Context. Changes in global climate and evidence of species’ responses to these changes have increased interest in relationships between climate variables and species demography and distributions. Although an important tool for many ecological questions, large-scale climate indices fail to provide the spatial resolution necessary to investigate drivers of change across small spatial scales. Climate variables that describe yearly climate variation at large spatial extents and small spatial grain are needed.
Aim. Here we develop a model for snow depth using snow water equivalent (SWE) data, which are readily available in a number of formats, to be included in a more general climate index. We use an existing winter severity index (WSI) for white-tailed deer to test the performance of the model.
Methods. We obtained data for 13 weather stations from north-western Canada, reporting both SWE and snow depth. We accumulated a snowpack from daily SWE of snowfall and then tested two methods for converting the SWE of the snowpack into the snow portion of the WSI. We then generalised the model for application to the northwest forest climate region.
Key results. Coefficients of determination (R2) relating the actual and predicted snow depth portion of the WSI ranged from 0.41 to 0.78, with only three stations being below 0.50. Coefficients of determination (R2) relating the actual and predicted WSI for the northwest climate region ranged from 0.58 to 0.88.
Conclusions. The SWE model predicts the snow portion of the WSI well for most stations and, when incorporated into the full WSI, provides a good measure of relative winter severity across space and time for most stations.
Implications. The method developed here could be applied elsewhere, where snow depth is an important factor in species ecology. The benefit of this approach is a comparatively simple method that maximises the use of widely available SWE data in place of snow-depth data.
