Density and diameter distributions of saplings in naturally regenerated and planted coniferous stands in Québec after various approaches of commercial thinning

Abstract

Key Message: A model describing species composition, density and diameter distribution of saplings was developed from operational inventory data. It could be used as an input into growth models calibrated exclusively with merchantable trees to correct some recruitment bias. Important differences in distributions were found between plantations and naturally regenerated stands. Longer-term monitoring would be required to observe the effects of thinning treatments on saplings. Context: Saplings play important ecological and structural roles in forest stands. They also constitute the pool of candidate trees that are responsible for recruitment of merchantable sized trees. However, sapling information is often absent from regular inventory measurements (e.g. where no saplings are measured) even though they could be used as inputs in predicting recruitment in merchantable trees. Aims: The objectives were to develop models describing density and diameter distribution of saplings from operational inventories, e.g. having only merchantable tree inventory, and to evaluate how stand type (naturally regenerated stands and plantations) and various thinning treatments influence these distributions. Methods: Using data from both white spruce (Picea glauca [Moench] Voss) plantations and naturally regenerated balsam fir (Abies balsamea (L.) Mill.) stands having been commercially thinned, a zero-inflated poisson regression was used to model the stand density and a gamma regression to predict the two parameters of the Weibull used to characterize the diameter distribution. Results: Despite the fact that the operational inventory data is often limited (e.g. species, dbh, height), the accuracy of the models was good and unbiased. It could be integrated into growth models to complete missing sapling distributions and help to correct some recruitment bias. Important differences in species composition, density and diameter distribution were found between plantations and naturally regenerated stands, but only a moderate response in diameter distribution was observed with thinning treatments. Conclusion: These models will enable managers to estimate saplings in intermediate aged softwood forests of eastern Quebec using harvesting inventories or National Forest Inventory. Characterization of differences between plantations and naturally regenerated stands will be useful for integrating intensive plantation silviculture with ecosystem-based management. Longer-term follow-up would be needed to better evaluate the effects of thinning treatments.