When we’re tracking tropical cyclones, we talk a lot about wind speed. But it turns out that wind speeds are tricky to predict, not to mention to record (given the intensity of these storms, which have a tendency to break instruments). To find out why, I talked with Dr. Tom Mortlock, senior analyst at Aon and adjunct fellow at the Climate Change Research Centre, University of North South Wales. Here’s our conversation, edited for brevity. |
What makes predicting cyclone wind speeds difficult? When we talk about predicting cyclone wind speeds, there are two ways we can do this. Either we are wanting to predict the wind speeds associated with a cyclone that is about to make landfall, or we want to predict the likelihood of wind gusts of a certain magnitude being exceeded at a given location. We are now quite good at the former, and there is still a lot of uncertainty associated with the latter. Cyclone wind speeds are closely related to the central pressure of the cyclone. Therefore, if we know the central pressure, we can pretty accurately estimate what the maximum open-terrain wind gust is likely to be. Regarding predicting the likelihood of certain wind gusts occurring, that really depends on the length of the observational record we have. To undertake what we call “extreme value analysis” we need to extrapolate from the observational record. The rule of thumb is that we can reliably extrapolate three times the record length, before uncertainties really start to blow out. So, if we have a record length of 20 years, then we can estimate the 60-year recurrence interval wind gust (the wind gust that on average would be expected to occur once every 60 years, or more accurately, has a 1.7 % (1/60 * 100) chance of occurring in any given year). So, our ability to predict the likelihood of extreme cyclone wind speeds is closely related to the length of instrumental record we have at hand. |