Publication (Technical report): 'Dirty water' models. Predicting community composition for streams in disturbed landscapes. Summary Report. ScD1

Publication Type:	Technical report / Consultancy
Publication Name:	'Dirty water' models. Predicting community composition for streams in disturbed landscapes. Summary Report. ScD1

Reference Information

Norris, R.H.; Coysh, J.; Linke, S.; Walsh, C.; Choy, S. (2000) 'Dirty water' models. Predicting community composition for streams in disturbed landscapes. Summary Report. ScD1. Summary report September 2000, CRCFE, Canberra.

Optional Information

Attached document(s):

- Dirty Water.pdf

Other information:

pdf version attached below.

'Dirty water' models. Predicting community composition for streams in disturbed landscapes. Summary Report. ScD1.

Assoc Prof Richard Norris, Julie Coysh, Simon Linke, Chris Walsh and Satish Choy

Background

A common element of a number of recently developed models for stream assessment is the use of reference sites (Norris & Thoms, 1999). Reference sites are analogous to control sites: they are similar to sites that will be assessed (‘test sites’), but are not influenced by disturbing factors that may be affecting the test sites. Assessment models, such as AUSRIVAS (Coysh et al. 2000) and others derived from RIVPACS (Wright, 1995), use sets of reference sites to predict the structure of biotic assemblages at test sites. This approach is difficult to apply to the assessment of streams in and around large cities.

Because humans tend to settle around lowland streams, few lowland streams orrivers exist that are free of urban disturbances. The nature of urban expansion around large cities makes it unlikely that a relevant set of reference sites could
be found in the same region as a city.

This project aimed to overcome these problems by building AUSRIVAS models using sites from metropolitan and surrounding rural areas, ranging in condition from highly disturbed to minimally disturbed. Instead of using only variables that are not influenced by human activities, as in normal AUSRIVAS models, the ‘dirty water models’ used predictor variables that can potentially be altered by catchment managers. Thus, the models will avoid the need for a large number of reference sites and enable prediction of the effect of management actions on stream community composition.

Aims:

The aims were:

1) to develop predictive models of community composition for macroinvertebrate and benthic diatom communities of the Melbourne, Brisbane and Canberra urban regions, using predictor variables representative of human impacts; and

2) to run scenarios by changing the values of the predictor variables to assess the possible value of such models for predicting the outcomes of management activities.

Search classifications

Administration