There is a lot of talk about hurricane Irma right now. Irma has been tracking across the Atlantic over the last few days with fluctuating intensity. Currently at category 3 (sustained winds of 115 mph at 0200 AST 4 Sep), Irma is officially a major hurricane. This blog post tackles some of the challenges in the forecast and there is no better place to gain an overview of the situation than the National Hurricane Centre’s (NHC) official 5 day forecast.
Hurricane watches have already been issued for parts of the Leeward Islands for later on Tuesday (tomorrow). Puerto Rico, Dominican Republic and the Bahamas are increasingly likely to see watches issued over the coming days. Already using the word ‘likely’ gives a hint at where this blog post is going. How likely is it that the above mentioned places will see impacts from Irma?
Starting with a forecast that gives us event probability, the realm of uncertainty can be tackled one step at a time. It is worth being aware that a hurricane does not need to track directly over a certain area to be problematic. A radius of damaging winds can extend hundreds of kilometres from the storm centre, this is why we developed tools to spatially identify where winds are likely to exceed various thresholds. One of the most useful thresholds is that of tropical storm strength (34 kt sustained wind). The following flow of images looks at Irma at 24 hour intervals over the next 5 days.
This distribution of probability leads us to the all important topic of likelihood and uncertainty. Communicating uncertainty is one of the biggest challenges a weather forecaster will face. The track, timing and intensity of a hurricane (like Irma) are all governed by processes within a chaotic yet vastly connected atmosphere – there are bound to be countless challenges. Ultimately, we want to know where Irma is going to go. The timing and track uncertainty go hand in hand – crucial timing of large scale processes can have large impacts on Irma’s path and speed. There are two main upper-level driving features that we are going to cover here, a trough over the eastern USA and the broad high over the North Atlantic.
Previous forecasts have suggested that Irma would take a more northerly track, brushing the Bahamas before curving along the Eastern Seaboard stirring some interesting conversations on social media. However, more recent guidance suggests that the ridge keeps a nose over the Sargasso Sea so that Irma is partially blocked from following a safe path into the North Atlantic. But what of the trough? Usually a pronounced upper feature such a trough would help push Irma away from a US landfall. Unfortunately, the fate of the trough is rather complicated. By the time Irma reaches the vicinity of the Bahamas, the trough may have become cut-off from the main flow under an anticyclonic wave break. Essentially a cut-off feature holds more uncertainty than a typical trough pattern. Competition in the dominant steering flow gives a number of outcomes. Take an ensemble of storm tracks for example:
A number of scenarios can be displayed in a single GEFS run and Irma’s tracks do that to a certain extent. The difference between each ensemble member comes down to how far west the storm centre reaches before the curve north begins. This partially falls down to the fate and timing uncertainty associated with large scale drivers like the aforementioned trough. The last (but certainly not least) uncertainty yet to be discussed is intensity.
Going back to the point about track and timing uncertainty going hand in hand… If the track governs the environment in which Irma will move through, then the intensity is hugely dependent on the track itself – a conditional uncertainty if you will. Uncoupled models like the GFS regularly resolve Irma to bottom out a central pressure well below 900 hPa. It is very likely running away with an over intensification but we can learn that the environment in which Irma is likely to move through has very little impeding characteristics (like shear, cooler sea surface temperatures or significant land) and this holds healthy for hurricane intensification along the forecast track.
It is best not to dwell on individual forecast runs for intensity (or track for that matter) so attempting to identify a trend in a forecast model can help put a single forecast run into context.
All images in this panel are valid for the same time taken from 4 different GFS forecast runs. Reading the images from top left to image to bottom right (oldest to latest forecast), notice that Irma is resolved consistently as a powerful hurricane and trends to the southwest. This is of great concern and has also been reflected in the ECMWF model. To compare the ECMWF to the GFS, the latest forecasts are impressively similar out at 1 week, spot the difference.
Clearly a lot to take in. To summarise the uncertainty of Irma is no easy task but hopefully this short blog post ties together a few concepts that underly the complicated forecast.
Best and latest information on official advisories, watches and warnings concerning Irma can always be found at the NHC’s website.
A wide range of models and forecasting tools can be found on our page: wxcharts.eu