The study of the soil seed bank can shed light on past and present states as well as determining a vegetation community's future ability to maintain itself through propagation and to respond to environmental change. It is the role of the seed bank in determining vegetation following mining of heavy mineral sands on North Stradbroke Island (NSI), which formed the basis of this thesis. The presence of many plant species in the soil seed bank of an area can lead to the successful establishment of those species at an early stage of community redevelopment and, in some cases, may be the only way to initially establish a diverse range of species. Up to 75% of all species that re-establish following heavy mineral sands mining on NSI are introduced from re-spread topsoil.
The density of the soil seed bank of the three native vegetation types most commonly disturbed by mining on NSI was low (<116 seeds m-2) in comparison to other Australian ecosystems. The composition of species in the topsoil varied substantially from the standing vegetation and so supplementation of some species in rehabilitation will be required. However, the proportion of species recorded in the soil seed bank that were also present in the wallum vegetation was high (66-80%) in comparison to other vegetation types on NSI and elsewhere. These and other findings reinforce the particular importance of establishment from topsoil collected from cleared wallum areas, where the seed bank is dominated by shrub species, in comparison to other ecosystem types.
Greater than 80% of all native seed was held within the 0-5 cm topsoil fraction for the three vegetation communities studied, although almost half of the seed reserve occurred in the 2.5-7.5 cm topsoil fraction at heath sites. Seed density and composition changed seasonally, with the maximum proportion and species richness of native shrub seeds occurring in October. The seasonal variability in soil seed stores was most likely a function of seasonal flowering patterns of the species of the respective communities. Furthermore, the lowest density and species richness of the seed store recorded at all sites in July, was most likely explained by the fact that much of the soil-stored seed had germinated, in the case of the transient seed bank, or in the case of the persistent seed bank, had either lost viability or been removed by granivores. Thus, topsoil stripping and re-spreading should be completed in late spring in order to coincide with the peak in native seed density and richness and also the imminent wet season.
A second investigation explored the relationship between fire history and the soil seed bank. The results suggested that the proportional species composition of the standing wallum vegetation returned to the approximate pre-fire state about eight years following fire. However, seed bank density remained significantly lower in all burnt sites than in the respective control sites, suggesting that the soil stored seed bank is substantially depleted by wildfire (> half) and that recovery is slow (> eight years). Three years after fire, the soil seed bank also had a greater proportion of seeds of exotic species and tree species. These findings suggest that collection of topsoil in the years following fire will result in a depreciation in both the quantity and the native species richness of the soil seed bank, undermining it's value as a source of seeds for rehabilitation.
A subsequent study, however, suggested that there is a window of opportunity lasting several months between fire and topsoil stripping, in which the germinable seed bank may be elevated The total density of the soil seed bank was significantly greater 44 days after a high intensity wildfire, during which period 142mm of rain was recorded, than either two days after the fire, or 480 days before the fire. While between-year differences in the soil seed bank may account for some variation between before cmd after samples, the difference in the soil seed bank between two and 44-days after fire suggested that the measurable soil seed bank of this wallum area is increased by fire in the short-term. This was most likely due to a greater portion of the existing soil seed bank being stimulated to germinate by in situ natural fire stimuli, as opposed to the artificial heat and smoke treatments that were applied to the samples collected prior to the fire. The results have implications for post-mining rehabilitation on NSI, in suggesting that partial burning of areas prior to vegetation clearing, coupled with immediate topsoil stripping and direct replacement in a previously mined area, may result in significantly greater establishment of fire responsive species, whilst retaining seed of fire sensitive species in the topsoil.
Fourteen recommendations are furnished that aim to maximise the return of propagules of native species following mining and also to highlight where future research should be directed.