The Comprehensive Regional Assessment (CRA) provides information about individual flora and fauna species and their habitats, forest ecosystems and communities, and threatening processes. It reviews existing information and the results of additional studies of priority taxa and communities.
The review of existing information has two main elements: an audit of biological records data so as to identify any major gaps in biodiversity information; and a review of information on species and forest ecosystems, the effects of threatening processes and existing or proposed management actions which address these. Chapter 3 discusses the approach to the data audit which was undertaken.
Analysis of data involves the following:
Information identifying survey intensity for flora and selected faunal taxonomic groups in relation to different environmental strata across the Region;
generation of maps of the current distribution of Ecological Vegetation Classes (EVCs) in the Central Highlands and analysis of their reservation status in relation to modelled pre-1750 distributions and current tenures; and
analysis of species and ecosystem responses to disturbance.
The CRA has focused primarily on the ecosystem and species levels of biodiversity because information about genetic variation within species is limited. Ecosystem biodiversity has been dealt with for flora only because there are at present no well-defined faunal ecosystems. Floristic ecosystems are dealt with in detail in the EVC mapping component of the CRA (see Chapter 4).
The biodiversity information presented here is intended to reflect the best understanding of the available information, including information obtained through data audit, expert scientific opinion, analysis of available data. It also points to deficiencies in existing information.
The data presented will be used in the development of the Central Highlands RFA, including configuration of the CAR forest reserve system, and in the formulation of management recommendations.
2.2 Limits to reliability of information
The utility of all scientific information is constrained by the reliability inherent in the method of its collection. The limitations imposed by incompleteness and/or a lack of replication of biological datasets are largely unavoidable, but their impact can be minimised if deficits are acknowledged and well circumscribed. The Chapter on data audit deals with a number of these issues. The following are other important factors relating to the reliability of assessment of biodiversity in the Central Highlands CRA. Many are generally applicable to forested regions of Australia as a whole:
For species assessments,
A lack of data of the biology, population and life history characteristics of taxa can lead to uncertainty in identifying the status of specific threatening processes and identifying remedial action.
The dearth of knowledge about the distribution and characteristics of invertebrate and non-vascular plant species, many of which remain undescribed, means that assessments are necessarily weighted towards the less cryptic elements of flora and fauna.
For Ecological Vegetation Class (EVC) mapping,)
The digital coverages were produced at a scale of 1:100 000. The minimum polygon size defined is approximately 25 hectares.
Vegetation associations tend to merge along a continuum, so that a line on the vegetation map often represents an ecotone rather than a discrete boundary. Discrete boundaries do, however, occur in some situations; for example, the boundary between closed forest and sedgelands.
Most of the vegetation boundaries are derived directly from the photo interpretation typing coverage, which is forest structure based on canopy height and canopy cover. Dominant floristics are attributed to each polygon on the basis of the site data present, expert knowledge, aerial photo-interpretation of forest types, and extensive field validation.
The pre-1750 vegetation reconstruction was conducted using the best available environmental modelling, remnant site data and expert knowledge. This component of the assessment was, however, impossible to validate in the field in most places.
Biodiversity assessment relies on having adequate information about the distribution of species. It is important to know whether or not surveys undertaken for species or groups of species have been adequately distributed across the range of environments represented within the region. As part of this assessment, analyses were undertaken to determine where surveys for biodiversity were undertaken in the Central Highlands region, which species were targeted, and whether survey sites are reasonably distributed to detect most species in most geographic or environmental components. The results of these analyses were used to highlight gaps in information and identify those areas which still require further survey work.
The data review process involves systematically working through databases to determine the adequacy of existing site-based biological data for identifying priority areas and data gaps to be filled through additional survey work. The data review relies on expert knowledge and professional judgment but is supplemented by explicit analyses where appropriate.
The first step in the data review process is to select only those survey data which meet required standards of accuracy, precision and reliability. This allows a degree of confidence when analysing the distribution of species.
The next step involves assessing environmental and geographic representation by sites from accredited data sets is to stratify the region. The environmental variables on which the stratification procedure should be based are those thought to either directly or indirectly influence the spatial distribution of species. These include solar radiation, temperature, terrain wetness, nutrient status, ground water, rainfall, elevation, slope, aspect and geology. The strata developed may represent either classes of single variables, such as temperature or rainfall, or may consist of environmental units developed from the integration of variables using objective or intuitive multivariate classification analyses.
The distribution of flora and fauna survey sites among strata can initially be analysed in terms of the size of each stratum and its geographic distribution. The density of survey sites in each stratum is calculated and strata with no sites or low site densities are identified as possibly requiring future field work. Ideally, the density of survey sites in each stratum should be a function of the stratum's total species richness and spatial heterogeneity. These parameters can be examined by using species data from existing sites to derive species accumulation curves and associated statistics for each stratum. Species accumulation curves are frequently used to assess sampling adequacy in a given area by graphically illustrating the rate of addition of new species to a sampling unit with repeated sampling events. Curves that show an asymptote indicate the full complement of species in the area being investigated have been sampled, assuming an unbiased distribution of adequately sampled sites.
