The process of modelling and replicating extreme fire behaviour like junction fire is essential for understanding the phenomena associated with extreme fires. Numerical simulations of junction fires, replicating laboratory-scale experiments, with no imposed wind, were performed for a shrub fuel bed with slopes ranging from 0° to 30°. The simulations of junction fires were conducted for two junction angles 30° and 45°. For each scenario, the sensitivity to a range of numerical and physical parameters was investigated. The rate of spread (ROS) is a key parameter for assessing risks from vegetation fires. Experimental spreading junction fires, conducted at laboratory scale at Coimbra University (Portugal), were simulated using FIRESTAR3D - a three-dimensional physics-based fire model. To ensure the robustness of simulations, sensitivity analyses were carried out by varying the grid resolution, domain size, different fuel characteristics, as well as some of the thermo-physical parameters. The corresponding simulations were carried out using a single and two descriptions levels of the shrub: the vegetation was represented using only one cylindrical-shaped solid-fuel type (excelsior fuel using the characteristic parameters for Erica shrub), or two fuel by adding the contribution of twigs of various diameters up to 6 mm while keeping the same packing ratio. Finally, the validation of FIRESTAR3D simulations was achieved through the comparison of predicted and measured ROS values. The experimental trends of the compared quantities were well reproduced by the simulations. Accelerating and decelerating propagation phases were observed in all simulations, with a dependence on the slope angle, while the maximum rate of spread depends critically on the junction angle. As it was the case of other wildfires simulated by FIRESTAR3D, it was found that this fully-physical model is capable of simulating junction fire propagation. There are several processes associated with the development of a junction fire behaviour, in which dramatic changes in fire behaviour can occur with little change in various fuel, weather and topographical parameters. In a subsequent study, we aim to develop an understanding of junction fire behaviour taking into account essential parameters that affect the behaviour, namely: slope, junction angle, and driving wind velocity.