Total average bioerosion rates across Eva and Fly reefs was 0.152 ±0.012 kg m-2 yr-1, with macroboring being the dominant agent of bioerosion (average = 0.084 ± 0.008 kg m-2 yr-1). This is below typical average macroboring rates for reefs globally (0.314 ± 0.333 kg m-2 yr-1: Lange et al. 2020). However, these rates are comparable to some studies at similar reef sites, and studies that assessed macroboring through the use of microCT with experimental Porites blocks. For example, Tribollet and Golubic (2005) assessed rates of bioerosion at two inshore island reefs (Snapper Island and Low Isles) of the Great Barrier Reef (GBR) using 2D image analysis of experimental Porites blocks. These sites have similar fringing reef structure and turbidity levels to Eva and Fly and recorded total bioerosion rates of 0.27 kg m-2 yr-1 at Snapper Island and 0.18 kg m-2 yr-1 at Low Isles (0.13 and 0.01 kg m2 yr1 due to macroboring, respectively). Silbiger et al. (2015, 2016, 2017) used microCT of experimental Porites sp. blocks and also found similar endolithic bioerosion rates ranging between 0.072 and 0.15 kg m-2 yr-1 at sites in Hawai’i. Further, Enochs et al. (2016) employed microCT with blocks of Porites sp. and recorded rates ranging from 0.026 to 0.13 kg m-2 yr-1 in Papua New Guinea. Global averages in bioerosion rates may be inflated by measures reported in older studies due to less accurate methodologies, or due to a shift in dominant bioeroding taxa populations over time. For example, early work on bioerosion in the Caribbean reported macroboring rates of 0.443 kg m-2 yr-1 (Rutzler 1975; Lange et al. 2020). This is a magnitude higher than average macroboring rates gathered for the Caribbean in the past decade (~0.062 kg m-2 yr-1; Perry et al. 2014; Murphy et al. 2016), and is expected to be due to a significant decline in bioeroding taxa as result of a rapid decline in coral cover and structural complexity. More so than macroboring, grazing pressure on Caribbean reefs has also declined as these previously complex reef systems continue to shift to less complex macroalgae dominated systems, similar to that of Eva and Fly reefs.
Low grazing rates are typical of turbid inshore reefs characterised by low populations of parrotfish and urchins. As such, Eva and Fly reefs recorded low rates of external erosion on experimental blocks (average = 0.016 kg m-2 yr-1 at Eva and 0.031 kg m-2 yr-1 at Fly) compared to clear water reefs (e.g., up to 4.31 ± 0.43 kg m-2 yr-1 at Reunion Island, Indian Ocean: Chazottes et al. 2002). Here we relate the external erosion observed on the blocks to be due to herbivorous fish, such as parrotfish, given that the blocks were not accessible to urchins. Additionally, we acknowledge that we cannot be sure that all external erosion measured on experimental blocks was due to fish grazing, Although we observed distinct scrapings on the block surface, some external erosion may be the result of physical erosion. Across Eva reef herbivorous fish had a low mean abundance of <10 per 1000 m2, with the only grazer species recorded being Scarus ghobbin with a mean biomass of 15.53 ± 7.33 kg per 1000 m2 and estimated average bioerosion rate of 0.059 ± 0.030 kg m2 yr-1. This estimate is almost four times the average rate of external erosion captured on experimental blocks (0.016 ± 0.003 kg m-2 yr-1). This variation displays the uncertainty of capturing grazing with small experimental units, especially in algae dominated settings where herbivorous grazers have an abundance of food across the reef substrate.
Previously, the only research into bioerosion (external and internal) from north Western Australia has been from Ningaloo Reef where studies focused on external grazing of parrotfish and urchins using snapshot abundance data combined with offsite grazing rates from previous GBR studies (Bellwood 1995; Bonaldo and Bellwood 2009). Across Ningaloo, estimates of parrotfish (Chlorurus microrhinos and Chlorurus sordidus) erosion ranged from 1.18 kg m-2 yr-1 to 2.30 kg m-2 yr-1 (Johansson et al. 2010). Additionally, Johansson et al. (2010) found that urchin grazing rates (Diadema sp. and Echinometra mathaei) at Ningaloo ranged from 0.00 to 0.55 kg m-2 yr-1 with a later study by Langdon et al. (2013) suggesting that rates were as high as 1.00 - 4.50 kg m-2 yr-1. Although the grazing rates we have recorded for Eva and Fly reefs are a magnitude lower than that recorded at Ningaloo, they are comparable to similar inshore sites of the GBR that employed the use of experimental substrates (<0.01-0.02 kg m-2 yr-1 at Snapper Island and <0.01-0.06 kg m-2 yr-1 at Low Isles; Osorno et al. 2005; Tribollet and Golubic 2005). Furthermore, there is also the potential for over-estimation of grazing bioerosion rates based purely on a snapshot measure of taxa abundance and biomass (see Perry et al. 2017 and Yarlett et al. 2020 for further discussion). With such low grazing rates, the relatively low bioerosion rates on these marginal inshore reefs is the result of endolithic borers. This suggests that even with lower coral cover and consequently low gross carbonate production rates, these marginal reefs may display stable rates of framework accretion.
