Mesoscale numerical forecasting of wind changes in support of fire-fighting operations

Southeastern Australia suffers extreme fire weather in the summer months, and a particular problem for fire weather forecasters is the timing of "cool changes", or shallow cold fronts. Most of the extreme events historically have been in association with such cool changes, and massive loss of life and property damage can occur. Seventy-five people were killed in the Ash Wednesday fires of 14 February 1983. Problems for fire-fighters are caused by three main meteorological consequences of these cool changes. First, the atmosphere is usually very hot and very dry ahead of these changes, and extreme low-level winds can also be experienced associated with the low-level jet ahead of the cold front. Temperatures above 40 C, humidities below 10%, and wind speeds above 50 knots are not uncommon. The second problem is the timing of the change, so that life and property can be protected, at the abrupt change of wind direction (usually greater than 90o). Third, there is frequently a period of strong, gusty winds following the change, which again makes control operations very difficult, and forecasting the speed and duration of this wind surge is difficult. The Bureau of Meteorology provides operational forecast to fire agencies. Much of the knowledge of the structure of these cool changes has come from a series of field experiments carried out in western Victoria in the 1970's, and idealised modelling studies based on these data. However, understanding was still limited by the data sets available, and the detailed structure of the winds ahead of and behind the fronts could only be guessed. Since the summer of 1996-97, mesoscale NWP forecasts have been available to forecasters, at a horizontal resolution of 25 km. Marked improvements in timing of forecast wind changes has resulted, and some of the wind structure associated with these changes began to be resolved. For days of significant fire behaviour problems, these forecasts, and research model runs at 5km resolution have been analysed in depth, and these results will be presented at this conference. Coastal frontogenesis, coastal surging of cold fronts, and flow blocking by relatively low coastal mountain ranges which could only be conjectured from the observations have been revealed. This has lead to an improved understanding of the frontal development and movement behaviour over southeastern Australia. Indeed, the model forecasts provide a dynamically consistent framework in which some "rogue" observations can be explained. The sensitivity of forecasts to surface moisture availability, and so to season-to season behaviour of these fronts will also be shown. As the operational models are soon to be upgraded to 12.5km horizontal resolution, many of these research developments will be available to forecasters in real-time for the 1999-2000 fire season.