<p>Coastal benthic communities represent one of the most biodiverse ecosystems in the marine environment and perform numerous important ecological functions, including the cycling of carbon and other nutrients in the water column. However, the benthic communities of temperate ecosystems are far less studied than their tropical counterparts, especially beyond the top 30 m of water into the so-called temperate mesophotic zone (30 – 150 m), which remains largely unexplored globally. The lack of information regarding the structure and subsequent ecological functions of benthic communities in both shallow and temperate mesophotic ecosystems (TMEs) is likely to be imposing critical limits on our understanding of how temperate coastal ecosystems function generally, and the ecological services they provide. Marine sponges are often one of the most abundant organisms occurring in shallow temperate benthic habitats, and are likely to be performing important ecological functions, including transferring carbon from the water column to the benthos via their feeding activities. However, despite their high abundance, temperate marine sponges remain generally overlooked, where almost nothing is known about their distributions and potential ecological importance in both infralittoral and mesophotic habitats. The overall aim of this thesis was to address this knowledge gap by describing how sponges are distributed through the infralittoral and mesophotic zones of rocky reefs in New Zealand, and assessing the role of trophic relationships between sponges and microbial food sources in determining sponge distribution, population dynamics, and carbon retention in these habitats. In Chapter 2 I addressed a key knowledge gap of how the composition and abundance of benthic communities (including sponges) change from the infralittoral zone to the almost entirely unexplored mesophotic zone at selected sites in New Zealand. I describe quantitative changes in the benthic community composition of rocky reefs from 5 to 120 m at six locations across New Zealand, including the Fiordland Marine Area and the Poor Knights Marine Reserve. Benthic community data were analysed from videos and photographs collected using SCUBA (30 m) and a remotely operated vehicle (ROV) (>30 m). I found significant changes in community composition with depth at all locations, suggesting that temperate mesophotic ecosystems (TMEs) provide habitats different from those in shallower water. I show that the morphological composition of these sponge assemblages changes with depth at all locations. I created a sponge assemblage complexity metric to describe how changes in sponge morphologies with depth have potentially important ramifications in relation to the provisioning of habitat complexity. The significant variation in both sponge abundance and the presence of free-substrate space were left unexplained in Chapter 2, where no correlations with multiple environmental variables could be found. In Chapter 3 I suggest that food availability is the most likely driver of these patterns. Here, I first describe the composition and distribution of the food pool within the size range potentially available to sponges, from the surface down to 120 m. I collected water samples from the Poor Knights and four sites in Fiordland from the innermost to the outermost locations of Doubtful Sound to accommodate the strong environmental gradient found in this region. Using flow cytometry, I identified and quantified the microbial community components available to sponges within the particulate organic carbon (POC) pool and show how the abundance of these components changes significantly with depth and across locations. I also show how dissolved organic carbon (DOC), as another potential resource to sponges, changes significantly along these same depth gradients. Using a combination of cell sorting and scanning electron microscopy, I confirmed specific previously identified microbial populations, and corrected a previous misclassification of a cell population now reclassified as picoeukaryotes. I found strong positive correlations between sponge distributions and food availability when data from all Fiordland sites was combined, and some smaller-scale patterns at the Poor Knights. These observations provide fundamental ecological information about the composition and distribution of resources at the foundation of marine trophic structures in New Zealand’s infralittoral and mesophotic habitats, and how the distribution of temperate sponges maybe determined by bottom-up effects. In Chapter 4, I address the substantial knowledge gaps that remain in the potential dietary range and feeding preferences of temperate sponges, and how this might determine their population dynamics. I determined the diets of seven common sponge species occurring on shallow temperate reefs at three sites in New Zealand in situ. I assessed the potential for active food selection and interspecific differences in food preference to support resource partitioning and trophic plasticity. For the first time on shallow temperate reefs outside of the Mediterranean, I measured the uptake of multiple pelagic microbial communities as identified in the previous chapter, as well as DOC. Sponges showed active selection for different POC groups as well as between POC and DOC, although only two species (P. penicillus and Polymastia sp.) showed significant DOC retention. I found that retention efficiencies of specific POC groups were consistently high in all species that fed exclusively on POC. However, the consumption of DOC by only P. penicillus and Polymastia sp. coincided with lower retention efficiencies of POC groups and was entirely responsible for inter-specific differences in food selectivity and therefore resource-partitioning. Correlations between DOC availability and DOC consumption, and DOC selectivity indicate trophic plasticity in the study species. The results from this chapter suggest that sponges can ‘switch’ between food types based on relative food availability as an active rather than passive response. I found limited evidence for niche partitioning within the POC food pool, but propose that trophic plasticity, generalist feeding strategies, and DOC consumption might help explain the high abundance of sponges relative to other benthic invertebrate groups in resource-poor environments. The quantification of carbon retained by sponge assemblages is of considerable importance to our understanding of the ecological dynamics of benthic habitats generally. In Chapter 5, I combine multiple components from my previous research to estimate the potential range in quantity of carbon sponges are retaining at the assemblage scale as a result of their feeding activity on New Zealand reefs in the infralittoral and mesophotic zone. I determined the pumping volumes of five particularly common sponge species occurring on the Wellington South Coast in situ using SCUBA. I assessed potential correlations between multiple sponge biometrics (mass / number of oscula / size of oscula / total oscula area / pumping velocity) and pumping volume, to determine the most accurate and efficient way to standardize and extrapolate pumping volumes to entire assemblages. I found total oscula area (OSA) to be the best predictor of sponge pumping volume, and that the ratio of total oscula area to sponge size (~ 6%) to increase allometrically with sponge size, without any inter-specific variation. I used a range of potential OSA-specific pumping volume estimates, in combination with a range of POC retention efficiency estimates of different pelagic microbial groups, (determined in Chapter 4), to determine the range of total carbon mass retained by sponges. I then extrapolated this information to entire sponge assemblages at the Poor Knights Marine Reserve and in Doubtful Sound using sponge distribution information from the infralittoral and mesophotic zones of these regions (determined in Chapter 2). This study confirms the efficacy of applying OSA-specific pumping volumes to population scales, and demonstrates the significant contribution sponges make to the transfer of carbon (> 100% of available carbon in the benthic boundary layer per hour) to the benthos through the infralittoral and mesophotic zone of New Zealand reefs. The dominance of temperate sponges throughout both shallow and mesophotic zones and the large proportion of available carbon they transfer from the water column to the benthos is likely to have substantial ecological implications. This includes the regulation of the availability of microbial communities which are fundamental to other important ecological processes, such as primary production, and the microbial loop. Furthermore, while numerous factors are likely to determine the presence and proliferation of sponges in temperate environments (e.g. habitat availability, resource competition, environmental variables), the quantity of carbon accumulated by heterotrophic sponges fundamentally dictates the production potential of sponge assemblage biomass itself (as well as all other energetic outputs), and therefore forms the foundation on which all other ecological functions performed by sponges rely. The information provided by this thesis suggests that as the most dominant benthic organism throughout both infralittoral and mesophotic habitats, these ecological functions are likely to be of particularly high importance, especially in deeper habitats where the wider benthic community becomes increasingly depauperate relative to sponges. As such, sponges are likely to be one of the most important functional components of temperate coastal ecosystems, where their consideration across the full habitat ranges available to them, from infralittoral to mesophotic zones, is essential in developing a holistic understanding of coastal temperate ecosystem dynamics generally. </p>
