Predicting tree-seedling height distributions using subcontinental-scale forest inventory data

•A framework for prediction of tree-seedling height distributions is proposed.•Weibull distribution parameters depend on various stand/site characteristics.•Results suggest that light availability is a key driver of seedling height distributions.•Models can add-value to inventory data containing only seedling count information.

The importance of tree seedlings in determining future stand composition and structure is well-documented in forestry literature. When planned or unanticipated overstory removal events occur, subsequent regeneration success is often linked to the number of seedlings and their height distribution. Yet, in most forest inventories, only counts of seedlings are obtained as it is too time-consuming to measure individual seedlings. To better understand the expected height distribution, models were developed to predict Weibull distribution parameters based on seedling abundance information and stand/site characteristics. A number of these characteristics were found to be statistically significant predictors of the distribution parameters; however, a more parsimonious model using stand basal area, stand age, number of seedlings, and latitude provided essentially the same fit statistics. Models were fitted for all species and for selected species subgroups, but there was generally insufficient data at this time to develop species-level analyses.