Search

• We fitted growth models suitable for both continuous cover and rotation forestry stands. • Sensitivity to climate was restricted by using only temperature sum. • Species-specific symmetric and asymmetric competition affected tree growth. Forest growth models employed in Fennoscandia have been generally targeted at rotation forestry (RF) stands, relying on age as a key predictor. Uneven aged, irregular stands, such as the ones managed with continuous cover forestry, are becoming of increasingly common. New models suited for all kind of management approaches (i.e., age-independent) have been developed in Fennoscandia. Although the ongoing climate change is projected to strongly affect tree growth in boreal regions, climatic variables included in current models are usually restricted to temperature sum averages with simple links. The objectives of our research were: 1) fitting a new age independent empirical tree basal area increment model (B2023) with inclusion of additional climatic variables for the main Nordic tree species (Norway spruce, Scots pine, and birches); 2) using independent data to validate both the new model and other two age-independent published empirical models (P2013, P2021); and 3) investigating the sensitivity of growth predictions of all the empirical models to climate change. Our results showed that the new model B2023 was as accurate as P2013 when independently validated. Both models performed well in different forest structures and management alternatives (namely rotation forestry, continuous cover forestry, two-storied stands, and old-growth natural forests), although with few differences, and on average slightly better than P2021. At plot level, the new model B2023 showed slight underprediction for the overstorey pine layer in continuous cover forestry and two-storied stands. The predicted climate change scenarios increased simulated growth in all models, although P2021 showed very high values for spruce. We failed to include additional climatic variables than temperature sum in B2023, thus not improving much its accuracy under historical data, nor its sensitivity to future climate. Concluding, the individual tree models here presented can be applied to a wide range of forest structures and managements in Fennoscandia. For long-term simulation scenarios, different approaches to improve the climate sensitivity of empirical, individual tree model should be explored. [ABSTRACT FROM AUTHOR]