Within this study, the majority of bioerosion measures across Eva and Fly reefs were from macroboring polychaete and sipunculan worms. Given that previous studies have suggested that the substrate should be exposed for a minimum of three years to allow for full borer succession (Kiene and Hutchings 1994), the lack of macroborer diversity may be due to the short (1 year) deployment of the experimental blocks. Bioeroder successions patterns are variable over time (Chazottes et al. 1995; Osorno et al. 2005; Tribollet and Golubic 2005; Carreiro-Silva and McClanahan 2012), as well as over small spatial scales (Davies and Hutchings 1983; Sammarco and Risk 1990; Kiene and Hutchings 1994; Hutchings and Peyrot-Clausade 2002), likely due to both variable biotic (e.g., reproduction) and abiotic factors (e.g., substrate density, environmental variables). Typically, within Australia, microborers inhabit dead coral substrate rapidly, followed by short lived polychaete species such as Polydora (Spionidae) and Fabriciniids (Sabellidae) (Hutchings 2011). Within two to six months sipunculans bore into cracks made in the substrate by polychaetes (Hutchings 2011). Kiene and Hutchings (1994) suggest that it is not until three years of exposure that bivalves and sponges populate available substrate. However, Tribollet and Golubic (2005) recorded macroboring of sponges and bivalves in experimental substrate that had been deployed in situ for 1 year at multiple sites across the GBR. Additionally, Wizemann et al. (2018) recorded bivalves in substrate after two months of exposure at reefs off Costa Rica. These more recent studies demonstrate that one year of substrate exposure may be adequate to detect other boring organisms. Furthermore, low coverage of sponges across both Eva and Fly reef substrate have been observed (average 0.42%; Dee et al., 2020) suggesting that sponge borings may not be observed here even after three years of exposure. Therefore, endolithic boring rates recorded here are potentially representative of long-term bioerosion, although these rates could fluctuate with environmental change.
Across Eva and Fly reefs, we saw spatial variation in microboring and macroboring activities with greater microboring rates at Eva reef and greater macroboring rates measured at Fly Reef. Eva Reef was characterised by higher light levels whereas Fly Reef had greater chlorophyll-a. These spatial differences in micro and macroboring at each reef suggests that these groups of bioeroders may be driven by different environmental conditions particularly given that the habitat between reefs is comparable. Linear regression further found that macroboring was negatively correlated with both temperature (r2=0.24 P=0.008) and light (r2=0.16 P=0.037). The negative effects of temperature on macroboring are potentially due to temperature anomalies (up to 3.6°C) recorded during the summer months, which may have caused thermal stress and decreased the activity of macroboring taxa. We offer two theories for the negative relationship found between macroboring and light; firstly, Hutchings et al. (2005) suggest that sedimentation at inshore reef sites may inhibit the settlement and development of epilithic and endolithic algae, which limits grazing, and in turn facilitates recruitment of macroboring taxa. Secondly, sites with lower light levels may be associated with higher levels of chlorophyll-a and other nutrients that are predated on by filter feeding macroborers (Le Grand and Fabricius 2011), increasing their abundance and activity. The lack of a relationship between chlorophyll-a and macroboring is likely due to the fact that chlorophyll-a data (as well as turbidity, salinity, pH) were collected at a lower spatial (one site) and temporal (monthly) resolution. Given that macroborers were significantly correlated with light and temperature collected at a high spatial and temporal resolution, we suggest that future studies attempt to collect all environmental data at this resolution as a means of further developing environmental proxies for estimating changes in bioerosion rates over time. The latter likely also applies to microboring given that we did see significant spatial differences between reefs and sites.
