Abstract

1. Methods used to estimate home ranges in restricted environments such as rivers are currently based on standard analyses developed for unrestricted environments. These methods are limited in restricted environments, often overestimating home range area through the inclusion of habitats outside the restricted area. 2. Location data from 23 radio-tagged northern pike (Esox lucius) were used to examine the most appropriate method for investigating differences in home range estimation. The example used was seasonal home range area. Standard home range estimators were tested in addition to a new technique of clipping away the area of home range outside the river. Distances travelled by each individual were calculated using a standard technique and through the use of a midline to ensure the distance calculated followed the path of the river. 3. Pike home ranges and distances travelled were found to vary seasonally. Increased activity for spawning in spring increased home range area and reduced activity in autumn lead to a reduced home range. However, the method used to estimate the home range did influence the results. 4. Cluster analysis, minimum convex polygons (MCPs) and kernels were found to be robust range estimators, each favouring the prediction different range structures such as core area or excursive activity. 5. Clipping reduced out-of-bank error and hence bias generated by this error, aiding the clarity of the results. While use of a midline increased variability this was biological variation that had been masked by the previous method as it did not follow the shape of the river. Thus, excursive movements had been underestimated. 6. Our results show that no single home range estimator always produces the strongest results. Selection of the most appropriate estimator depends on the particular biological question being posed. Clipping and use of a midline can be used to adapt standard methods to better suit restricted environments.