Evaluation of growth models for mixed forests used in Swedish and Finnish decision support systems

Interest in mixed forests is increasing since they could provide higher benefits and positive externalities compared to monocultures, although their management is more complex and silvicultural prescriptions for them are still scarce. Growth simulations are a powerful tool for developing useful guidelines for mixed stands. Heureka and Motti are two decision support systems commonly used for forest management in Sweden and Finland respectively. They were developed mostly with data from pure stands, so how they would perform in mixed stands is currently uncertain. We compiled a large and updated common database of well-replicated experimental research sites and monitoring networks composed by 218 and 1,160 plot-level observations of mixed stands from Sweden and Finland, respectively. We aimed to evaluated the accuracy of Heureka and Motti basal area growth models in those mixed-species stands and to detect any bias in their short-term predictions. Basal area growth simulations (excluding mortality models) were compared to observed stand-level values in a period-wise process with update of the start values in each period. The residual plots were visually examined for different stand mixtures: Norway spruce (Picea abies Karst.)-birch (Betula spp), Scots pine (Pinus sylvestris L.)-birch and Scots pine-Norway spruce. We observed that the basal area growth models in both decision support systems performed quite well for all mixtures regardless of the proportion of species. Motti simulations overestimated growth in Scots pine-Norway spruce mixtures by 0.063 m2·ha−1·year−1 which may be acceptable for practical use. Therefore, we corroborated that both decision support systems can be currently utilized for short-term forest growth simulation of mixed boreal forests.