An assessment of 31 global studies found that both temperature and chlorophyll-a were significantly influencing macroboring rates. Specifically, as temperature increased by 1°C macroboring decreased by 0.008 kg m-2yr-1. This relationship is somewhat surprising as temperature increases usually lead to an increase in metabolism, and therefore activity. Regionally it is expected that warm-water bioeroding taxa such as urchins and bivalves will increase in abundance, intensifying bioerosion in areas experiencing warming (Schiel et al. 2014). However, it is also predicted that higher ocean temperatures will ultimately lead to stressful conditions and depressed activity and survival of bioeroding taxa, as displayed in bioeroding sponges (Achlatis et al. 2017; Davidson et al. 2018). As studies in our analysis recorded temperatures up to almost 30°C (including our study with <3.6oC temperature anomoly), we may have captured the turning point from increased activity to depressed activity with increasing temperature, as observed in this study. However, given the physiological tolerance to temperature will vary between taxa and regions, there is a need for more controlled empirical studies on the influence of temperature before general predictions can be made. In contrast, chlorophyll-a had a positive influence on on macroboring globally, with in an increase of chlorophyll-a by 1 ug L-1 increasing the rate of macroboring by 0.269 kg m-2yr-1. This relationship has been recorded in multiple studies. For example, Carreiro-Silva and McClanahan (2012) found that macroboring activities were positively correlated to chlorophyll-a concentrations. While Le Grand and Fabricius (2011) recorded bioeroder densities increased 650-fold per additional 1 ug L-1 of chlorophyll-a. Furthermore, given the scale of increase in macro-boring per unit of environmental change, these analysis suggest that chlorophyll-a has a considerably larger influence on macro-boring rates than temperature. Interestingly, no significant correlation was recorded between macroboring and nutrients such as nitrates and phosphates, which have previously shown to influence bioerosion rates (Chazottes et al. 2002; Webb et al. 2017; Rice et al. 2020).
We found no significant influence of pH or aragonite saturation on macroboring within our study, or from global assessments. The range in pH values across studies involved in global analysis were 7.5 - 8.2 which represents a similar range to that measured in most individual studies. Multiple studies have shown that ocean acidification can increase bioeroder activity through weaking carbonate framework density and, on a broader scale, through increasing coral mortality (DeCarlo et al. 2014; Silbiger et al. 2014; Enochs et al. 2016; Schönberg et al. 2017; Webb et al. 2017). DeCarlo et al. (2014) further demonstrated that ocean acidification can lead to increased macroboring rates, with these impacts being significantly increased under high-nutrient conditions. This finding demonstrates that it is likely a combination of environmental factors (especially in marginal environments) that is driving bioerosion rates. We aimed to assess relationships between bioerosion and interacting environmental factors but found there is currently not enough environmental data to run a robust and comprehensive global analysis. As data increases, multivariate analysis and modelling will be an important future research direction to increase our understanding of how environmental interactions are influencing bioeroding functional groups across regions.
Within this study, we provide the first rates of endolithic bioerosion for marginal reef systems in Western Australia, where data for both Western Australia and marginal reefs are severely lacking. We found lower than expected rates of bioerosion compared to average global rates, yet rates were comparable to other studies on inshore turbid reefs of the GBR, and studies adopting microCT analysis. Macroborers were the dominant drivers of bioerosion at Eva and Fly reefs and were also more sensitive to environmental change than microborers or grazers. Further, this study provides empirical relationships between key environmental parameters of temperature and chlorophyll-a with macroboring that can be used to track changes in endolithic bioerosion overtime with changing environmental conditions. Lower rates of bioerosion across marginal inshore reefs is encouraging as these reefs may maintain a positive carbonate budgets despite lower carbonate production rates. However, as our findings showed global macroboring rates were significantly influenced by chlorophyll-a, the control of nutrient loads may be more important to the survival of inshore reefs than temperature. With changing climates and increasing anthropogenic pressures, the importance of understanding relationships between environmental parameters and ecosystem processes (i.e., bioerosion) has increased exponentially. As we were only able to find seven studies that reported environmental data coupled with bioerosion rates, it is apparent that this is an important knowledge gap. Therefore, an increase in data for the development of empirical relationships between bioerosion activities and environmental parameters, will allow for high resolution (site specific) environmental data to act as reliable and rapid proxies of bioerosion rates and facilitate the development of ecological models that aim to predict reef carbonate loss and net reef accretion.