes, some may stay as gently sloping ramparts or remain in depressions, potentially interfering with reef recovery (Scoffin 1993). A study in Hawaii found that rubble fragments transported and deposited on the reef slope remain unconsolidated after 20 years (Dollar and Tribble 1993). Rubble can also accumulate and cover large areas on reef flats, but will often undergo erosion into smaller pieces and be transported shoreward to form sand cays and reef islands in lagoons (Hughes 1999; Thornborough and Davies 2011). On steeper reef slopes (> 45°), "avalanches", which refers to the offshore and downslope movement of rubble, can take place and destroy reef communities in their path (Harmelin-Vivien & Laboute 1986; Scoffin 1993). However, on gentler parts of the slope, rubble will likely accumulate or be transported shoreward depending on the hydrodynamic properties of the environment (Hughes 1999). The profile of the reef slope dictates where rubble accumulates as rubble preferentially deposits in groove-like structures, which can act as channels to facilitate wave propagation and allow rubble transportation onto these specific zones (Shannon et al. 2012). Hydrodynamic properties and disturbance regimes, despite also influencing the distribution of rubble fields at a smaller scale, are more strongly tied to rubble formation and stabilisation processes than rubble accumulation (Harris and Vila-Concejo 2013; Rasser and Riegl 2002; Viehman et al. 2018). Hence, slope and profile are the key drivers in rubble accumulation, and a detailed spatial analysis should help identify vulnerable areas on a reef. Based on these key drivers, this study assumes that relatively flat and groove-like locations are vulnerable to rubble accumulation which may then impair reef recovery. Whether accumulated rubble is able to stabilise will depend on several factors, including agitation from wave-related energy, and is beyond the scope of this study. Our key focus is to identify areas where rubble might collect.

Here, we present a simple modelling approach to identify reefs that are likely to be susceptible to rubble accumulation once large areas of coral have died. The method uses newly-available bathymetric and geomorphic datasets across mid- and outer-shelf reefs of the Great Barrier Reef (GBR), which have sufficient water clarity for optical remote sensing. By analysing reef bathymetric profiles to extract rubble-prone areas, this study aims to: 1) estimate the extent of reef slope that is susceptible to rubble accumulation; 2) identify reefs that are at the highest risk of rubble accumulation as well as their spatial distribution; and 3) examine the effects of aspect and depth on rubble accumulation. We point out upfront that our mechanistic understanding of rubble retention is not sufficiently well developed to estimate the absolute amount of rubble accumulation. Thus, our goal is to estimate the relative vulnerability of reefs to rubble accumulation based on morphological features. We hypothesise that the potential rubble cover will vary with depth and aspect as hydrodynamic regimes largely shape reef geomorphology (Hopley et al. 2007; Montaggioni et al. 2009).

2.1 Overview

The model was developed in ArcGIS Pro 2.8.0 (ESRI 2022), MATLAB 9.11 (MATLAB 2021), and R 4.2.1 (R Core Team 2021), to characterise the susceptibility of coral reefs to rubble accumulation (Fig. S1). A one-dimensional algorithm was developed to evaluate the risk of rubble accumulation, expressed as the potential rubble cover, at a transect level. Bathymetric profiles were extracted and analysed from transects sampled throughout the Great Barrier Reef Marine Park (GBRMP) to determine the extent of rubble accumulation. Each transect was orientated roughly perpendicular to the reef, extending from the shallows towards deeper areas (Fig. 2). We then calculated reef-scale metrics, including the percentage of transects in which rubble was predicted to exceed defined thresholds. To create a GBR-scale metric, we examined the number of transects meeting a particular rubble criterion, and expressed their number as the linear distance, in kilometres, along the reef axis (i.e., perpendicular to the transects and at constant depth). Given that several of the algorithm’s components are uncertain, we explored the sensitivity of our conclusions to model assumptions.

2.2 Site description and sources of data

The study area is approximately 344,000 km², spanning 14 degrees of latitude from 10° 20′ S in the north to 24° 30′ S in the south (Brodie and Waterhouse 2012). The forereef slopes of 1834 offshore shallow-water reefs were sampled across the GBRMP for modelling rubble accumulation. Nearshore reefs were excluded due to the limitations in their visibility and the coverage of the available data sources. Following the removal of data inconsistencies (see section 2.3), 1706 reefs were ultimately chosen for susceptibility mapping. Reef slopes, both exposed and sheltered, were isolated from individual reefs and selected as areas of interest due to their higher potential for rubble accumulation (Dollar and Tribble 1993; Highsmith et al. 1980; Rasser and Riegl 2002).

Publicly available, open-source data layers depicting the reef boundaries, bathymetry, and geomorphic zonation in the GBRMP (Table 1) were analysed using ArcGIS Pro Version 2.8.0 (ESRI 2022). The indicative boundaries of reefs in GBRMP were extracted from the “GBR Features” polygon layer acquired from GBRMPA Geoportal (GBRMPA 2020). Islands, cays, rocks and dry reefs were excluded from the layer during the extraction. The layer was used to group results by reef boundaries in statistical analyses and data visualisation. The bathymetry data layer was derived from optical Sentinel-2 satellite imagery at 10 m resolution, with a vertical accuracy of ± 1 m and a maximum depth of 25 m corrected to mean sea level (Hedley et al. 2018; Roelfsema et al. 2021). The source bathymetry data were divided into four parts based on management areas (GBRMPA 2021). The four parts were mosaicked and collated into a common geographic coordinate system to ensure data integrity and continuity during transect sampling and rubble modelling. The geomorphic zonation layer was mapped by the Remote Sensing Research Centre using a machine learning approach which combines satellite imagery, environmental attributes, and occurrences of geomorphic zonation (Roelfsema et al. 2021). According to Roelfsema et al. (2021), the data were mapped at 10 m resolution for depths ranging from 0 to 20 m (corrected to mean sea level) and has an overall accuracy of 68%. The layer was used to identify and isolate the reef slopes in preparation for transect sampling.

While rubble is extensively surveyed as a substrate type throughout the GBRMP, rubble cover is often under-reported, and there are no baseline data for rubble parameters, such as size, volume, and density of accumulations (Ceccarelli et al. 2020). Most of the literature focuses on rubble formation while the consequences of rubble accumulation remain relatively understudied (Kenyon et al. 2020; Rasser and Riegl 2002; Thornborough 2012).

