In the initial phase of the study we focussed on examining temporal patterns in species abundance and the relationship with physical drivers such as temperature. For many species there was no clear relationshipevident, as the time-series of observations were, as yet, generally insufficient through time to detect relationships with changing environmental variables such as mean monthly temperature. The 20 year dataset from Maria Island proved to be the most meaningful in this context, and could readily be matched with oceanographical variables derived from a nearby CSIRO monitoring station. While few individual species in this dataset could be clearly determined to be responding to climate signals through time, a range of community level metrics did show significant trends when examined for the fish assemblage. Signatures of a warming trend could be seen in metrics such as functional trait richness, and functional diversity, reflecting increasing abundances of warm affinity species and species traits such as herbivory. It is this latter trait that may have one of the largest initial impacts in the SE region of Australia, as, prior to recent warming, herbivorous fishes were relatively rare in the cool temperate zone, thus their increasing biomass may reflect a significant change in system function through time.
One notable feature was that in some metrics, such as thermal affinity, there was a differing response to warming between the unfished sites in the Maria Island marine reserve and adjacent fished reference sites. These differences reflect “resilience” of the reserve to some aspects of climate change. The primary mechanism underlying this appears to be related to increased top down control of sea urchins within the reserve (via lobster predation) reducing the extent of urchin barren formation that in turn provides habitat for many warmer affinity species. The message from this is that MPAs can provide increased “resilience” to climate chance effects, particularly when these are driven by an ecosystem engineer such as the Long spined urchin Centrostephanus rodgersii. However, this resilience is context dependent, as in many areas such lobster/urchin interactions may not be the primary drivers of ecosystem function on reefs, or where they are, resilience can, and should, be enhanced in off reserve areas as well, by appropriate changes in fishery management. Ultimately this management needs to be informed by long-term studies examining differences between fished and protected areas at representative locations along our coastline, building on existing studies to extend that time series over future years of warming.
In the second phase of the study we modelled the latitudinal species abundance curves of a wide range of fish and mobile invertebrate species in order to identify the current shape of the curves and their abundance centres, and use these distributions to predict both likely future distributions and the relative contributions of individual species under possible climate change scenarios. The use of Reef Life Survey (RLS) data was essential for this modelling, as existing data from MPA and reef health monitoring programs was too sparse to identify both core abundance areas and the spatial extent of rarer abundances in the tails of species distributions. In addition, in many cases, knowing the upper thermal limit of distributions is important for refining models and examining likely losses at northern extent of ranges, and the RLS dataset was unique in providing abundance data across that range. Overall, the modelled distributions are invaluable for estimating the extent that some species will extend their central maximum abundance distributions into parts of SE Australia, or to the south of Tasmania and hence be lost, or simply increase/decrease marginally in influence if the distribution has a long tail around a central peak. The predicted likely emergent community at any location is clearly dependent on site (exposure regime etc), likely temperature increase through time, and the time for communities to come to equilibrium. Recent research suggests there will be a 2 deg C increase in temperature in the SE region by 2060, under the A1B scenario of the IPCC (Oliver et al. 2014). Under that basis we can determine likely assemblages based on our distribution data, and use that to inform discussions by the biological and resource management community as to future adaptation options, both with respect to conservation and fishery management outcomes. We have some confidence that our species distribution models are likely to predict the general species distribution following warming, as an additional study undertaken as part of this project determined that during the previous period of warming in this region, the range expansion of many species closely tracked the climate warming velocity. That change was surprisingly irrespective of individual species traits, such as dispersal capability via adult or larval movement.
The species distribution models predict significant changes in the assemblages of fishes and mobile invertebrate species in the SE region, although for many regional species this change was not at an order of magnitude level, and the influx of warmer water species meant that overall levels of diversity would increase. Few species were predicted to be lost, and with one exception (the Real bastard trumpeter), all were introduced species with a localised distribution. The major predicted change of consequence to ecosystem function was a doubling of Centrostephanus abundance in eastern Tasmanian waters, and extending to the south coast in significant numbers. This was coupled with a predicted decline in Southern rock lobster numbers in this region (in the order of 20%), such that the key predator of
Centrostephanus will be declining at a time when increasing numbers are needed to arrest likely barren formation.