DISCUSSION
Climatic changes are leading to species redistributions and consequent impacts on ecosystem function and services are starting to emerge in some regions (Pecl et al. 2017). In our study system – shallow reefs within a warming tropical-temperate transition zone where a major habitat, kelp forests, is declining - we found clear shifts in the abundance and trophic composition of a fish community. Our results show overall increases in fish abundance for most trophic groups, including both tropical and temperate species. Tropical herbivores increased most dramatically, whereas temperate planktivores were the only trophic guild that declined, with no corresponding increase in tropical planktivores observed. Our results here suggest that changes in pathways of energy flow may be one of the main impacts of climate change for tropicalised systems, with planktonic inputs becoming less important, while a higher proportion of algal productivity gets consumed locally by increasingly abundant herbivores.
Our 17 year study period encompassed the loss of kelp as one of the dominant foundation species on these tropicalised reefs (Vergés et al. 2016). We therefore expected to observe declines of temperate fish species that use kelp as a resource (directly or indirectly) for food, cover and/ or shelter. However, 22 of the 26 species that changed in abundance showed an increase in probability of occurrence, and this included both tropical and temperate species. Overall, this suggests that tropicalised reefs that have lost kelp can sustain greater fish biomass than previous kelp-dominated states. These results are consistent with observed patterns along the latitudinal coast of Eastern Australia, which show that total fish biomass in shallow reefs is an order of magnitude higher in warmer subtropical reefs than in higher latitude reefs which are dominated by kelp (Holland et al. 2020).
What remains to be established is what is fuelling increases in fish abundance, as well as an increase in diversity in these tropicalised reefs? One potential explanation is that new ecological niches are becoming available as foundation species like kelp decline. However, we found no evidence that tropical fishes increasing in abundance fill underrepresented niche spaces. Despite a shift from kelp to turf, neither tropical nor temperate fishes showed a distinct set of traits that may facilitate success in a shifted environment. For example, within the herbivore group, tropical fishes that have dramatically increased over time do not fill a distinct trait space when compared with temperate herbivores based on their morphometric traits. Morphometric analysis also did not show any evidence for an increase in generalist species, a proposed mechanism for successful establishment of invaders in the more general field of invasion biology.
An alternative explanation for the observed increase in fish biomass is that the bottom-up resource base in the system has increased. In particular, as reefs become tropicalised, it is plausible that a higher proportion of the primary production is consumed and retained within the local reef system through the observed increases in the abundance of herbivorous fish. In temperate kelp forests about 80% of kelp production is exported as detritus (Krumhansl & Scheibling 2012), whereas in tropical systems, herbivorous fish consume 50-100% of turf primary production produced in situ (Hay 1991). As fast-growing turf algae replace canopy forming kelp, the turnover rates of primary production (time taken to replace biomass) accelerates and the proportion of primary productivity consumed locally may be increasing, leading to higher fish biomass (Vergés et al. 2019). Increasing abundance of herbivorous fish in tropicalised reefs may also lead to more localised recycling of nutrients, as more local consumption also means greater supply of nutrients via excretion (Allgeier et al.2017). In turn, this may lead to a higher production of nutritious detritus, an important food source for many nominal herbivorous fish (Wilson et al. 2003).
A third possibility for the observed increase in richness and abundance with no addition of unique niches, also related to the resource base of the system, is that the fish community studied here is currently well under the carrying capacity of the system. Thus, despite a phase shift, increased high fish biomass is still supported making competitive exclusion currently less important for predicting species change in occurrence. Perhaps the increasing abundance of range expanding fishes has a low net effect on the resident community because the loss of structurally complex kelps has not significantly reduced the carrying capacity of these reefs, which may have alternative types of microhabitats and resources. An increase in richness and abundance could be a combination of both an increase in retained primary productivity and a fish community at abundance levels not limited by competition. However, while this may explain an increase in tropical species over the study period, a cumulative increase in abundance of temperate plus tropical species is less clear.
It is important to note that many of the species observed increasing in abundance have cosmopolitan distributions throughout Australia. While they contribute in our samples to biodiversity enhancement at the local scale, they may in fact be contributing to broader scale homogenisation of biodiversity. Scaling up local losses of rare or endemic species results in a net decrease in regional scale diversity. While in these tropicalised reefs we found that net gains in species seem to outweigh net losses, Finderup Nielsen et al. (2019) show that localised increases in species richness can mask broader patterns of biotic homogenisation where rare, range-restricted species are almost always disproportionally the losers. We found evidence of that here with a mix of tropical and temperate fishes increasing and only temperate species showing a decline. Although BRUV surveys can underestimate richness and diversity metrics, roving underwater visual censuses undertaken by divers in the Solitary Islands during 2001-2017 also reported strong increasing patterns of species richness for tropicalised mid-shelf reefs (Malcolm & Ferrari 2019).
