Development and evaluation of refined annualized individual tree diameter and height increment equations for the acadian variant of the forest vegetation simulator: implication for forest carbon estimates

Tree diameter increment (delta DBH) and total tree height increment (delta HT) are key components of a forest growth and yield model. A problem in complex, multi-species forests is that individual tree attributes such as delta DBH and delta HT need to be characterized for a large number of distinct woody species of highly varying levels of occurrence. Based on more than 2.5 million delta DBH observations and over 1 million delta HT records from up to 60 tree species and genera, respectively, this study aimed to improve existing delta DBH and delta HT equations of the Acadian Variant of the Forest Vegetation Simulator (FVS-ACD) using a revised method that utilize tree species as a random effect. Our study clearly highlighted the efficiency and flexibility of this method for predicting delta DBH and delta HT. However, results also highlighted shortcomings of this approach, e.g., reversal of plausible parameter signs as a result of combining fixed and random effects parameter estimates after extending the random effect structure by incorporating North American ecoregions. Despite these potential shortcomings, the newly developed delta DBH and delta HT equations outperformed the ones currently used in FVS-ACD by reducing prediction bias quantified as mean absolute bias and root mean square error by at least 11% for an independent dataset and up to 41% for the model development dataset. Using the revised delta DBH and delta HT estimates, greater prediction accuracy in individual tree aboveground live carbon mass estimation was also found in general but performance varied with dataset and accuracy metric examined. Overall, this analysis highlights the importance and challenges of developing robust delta DBH and delta HT equations across broad regions dominated by mixed-species, managed forests.