Using Ensemble Techniques in Fire Growth Modelling

This paper presents the use of ensemble techniques in fire growth modelling. Errors in meteorological parameter streams – temperature, humidity, wind speed and direction and precipitation – are introduced into a deterministic eight-point fire growth model to produce an ensemble of possible final fire perimeters. Errors, whether they be from forecasts or observations, take the form of systematic or random errors and the probabilities associated with these errors are calculated and presented in the final perimeter fields. Detailed weather streams are examined to show the nature and extent of these observational errors. Finally, this approach is demonstrated on several observed forest fires in Wood Buffalo National Park.