Differing effects of catchment land use on water chemistry explain contrasting behaviour of a diatom index in tropical northern and temperate southern Australia Bruce C. Chessman a, *, Simon A. Townsend b a New South Wales Department of Environment, Climate Change and Water, PO Box 3720, Parramatta, NSW 2124, Australia b Charles Darwin University, Darwin, NT 0909, Australia 1. Introduction The raw data collected in freshwater biomonitoring are often summarised as numerical indices, dozens of which have been created worldwide (Ziglio et al., 2006). Biological indices provide at least a semblance of objectivity and quantification, and are easier to communicate to non-biologists than lists of taxa and their abundances. However, the valid interpretation of index scores is not always straightforward because the same score can sometimes arise through quite different circumstances. For example, the macroinvertebrate-based O/E (observed-over-expected) index of the Australian River Assessment System (AUSRIVAS) was devel- oped in order to enable standardised, Australia-wide bioassess- ment of rivers (Davies, 2000). Lower scores of this index have been associated with a range of human-generated or human-modified stressors such as high concentrations of metals in water and sediment (Sloane and Norris, 2003), the presence of dams immediately upstream (Marchant and Hehir, 2002), and observed or modelled accumulation of fine riverbed sediments (Metzeling et al., 2006; Harrison et al., 2008). However, low scores have also been attributed in particular circumstances to natural influences such as spates, droughts, cyclonic activity, naturally saline conditions and naturally low productivity (Sheldon, 2005; Halse et al., 2007). Conversely, high scores can be associated with near- pristine conditions, but can also occur despite substantial exposure to anthropogenic stressors (Chessman et al., 2006; Halse et al., 2007). In small-scale applications, the range of potential influences on index scores is likely to be constrained, and resources may be sufficient to obtain a good understanding of index behaviour, leading to confident interpretation. For example, Besley and Chessman (2008) showed that for a series of small, nearby basins, the SIGNAL biotic index for stream macroinvertebrates responded in a predictable manner to the interaction between cessation of direct sewage discharges and persisting urban drainage. Therefore this index can be used as a biological indicator of the success of pollution abatement within this spatial context. However, in very large-scale applications, the likelihood of multiple cause-effect pathways and limited capacity to acquire contextual information can make the interpretation of indices less definitive. Across the whole of Australia, high SIGNAL scores are characteristic of cooler streams with low concentrations of suspended and dissolved substances, but low scores can indicate many kinds of natural or anthropogenic physical and chemical enrichment or contamina- tion (Chessman, 2003). In this paper we explore the large-scale behaviour of a recently developed biotic index, the Diatom Species Index for Australian Rivers (DSIAR: Chessman et al., 2007). DSIAR scores are calculated as averages of numerical ‘sensitivity values’ of the diatom species Ecological Indicators 10 (2010) 620–626 ARTICLE INFO Article history: Received 9 March 2009 Received in revised form 15 October 2009 Accepted 18 October 2009 Keywords: Diatom DSIAR Index Indicator Stream Water quality ABSTRACT The DSIAR biotic index for freshwater diatoms, regarded as a potential indicator of impact from agricultural and urban land use on rivers in temperate south-eastern Australia, did not correlate significantly with an index of catchment condition in a tropical region of northern Australia. However, the relationships between the index and water chemistry, especially pH, salinity and concentrations of nitrogen and phosphorus, were consistent in both regions. The variable relationship between the index and catchment conditions can be explained by differing effects of catchment land use on stream-water chemistry in northern and southern Australia. In the south, land use has commonly resulted in increases in stream pH, salinity and nutrients, whereas in the north its impact on pH and salinity appears weak. These findings emphasise the need to interpret biological and ecological indices in the context of the varying causal pathways by which human activities affect stream ecosystems in different circumstances. ß 2009 Elsevier Ltd. All rights reserved. * Corresponding author. Tel.: +61 2 9895 7154; fax: +61 2 9895 7834. E-mail addresses: Bruce.Chessman@environment.nsw.gov.au (B.C. Chessman), Simon.Townsend@cdu.edu.au (S.A. Townsend). Contents lists available at ScienceDirect Ecological Indicators journal homepage: www.elsevier.com/locate/ecolind 1470-160X/$ – see front matter ß 2009 Elsevier Ltd. All rights reserved. doi:10.1016/j.ecolind.2009.10.006