Table 1

Data sources for the spatial analysis.
Data layer name	Description	Source
GBR Reef	Vector layer showing indicative boundaries of shallow water reefs in GBRMP.	GBRMPA (2020)
GBR10 Bathymetry	Raster layer combining four parts of the bathymetry data of different management areas and projected onto the same geographic coordinate system, WGS 1984.	GBRMPA (2021b)
GBR10 Geomorphic	Raster layer showing the geomorphic zones in GBRMP.	(GBRMPA, 2021a)

2.3 Transect sampling

Transects (n = 32,603) were generated along reef slopes at a separation distance of 500 m and aligned perpendicular to the reef axis to capture the variations in depth down the slope. Reef slopes isolated from the GBR10 Geomorphic raster layer were converted into polygons (Conversion toolbox) and simplified (Cartography toolbox) to remove potential errors during transect generation. The conversion process was necessary for the calculation of the centrelines for generating transects. To reduce data noise while maintaining data integrity, all polygons smaller than 0.01 km² were removed after the polygons were simplified using the "retain critical bends" algorithm with a simplification tolerance of 500 m. The simplification process did not alter the values of the bathymetry layer but eliminated inconsistencies along the boundaries of the reef slopes. Centrelines were generated using the Topographic Production toolbox within the polygons to indicate the line along which perpendicular transects can be created. Some centrelines may not exactly reflect the alignment of the slope when the reef is irregularly shaped. These centrelines were manually removed to minimise inconsistencies during transect generation. Transects with a length of 500 m were generated along the centrelines using the Data Management toolbox to ensure they cross the entire reef slope. Both ends of the transects were truncated at a depth limit of -2 m to avoid sampling reef crests and reef flats (where the minus sign indicates below sea surface). The depth limit was selected based on the mean depth values for reef crests (M = − 1.32 m, SD = 0.5 m) and outer reef flats (M = − 1.49 m, SD = 0.87 m). The truncated transects were grouped by individual reefs on the GBR reef layer using Spatial Join in the Analysis toolbox. Transects were allocated to their closest reefs within a 1000 m searching distance radius. This step ensures that all transects were allocated correctly and accounts for any discrepancies between the GBR10 Bathymetry and GBR Reef layers. The average length of generated transects was 287.2 m (SD = 173.8).

To account for transects facing different directions, their mean aspect was calculated using a series of ArcGIS Pro tools, with reference to the python script tool by Beyerhelm (2013). The Aspect tool (Spatial Analyst) was used to generate the aspect from each cell of the bathymetry layer. Using the python script tool, a buffer of 10 m was created around each transect to calculate the zonal mean aspect. The following formula shows the fundamental calculation in the python script tool:

$$\left[360+ \text{atan2(s, c)}\left(\frac{\text{180}}{\text{π}}\right)\right]\text{mod360}$$

where s and c are the zonal mean values of the sine and cosine of the aspect, respectively. atan2 represents the two-argument arctangent function. The result is an angle in unit degrees that represents the average aspect of each transect and was classified into the four directions of North, East, South, and West based on its value.

Depth values were extracted from the transects and compiled into a data table for further analysis in the rubble accumulation algorithm. The extraction of bathymetric profiles required the Stack Profile tool in the 3D Analyst toolbox, which allows bilinear interpolation of depth values on the GBR Bathymetry layer for the grouped transects. The output of the tool was a table denoting the bathymetric profiles of the transects, with depth values recorded at each point on the transect spaced by 10 metres. The output was later cleaned and prepared in R version 4.2.1 for further analysis (R Core Team 2021). Duplicated records and missing values were also removed to maintain data consistency. After data cleaning, 22,562 transects which accounted for 1706 reefs were left for subsequent analyses.

2.4 Rubble accumulation algorithm

The algorithm was designed to locate rubble-prone areas on transects and calculate the percentage of potential rubble cover in different depth regions. Sections of transects were classified as either “shallow” or “deep”, based on their depths. "Shallow" sections are less than 10 m in depth; otherwise, they are considered "deep". The algorithm scans and evaluates each transect point to determine whether rubble is likely to accumulate, based on four input parameters – Moving window, Rubble height, Depth range within window that causes rubble to roll, and Depth change within window that is still considered flat (Fig. 3). A detailed description of each model parameter is given in Table 2. A base case of parameter values was determined using the distribution of transect lengths and critical slopes, in addition to expert opinion. The base case represents the most realistic scenario and acts as a reference point for the sensitivity analysis.

The ability of rubble to accumulate is determined within a floating horizontal window that passes along each transect. The width of the window is typically 5 pixels (Table 2). Rubble will accumulate in local depressions that exceed a depth of 0.5 m and on ‘flat’ sections of reef, defined as have a variation in depth of ≤ 0.5m. There are some exceptions to this. For example, rubble cannot accumulate on a local topographic high or if the ‘flat area’ is proximal to a steep slope, where rubble is likely to roll away. Proximity and slope are defined by the variation in depth ≥ 5 m within the evaluation window (Table 2). The parameterisation is somewhat subjective and based on two considerations. First, personal observations of one of the authors (PJM) studying the accumulation of rubble on reefs after cyclone impacts in Belize (Mumby 1999), Moorea (Mumby et al. 2016), and Palau (Roff et al. 2015). In several cases rubble accumulated for at least a decade, yet it rolled off reef structures that exceeded 5 m in height from surrounding areas. Second, the use of satellite-derived bathymetry, while being the only means of obtaining continuous data across the GBR, does have limited sensitivity. Despite the relatively high radiometric resolution of the Sentinel II imagery, bathymetry algorithms cannot resolve small-scale changes in depth, such as 0.1 m (Green et al. 2000). Therefore, we felt it appropriate to specify minimal depth differences of 0.5 m, which are resolvable in the data. We note that rubble accumulations of < 0.5 m are able to inhibit coral recruitment (Mumby, pers. obs.), but such levels could not be specified accurately in a GBR wide algorithm.

Table 2. Input parameters for the rubble accumulation model.

