Influence of past mortality and measurement thresholds on tree-ring inferred trends in tree distribution and size-growth relationships at a Populus tremuloides stand in the Northwest Territories, Canada

Abstract

Tree size distribution, size-growth relationships, and the spatial distribution of live trees were examined in an even-aged Populus tremuloides stand in the boreal shield ecozone of the Northwest Territories, Canada. A 26-year, annual resolution time series (1985–2010, stand age 48–73 years) was reconstructed from tree-ring derived size, growth and mortality data. Large trees grew disproportionately more than their size, and dominated growth. Contrary to expectation, this resulted in decreasing size inequality over time due to mortality in small trees. Live trees were initially random at all distances, shifting over time to regularity at distances corresponding approximately to mean tree spacing and mean estimated crown width. Restricting the analysis to the subset of trees alive at sampling and further to only trees >10 cm in diameter at breast height (1.3 m) strongly influenced the results, relative to when data on live and dead trees of all sizes were used. These are common sampling restrictions in the literature on tree size distributions and size-growth relationships, and as a result, some published studies may be unreliable. Many others, however, are still reliable even with these restrictions, if inferences are restricted to recent periods only, when size restricted sampling of live trees had less influence on the calculated values.