Freshwater Protected Area Resourcbook

5.3 Inventories and reserves:

To what extent are representative examples of Australian freshwater ecosystems protected within existing networks of protected areas? This is an important question, and one of the key questions behind any process for freshwater reserve identification and selection. It is important to note that terrestrial protected areas do not always protect imbedded freshwater ecosystems – for example the Snowy Mountains Hydroelectric Scheme lies in part within Kosciusko National Park. Other key questions relate to feasibility: land ownership and control, catchment land use, and the presence of threatening processes and possibilities for their management.
We know where our protected areas are (national parks, for example) – but how are different types of freshwater ecosystem distributed across the Australian continent, and how are they distributed in relation to the reserve network? Comprehensive inventories need to be developed covering all freshwater ecosystems to answer this question.
Reserves also form a layer in the ‘value’ information held within inventories of freshwater ecosystems. For example, Victoria’s 11 Ramsar sites have a surrogate ‘highest value’ (international importance) rating amongst 159 designated wetlands of ‘national importance’ – which themselves sit within a larger dataset of the State’s 13,114 listed wetlands. Victoria’s planning framework takes these different levels of value into account when assessing development applications⁹⁹.

5.4 Inventory construction

At present there are no accepted national frameworks (either funding or theoretical) which seek to provide consistency across the Australian continent in regard to the development of comprehensive freshwater ecosystem inventories.
Inventories generally use methods of classification, or ways of allocating different ‘types’ to different ecosystems (or – at a lower level of detail – habitats). Classification theory depends on the assumption that areas can be grouped which are alike; ie: areas within each group are more similar to each other than they are to areas which have been placed in different groups. Measures of similarity and difference are made by examining attribute values (water depth, for example). Wetland attribute values, at a particular site, generally fall within predictable ranges. Typically, Australian’s highly variable climate results in characteristic variations in attribute values over time, at any particular site.
Ecosystem classification is a tool for studying, managing, and communicating information about particular types of ecosystem. It typically involves defining ecosystem types, to which individual ecosystems can be allocated. Classification is a fundamental component of inventory; underpinning mapping and reporting of ecosystem occurrence by type.
Various Australian authors have reviewed classification and inventory issues for wetland environments. Notable examples are Barson and Williams (1991) and Pressey and Adam (1995). More recently Duguid et al. (2003) have reviewed these issues with particular reference to arid zone wetlands. The following summary of some of the issues comes from Duguid et al. (2003).
Pressey and Adam (1995, p.87) included as classification “any attempts, intuitive or numerical, to group wetlands with common characteristics or to identify the types of environments and biota they contain”. They stated the importance of seeing classifications “in two ways: (1) as hypotheses about the way in which features of wetlands are arranged in space and time; and (2) as responses to the need for particular types of information for particular purposes, dependent also on the geographical scale of the study and the variability of the wetlands.” (Pressey & Adam, 1995, p.95).
Similarly, Barson and Williams (1991) listed the following uses of wetland classifications:

• 
  description of ecological units – with certain homogeneous natural attributes;
  
• 
  aiding resource management;
  
• 
  inventory and mapping; and
  
• 
  aiding communication by promoting consistent terminology.
  

Methods of classification depend on the availability of information about each ecosystem. More detailed knowledge can support more detailed classification approaches. Typically, such knowledge is not uniformly available. We may know a great deal about highly visible ecosystems near centres of population, for example, but little about remote and inaccessible ecosystems.

The traditional approach to this dilemma is to use nested hierarchies of classification approaches. As more information becomes available, more detailed classifications are invoked. For example, a first cut may simply be to divide aquatic ecosystems into five broad categories: (a) rivers and streams, (b) inland wetlands, (c) estuaries, (d) shallow marine systems, and (e) aquifers (or subterranean ecosystems). To continue the example, rivers and streams could then be subdivided into five categories which take account of key ecosystem variables: tidal, lower catchment perennial, upper catchment perennial, undefined catchment perennial, and intermittent. In turn, each of these categories may be subdivided – for example by substrate type or dominant vegetation type.
The key environmental attributes that are generally used to classify the variety of wetland environments are:

• 
  geomorphic – landform, size and substrate;
  
• 
  hydrological regime (permanency, frequency, duration and depth of inundation);
  
• 
  water and soil salinity;
  
• 
  vegetation type and/or characteristic species.
  