The positive correlation between kelp decline and overall fish abundance and species richness may be reflecting a transitional stage, with further responses lagging and yet to manifest. For instance, a recent global synthesis of assemblage changes associated with terrestrial forest loss has shown that despite evidence for short-term positive effects on species diversity, there can be lagged responses of up to decades post tree-loss disturbances, particularly when assemblages include species with long generation times (Daskalova et al.2020). Similarly, lag effects have been identified in response to reduced structural complexity following mass bleaching in coral reefs where, despite initial maintenance of biomass, evidence from size structure analysis revealed that adult fish lost through natural mortality are not replaced by juveniles in the long term (Grahamet al. 2007). In a similar way, there may be a time-lag response to the loss of kelp for long-lived fish species that depend on this habitat for recruitment or during their juvenile life stages, but are less dependent on kelp during their adult stages. This fish community is likely still in transition and longer-term effects of a climate mediated phase shift may not yet be fully realised. Continued long-term monitoring of assemblages following disturbances will be key to fully understand the impacts of habitat change on biodiversity.
Temperate planktivores were the only group that declined over time. The decline in planktivores observed was largely driven by a decline in the Australian Mado (Atypichthys strigatus ), an abundant species on subtropical and temperate rocky reefs in south-eastern Australia. Small schooling planktivores such as A. strigatus can be considered outliers in BRUV analyses because they form large groups with highly stochastic distribution; thus, these results need to be interpreted with caution. Our results suggest that declines of this broad trophic group were unlikely due to direct negative competitive interactions with tropical planktivores, which did not change throughout the study period. It is interesting to note that an expansion in the distribution ofA. strigatus towards the cold edge of its thermal range and into Tasmania has been documented (Last et al. 2011). This may suggest that the decline of this species may be driven by physiological effects of warming driving a range contraction at the warm edge rather than shifts in ecological interactions, although complex indirect effects not quantified here could also explain such declines. Given the dominance and important role of planktivores in fuelling food webs in shallow coastal reefs across Australia and New Zealand (Truong et al.2017), a potential range contraction of this species may have important ecological implications.
In contrast, the decline of the two temperate wrasse species (Ophthalmolepis lineolatus and Pictilabrus laticlavius ) observed in our study may be directly linked to the decline of kelp, as these species are known to display a behavioural preference for this macroalgal habitat (Barrett 1995; Edgar et al. 2004; Tuyaet al. 2009; Fulton et al. 2016). Kelp canopies may be important for both wrasses through the provision of shelter and refuge from predation (Tupper & Boutilier 1997); and structurally complex kelp holdfasts harbour high invertebrate densities that provide a profitable food source for wrasses (Morton et al. 2008). In the case of the other species observed declining, the piscivore Dinolestes leweni(Long-fin Pike), we note that this species occupies a relatively unique niche by the standards of our model, i.e it occupies a large area within the convex hull and is not clustered towards the mean. This may indicate niche requirements that are not shared by others in the guild, and may now no longer be available, e.g. a specialist diet of a declining species. In their core range, long-fin pike are abundant and feed on small bodied planktivores including A. strigatus (Truong et al. 2017).
Shifts in predator-prey interactions due to climate change and species redistributions are largely unexplored. Here we briefly consider that the decline of the four temperate species identified in this study could be partly caused by an overall increase in piscivory, given that ten out of the 22 species observed increasing in abundance are piscivores with only one temperate associated piscivore showing a decline. While in-depth dietary analysis would be required to test this, loss of canopy cover of kelp may also make some species more vulnerable to predation. A factor that may have potentially contributed to increased predatory fish abundance is the establishment of the Solitary Islands Marine Park in 2002. Although none of the study sites are within sanctuary zones where fishing is totally prohibited, it is possible that increases in predator abundance are influenced by a spill over effect, i.e. the emigration of juveniles and/ or adults from protected sanctuary zones to nearby areas. Most studies find however, that spill over occurs at distances within 200m from no-take zones (Di Lorenzo et al. 2020), and our study sites were located further away. We therefore consider it unlikely that fishing regulations are influencing changes in fish abundance in our study sites, with the notable exception of the observed increases in the threatened black cod (Epinephelus daemelli ) which are likely to be a response to targeted management recovery plans for this species (Harasti & Malcolm 2013).
To summarise, our results suggest that although tropicalisation and associated loss of kelp can lead to increases in fish diversity and overall fish abundance, it is also leading to the loss of some temperate species and is causing major changes in pathways of energy flow, as reflected by rapid shifts in trophic composition. We hypothesise that continued warming and tropicalisation will favour range expanding species and may lead to increases in fish abundance and biomass. Our data here also suggests that this may be to the detriment of some temperate associated species. In particular, we observed the decline for four temperate fishes (Australian mado, southern maori wrasse, senator wrasse and long fin pike), which may be the first species identified as vulnerable to changes associated with tropicalisation for this region. The continued growth of tropical fish populations on high latitude reefs means that losses of temperate associated fishes may go undetected particularly if these are not target fisheries species.