Australian River Assessment System

The Australian River Assessment System (AUSRIVAS) was developed in Australia as a tool to assess the health of freshwater rivers and streams.

By looking at the water bugs or macroinvertebrate community, the health of a stream can be assessed as being in near natural condition or impacted. A range of water bugs live in a healthy stream, a so-called reference stream.

Looking at the water bugs of streams can reveal if the community is as diverse as a healthy stream. Such a stream is classed as being in ‘good health’ because it is similar to the reference condition.

In streams where there is a smaller variety of water bugs, the stream is described as ‘impaired in comparison to the reference condition’.

This impairment is commonly due to human activity leading to either pollution or habitat loss for the water bugs. Sampling of water bugs in combination with measuring nutrients, oxygen and other physico-chemical parameters, and a knowledge about the catchment, can give an indication of the reasons for stream impairment.

Using water bugs and therefore biological monitoring provides a more comprehensive assessment of stream health than physico-chemical measurements alone as the water bugs react to all different types of impact.

For example, one stream receives nutrients from the outflow of a sewage treatment plant, and another one receives heavy metal runoff from an old mine site. At a third stream all of the surrounding vegetation was cleared for pastoral land use. 

All of the three streams will show a changed water bug community for those different reasons. A fourth stream now is subject to all of the three impacts hence the water bug community will be more disturbed than in the first three streams.

Many models for reference condition were developed for each state and territory. 

The model used in the Darwin Harbour catchment is called the ‘NT Darwin-Daly region Genus-level Early-dry-season Edge’ model. 

The model is accessible through the internet and maintained by the AUSRIVAS team based in Canberra.

Go to the AUSRIVAS website for more information.

When testing the aquatic health of a stream, the model predicts the reference condition for this stream based on the water bug community. 

Scoring model

The model then compares this expected condition with the observed condition and calculates a score, which is called the observed to expected ratio or OE50.

An OE50 score around 1 means that the observed water bug community is similar to the expected one and therefore equivalent to that of a reference or undisturbed stream. 

A score lower than 1 means that less water bugs were observed than expected and that the community is impoverished when compared to a reference site.

To enable easier reading of OE50 scores, a banding system is used where the Band A is given to OE50 scores around 1, whilst Bands B, C and D are given to respectively lower OE50 scores.

It is possible to obtain a score higher than 1, which means that the observed water bug community is richer than the predicted or expected reference community. 

These are possible ‘hot spots’ of high biodiversity, but can also indicate low level organic pollution in streams naturally very poor in nutrients. 

When a very small increase in nutrient levels occurs, this enables more species to establish.

The definitions of the bands and their interpretations according to the AUSRIVAS user manual are given below.

These same definitions apply to AUSRIVAS models that do not use family taxonomic levels, such as genus-level models.

BandDescriptionO/E TaxaO/E Taxa Interpretations
XMore biologically diverse than referenceO/E greater than 90th percentile of reference sites used to create the model.

More families found than expected.
Potential biodiversity "hot-spot" or mild organic enrichment.
Continuous irrigation flow in a normally intermittent stream.

ASimilar to referenceO/E within range of central 80% of reference sites used to create the model.Expected number of families within the range found at 80% of the reference sites.
BSignificantly impairedO/E below 10th percentile of reference sites used to create the model. Same width as band A.Potential impact either on water and/or habitat quality resulting in a loss of famlies.
CSeverely impairedO/E below band B. Same width as band A.

Many fewer families than expected.
Loss of families from substantial impairment of expected biota caused by water and/or habitat quality.

DExtremely ImpairedO/E below band C down to zero.

Few of the expected families and only the hardy, pollution tolerant families remain.
Severe impairment.

Last updated: 12 October 2017