Parameter name	Description	Base case value	Unit
Moving window	Spatial scale over which reef slope is evaluated around a focal point. Helps determine whether rubble will naturally roll down to a deeper location	5 pixels = 50 m	Metres
Rubble height	Minimum thickness of accumulated rubble that causes problems for coral recovery. Indentations on the reef surface that are shallower than this may accumulate rubble but are not considered problematic	0.5	Metres
Depth range within window that causes rubble to roll	The magnitude of change in depth within the moving window that decides if the point is on or close to a steep slope where rubble may roll off. Rubble cannot accumulate on the focal point if the depth change within the moving window is greater than this	5	Metres
Depth change within window that is still considered flat	The maximum permitted value of the change in depth within the moving window that would allow the rubble to remain on the focal point	0.5	Metres

2.5 Calculating reef susceptibility

We defined susceptibility as the percentage of transects per reef having a potential rubble cover higher than a given threshold. This approach highlights problematic areas within reefs that have a high potential of rubble accumulation. Since there is limited literature on the critical value of rubble cover at which the recovery of the reef would be impaired, we chose values in the range observed in the field, following Typhoon Bopha in Palau, where accumulated rubble represented up to 50% of the reef. We therefore used thresholds of 30%, 40%, and 50% cover (though a full range of 10–90% are shown for some results). The distribution of percentages was plotted using the “ggpubr” package in R to determine suitable thresholds for visualising the spatial variability of susceptibility to rubble in ArcGIS Pro (Kassambara 2020). The number of transects exceeding the thresholds was multiplied by the interval between transects to estimate the extent of reefs (in kilometres) that are thought to be susceptible. For example, if there were ten susceptible transects in a reef and the transects were placed 500 m apart, the linear distance would be 5 km.

A depth-stratified analysis was performed to compare the susceptibility of reef slopes facing various directions in different depth regions using a beta regression model in R 4.2.1. The R packages “glmmTMB”, “DHARMa”, and “betareg” were used for the analysis (Grün et al. 2012; Hartig 2022; Mollie et al. 2017). A beta regression model was used to test the fixed effects of depth and aspect on the proportion of potential rubble cover of the transects. The proportion data were linearly transformed and compressed to avoid zeros and ones using the methods discussed by Smithson and Verkuilen (2006). A two-component mixture model was used to fit the data due to the bimodal distribution of residuals in the data when fitting a generalised linear mixed model. The mixture model assumes that the data can split into two groups that fit two different beta distributions to reflects its bimodality.

2.6 Sensitivity Analysis

The one-at-a-time method was utilised to assess the sensitivity of the rubble accumulation algorithm to the four input parameters (Hamby 1994). Parameters were changed by ± 20% relative to the base case in each run, yielding three scenarios for each parameter. The analysis evaluated two output variables: the susceptible reef count and the reef’s susceptibility ranking. The sensitivity of the susceptible reef count was expressed as the mean percentage change in the number of reefs containing transects exceeding the critical thresholds relative to the change in input parameter values. For the sensitivity of reef ranking, the top 10% of highly susceptible reefs were chosen and ranked based on their percentage of transects with potential rubble cover over critical thresholds. The greater the percentage, the higher the ranking of the reef. This variable assesses the algorithm's consistency in the selection of highly susceptible reefs when parameter values are varied. The absolute change in reef ranking of the top 10% of reefs against the ± 20% change in parameter values was averaged for every threshold category to compute the overall sensitivity of the model’s decision. The parameters were then ranked based on sensitivity analysis results to determine which parameters contribute the most to the model’s variability.

3.1 Linear distances and distributions of rubble accumulation

Approximately one quarter of reefs (404/1706, 23.7%) were predicted to have no rubble accumulation potential. As the threshold of rubble accumulation per transect was increased from 30–50% there was a sharp reduction in the number of reefs affected and the total length of problematic reef (Fig. 4). For example, if we consider that ≥ 30% of a forereef profile (transect) gathering rubble is problematic, then 336 reefs contain problematic transects. If we set a higher threshold of 50% rubble cover, then the number of reefs drops to 46 (Fig. 4), and their spatial distribution is more limited. The latitudinal range of susceptible reef under the 30% rubble cover threshold category spans 13 degrees, from 10° 40′ S to 23° 40′ S. In the case of the 50% threshold, the range is more restricted, covering only 6 degrees from 17° 10′ S to 23° 40′ S, with the exception of one outlier reef located at 11° 30′ S. Reefs with the higher percentages of problematic transects mostly occurred in the central to the southern GBR, including the Townsville/Whitsunday and Mackay/Capricorn management area, regardless of the threshold category (Fig. 5).

There were two notable observations regarding the positively skewed distributions of rubble accumulation data grouped by individual reefs, depth, and aspect. First, few reefs have a high percentage of transects with rubble exceeding the threshold (Fig. 6). With the exception of the lowest threshold category (10%), the median percentage of transects per reef with rubble greater than the threshold is zero for all threshold categories (i.e., the proportion of the average reef with problematic rubble typically falls in the lower 50th percentile). The median percentage of transects per reef with minor rubble (10% threshold) was 11.1%, and the interquartile range was 33.3%. Secondly, when assessing potential rubble cover on transects stratified by depth and aspect, the majority of transects had a low potential rubble cover, with only a few extremities completely covered by rubble regardless of depth and aspect (Fig. 7). The distributions of percentage rubble cover on transects of various depths and aspects had a median of 0% potential rubble cover and similar ranges. All boxplots in Fig. 7 had an upper quartile value that was less than 10%, indicating that at least 75% of the transects had less than 10% potential rubble cover no matter the depth or aspect. These observations suggest that the risk of rubble accumulation is both patchy and localised, as the overall percentage of reef areas containing transects with high rubble cover is remarkably low.

