Unrealistic animal movement rates as behavioural bouts: a reply
Chris J. Johnson, Ecosystem Science and Management Program, University of Northern British Columbia, 3333 University Way, Prince George, BC, V2N 4Z9, Canada. E-mail: firstname.lastname@example.org
- 1Johnson et al. (Journal of Animal Ecology, 2002, 71, 225–235) proposed a technique for stratifying the movements of ungulates into small- and large-scale behaviours. They identified movement paths for woodland caribou and fitted a nonlinear curve to the log-frequency of movement rates. They assumed that slow small-scale movements were correlated with foraging activities in patches and faster large-scale movements occurred when caribou moved between patches.
- 2Nams (Journal of Animal Ecology, 2006, 75, 298–302) reviewed the assumptions and tested the technique presented by Johnson et al. (2002). Simulated animal movements resulted in rates inconsistent with the data of Johnson et al. (2002) and the distribution necessary to fit the nonlinear curve. Nams (2006) challenged animal movement as suitable for the technique and concluded that sampling interval would confound results.
- 3We evaluated Nams's (2006) criticisms with movement data collected for caribou, moose and mountain goat. All three species demonstrated the required distribution of movement rates and sampling interval had little influence on the criterion used to identify scales of movement for a range of woodland caribou data. In addition, we tested the sensitivity of the curve-fitting model to the width of the frequency interval for the log-frequency plot of movement rates. We noted bias in the rate criterion, but the scalar relationship was consistent among interval widths.
- 4The discrepancy in movement data presented by Nams (2006) and Johnson et al. (2002) is likely the result of different movement processes. The movements of simulated animals did not encompass the full range of behaviours typically observed for ungulates. Our analyses and those of Nams (2006) provide little evidence to universally reject the nonlinear curve-fitting model and the results of Johnson et al. (2002). However, we caution against blind application of the technique, as not all movement processes are suitable and the scale of movement must be consistent with the scale of the behaviour.