A very intelligent friend of mine asked me a reasonable question. Is Laura the most difficult storm ever to forecast? My response was not really, but I knew why the question was asked. For someone paying close attention to the storm, the hyper-awareness or “need to know” urgency may reveal aspects of hurricane prediction or messaging quite familiar to meteorologists but confusing to the public. As typically happens with these storms, weather prediction models tend to gain a better handle on hurricanes within three days of landfall. IBM/The Weather Company meteorologist Dr. Michael Ventrice tweeted this morning, “Still some discrepancies in timing and precise location but the big three deterministic models all agree #Laura will intensify into a strong-major hurricane prior to striking the U.S. somewhere between Galveston TX and Lake Charles LA.” Here’s the latest information on where Laura is going, how strong it will be, and some “101” about the forecast process too.
As of Tuesday morning, the National Hurricane Center predicts that Laura will become a major hurricane (category 3 or higher) just before landfall along the Gulf coast. Just before going to press, the National Hurricane Center upgraded Laura to a hurricane.The storm is currently on track to make landfall somewhere near the Texas-Louisiana border early Thursday morning. Laura is a larger storm than Marco, which fizzled near the eastern Louisiana coast, so wind, rain, storm surge, and tornado threats will be featured over a broad area. The National Hurricane Center morning discussion notes, “Laura is forecast to move over the very warm and deep waters of the Gulf of Mexico, with similar or lighter shear conditions through the next couple of days.” This is a recipe for rapid intensification, as it does not appear that Marco churned up to much cold water to impede Laura’s progress. The National Hurricane Center ends the discussion with this caution, “ Users are again reminded not to focus on the exact details of the track or intensity forecasts as the average NHC track error at 48 h is around 80 miles and the average intensity error is close to 15 mph.”
I want to spend a little time on that last statement because it as the heart of my friend’s question. Meteorologists are very familiar with the graphic below. It is a clear indication that hurricane track forecast skill improves as we get closer landfall. Let’s focus on the blue line, which is the average forecast track error at 120 hours (5 days) out. It is apparent that skill has improved from the early 2000s to today. A decreasing line indicates lessening error (in nautical miles) in the position of the storm. The other revealing fact here is that forecast skill gradually improves at 4 days out, then again at 3 days out, and so on. The red line (1 day) has the best skill. It is interesting to note that a 3-day forecast today is as good as about as good as a 1-day forecast in 1970.
This information is useful for the public because in the social media era. Discussions about hurricane threats begin days (and days and days and days and days and days) in advance. Once the storm is named, you start to see track and intensity forecasts. The problem is that many people don’t understand that not only will those forecasts likely change, we fully expect them because of the aforementioned discussion. As someone who lives in the South, I see a similar phenomenon with snow forecasts. Here in Atlanta, a forecast on Monday might call for 3 inches of snow on Friday. Everyone starts buying bread. However, snow forecasts are notoriously uncertain in this region so I caution people to watch the evolving forecast during the week rather than locking in on what they saw Monday. This applies with hurricane forecasts too.
With Laura, there has been some uncertainty about whether the landfall point will be in Louisiana or Texas. To the public that may seem to indicate that meteorologists don’t know what is going on. However, the reality is that the likely landfall point has be in the cone of uncertainty for the last several days. Brian McNoldy explains the cone of uncertainty in a great blog referenced in one of my previous Forbes articles. McNoldy, a meteorologist at the University of Miami, writes, “ The cone is designed to enclose 2/3 of the historical track forecast errors… meaning that on average, there is still a 1/3 probability that the center of the storm could track outside of the cone….Some scenarios in nature are inherently less predictable than others.” Here is another mismatch between public expectations and science. The public wants an exact line. If the forecast is not on that line, there is often a perception that it was wrong. The forecast cone is inherently designed to capture uncertainty associated with model limitations, nature, and so forth. If Laura makes landfall in Houston or Lake Charles, both are in the cone right now. I know for many people that is not precise enough and makes planning a challenge, but “it is what it is.”
Here is one final caution as Laura approaches the U.S. Hurricane intensity forecast skill lags track forecasting skill. While forecasts bring Laura to major hurricane status, there is still wiggle room on what the ultimate intensity category will be. For more on why intensity forecasts have lagged track forecasting, this resource goes deeply into the topic.