3.2 Effects of aspect and depth on rubble accumulation

The relationship between depth, aspect, and the potential rubble cover was determined by the regression group to which the transect was assigned (Table 3). The regression model assigned 7,703 transects to the first group while 31,919 transects were assigned to the second group based on the values of posterior probabilities. No significant interaction effect was found between the independent variables in either of the groups (p = 0.257 for group 1 and p = 0.805 for group 2). There is moderate evidence showing that deeper regions (> 10 m) in the first group have a potential for rubble accumulation that is 39% higher than that of shallower regions (deep = 74 ± 16% rubble cover; shallow = 35 ± 16% rubble cover; z = 1.95). The level of statistical significance is marginal (p = 0.051). In the second group, the mean percentage cover of rubble on transects in deeper regions was significantly higher than in shallower regions by 0.2% (deep = 5.2 ± 0.051% rubble cover; shallow = 5.0 ± 0.072% rubble cover; z = 3.37; p < 0.001). In comparison to those facing north (5.0 ± 0.07% rubble cover), transects facing east or south have a significantly lower rubble potential (4.5 ± 0.062% and 4.4 ± 0.058% rubble cover; z = − 8.60 and − 10.55; both p < 0.001), while there is moderate evidence suggesting that reefs facing west have the highest risk of rubble (5.2 ± 0.075% rubble cover; z = 2.29; p = 0.022). Although statistical significance was observed in both depth and aspect results, the effect sizes are generally small in the second regression group.

Table 3

Significant results of the 2-component beta regression mixture model for each variable and their levels. (Significance codes: *0.05, **0.01, ***0.000)
Component	Variable	z value	p value
1	Depth - Deep	1.95	0.051*
2	Depth - Deep	3.37	0.001**
	Aspect - East	-8.60	0.000***
	Aspect - South	-10.55	0.000***
	Aspect - West	2.29	0.022*

3.3 Model sensitivity to input parameters

The algorithm was most sensitive to value changes in the Moving window, followed by parameters Depth change within window that is still considered flat, Rubble height, and finally Depth range within window that causes rubble to roll (Fig. 8). A 20% change in moving window resulted in an average change of 58.2 ± 14.4% in the reef count for containing problematic transects and an average absolute change of 14 ± 0.49 in the ranking of the top 10% reefs across threshold categories. In order to provide some perspective, 336 reefs had transects with a rubble cover exceeding 30% in the base case scenario; however, when the moving window value was increased by 20%, the reef count decreased to 235. Briggs Reef (No. 1) was initially ranked the highest for having 75% of its transects exceed the 40% rubble cover threshold in the base case scenario, but when the moving window value was increased by 20%, only 25% of its transects exceeded the threshold, causing its ranking to drop to fifth place. The reef ranking could fluctuate by a maximum of 17.5% within the range of ranks for all parameters. This implied that few reefs fall outside of the top 10% and that even when the parameters were altered, the algorithm would still place the same reefs as the most vulnerable to rubble accumulation. As shown in Fig. 8, the sensitivity of the model output varied greatly depending on the threshold category. For example, when the threshold category is 50%, a ± 20% change in the value of the parameter, moving window, resulted in a 137% change in the number of reefs that are considered problematic. On the other hand, when the threshold category is 90%, the same changes in the parameter value does not affect the number of problematic reefs at all. Although the change in parameter values could result in variable potential rubble cover, the algorithm was consistent in assigning the risk of rubble accumulation by selecting the same top 10% reefs.

Approximately one quarter of the GBR appears to be unlikely to develop persistent rubble problems whereas 20% of reefs have the potential to accumulate ~ 30% of rubble across their surface. These reefs can be found throughout the GBR, spanning 13 degrees of latitude. However, rubble-prone reefs are concentrated in the central and southern GBR, which may be attributed to the considerable variation in reef type and geomorphology by latitude (Hopley et al. 2007). The northern region of the GBR has a high density of narrower reefs with steeper reef slopes (Hopley et al 2007) where rubble may roll off instead of accumulating. Shelf-edge ribbon reefs with steeper slope profiles are limited to the northern region. As the continental shelf gradually widens south of Cairns, reefs are set back from the shelf edge and have gentler slopes, allowing rubble to persist (Hopley et al. 2007).

We found significant effects of aspect in the second group with 31,919 transects, which is consistent with hydrodynamic forcings influence reef geomorphology (Hopley et al. 2007; Montaggioni et al. 2009; Scoffin 1993). Windward, exposed reef slopes are generally steeper, resulting in rubble accumulating as talus at the slope's foot in deeper waters beyond the depth limits of this study (Montaggioni et al. 2009). Reef slopes facing north, or west are typically shielded from the south-easterly trade winds-induced currents that affect the GBR for nine months of the year, such that they tend to have gentler profiles and accumulate more rubble.

Spur-and-groove systems on reef slopes with high wave exposure, which have been frequently linked to significant rubble accumulation (Duce et al. 2014; Kan et al. 1997; Munk and Sargent 1954), cannot be differentiated in our analysis due to resolution constraints in the bathymetry and geomorphic data layers. Despite some mapping efforts at the Capricorn Bunker Group (Duce et al. 2014), there is currently insufficient high-resolution data on the spatial distribution of spur and groove structures at a GBR scale. As a result, rubble found in grooves may remain undetected in our analysis, potentially underestimating the reef's susceptibility to rubble accumulation. Future work could incorporate spur and groove information to enhance the model’s accuracy once it becomes available, allowing researchers to better evaluate the effects of various factors on rubble accumulation.

Despite the uncertainty in current knowledge with which to improve the parameterisation, the approach still serves as a first step in evaluating the vulnerability of the GBR to rubble-induced retardation of coral recovery. The one-dimensional transects were a computationally efficient strategy for sampling reef bathymetric profiles on a large scale, which would otherwise demand a substantial amount of resource. Although the model was moderately sensitive to several input parameters – especially the size of the window used to estimate slope – we found that the ranking of reef susceptibility was pretty robust, with most of the more vulnerable reefs remaining within the top 10%. This is important because it provides grounds to embark on preliminary risk analysis for the GBR and help target large scale validation surveys. In future, the accumulation model can be combined with geospatial data on rubble generation risks and mobilisation rates to develop a comprehensive prediction for reef recovery on a large scale.

