Current issue: 53(3)
Under compilation: 53(4)
This study compares the responses of two Swedish 5-year predictive stand-level functions with the observed responses in 721 fertilization experiment plots in Norway fertilized with nitrogen (N). All plots are single-species consisting of Norway spruce (Picea abies (L.) H. Karst.) or Scots pine (Pinus sylvestris L.) fertilized with ammonium nitrate (AN) or urea. The correlations between the observed and the two predicted responses were 0.34–0.40 for all plots taken together. One response function performed well on average, but underestimated the response in pine plots and overestimated the response in spruce plots. The second function overpredicted the response on the full dataset, in spruce plots and old forest, but performed well in pine plots. Both functions overestimated the growth response in high-productive plots. Higher N deposition in Norway than in Sweden may count for parts of the deviations. Testing of fertilization functions on new datasets is rare, but important part of the evaluation of functions. As the functions are not well fit for predicting the growth response in spruce and high-productive plots in our sample, new functions that include N deposition are welcome.
The usefulness of forest sector models in forest policy analysis is discussed, mainly based on experiences from Norway. Forest sector modelling is contrasted to two alternative approaches: (i) Intuitive, verbal analysis, and (ii) econometric models. It is concluded that forest sector models, properly developed in contact with the policy makers, should be of considerable value in forest policy analysis.