Regeneration
Simulating the ingrowth of young trees – regeneration – is a key aspect of stand dynamics and is critically important to projections involving multiple rotations, partial thinning and natural disturbance. ESSA has worked with the BC Ministry of Forests and the USDA Forest Service to study and develop a number of approaches to simulating regeneration in FVS/Prognosis.
Beginning in 1998 we evaluated the applicability to the PrognosisBC model of the Northern Idaho variant regeneration model. That analysis clearly showed that the NI regeneration model is inappropriate to most ecosystems in southeastern British Columbia. To begin to address this gap we developed a simple system in 1999 for use by the Prognosis Environmental Indicators model, based on the FVS Event Monitor system. This approach uses expert-based estimates of regeneration based on ecosystem, basal area after disturbance/removal and overstory composition.
In 2004 we began work with colleagues in the USDA Forest Service to adapt an existing logistic regeneration model developed for the Blue Mountains of northeastern Oregon, so that it became possible to operate the model – a standalone executable program which we also developed – from within the FVS simulation. The methods we developed to create this close and dynamic linkage are quite general, and were subsequently and almost simultaneously applied to the development of a regeneration model being developed for British Columbia.
In 2003 we began working with the University of BC and the BC Ministry of Forests to implement a database-driven solution to predict regeneration. This approach is founded on an extensive set of observations of overstorey and regenerating trees collected across a wide range of stand conditions. Overstorey and regeneration variables from stands in this reference collection are correlated with the local conditions of the simulated stand to choose a match from the reference set, using a canonical correlation procedure known as Most Similar Neighbour (MSN) to select the closest match. This statistical procedure, first developed by Moeur and Stage, allows PrognosisBC to locate the most similar stand from the reference collection and then assign the regeneration seen in that reference stand to the simulated stand.
This partnership has so far led to the development of two versions of MSN models, testing the stand selection procedure using different sets of model variables. Over the five year course of this work, UBC forest scientists and graduate students have so far measured regeneration in stands from the Interior Cedar Hemlock, Interior Douglas-fir and Montane Spruce biogeoclimatic zones of BC. As of early 2007, almost 1,400 stand examinations have been included in the database. Besides recording site information and disturbance history, the stand examinations have measured about 24,000 large trees (including snags), 14,000 smaller trees and 47,000 regenerating trees.
Additional Information at Related Sites:
UBC PrognosisBC site
Most Similar Neighbour Software
Regeneration Reports and Publications
