• Australia;
  • biodiversity;
  • bird atlas;
  • birds;
  • catchment;
  • multivariate correlations;
  • watershed


Aim  We assessed whether different patterns of land use within similar physiognomic catchments (= watersheds) produced discernible effects on avian assemblages and, if so, whether such effects were related to particular land-use activities (e.g. extensive cropping).

Location  Murray–Darling Basin in south-eastern Australia.

Methods  We used a recently (2007) published physiognomic classification of catchments based on different stream orders as our template. We used a subset of data from the second Birds Australia atlas to calculate reporting rates for each species in each subcatchment. We linked these two sets of data with proportions of major land uses within catchments to identify whether differences in proportions of land uses altered the expected avifauna for catchments of the same nominal physiognomic class.

Results  A significant proportion of the variation in bird reporting rates was explained by the physiognomic classification. Additional explanatory power resulted from including an interaction matrix of land-use covariates. Livestock grazing was a major explanatory variable in classes characterized by more mountainous catchments. Cropping affected avifaunas consistently by producing a more uniform assemblage.

Main conclusions  The physiognomic template was an important determinant of avifaunal composition, but its interaction with land-use variation within physiognomic classes doubled the amount of variance explained. Within a physiognomic class, if one identifies the ‘ideal’ avifaunal composition for that class one can identify land-use mixes that are most likely to be beneficial for the avifaunas of that class and recommend directions for large-scale management objectives vis-à-vis mixtures of land-use types.