What Tropical Cyclone Owen Can Teach Us About Model Data

Tropical Cyclone Owen was a great example of why model data can’t always be used on face value. Here’s a classic case of what was forecast vs what actually happened!


Tropical Cyclones are notoriously difficult to forecast – we even saw TC Owen make an unexpected shift to the south before crossing the Gulf Coastline and this begs the question – how can we trust forecasts 6 days out, if forecast models can’t predict where a system will move 6 hours in advance?

The forecast model comparison here is via windy.com. 

The comparison forecast on the left was a forecast by the Euro forecast model for Monday the 17th of December (issued on Tuesday the 11th of December).  The forecast on the right was a forecast for Monday the 17th of December (issued on Sunday the 16th of December).

You might recall an image we posted on our Facebook blog that shows how a small change in movement could have a massive shift in where the system actually moves? And that’s exactly what happened – and this actually touches on a complex part of mathematics known as ‘chaos theory’ where a small change in initial conditions can create very different outcomes.  A common part of chaos theory might be planning your route to work each day.  Sometimes leaving 5 or 10 minutes earlier or later could have a significant impact on your commute time due to the unpredictable nature of traffic or accidents.

If we look at TC Owen’s path, it relied on the upper low to its south moving as expected (and upper lows are also notoriously difficult to predict).  So when you combine the interaction between tropical cyclones and upper lows then you have a highly dynamic system that is likely to change and evolve rapidly over time providing different outcomes.  TC Owen still produced a lot of rain down the east coast, even if it wasn’t a direct impact.  Rather, the moisture from Owen fed down into an unstable and cold upper atmosphere allowing for heavy storm activity and this saw floods and severe storms in Tasmania, Victoria, NSW and much of eastern Queensland!

It also shows the importance meteorologists can have too as they can use their experience to interpret the data. In the case of Weatherwatch, while we acknowledged it as a potential outcome, we didn’t suggest it to be a likely outcome in our forecasts because there was too much variability at play.