One of the greatest limitations of the study is the lack of information on what constitutes a critical cover of rubble to impede coral recovery. Since no study has ever experimentally quantified the critical thresholds for reef recovery, they were chosen based on observed levels on reefs that have exhibited little recovery over 10 years (Mumby, pers. obs.). Although aspect has significant effects on rubble cover in one of the regression groups, it is doubtful whether the magnitude of this effect is sufficient to influence ecological processes. Thus, we conclude that aspect is unlikely to be an important driver of the ability of reefs to accumulate rubble, though it might be important when considering the forces that prevent rubble from stabilising. Indeed, factors including hydrodynamic properties and disturbance regimes have an intricate relationship with rubble dynamics (Harris and Vila-Concejo 2013; Rasser and Riegl 2002; Viehman et al. 2018). The study assumed that these factors contribute more to processes of rubble formation and stabilisation than accumulation.

Coral reefs are experiencing accelerating disturbance (Hughes et al 2018; McWhorter et al 2022), and while management can mitigate some sorts of damage, such as predation from crown-of-thorns starfish (Castro-Sanguino et al 2023), major events like heatwaves (‘coral bleaching’) and cyclones cannot be mitigated directly. Thus, management tends to focus on facilitating the process of coral recovery, whether through improving conditions for coral recruitment and survival (Gove et al 2023; Mumby et al 2021) or restoration (Doropoulos and Babcock 2018). Rubble beds pose a threat to such activities because they can lead to persistent failure of coral recruitment (Dollar and Tribble 1993; Fox et al 2003; Riegl 2001). While we cannot quantify that absolute risk of developing problematic rubble without additional data on rubble stabilisation rates, wave forcing, and probabilities of rubble formation, our analysis does suggest that up to 20% of GBR reefs have the sort of bathymetric profile that might trap rubble and allow substantial levels to accumulate. Some areas of the GBR appear to be particularly vulnerable to this issue, including the Pompey Complex and the Capricorn and Bunker Group in the southern GBR. By highlighting that such an important proportion of reefs have the potential to develop problematic rubble we hope that further research will allow risks to be refined such that managers and reef users can better target where rubble stabilization might be targeted in future.

Funding

The authors declare that no funds, grants, or other support were received during the preparation of this manuscript.

Competing Interests

The authors have no relevant financial or non-financial interests to disclose.

Author contributions

Peter Mumby contributed to the study conception and all authors contributed to the study design. Coding for the MATLAB rubble accumulation algorithm was performed by Peter Mumby. Material preparation, data collection and analysis were performed by Shu Kiu Leung. The first draft of the manuscript was written by Shu Kiu Leung and all authors commented on previous versions of the manuscript. All authors read and approved the final manuscript.

Data Availability

The datasets analysed during the current study can be found online and are publicly available (see section 2.2). The datasets generated during the current study are available from the corresponding author on reasonable request.

