REVIEW OF VEGETATION CLASSIFICATION AND MAPPING SYSTEMS UNDERTAKEN BYMAJOR FORESTED LAND MANAGEMENT AGENCIES IN AUSTRALIA

This paper provides a detailed review of the major vegetation classifi cation and mapping systems used by the management agencies with primar y responsibilities for forested land in Australia. It focuses on the c larification of vegetation units and methodologies used. The paper als o provides a comparison of the different nomenclatures against a simpl ified standard to show how the different systems relate to each other. In Australia, different systems for classifying and describing forest vegetation have been developed by various forest land management agen cies to suit their own situations. Most vegetation classification syst ems reviewed are similar in using floristics and structure as the two primary elements in classifying vegetation types, and all use growth f orm (physiognomy) to distinguish vegetation units. The classification and mapping systems for wood production purposes differ from those for conservation and environment purposes in several aspects-wood product ion classifications emphasise commercial tree species and/or attribute s such as height, whereas conservation classifications emphasise ecolo gy, vegetation coverage, and the importance of understorey species. Th ere are three broad strategic approaches in the vegetation classificat ion programs being undertaken by the major forest land management agen cies in Australia: (1) conducting a single classification across the w hole of the agencies' land in a State; (2) conducting a vegetation cla ssification at the regional level, but using the same methods in each region; and (3) using different methods depending on the specific obje ctives of individual studies. This paper highlights the value of accur ate quantitative measurements in the field. For example, for the two k ey structural attributes of height and crown density, the measured raw data can be accommodated by a number of different classification sche mes whereas if the raw data consists of only records by predetermined classes, then such accommodation is difficult and loses precision.