Revealing fine-scale variability in boreal forest temperatures using a mechanistic microclimate model

Abstract

Fine-scale temperatures are important drivers of ecosystem functions and biodiversity in boreal forests. However, accounting for large thermal variability has been difficult due to the coarse spatiotemporal resolution of climate data that is commonly applied in studies of biodiversity and forest health. Here, we use a mechanistic microclimate model and geospatial environmental and weather data to reveal microclimate temperature variability in a broad macroclimatic gradient in boreal forest environments. We modelled hourly near-surface temperatures (0.15 m above ground) in May-August 2020 over three focus areas located in hemiboreal, southern boreal and northern boreal forest zone in Finland at a spatial resolution of 10 m x 10 m. A comparison against data from 150 microclimate stations showed reasonable agreement (root mean square error [RMSE] 2.9 °C) between the measured and modelled temperatures. RMSE for the three focus areas ranged 2.2 –3.2 °C, and the difference was found to be generally smaller under dense canopies compared to open areas. The modelling revealed substantial thermal variability over the landscapes; for example, seasonal near-surface temperature ranges varied 26.5 °C – 42.9 °C, with the variation being smallest in the hemiboreal landscape with multiple large waterbodies, and largest in southern boreal landscape with large wetland areas. These results demonstrate the great potential of mechanistic microclimate modelling to increase our understanding of the thermal characteristics of various boreal forest environments. Ultimately, high-resolution spatiotemporal microclimate data will permit better understanding of e.g., boreal species distribution under climate and land use change and fine-scale variability in disturbances, including insect pests and forest fires.