During the summer and autumn of 2003 the procedure described in the document 'An Operational Sampling Program to Determine the Effect of Site Variables on the Natural Regeneration Density of Lodgepole Pine after Harvest' (rev. June 2003) was implemented. The purpose of this project is to develop predictive relationships between a number of easily quantified field variables and the density of naturally regenerated lodgepole pine for sites in the SBSdk and SBSmc2 subzones in the Vanderhoof and Lakes Forest Districts. This required sampling combinations of site conditions in which subzone, humus depth, solar radiation levels, and seed loadings were all varied. The report documents departures in procedures. All departures served to improve the quality and utility of the data collected. A total of 315 - 2 m2 plots were established in 22 harvest blocks, with approximately equal numbers in each subzone. Plots were well distributed across the spectrum of combinations of the aforementioned variables. Charts are used to summarize the distribution of sampling. Some combinations of variables could not be found, either because they do not exist or were not part of the previous season's harvesting program. Challenges in implementing the procedures are summarized and these include issues such as diminishing returns on search efforts, ensuring within-plot homogeneity, and the effect of yearly harvesting program on the range of conditions available for sampling. Analysis of filled seed counts per cone showed that there were significant differences between harvest blocks. Some blocks had very high variance in the number of seeds per cone. The conclusion for anyone conducting ground cone surveys is that the minimum of 30 sample cones per block recommended under the procedures currently used in BC be observed, and it would be prudent to sample more cones per block. The high variability in seed per cone counts has implications for the development of regression formulae. The final form of the equations may require cone counts per plot as an independent variable, rather than estimates of seeds per plot. Poisson regression methods will be required because the dependent variable (seedlings at year 5) is a count variable with numerous low values. Torsten Kaffanke.
Kaffanke, Torsten, T. Kaffanke Consulting Ltd.. 2004. Establishment report: 2003 implementation of sampling program to determine the effect of site variables on the natural regeneration density of lodgepole pine after harvest. Forest Investment Account (FIA) - Forest Science Program. Forest Investment Account Report. FIA2003MR309