Once again NOAA has predicted “above normal” activity for hurricane season
It seems that they always predict above normal. Is there a place somewhere where we can find historic predictions (including the statements ‘above’ or ‘below’ normal).
I’m not looking for tracking and strength predictions of specific hurricanes. I’m looking to see how accurate NOAA has been in predicting above ‘normal’ and the metric they use for those statements.
Start here: National Hurricane Center Forecast Verification. NHC is very big on tracking and refining their accuracy.
IMO your statement that “they always predict above normal” is false.
yes, you might be right on your questioning my statement, but honestly, I don’t think I’ve ever heard them predict a ‘below’ normal hurricane activity for a particular year.
NOAA has a forecast verification page for annual predictions and for individual storm tracks
As for below normal activity, that’s the summary statement and perhaps oversimplified. In reality, it’s more of a probabilistic model. For example, the 2025 updated forecast has it as a 50% above normal chance, 35% near normal, and 15% below normal.
And, as a matter of fact, back in 2018, they did initially predict a below normal season, which was broken down as 60% chance of below normal, 30% near normal, and 10% above normal. Granted, that had to be revised downward when the season ended up less busy than expected, though still a very busy season. Climate change and all. Makes things hard to predict based on historic averages.
It usually it is because media does not report well when they predict less activity, this nice graph shows how NHC is actually conservative with their forecasts, Turns out that when they do report “above normal”, it is usually worse than expected.
As a commenter put it:
[Unadvantaged]
”Right? Looks like odds are pretty high they’ll be proven right, and if they’re wrong it’ll be because the season was even worse than expected.”
Yes, kind of a “dog bites man” thing. It’s not interesting news and doesn’t stick in memory when it happens.
Also, if it works the same way as most climatological references, “normal” is defined as “how things were in around 1950”. If you’ve been an adult for less than 75 years, it’s hardly surprising that you’d remember them as almost always predicting “worse than normal”, due to climate change.
I’m sure that’s part of it. I know that temps predictions are normalized over the previous 30 years. Are hurricane numbers also normalized over that same 30 years? If so, I’m pretty confident that would be a root of the phenomena.
If you check the page I linked, they’ve got all the methodology. It’s based on numbers from the last 30 years.
And, as noted, there are years they predict ‘normal’ and ‘below normal’ on average, though these are given as percentages (again, the updated August 2025 guidance is 50% above normal, 35% near-normal, and 15% below-normal).
One issue is that people don’t pay much attention when a below average forecast comes to pass - unless they get hit by one of the storms. For example, 1992 was a significantly below-average hurricane season (and correctly predicted to be so!) with only 1 major hurricane. That hurricane happened to be Andrew, and the folks who lived in southern Florida at the time don’t think of 1992 as a below-average hurricane season, though it really was.
This is helpful information, thanks. While I was also under the impression that NOAA generally forecasts worse than normal hurricane seasons, it also appears that at least in the past 20 years or so, they’ve been pretty accurate.
If 1950 is taken as “normal”, then it would be unsurprising that recent seasons have been above normal. Hurricane/storm/tropical depression tracking back then depended highly on reports from land-based stations, ships and reconnaissance flights. Detection of short-lived tropical depressions and storms was not as sensitive in 1950.
Apparently, “normal” is now defined as the average from 1991-2020.
I see I was partly ninja-ed while I was typing. Oh well …
The 30-year lookback defines “normal” for hurricanes too. But it isn’t 30 years from the current year. It’s really a 3-decade lookback from the current decade. The 3-decade range updated in 2020 and will again in 2030.
So far both theory and practical experience show that global warming does not increase the quantity of tropical cyclones. It does increase their severity. And most ominously, it increases their rate of intensity and size growth even more.
Which has the effect of turning some of what might previously have been tropical depressions into tropical storms, and some some of what might previously have been tropical storms into minor hurricanes, and some of what might previously have been minor hurricanes into major hurricanes.
As well, our ability to detect minor storms or storms at sea has only gotten better since the advent of satellites in the mid 1960s. So comparisons of storm quantity need to account for that difference.
Yeh. One of both NHC’s and local emergency people is that it doesn’t matter if the season is above or below normal if you get hit. “It only takes one” is one of their buzz phrases.
Are you looking at the NOAA site directly, or noticing various news organizations’ reports of what NOAA is saying?
Because ‘bad hurricane season coming, says NOAA!’ makes a much better headline than ‘mild hurricane season coming, says NOAA’ and is therefore, I suspect, a lot more likely to happen.
– or what @GIGObuster and @Great_Antibob said.
