Predicting barriers to the movement of Macquarie perch (Macquaria australasica)

Abstract

Barriers to the migration of fish limit access to suitable habitat and have been identified as a major contributor to the decline of freshwater fish species. GIS and fine scale elevation data were used to create a model to identify physical barriers to fish migration on the Cotter River in the ACT, Australia. The effect of spatial resolution on the efficacy of the model was also examined. Elevation data gridded at 1m, 5m, 10m and 20m were used to simulate the terrain. Two types of models were analysed for barrier identification: logistic regression and non-hierarchical clustering. Logistic regression based on 1m resolution terrain data was found to create the best models. The 1m model using transect averages predicted 88% of the barriers identified in the field. The relative steepness parameter was found to be a key variable for predicting barriers. A general trend of increased accuracy of barrier prediction of field surveyed barriers with a decrease in grid cell size was found for grid spacing less than 20m. The model allows for a semi-automated process to rapidly assess catchments for instream barriers to fish passage. The predictions from the model provide the means for prioritizing areas containing possible physical barriers in the Cotter River. It is intended to be used in conjunction with traditional barrier identification. Show more