Forecasting the extreme rainfall, low temperatures and strong winds associated with the northern Queensland floods of February 2019

The weather and associated floods that occurred in northwest Queensland in February 2019 were responsible for the death of 625,000 head of cattle, 48,000 sheep and flooding of over 3,000 homes. The monthly rainfall outlook for February that was issued on January 31st predicted a 45-55% chance of exceeding the median monthly rainfall. There was less than a 15% chance of more than 300 mm in February. More than 600 mm of rain fell over 10 days in inland locations.

Researchers at the University of Southern Queensland and the Bureau of Meteorology have investigated the climate drivers behind the event and how the forecast could have been improved. Typical drivers of climate in Australia, such as El Nio and the Indian Ocean dipole did not play a role in this event. The most predictable driver showing a clear influence was the Madden- Julian Oscillation (MJO). The MJO is a large-scale eastward moving 'pulse' of cloud and rainfall near the equator which leads to higher rainfall over northern Australia when located over southeast Asia and the western Pacific in summer.

The communicated forecast from the 26th of January was initialised up to 8 days beforehand and did not predict the event. However, applying the same methods closer to the time of the event improved the prediction. Initialising a week prior to the event resulted in a forecast with a 70% chance of high rainfall and similarly improved predictions of low maximum temperature and high windspeeds. Two weeks in advance the prediction was also improved, but to a lesser extent. The results suggest that multi-week forecasts can fill the gap between daily weather forecasts and monthly-seasonal outlooks, providing critical lead time for farmers to prepare for extreme weather events.