Seagrass resilience in Port Phillip Bay: final report to the seagrass and reefs program for Port Phillip Bay

Seagrass in Port Phillip Bay is dominated by the eelgrass, Zostera nigricaulis, which occurs around the margin of the bay from the shallow subtidal zone to depths of up to 8 metres. Zostera provides crucial ecosystem services such as stabilising sediments and improving water quality, reducing coastal erosion, and increasing biological productivity for the marine food chain as well as providing nursery habitats for key recreational and commercial fish species. At the scale of the whole bay our research shows that the broad distribution of Zostera can be explained by two main factors, wave exposure (Zostera does not occur in areas of high wave energy) and depth (a proxy for the light reaching seagrass for growth). Within the broad area where Zostera can grow, however, there are regional areas of presence or absence and variation over time that cannot be explained by this simple model, and are likely influenced by limiting factors such as nutrients, turbidity affecting light, and sediment movement. Zostera in Port Phillip Bay can obtain nutrients from a number of sources. Nutrients can be obtained directly in dissolved form from the water column or sediments (pore-water). These nutrients can come from sources such as river catchment, sewage treatment, Bass Strait and the atmosphere. Our experiments showed that nutrients can be taken up above ground (leaves) and below ground (roots and rhizomes) and can move between these compartments. Nutrients can also be taken up from the sediments indirectly through the bacterial breakdown of detritus that releases nutrients for uptake by seagrass. This detritus does not only come from dead seagrass but also algae, including phytoplankton (single-cell plants floating in the water column) that are “trapped” by the seagrass bed (reduced waves and currents in seagrass beds lead to phytoplankton settling into the bed). Finally, it is possible for seagrass to utilise Nitrogen in the gaseous form when bacteria associated with the root/rhizome system ‘fix’ atmospheric Nitrogen (although this pathway only meets a minor part of the plants’ nutrient needs). Our studies on the influence of nutrients and sediments using modelling, chemical (stable isotope) analyses and experiments have shown that there are three broad categories of Zostera habitat within the bay. Areas of Zostera that are protected from current and wave exposure, and relatively isolated from the catchment, bay such as Swan Bay and Corio Bay, have relatively stable cover of seagrass over time (‘persistent’ seagrass beds). These seagrasses live in fine, muddy sediments, and most of their nutrients come from internal breakdown and recycling of detritus. In contrast, seagrasses living in more exposed parts of the bay, particularly the Bellarine Bank and the southern areas of the bay, have shown major increases and declines since the middle of last century (‘ephemeral’ seagrass beds). Our field and experimental studies indicate that these seagrass beds are nutrient limited, and major losses of Zostera occurred in these areas during the Millennium drought when catchment inputs of nutrients were low. Our studies also show that these areas have dynamic sediment movements and these may have also changed in response to climatic shifts during the drought, affecting seagrass distribution. The third category of seagrass habitat occurs along the north-west coast of the bay where nutrients are derived from the Western Treatment Plant and are unlikely to be limiting, but the combination of fine sediments and wave exposure means that turbidity is often relatively high and limiting for seagrass growth. These different categories of Zostera habitat have differing levels of resilience to changes in water quality parameters such as nutrients and sediments. ‘Persistent’ beds are largely independent of changes to catchment and other inputs, and sediment transport processes, and as such are relatively resilient. In contrast ‘ephemeral’ beds are quite sensitive to changes in catchment inputs and sediment transport processes and will be expected to continue to show major variability over time in relation to climate and other factors. The demography of Zostera also varies around the Bay, and we found evidence of separation of seagrass populations around the Bay. Demographic variation (and population segregation) are a mixture of regional differences and more localised, site specific ones. Page 2 Seagrass Resilience in Port Phillip Bay At a broad scale, seagrasses responded to small-scale disturbances, including loss of leaves, loss of leaves and below-ground parts, in a broadly consistent way. Leaf regrowth was rapid, as was the extension of rhizomes from neighbouring areas into the disturbed area. When regrowth was prevented, recovery slowed dramatically, as we saw few signs of successful colonisation by seeds or drifting fragments. While we saw this general pattern everywhere, there were big differences between individual sites in the speed of this recovery, suggesting that some areas are less resilient than others. These differences did not match the broad regional patterns that we saw elsewhere. As a small side project, we also investigated, for two quite different seagrass sites, whether small disturbances caused the loss of “Blue Carbon”, but we saw little change. Seagrass reproduction occurred from mid-spring through to early summer, with peak flowering in October. There were small differences around the bay in the timing of flowering, but big differences in the amount of flowering, numbers of seeds produced, and numbers of seeds buried in the sediment. We saw consistent, extensive flowering in the Geelong Arm, consistent, but less extensive flowering in Swan Bay, and limited flowering at the north end of the Bay. In the southeast, a seagrass meadow at Blairgowrie flowered extensively, and we saw some build-up of the seed bank. We used genetic tools to examine patterns of sexual and asexual reproduction within seagrass meadows and also to infer dispersal between different areas. We saw considerable differences around the Bay, consistent with patterns of flowering, but also indicating that dispersal at large scales may be uncommon. In the Geelong arm, we saw genetically diverse meadows, with evidence of connections between different sites. Swan Bay was also diverse, but distinct from the Geelong sites. At the north of the Bay, we saw more reliance on asexual reproduction, with reduced genetic diversity. Some sites seem quite isolated, with limited sexual reproduction, low diversity, and few obvious connections to other seagrass beds. This pattern is particularly clear at Ricketts Point and Point Lonsdale. In the south east of the Bay, we saw diverse sites that seemed to have some connections to other seagrass areas. The genetic results do suggest that several parts of the bay are relatively discrete. The most likely dispersal pathway is by drifting fragments. Seeds are not likely to disperse, as they sink, but we identified an unusual dispersive fragment that was produced by plants. These fragments remain healthy for up to 6 weeks, and may be an important means of recolonisation and genetic exchange. We found poor survival of these fragments in sediments, and remain uncertain about whether these fragments are produced in vast numbers and only a few ever establish, or if there are particular circumstances that allow them to become established. Seeds also are a paradox. Their large numbers (>10,000 per m2) in some cases suggests an important ecological role, but there are questions about what causes them to germinate and how easily seedlings become established. We induced germination by changing salinity, but only achieved low germination rates. There are questions about whether there is a set of environmental cues that cause most seeds to germinate. A technique that involved high rates of fragment or seedling establishment would have potential for small-scale restoration or remediation projects. The combined results from individual projects were used to develop a broad conceptual model for seagrasses in Port Phillip Bay. The conceptual model, illustrated using the Bellarine Bank, can be used to clarify expectations about “normal” seagrass dynamics around Port Phillip Bay and to highlight important drivers of seagrass. Identifying these drivers in a particular area allows identification of further work is needed, and is central to generating predictions of how seagrasses may respond to future Bay environments.