Alvarez-Filip L, Dulvy NK, Gill JA, Cote IM, Watkinson AR (2009) Flattening of Caribbean coral reefs: region-wide declines in architectural complexity. Proceedings of the Royal Society B 276(1669):3019-3025. https://doi.org/10.1098/rspb.2009.0339
Beeden R, Maynard J, Puotinen M, Marshall P, Dryden J, Goldberg J, Williams G (2015) Impacts and recovery from severe tropical cyclone Yasi on the Great Barrier Reef. PLOS One 10(4):e0121272. https://doi.org/10.1371/journal.pone.0121272
Beyerhelm C (2013) A Python Script Tool to Calculate Zonal Mean Aspect [Source code]. Coconino National Forest. https://community.esri.com/t5/arcgis-spatial-analyst-questions/calculate-mean-aspect-for-polygons/td-p/67951.
Biggs BC (2013) Harnessing natural recovery processes to improve restoration outcomes: an experimental assessment of sponge-mediated coral reef restoration. PLOS One 8(6): e64945. https://doi.org/10.1371/journal.pone.0064945
Brodie J, Waterhouse J (2012) A critical review of environmental management of the ‘not so Great’ Barrier Reef. Estuarine, Coastal and Shelf Science 104-105:1-22. https://doi.org/10.1016/j.ecss.2012.03.012
Brown BE, Dunne RP (2009) The Environmental Impact of Coral Mining on Coral Reefs in the Maldives. Environmental Conservation 15(2):159-165. https://doi.org/10.1017/s0376892900028976
Cameron CM, Pausch RE, Miller MW (2016) Coral recruitment dynamics and substrate mobility in a rubble-dominated back reef habitat. Bulletin of Marine Science 92(1):123-136. https://doi.org/10.5343/bms.2015.1030
Castro-Sanguino C, Bozec Y-M, Condie SA, Fletcher CS, Hock K, Roelfsema C, Westcott DA, Mumby PJ (2023) Control Efforts of Crown-of-Thorns Starfish Outbreaks to Limit Future Coral Decline across the Great Barrier Reef. Ecosphere 14(6):e4580. https://doi.org/10.1002/ecs2.4580
Ceccarelli DM, McLeod IM, Bostrom-Einarsson L, Bryan SE, Chartrand KM, Emslie MJ, Gibbs MT, Gonzalez Rivero M, Hein MY, Heyward A, Kenyon TM, Lewis BM, Mattocks N, Newlands M, Schlappy ML, Suggett DJ, Bay LK (2020) Substrate stabilisation and small structures in coral restoration: State of knowledge, and considerations for management and implementation. PLOS One 15(10):e0240846. https://doi.org/10.1371/journal.pone.0240846
Cheal AJ, MacNeil MA, Emslie MJ, Sweatman H (2017) The threat to coral reefs from more intense cyclones under climate change. Global Change Biology 23(4):1511-1524. https://doi.org/10.1111/gcb.13593
Chong-Seng KM, Graham NAJ, Pratchett MS (2014) Bottlenecks to coral recovery in the Seychelles. Coral Reefs 33(2):449-461. https://doi.org/10.1007/s00338-014-1137-2
Clark S, Edwards AJ (2013) Coral transplantation as an aid to reef rehabilitation: evaluation of a case study in the Maldive Islands. Coral Reefs 14(4):201-213. https://doi.org/10.1007/bf00334342
De'ath G, Fabricius KE, Sweatman H, Puotinen M (2012) The 27-year decline of coral cover on the Great Barrier Reef and its causes. Proceedings of the National Academy of Sciences USA 109(44):17995-17999. https://doi.org/10.1073/pnas.1208909109
Dollar SJ (1982) Wave stress and coral community structure in Hawaii. Coral Reefs 1(2):71-81. https://doi.org/10.1007/BF00301688
Dollar SJ, Tribble GW (1993) Recurrent storm disturbance and recovery: a long-term study of coral communities in Hawaii. Coral Reefs 12(3-4):223-233. https://doi.org/10.1007/bf00334481
Done TJ (1992) Effects of tropical cyclone waves on ecological and geomorphological structures on the Great Barrier Reef. Continental Shelf Research 12(7-8):859-872. https://doi.org/10.1016/0278-4343(92)90048-o
Doropoulos C, Babcock RC (2018) Harnessing connectivity to facilitate coral restoration. Frontiers in Ecology and the Environment 16(10):558-559. https://doi.org/10.1002/fee.1975
Doropoulos C, Roff G, Bozec Y-M, Zupan M, Werminghausen J, & Mumby PJ (2016) Characterizing the ecological trade-offs throughout the early ontogeny of coral recruitment. Ecological Monographs 86(1):20-44. https://doi.org/10.1890/15-0668.1
Duce S, Vila-Concejo A, Hamylton S, Bruce E, Webster JM (2014) Spur and groove distribution, morphology and relationship to relative wave exposure, Southern Great Barrier Reef, Australia. Journal of Coastal Research Special Issue 70:115-120. https://doi.org/10.2112/si70-020.1
Edmunds PJ, Bruno JF, Carlon DB (2004) Effects of depth and microhabitat on growth and survivorship of juvenile corals in the Florida Keys. Marine Ecology Progress Series 278:115-124. https://doi.org/10.3354/meps278115
Elliott JA, Patterson MR, Staub CG, Koonjul M, Elliott SM (2018) Decline in coral cover and flattening of the reefs around Mauritius (1998-2010). PeerJ 6:e6014. https://doi.org/10.7717/peerj.6014
ESRI. (2022). ArcGIS Pro (Version 2.8.0) [Computer software]. Retrieved from https://www.esri.com/en-us/arcgis/products/arcgis-pro/overview.
Fabricius KE, De'ath G, Puotinen ML, Done T, Cooper TF, Burgess SC (2008) Disturbance gradients on inshore and offshore coral reefs caused by a severe tropical cyclone. Limnology and Oceanography 53(2):690-704. https://doi.org/10.4319/lo.2008.53.2.0690
Fox HE, Caldwell RL (2006) Recovery from Blast Fishing on Coral Reefs: A Tale of Two Scales. Ecological Applications 16(5):1631-1635. https://doi.org/10.1890/1051-0761(2006)016[1631:Rfbfoc]2.0.Co;2
Fox HE, Pet JS, Dahuri R, Caldwell RL (2003) Recovery in rubble fields: long-term impacts of blast fishing. Marine Pollution Bulletin 46(8):1024-1031. https://doi.org/10.1016/s0025-326x(03)00246-7
[Dataset] GBRMPA. (2020). Great Barrier Reef Features [Feature layer]. https://eatlas.org.au/data/uuid/ac8e8e4f-fc0e-4a01-9c3d-f27e4a8fac3c
[Dataset] GBRMPA. (2021a). GBR10 GBRMP Geomorphic. [basemap]. https://www.arcgis.com/home/item.html?id=2f3ee0ffc3f64e9bbb15530735501907
[Dataset] GBRMPA. (2021b). GBRMP Bathymetry Map Data [basemap]. https://www.arcgis.com/home/item.html?id=f644f02ec646496eb5d31ad4f9d0fc64
Glynn PW (1976) Some Physical and Biological Determinants of Coral Community Structure in the Eastern Pacific. Ecological Monographs 46(4):431-456. https://doi.org/10.2307/1942565
