Modeling individual tree and snag dynamics in the mixed-species Acadian Forest.

摘要

Forest growth modeling has a long tradition of development and application in even-aged stands targeting single-species plantations. Modeling efforts in mixed-species stands that contain uneven-aged stand structures are much more recent. Serving as a transitional zone between the boreal and eastern broadleaf deciduous forest types, the Acadian Forest found throughout Maine and the Canadian Maritime Provinces is host to a wide variety of tree species that form complex stand structures. This study validated existing and developed component equations that comprise a widely-used individual tree growth and yield model in the northeastern US and Canadian Maritime provinces. An assessment of deadwood stocking was conducted and models were developed to improve our understandings of standing deadwood dynamics as they relate to silvicultural treatment, species, and stand conditions in these forests.;Three key submodels of the Northeastern variant of the Forest Vegetation Simulator (FVS-NE) were benchmarked and calibrated using remeasurement data obtained from a national forest inventory, suggesting improvements that could be made in model structure and methodologies. Using 29 years of remeasured tree data from the US Forest Service Penobscot Experimental Forest (PEF), long-term projections suggested that modeling diameter (dbh) increment as opposed to basal area increment reduced root mean square error by up to 24% for the primary species in the region. Advances in methodologies for fitting individual-tree increment equations in mixed-species stands were made by including species as a random element of the regional equations. Using an extensive regional database compiled with over 1.15 million dbh remeasurements, dbh and maximum height (ht) increment submodels were fit using nonlinear mixed-effects models that employ tree species as a random effect. Predictions of dbh and ht increment represented an improvement over currently-used models in FVS-NE and reduced the complications of portraying growth dynamics in mixed-species stands with multi-cohort stand structures.;Snag measurements totaling 2,751 observations collected across eight silvicultural treatments on the PEF indicated the highest volume in standing deadwood occurred in a nonharvested reference area (23.6 m3ha -1) and lowest volume in a 5-year selection cutting (5.2 m 3ha-1). Methodologies highlight the effectiveness of models that relate standing deadwood abundance variables common to traditional forest inventories. Results provide insight into snag survival and decay dynamics for the species in the region and further our knowledge about the roles that deadwood dynamics play in the regional forest carbon cycle.