Office for Environmental Programs - Theses

Permanent URI for this collection

Search Results

Now showing 1 - 1 of 1
  • Item
    Thumbnail Image
    Multiweek-to-seasonal prediction of fire weather in south-eastern Australia
    Sibbing, Joshua ( 2021)
    Weather and climate prediction play a vital role in preparing the public and emergency management authorities for extreme events. In this study, the Australian Bureau of Meteorology’s seasonal forecasting system, ACCESS-S1, is tested on its ability to predict severe fire weather in south-eastern Australia. For the severe event occurring on 07 February 2009 (“Black Saturday”), a hindcast dataset is used to compare model predictions of fire-weather-relevant variables to reanalysis values at lead times of 6, 14, 23, and 69 days. At the longest lead time (69 days), predictions were also compared with those from other years and correlated with the El Niño Southern Oscillation (ENSO) index NINO3.4. At sub-seasonal lead times (6, 14, and 23 days), ACCESS-S1 is found to skilfully predict Tmax values but overpredict rainfall. At the longest lead time (69 days), ACCESS-S1 predicts ordinary counts for hot and dry days during the 2008-2009 summer period, despite this period showing anomalously high counts in the reanalysis. For all hindcast years(1990-2011), model predictions of hot and dry day counts correlate strongly with NINO3.4 values, suggesting an overdependence by the model on the state of ENSO when predicting these variables. Other model biases, such as a warm bias in central Victoria and a wet bias in New South Wales, are suggested to exist due to the routine overprediction of hot day counts (central Victoria) and underprediction of dry day counts (New South Wales) found at these locations. Further investigation is needed to determine the reasons contributing to the model biases identified here and to gain a comprehensive understanding of the model’s abilities.