Gove JM, Williams GJ, Lecky J, Brown E, Conklin E, Counsell C, Davis G, Donovan MK, Falinski K, Kramer L, Kozar K, Li N, Maynard JA, McCutcheon A, McKenna, SA, Neilson BJ, Safaie A, Teague C, Whittier R, Asner GP (2023) Coral reefs benefit from reduced land–sea impacts under ocean warming. Nature https://doi.org/10.1038/s41586-023-06394-w
Graham NAJ, Wilson SK, Jennings S, Polunin NVC, Bijoux JP, Robinson J (2006) Dynamic fragility of oceanic coral reef ecosystems. Proceedings of the National Academy of Sciences 103(22):8425-8429. https://doi.org/doi:10.1073/pnas.0600693103
Green EP, Mumby PJ, Edwards AJ, and Clark CD (2000) Remote sensing handbook for tropical coastal management. Coastal Management Sourcebooks 3, UNESCO, Paris.
Grün B, Kosmidis I, Zeileis A (2012). “Extended Beta Regression in R: Shaken, Stirred, Mixed, and Partitioned.” Journal of Statistical Software, 48(11):1-25. https://doi.org/10.18637/jss.v048.i11
Hamby DM (1994) A review of techniques for parameter sensitivity analysis of environmental models. Environmental Monitoring and Assessment 32(2):135-154. https://doi.org/10.1007/BF00547132
Harmelin-Vivien ML (1994) The Effects of Storms and Cyclones on Coral Reefs: A Review. Journal of Coastal Research 12:211-231.
Harmelin-Vivien ML, Laboute P (1986). Catastrophic impact of hurricanes on atoll outer reef slopes in the Tuamotu (French Polynesia). Coral Reefs 5(2):55-62. https://doi.org/10.1007/bf00270353
Harris DL, Rovere A, Casella E, Power H, Canavesio R, Collin A, Pomeroy A, Webster JM, Parravicini V (2018) Coral reef structural complexity provides important coastal protection from waves under rising sea levels. Science Advances 4(2):eaao4350. https://doi.org/10.1126/sciadv.aao4350
Harris DL, Vila-Concejo A (2013) Wave transformation on a coral reef rubble platform. Journal of Coastal Research, 65:506-510. https://doi.org/10.2112/si65-086.1
Hartig F (2022) DHARMa: Residual Diagnostics for Hierarchical (Multi-Level / Mixed) Regression Models. R package version 0.4.6, <https://CRAN.R-project.org/package=DHARMa>.
Hedley JD, Roelfsema C, Brando V, Giardino C, Kutser T, Phinn S, Mumby PJ, Barrilero O, Laporte J, Koetz B (2018) Coral reef applications of Sentinel-2: Coverage, characteristics, bathymetry and benthic mapping with comparison to Landsat 8. Remote Sensing of Environment, 216:598-614. https://doi.org/https://doi.org/10.1016/j.rse.2018.07.014
Highsmith RC, Riggs AC, D'Antonio CM (1980) Survival of hurricane-generated coral fragments and a disturbance model of reef calcification/growth rates. Oecologia 46(3): 322-329. https://doi.org/10.1007/BF00346259
Hoegh-Guldberg O, Mumby PJ, Hooten AJ, Steneck RS, Greenfield P, Gomez E, Harvell CD, Sale PF, Edwards AJ, Caldeira K, Knowlton N, Eakin CM, Iglesias-Prieto R, Muthiga N, Bradbury RH, Dubi A, Hatziolos ME (2007) Coral reefs under rapid climate change and ocean acidification. Science, 318(5857):1737-1742. https://doi.org/10.1126/science.1152509
Hopley D, Smithers SG, Parnell K (2007) The Geomorphology of the Great Barrier Reef: Development, Diversity and Change. Cambridge University Press. 92-165. https://doi.org/DOI: 10.1017/CBO9780511535543
Hughes TP (1999). Off-reef transport of coral fragments at Lizard Island, Australia. Marine Geology 157(1-2):1-6. https://doi.org/10.1016/s0025-3227(98)00187-x
Hughes TP, Kerry JT, Alvarez-Noriega M, Alvarez-Romero JG, Anderson KD, Baird AH, Babcock RC, Beger M, Bellwood DR, Berkelmans R, Bridge TC, Butler I R, Byrne M, Cantin NE, Comeau S, Connolly SR, Cumming GS, Dalton SJ, Diaz-Pulido G, Eakin CM, Figueira WF, Gilmour JP, Harrison HB, Heron S F, Hoey AS, Hobbs JA, Hoogenboom MO, Kennedy EV, Kuo CY, Lough JM, Lowe RJ, Liu G, McCulloch MT, Malcolm HA, McWilliam MJ, Pandolfi JM, Pears RJ, Pratchett MS, Schoepf V, Simpson T, Skirving WJ, Sommer B, Torda G, Wachenfeld DR, Willis BL, Wilson SK (2017) Global warming and recurrent mass bleaching of corals. Nature 543(7645):373-377. https://doi.org/10.1038/nature21707
Kan H, Hori N, Ichikawa K (1997) Formation of a coral reef-front spur. Coral Reefs 16(1):3-4. https://doi.org/10.1007/s003380050052
Kassambara A (2020). ggpubr: 'ggplot2' Based Publication Ready Plots. R package version 0.4.0, <https://CRAN.R-project.org/package=ggpubr>.
Kenyon TM, Doropoulos C, Dove S, Webb GE, Newman SP, Sim CWH, Arzan M, Mumby PJ (2020) The effects of rubble mobilisation on coral fragment survival, partial mortality and growth. Journal of Experimental Marine Biology and Ecology 533:151467. https://doi.org/10.1016/j.jembe.2020.151467
Knutson T, Camargo SJ, Chan JCL, Emanuel K, Ho CH, Kossin J, Mohapatra M, Satoh M, Sugi M, Walsh K, Wu L (2020) Tropical Cyclones and Climate Change Assessment: Part II: Projected Response to Anthropogenic Warming. Bulletin of the American Meteorological Society 101(3):E303-E322. https://doi.org/10.1175/BAMS-D-18-0194.1
MATLAB (2021) 9.11 (R2021b) [Computer software]. Natick, Massachusetts: The MathWorks Inc.
McWhorter JK, Halloran PR, Roff G, Skirving WJ, Perry CT, Mumby PJ (2022) The importance of 1.5 degrees C warming for the Great Barrier Reef. Global Change Biology 28(4):1332-1341. https://doi.org/10.1111/gcb.15994
Brooks ME, Kristensen K, van Benthem KJ, Magnusson A, Berg CW, Nielsen A, Skaug HJ, Maechler M, Bolker BM (2017) glmmTMB Balances Speed and Flexibility Among Packages for Zero-inflated Generalized Linear Mixed Modeling. R J 9(2):378-400.
Montaggioni LF, Braithwaite CJR, Braithwaite C (2009) Quaternary Coral Reef Systems (Vol. 5). Elsevier Science & Technology, pp 223-321.
Mumby PJ (1999) Bleaching and hurricane disturbances to populations of coral recruits in Belize. Marine Ecology Progress Series 190:27-35.
Mumby PJ, Steneck RS, Adjeroud M, Arnold SN (2016) High resilience masks underlying sensitivity to algal phase shifts of Pacific coral reefs. Oikos 125:644-655.
Mumby PJ, Mason RAB, Hock K (2021) Reconnecting reef recovery in a world of coral bleaching. Limnology and Oceanography Methods 19:702-713. https://doi.org/10.1002/lom3.10455
Munk WH, Sargent MC (1954) Bikini and nearby atolls, Marshall Islands; oceanography (physical): Adjustment of Bikini atoll to ocean waves [Report]. Geological Survey Professional Paper 260-C:275-280. http://pubs.er.usgs.gov/publication/pp260C
