The survival probability of a tree exposed to surface fire varies widely depending on its bark. To advance the understanding of insulation properties of bark, mean thickness (BT), moisture content (MCb), surface structure (BS) and density (ρb) of bark samples of Mongolian oak (Quercus mongolica) (n=395) for four diameter classes were investigated. In addition, data from 158 heating experiments simulating low-intensity surface fires in the laboratory were used to assess the relative importance of these properties affecting thermal insulation and evaluate the applicability of two stem heating models, an analytical, one-dimensional model and the FireStem2D model. Overall, BT is the best predictor of bark insulation capacity and MCb only contributes significantly to explain the residence time of cambial temperature >60°C (τ>60), whereas ρb and BS have minor effects. Although the two stem heating models overestimate the time required for cambium temperatures to reach 60°C (τ60), FireStem2D performed better than analytical model. Furthermore, FireStem2D provides good predictions of τ>60 and maximal cambial temperature (Tmax). In addition, errors in FireStem2D may be driven mainly by the errors in temperature measurement and the limitation of a two-dimensional model. The study provides a better knowledge of interactions between bark properties and heat transfer, which may improve the predictability of fire-caused stem injury for Mongolian oak and other species with similar bark properties.
