2023 Hurricane Season Forecast

Spring. It’s a beautiful time of year. Well, except pollen is in the air bringing with it sneezing, washing the car and solar panels every other day, and other annoyances like the annual Hurricane Season Forecast Season 😛 . Unlike pollen, which is of course essential for plant and insect life, seasonal hurricane forecasts aren’t really that helpful other than as a reminder hurricane season is coming. With that, here’s what you need to know.

First, never change your plans based on the seasonal forecasts unless you are investing in insurance futures or commodities. For personal, institutional, and business planning this is really simple: treat every year as if you are going to get hit. That means revisiting your checklist (link to FEMA for some good starting ideas), starting with insurance. If there is any chance at all of flooding, sign up for the National Flood Insurance Program. Not on the coast? Don’t be so sure hurricanes can’t hurt you, especially when it comes to flooding, so treat spring as a good opportunity to take a look at your situation. Be it for evacuation from a huricane or if your home is unlivable due to a tornado, earthquake, chemical spill, whatever, think through where you will go, how to keep in touch with family, taking care of pets, etc. Doesn’t take long to come up with a plan – and coming up with a plan in a crisis is often too late.

So, you’ve done all that and you still want to know how likely it is you will need to use all that great planning, or if by September you’ll be eating all the Cheetos you’ve stashed. The biggest driver of seasonal hurricane activity is the El Niño/Southern Oscillation cycle, or ENSO. The ENSO cycle is a major driver of northern hemisphere weather patterns (link to NOAA discussion). It also has a strong influence on hurricane activity in the Atlantic. There are three phases to ENSO, called El Niño, Neutral, and La Niña. The names were given to the phases by fisherman in the eastern Pacific, who noticed the cycle long before meteorologists did. In general, during an El Niño summer/fall, there tends to be fewer hurricanes in the Atlantic. This is due to stronger wind shear and stability, which tends to make it harder for storms to form. La Niña has the opposite effect – conditions are usually favorable. Neutral years are not, as you might expect, in the middle, but actually tend to have more storms than the long term average. This is because El Niño years really drag down the long term average number of hurricanes, especially at some sites like Savannah (this effect is known as the “Flaw of Averages“). El Niño years aren’t just quieter than average, they tend to be a lot quieter than average.

At this point it’s worth discussing one issue in analyzing hurricane trends that often goes unmentioned but is vital: problems with the historical record. Periodization isn’t just a problem for historians or foreign policy analysts, it’s a problem for scientists as well. The further back in time we go, the more issues there are with the historical record, such as mischaracterizing storms or missing them altogether (especially over the oceans), changes in our observing technology (better sensors), and so forth.

If you used the period 1886 to 2022, the average number of tropical cyclones per year was 12.37. However … if you only use 1980 to the present, it is 15.97! CSU uses 1991 and after for their averages. So … when comparing various sources, be careful they are using the same period of time for their underlying data, and what they mean by “active” or “quiet.”

So with those caveats … using data from 1980 to the present (to get enough years for statistical stability without going so far back in time to create insurmountable issues with classification), in an ENSO Neutral year we expect 17 events, a La Niña year 20, but an El Niño year should only have 12.

The last few years have been either La Niña or neutral. While we are currently in an “ENSO Neutral” status, the forecast is by August we will be in a full blown El Niño, and will remain in a strong El Niño through the end of the year. That’s good news in terms of the raw number of storms to expect. However, and it’s a big “however,” the number of storms doesn’t really tell you that much about your risk. After all, the classic example is 1992: it was a quiet year, except for Hurricane Andrew, which made a mess of South Florida and, critically, disrupted distribution from the vital Tabasco Hot Sauce factory on Avery Island, Louisiana.

And there are other factors that influence hurricane activity. We are currently experiencing warmer than normal sea surface temperatures across the Atlantic/Caribbean and Gulf. That means that if a storm does form, it has a lot of energy to draw on. Other cycles in the climate also impact storm formation probabilities. But, in general, ENSO is the big driver, and an El Niño is good news (well, for hurricane risk). NOAA has not yet released their 2023 forecast, but the usual suspects have. Colorado State is predicting a below average season in terms of raw numbers. That Weather Channel can’t bring themselves to say it will be below average, only average.

But, again remember that the number of storms doesn’t say much about your odds of being hit, or how bad it might be if you are. Another factor is that a direct hit from a weak storm can cause the same damage as a more distant miss by a stronger event. There are a lot of different ways of answering the question “how bad is hurricane season going to be.” Here I’ll explore two of them. The first is “what are my odds of experiencing hurricane force winds?” The second is “What winds do I have a 5% (1 in 20) chance of experiencing? I’ll look at that for Savannah, Georgia, but will link to a Google Earth file that will let you explore the rest of the US, Canada, and Caribbean at each of the National Hurricane Center’s coastal warning locations.

Here is a plot of the return period curves for the mouth of the Savannah River, Savannah, Georgia (this is the view you get when clicking on any of the icons in the KML):

As you can see, activity is very depressed in an El Nino year (red curve). That is why our long term average is to see hurricane force winds from a hurricane only once in 21 years – however, since more than 2/3 of years are either “Neutral” or “La Nina” (green and blue curves), our risk in an average year is more like one in 12.

So given this data, what is our risk? Well, we have a 1 in 37 chance of experiencing hurricane force winds. That is down from the last few years, where it has hovered between 1 in 11 and 1 in 15. Another way of looking at it is there is only a 5% chance of experiencing strong tropical storm winds or higher (50 knots, 58mph) this year. That’s not too bad.

A lot of places have similarly lower risks this year, but some also have higher risks than average. Notice how many areas still have a 5% or better chance of hurricane force winds. If you have Google Earth or other GIS software that can display KML, the KML file with all this is located at this link. Clicking on an icon pulls up the graph for that location. Note some areas in the far southern Caribbean don’t have much risk in any case, and for this data set the west coast of Central America is only showing the chances based on an Atlantic storm crossing over.

Here are some static maps (embiggenable of course) showing the 5% chances of various Saffir-Simpson category events. There are four icons, thunderstorm (below tropical storm), weak tropical storm, strong tropical storm (over 50 kts/58mph), and hurricane, with the hurricane icons having the Saffir-Simpson scale indicated:

So as noted above, prepare for a hurricane, then at the end of the season you can eat your snacks and enjoy if it turns out to be a quiet one!

---

Thanks to Dr. Mark Johnson, Professor Emeritus of Statistics at the University of Central Florida, for his work on the underlying statistical algorithms that make these kinds of analyses possible.,