Weather plays an important role in all stages of wildfire development. The term fire weather is used to describe the meteorological conditions that affect wildfire potential, behavior and suppression. Virtual Fire provides the ability to monitor and survey several forecasting and real-time fire weather information, both in spatial and non-spatial form, achieving either an early warning or knowledge of the current real-time conditions. The weather forecasting system is based on the SKIRON weather prediction model4 (Kallos et al., 1997) and supplies the necessary input weather data for visualizing weather maps. In Virtual Fire, a total of 85 forecasting maps of temperature, relative humidity, cloud cover, rain and wind are created once every day (at 10:00 am local time) with a prediction range of 3 days and a time interval of 3 h. These maps are prepared with horizontal resolution of 0.05 × 0.05° (~ 5 × 5 km) for an area that covers the entire North Aegean Region. This daily generation of the weather maps is created automatically. The daily sequence of input weather files is downloaded into the Virtual Fire server according to a time scheduler. Then the sequence of the output files is generated and the ArcSDE database that stores the weather maps is updated. Following the update, end users have the ability of displaying weather maps by simply selecting the desired weather parameter (e.g. wind) and the corresponding time. The system also supports real-time data monitoring from a network of remote automatic weather stations (RAWS). In Lesvos Island, four RAWS have been installed in characteristic parts of the island to achieve an optimum geographical coverage. These weather data are also available on an external website5 developed with Campbell's Loggernet software. A modern graphical interface visualizes all the weather data in an easily comprehensible form (Fig. 3).