The three decade baseline is the standard in meteorology to assess versus annual predictions for large weather events. However, that is being challenged by dramatic shifts in trends of mean surface, ocean surface, and ocean subsurface temperatures which not only have upward slopes but the increase in slope is also trending upward in an apparently geometric progression, which is expected because it represents not only increases in ocean surface temperature but the increasing depth of elevated temperatures providing both a heat sink and source of water vapor to drive a tropical storm from a ‘depression’ to a hurricane or cyclone. Whenever ocean surface and near-surface (down to 50 meters) get to a temperature of 26.5 ℃ or higher they have the potential to feed and intensify tropical storms, and tracking these temperatures and ocean circulating current oscillation cycles like the ENSO provides the basis for predictions of likelihood and number of storms that can be spawned in a season.
The ‘pre-industrial baseline’ of 1850-1900 is used to assess the mean global surface temperature anomaly is because it is representative of a period prior to massive carbon dioxide and other greenhouse gas emissions by anthropogenic sources. NASA, however, uses a 1951-1980 baseline because it is in line with meteorological methodology even though there is clearly a ~0.2 ℃ increase in mean temperature between the pre-industrial baseline and the period NASA uses, and has a pretty flat regression trend even though there are significant periods of increase or decrease within that. This difference is why you sometimes see people arguing about whether or not we have exceeded the 1.5 ℃ Paris Agreement threshold because the agreement didn’t actually define what the baseline is, or over what period to measure a ‘current’ temperature anomaly. However, based on the 1850-1900 baseline and using a standard 5 year weighted least squares regression we clearly see an exceedance in mean global surface temperature of 1.5 ℃, although depending on whose temperature model you use the lower uncertainty bound may be significantly below that threshold.
In summary, we are seeing pretty consistent “above normal” hurricane predictions because the surface and subsurface ocean temperature conditions keep increasing, meaning that the baseline over any arbitrary 30 year period is higher than any measurements before it. Furthermore, because of concentrations of heat in the oceans and more active and perturbative oscillation cycles, we are also seeing larger storms forming, and occurring more frequently over shorter intervals in the periods of high storm activity, with some storms even moving to higher latitudes and further inland compared to prior experience. The more heat the ocean absorbs, the more frequent such storms will be, leading to successive “above normal” predictions.
Stranger
Yes, I’m sure the headline of bad hurricane season makes more headlines. But I still seem to have only seen every year its a bad predicted year. One nuance is how the media spins it.
I didn’t realize there were multiple parameters….#named storms, #hurricanes, #hurricanes above Cat.3, speed of growth, intensity, updated predictions, etc. Such that maybe half of the parameters are predicted to be lower and half predicted to be higher (just for example) and the media focuses on those that fit the narrative.
And turns major hurricanes into things that cause hurricane-trackers to consider creating a new “Category 6”.
True.
There’s also a growing realization that the Saffir-Simpson Hurricane Wind Scale AKA “Category” system is utterly broken for communicating hazard levels to the general public. The peak wind speed is not what kills people or wrecks stuff. Or at least not much. The great bulk of death and destruction comes from storm surge, rainfall rate, and rainfall cumulative totals. Wind just stirs up the wreckage and lets Jim Cantore look cool with his jacket flapping madly.
Another hazard dimension not addressed by the category system is how large the region of impact is expected to be. A storm that leaves a 20 mile wide swath of destruction is very different from one leaving a 100 mile wide swath. But if they’re both blowing at 105mph, they’re both Cat 2.
And from the perspective of somebody generally in the path, the center of a 20-mile swath has to miss you by only 10 miles before you’re spared. Whereas the center of the 100 mile swath has to miss you by 50 miles to be spared. Big difference.
Indeed. If you look at last year’s Hurricane Helene, some of the most devastating damage was in North Carolina; by the time it reached that area, it was “only” a tropical storm, which still dumped absolutely massive amounts of rain.
That, and we may be near the top of wind speed anyway. A few storms may get just over the 200mph mark but that may be as much as the physics actually allows. Adding a Cat 6 would be akin to a Spinal Tap-esque “11” on an amp.
But it’s hard to develop a meaningful scale that is both sufficiently descriptive of destructive potential yet simple enough for the average person to understand. To mix metaphors from a different context, there’s significant overlap between the smartest bears and the dumbest humans.
Hurricane Camille in 1969 dropped “only” 7-10 inches of rain in the area of its landfall in Mississippi. By the time it reached western and central Virginia it was a tropical depression, but a much bigger rainmaker. "…rainfall was so heavy that reports were received of birds drowning in trees, cows floating down the Hatt Creek and of survivors having to cup hands around their mouth and nose in order to breathe through the deluge. Though the official rainfall was recorded as 27 inches (69 cm), unofficial estimates are much greater. Some estimate that more than 40 inches (100 cm) of rain fell at Davis Creek…So much rain fell in such a short time in Nelson County that, according to the National Weather Service at the time, it was “the probable maximum rainfall which meteorologists compute to be theoretically possible.”