Because most, if not all, strata will be made up of numerous geographically discrete areas (substrata), it is necessary to also examine the distribution of sites between substrata within strata. Sites should be replicated across the geographic extent of each stratum. Where this is not the case, a geographically representative sample of substrata may be identified for further survey work (Cocks & Baird 1991). In the case of very large substrata, the distribution of existing flora and fauna survey sites should be examined for spatial biases resulting from the design and objectives of the original surveys and logistic constraints (for example, sampling along roads).
A data audit methodology toolkit was developed by the Environment Forest Group within the Department of the Environment, Sport and Territories to assist assessment of the quality of data to be used in regional biodiversity assessments. The toolkit has been developed as an ARC/INFO geographic information system application with a menu interface that incorporates ARC/INFO advanced macro language scripts menus and functions, in addition to system scripts and other programs. The methodology helps users to:
identify spatial, environmental and temporal biases in the survey data, and
ascertain sampling adequacy for species groups within a region.
The toolkit is designed to perform the following tasks:
develop a regional environmental stratification;
create ARC/INFO point coverages from site text files and add species attributes;
intersect sites with a regional environmental stratification and calculate statistics;
generate cumulative species curves and predicted species richness statistics;
create a histogram showing the proportion of total land area and the proportion of total sites of each stratum;
produce maps of the regional environmental stratification and survey intensity; and view and print graphs and maps.
Process of developing the stratification
The process for identifying potential stratification variables and deciding on a final set of variables and their respective cut-off points was achieved through a joint Victoria-Commonwealth Workshop involving both flora and fauna specialists. This process was aided by the examination of hard-copy colour A3 maps of all of the variables being considered for the stratification and summary statistics and frequency histograms relating to each of the variables. Ideally, a number of different stratifications could be produced and assessed. However, for the Central Highlands, only one stratification was developed for the region as a tool for assessing spatial bias in the available flora and fauna site data. The stratification of the Central Highlands region was based on spatial estimates of climate and substrate (lithology). The sources and derivation of these data are outlined below.
Methods have been developed to estimate climate at any point in a landscape, given the availability of topographic and meteorological data. ‘Climate surfaces’ fitted to a Digital Elevation Model provide spatially reliable estimates of mean monthly climate attributes derived from long-term meteorological station records for any given longitude, latitude and elevation (Hutchinson and Bischof, 1983; Hutchinson et al., 1984; Hutchinson, 1989, 1991a, 1991b). Currently, the estimated standard errors are 0.5o Celsius for monthly mean temperature and less than 10% for mean monthly precipitation (Hutchinson, 1984; Hutchinson et al., 1992).
Key climatic attributes which describe the range, seasonality and extremes of climate (temperature, precipitation and radiation) of the Central Highlands region were calculated for each cell in the nine second elevation grid using the software package ANUCLIM (McMahon et al., 1995). Of the 24 climatic variables calculated for the Central Highlands region, mean annual precipitation (with a range of 559 to 1,802 mm), mean maximum temperature of the warmest month (17.4 to 29.4oC) and mean minimum temperature of the coldest month (minus 4.4 to plus 6.4oC) were selected for use in the stratification of the region. Each of these climatic variables was then divided into even intervals within the range exhibited in the Central Highlands (Table 3.1).
Lithology (rock type)
Using the Land Systems coverage of Victoria at a 1:250,000 scale , lithological types were aggregated in an attempt to lump those geological groups which showed the greatest effect on vegetation distribution, that is fertility, drainage and landform. Seven of these classifications were represented in the Central Highlands (see Table 3.1). An eighth category for areas of undescribed lithology was also included.
Deriving the regional stratification
The environmental stratification was based on the three climate attributes and one lithology attributes described above and estimated for each 250 x 250 metre grid cell. A total of 288 individual units or strata are possible when the four classes of annual precipitation, three classes of minimum temperature of the coldest month, three classes of maximum temperature of the warmest month and 8 classes of lithology are combined. Of the potential 288 strata, only 80 occurred in the Central Highlands, ranging in area from 4 to more than 100,000 hectares. Clipping strata classes with an overlay of a forest/non-forest classified coverage reduced the number of applicable strata to 68. The spatial arrangement of these strata across the Region is shown in Map 2S. This environmental stratification was subsequently used for the analyses of flora and fauna databases presented here.
Table 3.1: Attributes and classes used in the Central Highlands environmental stratification.
mean annual precipitation
Central Highlands range = 559 - 1802mm
Low = 559 - 870mm
Moderate = 871 - 1180mm
High = 1181 - 1491mm
Very high = 1492 - 1802mm
mean minimum temperature of coldest month
Central Highlands range = minus 4.6 - 6.4C
Low = minus 4.60 - minus 0.94C
Moderate = minus 0.93 - 2.74C
High = 2.75 - 6.40C
mean maximum temperature of warmest month
Central Highlands range = 17.4 - 29.4C
Low = 17.4 - 21.4C
Moderate = 21.5 - 25.4C
High = 25.5 - 29.4C
a = coarsely textured unconsolidated deposits: low fertility