Australia's threatened species-Box 4

	Published by Australian Academy of Science
Australia's threatened species Box 4 \| The mathematics of reserve systems

One possible way to conserve biological diversity is to create reserve systems that contain a representative sample of habitats. In recent years, various governments in Australia have stated their aim that at least 15 per cent of all habitats that existed before European settlement (often called 'pre-1750 vegetation') should be represented in conservation reserves.

But to do this we need to know, first, the range and extent of habitats that were present before colonisation, and, second, what habitats are present now and where they are located. We could go out and survey every square metre of land in the country, but that would take too long and cost far too much money – and it wouldn't tell us about those habitats that once occupied land which is now cleared.

A different approach is to use mathematics. By using information collected in surveys, it is possible to develop mathematical relationships between different habitats and the environmental factors that determine where they occur – such as rainfall, temperature, soil type and position in the landscape. These relationships, or models, can then be used to predict habitat distribution in areas that have not been surveyed. It's a case of using the known to predict the unknown.

CSIRO scientists in collaboration with the New South Wales National Parks and Wildlife Service, have done this kind of modelling in the forests of southeastern New South Wales. These forests have been subject to decades of dispute between conservationists, the timber industry and agriculturalists. As the political arguments have raged on, ecologists from different agencies have slowly been collecting data from large numbers of vegetation survey plots dotted throughout the forests.

The scientists brought together the results of over 10,000 of these plots. They then used a statistical technique called regression analysis to relate the presence or absence of 96 eucalypt species to a wide range of environmental factors. In this way, they produced computer-generated mathematical expressions of the relationships between the eucalypt species and the environment.

For the relationships to be useful in predicting the occurrence of habitats, one more tool was needed – a geographic information system, or GIS. This consisted of a computerised dataset containing information about the region, such as topography, soil and climate. The regression analysis showed the scientists the combination of factors that was needed in a location for a particular species of eucalypt to occur. They then used the GIS to find out where in the region such combinations of factors could be found.

Using this technique, they produced a map predicting the pre-1750 distribution of tree species associations – or habitats – in the southeastern forests. The map has since been used by the NSW National Parks and Wildlife Service to add some 100,000 hectares to reserves in the region, including a number of habitat types that were previously under-represented.

CSIRO scientists collaborated with the NSW National Parks and Wildlife Service, the Department of Land and Water Conservation, and local community groups to predict the pre-1750 distribution of vegetation in the central Lachlan catchment in western New South Wales. The map produced from the study could be used to aid remnant vegetation management and revegetation efforts in the region.

Further reading

Bending lines with the mind (by Alastair Sarre),
Ecos, No. 91, 1997, pages 34-35

Related site