• 
  scale and spatial arrangement (including complexity or uniformity); and
  
• 
  source of water;
  

Climate is usually excluded if analysis is conducted on a bioregional (or sub-bioregional) basis, on the assumption that climatic variation can be captured by protecting similar ecosystems across bioregions. It should be born in mind that bioregions defined according to the protocols of the Interim Bioregionalisation of Australia (IBRA) do not attempt to account for micro-climatic variation: there can be significant climatic differences on opposite sides of a mountain, for example. While the IBRA design principles attempt to capture regions of relatively homogenous climate, this may not always be achieved.

Continuing with examples, a freshwater permanent deep wetland could be subdivided into finer categories, depending on the biotic assemblages found in different locations. Faunal biota classifications might consider dominant or keystone species¹⁰⁰. Floral classifications may refer to species dominating energy or nutrient pathways.
A discussion of different approaches to wetland classification may be found in Finlayson (1999). This book describes an outline of an approach for wetland inventory that overcomes some of the difficulties of classification. It supports the basic water regime and landform categorisations, with other detail added as necessary. Using this approach, core data are collected for each wetland and arranged in a database, free of classification categories. This data can then be analysed as needed in a variety of classification formats (or outside these formats as needed for a particular application). This approach has been used as the basis for the Asian Wetland Inventory (see www.wetlands.org). Many of the features are also included within the draft Ramsar framework for wetland inventory (see papers available on www.ramsar.org).
These approaches are multi-scalar with a hierarchical data format. That is, depending on the scale and/or objective chosen for the particular study, the inventory can be undertaken within a linked framework with cascading data fields. It can operate either top-down or bottom-up.
The classification system used by the Queensland Wetland Inventory Program (Blackman et.al. 1992) was the Australian forerunner of this approach¹⁰¹, and is the best example of the use of this technique in Australia. The Queensland handbook describes both the theory behind the classification method, as well as techniques for field data collection. The Queensland Wetland Inventory, while not complete, is the most rigorous and comprehensive of any Australian State in terms of scope and structure.
An important question is: how large should a system of protected areas be to preserve most of a bioregion’s biodiversity? In other words, could 90% of the biodiversity be protected within a system of reserves holding 20% of the bioregion’s area? Information on the way in which biodiversity is distributed across the landscape is needed to answer this question. In this context biodiversity is difficult to measure directly¹⁰²; the usual approach is to use the finest level of ecosystem information available (as a surrogate for measuring biodiversity) – which is usually habitat attribute¹⁰³. Blackman’s work includes multivariate attribute analysis providing measures of difference between groups of wetland aggregations – a useful measure to address this issue.
The durability of reserves also needs to be considered. Island biogeographic theory predicts that small (and even medium sized) reserves will lose many species through local extinction events if they are isolated from similar habitat.
The NZ Department of Conservation has been undertaking studies of environmental differences for around 5 years now, where differences are mapped at a 30m pixel level using climatic and landform attributes. Again, these attributes (or groups of attributes) can be viewed as a surrogate for biodiversity. Such studies can indicate how biodiversity is likely to be distributed nationally, and with respect to the nation’s reserve framework (Department of Conservation NZ, 2001a, 2001b). Such data need to be checked against field surveys, of course. As a first step it provides a powerful tool for the strategic planning of biodiversity conservation measures.
If ecosystems within a bioregion are very similar, a high level of protection (for the region’s biodiversity) may be (theoretically) obtained by protecting a relatively small area. This is usually not the case, reinforcing the importance of off-reserve biodiversity protection measures.
Provisional classifications for Queensland wetlands and deepwater habitats (see Blackman reference above) are included at the close of this chapter. A list of different approaches to river classification can be found in Appendix 5 below.