Several studies investigated the regional temperature effects of afforestation or deforestation, but the impacts of different forest development stages or alternative forest management received limited attention. This is mainly due to challenges in representing area-limited forest dynamics in low-resolution climate models and the need for accurate forest parameters. This study investigates the impact of alternative forest development stages and composition on regional climate in Fennoscandia using a coupled regional climate model. By incorporating realistic and high-resolution forest maps, our modelling framework reduces biases in estimating surface temperature compared to default model runs. If today's forest composition of tree species is left to achieve a mature state (a proxy for the absence of harvesting), an annual mean reduction in 2 m air temperature is estimated, with a cooling peak in summer of -0.53 ± 0.20 °C (mean ± standard deviation) mainly induced by increased cloud cover. Conversely, undeveloped forests (a proxy for increased harvest) induce a contrasting seasonal response: a summer warming of 0.53 ± 0.15 °C (mainly caused by higher sensible heat fluxes), and a weak winter cooling of -0.14 ± 0.24 °C (mainly caused by a higher surface albedo). A transition from evergreen to deciduous forests shows a summer average cooling of -0.57 ± 0.28 °C, mainly attributed to changes in surface albedo. These temperature effects are equivalent to a relatively large fraction of the expected warming by 2050 in Fennoscandia (from 16 % to 70 %, depending on the specific scenario and season). Some modelling outputs appear inconsistent with observations and past modelling studies, such as the cooling effects in winter of more developed forests. Our results provide new insights into the complex relationships between forest dynamics and regional temperature, but modelling improvements are still needed to achieve a robust understanding of the regional climate effects of forest management.
