A fifty-year reconstruction of annual changes in the spatial distribution of Pinus banksiana stands:: does pattern fit competition theory?

We used tree-ring reconstruction data to study changes in the spatial pattern of live and dead trees at an annual resolution over a 50-year period at four unmanaged, even-aged fire origin jack pine (Pinus banksiana Lamb.) stands in Saskatchewan and Manitoba, Canada. Previous studies of the spatial pattern in P. banksiana have either looked at only a snapshot from a survey done at a single point in time, or repeated measurements of permanent plots taken at 10-year intervals. With annual data, we could examine detailed changes in spatial patterns and relate these to events during stand development and external disturbances. Trees were initially clustered at all sites, but at different distances at each site, most likely because of variability in seedbed distribution at stand initiation. Clustering disappeared over time at all sites, and at a similar mean tree spacing at each site. However, significant regularity only appeared sporadically at one site, indicating that competition with neighbours was not the only factor influencing changes in spatial pattern. At two of the four sites, clustering disappeared suddenly at the same time that mortality rate reached a peak, in one case also coinciding with a jack pine budworm (Choristoneura pinus pinus Freeman) defoliation event. Dead trees were also initially more clustered than the distribution of all trees, but at different distances than the clustering of live trees. This also disappeared over time so that dead trees were eventually a random sample from the distribution of all trees. After the peak of mortality had passed, factors other than competition were determining the dynamics of these forests.