Can an indicator of river health be related to assessments from a catchment-scale sediment model?

The accumulation of sand and fine gravel (<5 mm diameter) on riverbeds can adversely affect benthic macroinvertebrates, which are good indicators of the ecological health of rivers. The possibility arises, therefore, that predictions of sedimentation could form a useful proxy for indicating the health of a river. The Sediment River Network Model (SedNet) constructs sediment budgets to predict the depth of bed material accumulation (BMA) in each link of a river network. This study tests whether the predicted BMA depth was associated with spatial differences in macroinvertebrate community structure, in the Upper Murrumbidgee River catchment of southeast Australia. There was a significant, albeit limited, correlation. Riffle sites with low BMA depth (0–0.01 m) had a significantly different macroinvertebrate community structure compared to sites with medium (0.01–0.3 m) or high (>0.3 m) BMA depth. At these sites, taxa sensitive to habitat were in greater abundance when BMA depth was low. Additionally, riffle sites with high predicted BMA depth had lower values for three macroinvertebrate community structure measures—AUSRIVAS observed-to-expected (OE) taxa ratio, Ephemeroptera abundance and Plecoptera abundance. There was no significant difference in macroinvertebrate community structure between sites with medium and high levels of BMA depth. Possible reasons for this result are: (1) there may have been few sites in the high and medium categories to provide sufficient statistical power to detect a significant difference; (2) spatial variation in BMA depth within SedNet river links; or (3) only a minimal amount of BMA is required to change community structure. To further define spatial variation in biological damage from BMA, data are required on the spatial scale of variations in BMA depth and damage to macroinvertebrate community structure