For flora, the site-based biological data sets used in this assessment were drawn from the Flora Information System of Victoria and the Victorian Rare or Threatened Plant Population database (VROTPOP). A description of these databases is given in a separate report. The flora core data fields extracted were reference: (quadrat) number; date; latitude longitude; and species code. The latitude/longitude is accurate to 100 metres.
More than 4,300 sites (quadrats) have been sampled for vascular plants in the Central Highlands since 1979. The sites sampled have been collected in a consistent manner as part of a range of studies including: region-wide studies; pre-logging flora and fauna studies based on forest blocks (e.g. Brown et al. 1989, Lobert et al 1991) intensive sampling of experimental areas (e.g. Mueck 1987); and other studies based on targeted sampling of particular habitats, such as rainforests and heathlands. The quadrat sampling has been largely undertaken by DNRE for the purpose of classifying and describing the variation in native vegetation.
Summary information for each stratum, along with figures relating to the flora site density analysis discussed below, is presented in Appendix A. The flora survey intensity is shown in Map 1 and is discussed below in relation to the environmental strata of the region (Map 2S).
Of the 68 strata generated from the stratification, the 12 strata which occupied less than 100 hectares have not been evaluated in the following discussion. The remaining 56 strata were classified on the basis of flora survey intensity (none, low, moderate, high and very high - see Map 1). The geographic locations referred to in the discussion below relate primarily to the Geographic Representation Units identified in the Proposed Central Highlands Forest Management Area Plan.
3.2.2 Results and discussion
Strata with very high flora survey site density (> 100 sites per 10,000 hectares)
Strata with very high site densities occupied 72,489 hectares or 10% of the total forested land area. The largest of these strata were 1 (in the Yarra and Disappointment geographic units), 39 (in the Bunyip, La Trobe and Thomson geographic units), and 42 (in the Bunyip geographic unit).
Strata with high flora survey site density (40-100 sites per 10,000 hectares)
Strata with high site densities comprised 331,157 hectares or 46% of the total forested land area. The largest of these strata were 36 (throughout the region), 20 (in the Acheron, Marysville and Alexandra geographic units), and 45 (in the Bunyip, Acheron, La Trobe and Alexandra geographic units).
Strata with moderate flora survey site density (10-40 sites per 10,000 hectares)
Strata with moderate site densities comprised 257,778 hectares or 36% of the total forested land area. The largest of these strata were 10 (in the Acheron, Thomson, Matlock, Big, Marysville and Alexandra geographic units), 43 (in the Yarra, Bunyip, Acheron, La Trobe, Thomson, Matlock, Big, Alexandra geographic units), 15 (in the Acheron, Thomson, Matlock, Big geographic units) and 9 (in the Acheron, La Trobe, Thomson, Matlock/Big, Alexandra, Disappointment geographic units).
Strata with low flora survey site density (1-10 sites per 10,000 hectares)
Strata with low site densities comprised 46,639 hectares or 7% of the total forested land area. The largest of these strata is 8 (in the Matlock/Big, Alexandra, Disappointment geographic units).
Strata without flora survey sites
The strata without flora survey sites comprised 7,572 hectares or 1% of the total forested land area. The strata were generally small and scattered, occurring either in the drier, more remote parts of the region, or in the fragmented landscape of the private land/public land interface. The largest strata without flora survey sites were 5 (in the Disappointment geographic unit) and 70 (in the La Trobe geographic unit).
Survey density in largest strata (by area)
Of the 16 strata occupying more than 10,000 hectares each, all but one fell within the moderate, high or very high site density categories, the exception being stratum 8 (in the Matlock, Big, Alexandra, Disappointment geographic units) which fell in the low density category.
The results of the cumulative species curve analysis were expressed as a probability that the next species encountered for a stratum would not have already been encountered. A high probability therefore generally reflected relatively low sampling densities, while a low probability generally reflected relatively high sampling densities.
Although the results of the cumulative species curve analysis tended to mirror those of the site density analysis, the probability also strongly reflected the absolute number of samples collected. Thus a relatively small (in area) stratum with high sampling density but only a small number of samples would be likely to have a higher probability that the next species would be new than a large stratum with the same sampling density but many more samples. Unevenness in sampling of extensive and/or floristically diverse strata is also likely to produce higher probabilities.
The results obtained from this analysis broadly match these predictions, with 89% of the region, comprising 23 strata, having probabilities that the next species is new below 10%. If this threshold is raised to 20%, 95% of the region comprising 28 strata, are included. The largest strata with probabilities greater than 20% are stratum 2 (5,969ha, 40%) and stratum 41 (4,289ha, 66%).
The level of flora survey in the Central Highlands is high compared to other forested regions in Australia. The most intensively surveyed areas are those which have been specifically targeted - the higher rainfall mountain areas which have been the focus of pre-logging surveys, the Dandenong Ranges and the Baw Baw Plateau.
Those areas within the Central Highlands which have relatively low sampling density are in the north-east, including the Matlock and Big River geographic units, and around the public land / private land interface, especially in the Alexandra, Disappointment and La Trobe Geographic Representation Units.
In general, the distribution of flora sites, while highly clustered in timber production areas, is representative of the variation across the region, although additional survey effort in less well sampled environments would improve the overall value of the database.
The following chart shows the percentage of the region in the various sampling density categories for flora: