Regeneration and small tree height increment in the interior Douglas-fir zone of the Rocky Mountain trench

Data were collected in the Kootenay Dry Mild Interior Douglas-fir subzone variant (IDFdm2) of the Interior Douglas-fir biogeoclimatic ecological classification zone, Invermere Forest District. A total of 111 plots were sampled for plot-level data; 25 of these were subsampled for additional substrate information and spatially mapped. Objectives were to: 1) examine the explanatory ability of modelling techniques based on plot-level predictors and 2) examine the utility of spatial and substrate data for improving our knowledge of understory stand dynamics. The utility of imputation techniques as a method of predicting regeneration abundance was examined using the full data set. Imputation techniques performed moderately well. Subsampled data were used to examine regeneration abundance relative to substrate availability and spatial and aspatial competition and release indices. Substrate abundance was demonstrated to vary by moisture condition, as did substrate suitability for regeneration. Spatial indices showed interesting patterns between germinants and competition indices, and established regeneration and release indices. The spatial relationship between localized aggregations of regeneration and overstory trees, examined using Ripley's K(t) point pattern analysis, provided statistically significant results, but trends were generally inconsistent and could not be explained by available information. The utility of regression-based modelling as a method of predicting small tree height increment was examined using the full data set. Height increment models performed reasonably well. Small sample sizes were a limitation for some species. Subsampled data were used to examine spatial competition and release indices for model improvement. Spatial indices showed little potential for improvements over aspatial, plot-level indices, but provided an indication of the type of competition affecting small tree height increment. Both regeneration imputation and regression-based height increment modeling performed well, even with small data sets. Additional data for more sparsely sampled species are desirable. Further focus on substrate conditions and their effect on regeneration, within a multivariate context and for all species, is warranted. Subsurface properties are an unexplored avenue that merits investigation, since the analysis points to below-ground competition for resources. Finally, spatial tools may be more illustrative if data are gathered for both horizontal and vertical structures, or at larger scales.Forestry, Faculty ofGraduat