Ortiz JC, Wolff NH, Anthony KRN, Devlin M, Lewis S, Mumby PJ (2018) Impaired recovery of the Great Barrier Reef under cumulative stress. Science Advances 4(7):eaar6127. https://doi.org/doi:10.1126/sciadv.aar6127
Osborne, K, Dolman AM, Burgess SC, Johns KA (2011) Disturbance and the dynamics of coral cover on the Great Barrier Reef (1995-2009). PLOS One, 6(3):e17516. https://doi.org/10.1371/journal.pone.0017516
Pandolfi J (2002) Coral community dynamics at multiple scales. Coral Reefs, 21(1):13-23. https://doi.org/10.1007/s00338-001-0204-7
Perry CT, Alvarez-Filip L, Graham NAJ, Mumby PJ, Wilson SK, Kench PS, Manzello DP, Morgan KM, Slangen ABA, Thomson DP, Januchowski-Hartley F, Smithers SG, Steneck RS, Carlton R, Edinger EN, Enochs IC, Estrada-Saldívar N, Haywood MDE, Kolodziej G, Murphy GN, Pérez-Cervantes E, Suchley A, Valentino L, Boenish R, Wilson M, Macdonald C (2018) Loss of coral reef growth capacity to track future increases in sea level. Nature 558(7710):396-400. https://doi.org/10.1038/s41586-018-0194-z
Pratchett MS, Heron SF, Mellin C, Cumming GS (2021) Recurrent Mass-Bleaching and the Potential for Ecosystem Collapse on Australia’s Great Barrier Reef. In: Canadell JG, Jackson RB (eds) Ecosystem Collapse and Climate Change. Springer International Publishing, pp 265-289. https://doi.org/10.1007/978-3-030-71330-0_10
R Core Team (2021) R: A language and environment for statistical computing [Computer software]. R Foundation for Statistical Computing, Vienna, Austria. URL https://www.R-project.org/
Rasser M, Riegl B (2002) Holocene coral reef rubble and its binding agents. Coral Reefs 21(1):57-72. https://doi.org/10.1007/s00338-001-0206-5
Riegl B (2001) Degradation of reef structure, coral and fish communities in the Red Sea by ship groundings and dynamite fisheries. Bulletin of Marine Science 69(2):595-611
Roelfsema CM, Lyons MB, Castro-Sanguino C, Kovacs EM, Callaghan D, Wettle M, Markey K, Borrego-Acevedo R, Tudman P, Roe M, Kennedy EV, Gonzalez-Rivero M, Murray N, Phinn SR (2021) How Much Shallow Coral Habitat Is There on the Great Barrier Reef? Remote Sensing 13(21):4343. https://doi.org/10.3390/rs13214343
Roff G, Doropoulos C, Zupan M, Rogers A, Steneck RS, Golbuu Y, Mumby PJ (2015) Phase shift facilitation following cyclone disturbance on coral reefs. Oecologia 178:1193-1203
Rogers A, Blanchard JL, Mumby PJ (2014) Vulnerability of Coral Reef Fisheries to a Loss of Structural Complexity. Current Biology 24(9):1000-1005. https://doi.org/10.1016/j.cub.2014.03.026
Scoffin TP (1993) The geological effects of hurricanes on coral reefs and the interpretation of storm deposits. Coral Reefs 12(3-4):203-221. https://doi.org/10.1007/bf00334480
Shannon A, Power H, Webster J, Vila-Concejo A (2012) Evolution of Coral Rubble Deposits on a Reef Platform as Detected by Remote Sensing. Remote Sensing 5(1):1-18. https://doi.org/10.3390/rs5010001
Smithson M, Verkuilen J (2006) A better lemon squeezer? Maximum-likelihood regression with beta-distributed dependent variables. Psychological Methods 11(1): 54-71. https://doi.org/10.1037/1082-989X.11.1.54
Sweatman H, Delean S, Syms C (2011) Assessing loss of coral cover on Australia’s Great Barrier Reef over two decades, with implications for longer-term trends. Coral Reefs, 30(2): 521-531. https://doi.org/10.1007/s00338-010-0715-1
Thornborough K (2012). Rubble-dominated Reef Flat Processes and Development: Evidence from One Tree Reef, Southern Great Barrier Reef [PhD Thesis, The University of Sydney].
Thornborough KJ, Davies PJ (2011) Reef Flats. In D. Hopley (Ed.), Encyclopedia of Modern Coral Reefs: Structure, Form and Process 869-876. Springer Netherlands. https://doi.org/10.1007/978-90-481-2639-2_135
Victor S (2008) Stability of reef framework and post settlement mortality as the structuring factor for recovery of Malakal Bay Reef, Palau, Micronesia: 25 years after a severe COTS outbreak. Estuarine, Coastal and Shelf Science 77(1):175-180. https://doi.org/10.1016/j.ecss.2007.09.009
Viehman TS, Hench JL, Griffin SP, Malhotra A, Egan K, Halpin PN (2018) Understanding differential patterns in coral reef recovery: chronic hydrodynamic disturbance as a limiting mechanism for coral colonization. Marine Ecology Progress Series 605:135-150. https://doi.org/10.3354/meps12714
Wolfe K, Kenyon TM, Mumby PJ (2021) The biology and ecology of coral rubble and implications for the future of coral reefs. Coral Reefs 40(6):1769-1806. https://doi.org/10.1007/s00338-021-02185-9
Woodley JD, Chornesky EA, Clifford PA, Jackson JB, Kaufman LS, Knowlton N, Lang JC, Pearson MP, Porter JW, Rooney MC, Rylaarsdam KW, Tunnicliffe VJ, Wahle CM, Wulff JL, Curtis AS, Dallmeyer MD, Jupp BP, Koehl MA, Neigel J, Sides EM (1981) Hurricane Allen's Impact on Jamaican Coral Reefs. Science 214(4522):749-755. https://doi.org/10.1126/science.214.4522.749
Wulff JL (1984) Sponge-mediated coral reef growth and rejuvenation. Coral Reefs 3(3):157-163. https://doi.org/10.1007/bf00301960
Yadav S, Rathod P, Alcoverro T, Arthur R (2015) “Choice” and destiny: the substrate composition and mechanical stability of settlement structures can mediate coral recruit fate in post-bleached reefs. Coral Reefs 35(1):211-222. https://doi.org/10.1007/s00338-015-1358-z

No competing interests reported.

SupplementaryInformation.docx

Download PDF

Version 1

posted

You are reading this latest preprint version

Mapping the susceptibility of reefs to rubble accumulation across the Great Barrier Reef

Status:

Version 1

Abstract

Figures

1 Introduction

2 Materials and methods

2.1 Overview

2.2 Site description and sources of data

2.3 Transect sampling

2.4 Rubble accumulation algorithm

2.5 Calculating reef susceptibility

2.6 Sensitivity Analysis

3 Results

3.1 Linear distances and distributions of rubble accumulation

3.2 Effects of aspect and depth on rubble accumulation

3.3 Model sensitivity to input parameters

4 Discussion

Declarations

References

Additional Declarations

Supplementary Files

Status:

Version 1