Wednesday marks the official END of the 2022 hurricane season which began June 1, (interestingly we had no pre-season storms this year for the first time since 2014). The season proved pretty interesting with three systems before July 1, and then NOTHING until September! Thank you Saharan dust and dry air!
Of course, when the storms cranked up they revved big and Canada recorded its strongest hurricane in history with Fiona on Sept. 20. Then Ian in late September hit Florida as the deadliest storm since 1935 (at least 146 in that state alone perished) and racked up a $50 billion damage tab. Nicole in early November would decimate Florida beaches already tender from Ian’s wrath.
As we close the season, NO hurricanes hit the rest of the Gulf of Mexico states or Eastern U.S. Seaboard (two minor storms in June, Alex and Colin rained on Florida and South Carolina).
As we count it up, the 2022 hurricane season was about as “average” as it gets with 14 tropical storms of which eight became hurricanes and two became major. Compare that to the average of 14-seven-three. That despite all those forecasts calling for an above normal hurricane season. I personally think that Saharan dust was the outlier that threw off so many forecasts, but then again not every forecaster went big. In fact, the University of Arizona’s forecast was pretty much on-target. I’ve highlighted theirs below:
In fact, these are late-season forecasts when most everyone went up a bit (June to late August), and UA’s April forecast was actually for 14-seven-three. These were made by UA’s atmospheric science professor Xubin “Shubin” Zeng and former graduate student Kyle Davis who began their predictions in 2014, so they haven’t been in this game terribly long when you consider Colorado State University started this in 1984 and the others followed suit over the decades.
So what’s Zeng and Davis’ secret sauce? In a phrase: sea surface temperatures, or SSTs. Particularly for their April forecasts where they look at both Atlantic and Pacific Ocean water temperatures and assimilate past year’s data with ensemble forecasts from the European model. This year, those SSTs were fairly normal or just above and so their forecast followed that lead. Unlike other forecasters, Zeng doesn’t feel wind forecasts are as reliable for early season predictions.
For their June forecasts, Zeng does factor in upper level winds and the presence of La Nina/El Nino in the Pacific.
I sent him an email to ask a few more questions, including:
- Why do you think your forecast for this year was so much more accurate than the others?
“Every year, we re-calibrate our models using one more year of data (e.g., in 2021, we used data up to 2020 for calibration; for 2022, we used data up to 2021 for calibration). Our results are also overall consistent with our expectation that the results in 2022 would be around average (depending on which parameter to use) - similar to 2019. So we know our results would be good most of the time.”
- How long have you been making seasonal forecasts and what is your accuracy rate?
“Here is the summary of our record prepared in June 2022:
“Since we first started issuing hurricane outlooks in 2014, the average errors have stayed very close to what we reported in both Davis et al. (2015) and Davis and Zeng (2019). For hurricanes, our average prediction error is 1.9 hurricanes (2014-2021).
“Since 2017, when we started issuing forecasts for ACE and major hurricanes, our average error has been 30.0 units and 0.4 major hurricanes (2017-2021).
“For named storms, for which we started issuing predictions in 2019, our average error has been 5.7 named storms (2019-2021). This number is so high because we only have three years of errors and 2020 was a record-breaking year.
“Besides the excellent performance this year, our 2017 June prediction was amazingly accurate (which is particularly relevant to your area): we predicted a super-active hurricane season (when essentially all other centers/teams predicted an average season) - fully consistent with the reality (remember hurricanes Harvey hitting Houston, Irma, and Maria?)”
- Do you have any early guidance or what we should watch for going into 2023? SSTs, I’m guessing?
“We have to wait until April 2023 for our seasonal prediction. We have not done any seasonal outlook prior to April each year. but we might if our future efforts find some skill (for an outlook prior to April).”
So here’s to the end of the 2022 hurricane season and a big congratulations to the work of professor Xubin Zeng and Kyle Davis. I’ll have a close eye on your April predictions next year!