Ole Martin Bollandsås  presented a lecture on a predefined topic and defended his doctoral thesis on Friday 18, January 2008. The topic for the trial lecture was “Predicting the development of uneven-aged stands”  and the title of his thesis was “Uneven-aged forestry in Norway. Inventory and management models”.

Abstract

This thesis presents management models applicable for uneven-aged forest structures in Norway. The framework is a matrix model consisting of several sub-models estimated from National Forest Inventory (NFI). The matrix model operates on diameter classes, projecting future outcomes by means of transition probabilities. Within the thesis, models for diameter distributions, recruitment, diameter growth, mortality, and tree height were developed. These four latter models were implemented in a computer tool and organized as a matrix model.

Diameter distributions are not registered in ordinary forest inventories. Such distributions are necessary as input data for diameter class models such as the one presented in Paper IV of the current thesis. A model for retrieving diameter distributions from laser scanner data was developed in Paper I. The model was adapted to uneven-aged forest structures and was performing well within the range of the model development data. Even if the diameter distribution of other forest structures were not as well predicted, estimated stand volumes derived from the predicted diameter distribution were quite accurate irrespective of forest structure.

Usually, diameter distributions are limited to a lower diameter limit of for example 3 cm. However, information of the magnitude of the trees smaller than this limit may be useful for predictions of future states because the initial state is described in more detail. Paper II deals with identifying variables derived from laser scanner data suitable for explaining the variation in quantity and vitality of young seedlings in a size divers spruce forest. The results indicated that the relationships were strong enough so that prediction models can be developed.

Volumes of single trees are usually determined by means of volume equations that depend on diameter and height. To be able to compute volume of single trees described in a diameter distribution, tree heights have to be modeled. Paper III presents non-linear models that predict the height from diameter and stand variables and enabled prediction of tree heights in the study presented in Paper IV. The models were developed from National Forest Inventory (NFI) data and are therefore valid for all parts of Norway.

Paper IV presents a matrix model for forest management. The matrix model was tested by performing both a short term and a long term validation. The short term validation comprised tests that compared actual stand growth and model predictions on independent data. The results showed that the model were able to produce unbiased estimates of the number of stems. The long term model validation consisted in simulating stand growth without harvest for 1,000 years to see if the model predicted logical values with respect to steady state stand basal areas and species composition. The results indicated that the model produced estimates that reached steady states according to the hypothesis. The steady state was irrespective of the initial state of the stand being projected, as expected. The species composition was also logic producing large spruce volumes on the good sites and more pine on the poor sites.

Supervisors

Professor Erik Næsset
Associate Professor Terje Gobakken
Professor Tron Eid

Evaluation committee

Professor Annika Kangas (University of Helsinki, Finland),
Professor Göran Ståhl (Sveriges Lantbruksuniversitet, Sverige),
Professor Andreas Brunner (UMB)

Reference

• O. M. Bollandsås, “Uneven-aged forestry in norway: inventory and management models,” PhD Thesis, 2007.
  [Bibtex] [Download PDF]

  @PhdThesis{Bollandsaas2007c,
  Title = {Uneven-aged forestry in Norway: Inventory and management models},
  Author = {Ole Martin Bollandsås},
  School = {Norwegian University of Life Sciences},
  Year = {2007},
  Month = {PhD thesis 2007:30},
  Owner = {hanso},
  Timestamp = {2012.11.30},
  Url = {https://static02.nmbu.no/mina/forskning/drgrader/2007-Bollandsas.pdf}
  }