Wildland fires can take different forms such as surface fire or an elevated crown fire or combination of both. Crown fires are normally originated from surface fires spreading either along the bark of the tree trunks or direct flame contact to low branches with leaves and needles. In the past, surface fire (grassfire) spread simulations were conducted using physics-based models with fidelity. Here, we firstly seek numerically converged results for the burning of a single tree. Previously, numerical convergence for such physics-based fire simulations has been elusive. Subsequently, the linear and Arrhenius thermal degradation sub-models are appraised. For both thermal degradation sub-models grid convergence of the mass-loss rate is achieved with a 50 mm grid. The grid converged simulations also agree with experimental results of a single burning Douglas fir tree. A fire in a modelled tree plantation is then simulated using the linear thermal degradation sub-model. The aim of this part is twofold: one to demonstrate a good modelling practice; and secondly to assess the model capability to simulate transitioning from a forest floor fire to a crown fire leading to a quasi-steady rate of spread. The Kolmogorov–Smirnov test is used to rigorously demonstrate that the simulations on different grid and domain sizes have converged — that is, the results have become independent of the numerical parameters imposed upon the simulation.A physics-based model reproduces many observed features of surface fire to forest fire transition. The crown fire propagates with a quasi-steady rate-of-spread after an initial development period. Analysis of the volumetric heat release rate shows that a surface fire propagates under the crown fire and supplies energy to support the burning of the crown. Overall many features are qualitatively in agreement with other tree and crown fire studies.