Wildfires in mountainous terrain can be extremely challenging to address. Apart from the difficulty in reaching the fire itself, steep slopes and the characteristics of the fuel play a key role in this complex problem. Additionally, the vertical variability of temperature, humidity and wind speed may also be significant and therefore contribute to unpredictable fire behaviour [1,2]. In the framework of the FireStorm research project, four weather stations were installed, at different elevations, in the main ridge of Mainland Portugal, in a region prone to forest fires. Two additional weather stations, from the Portuguese Meteorological Service’s official network of surface observations, are also used in the project. These six weather stations provide an extremely useful set of data, as they can monitor the slope, from the valley up to the mountain top. In both cases, the difference in elevation, between the highest and lowest weather stations, is roughly 1000 m. The goal of this work within the FireStorm project is to monitor the area and assess which meteorological conditions enhance the likelihood of large vertical variability in near-surface weather data. The data will also allow for the verification of operational weather forecasts, namely from ECMWF and the Convective Scale Model AROME [3,4], which have, respectively, a horizontal resolution of 9 and 2.5 km. This verification will ultimately allow one to quantify the uncertainty that using this meteorological data has when used as an input to fire propagation models.