Let's discuss ACE as a metric

[Nuno]

I'm new to S2k generally, but I've frequented weather communities elsewhere over the last twenty years and I find that ACE has really come into popularity in amateur circles in recent years as a useful metric.

But I'm seeing a lot of debates pop up in other threads and perhaps a standalone thread to discuss the metric itself could yield an interesting discussion. I personally have a few issues with ACE being thrown around so frequently:

1) Much like the Saffir Simpson scale, it seems as if ACE itself is often incorrectly used compared to its original intention. It was created by Bill Gray and was part of his Hurricane Destruction Potential index, not as a measurement to describe seasonal activity or inactivity as it often is used now.

2) If ACE as a metric describes cyclonic energy, are there other metrics used to measure other kinds of energy in the atmosphere that are calculated in a similar fashion? Is such a metric even possible given atmospheric fluidity? Why not measure mid-lattitude storms? Or the energy accumulated by a typical rainstorm?

3) Is ACE even a useful descriptor of hyperactivity, or inactivity in the other extreme? I wonder how many are going to claim 2020 was not hyperactive because of the arbitrary measurement of ACE despite having had 17 named storms up to this point. Do seasons need a high ACE for hyperactivity status? Why if so? Surely any season that spins up that many storms, regardless of maximum intensity or duration, is going to be considered hyperactive independent of the factors that influence ACE?

Let me clarify, this is not an anti-ACE bash but I want to ask others why they feel it is the best indicator or measurement in comparing seasonal activity. I feel this obsession on ACE is impacting our own interpretations and attitudes towards storm tracking which in turn is causing people to argue over what seem to be trivial matters.

[SconnieCane]

I think ACE is good for determining at a glance whether or not a season was favorable for intense, long-tracking major hurricanes the caliber of Irma, Ivan, Isabel (I name overload in this department), Florence, Frances, Maria, Jose, etc. It doesn't really tell the story of storm count or impacts very well.

2005 was an interesting anomaly in that it produced extreme storm counts similar to 2020, but also an extreme ACE value because numerous relatively short-tracking hurricanes became very intense.

[weeniepatrol]

I strongly prefer track integrated kinetic energy (TIKE)

[Hammy]

It's a good in-depth look at a season beyond simply the number of storms, hurricanes, and majors, and is key for historical cataloging and comparison when those numbers alone don't tell the whole story. 1964 for example had six hurricanes which in itself doesn't seem like a lot, many of those had high ACE counts and contrasting with 2019, a more active season similarly with six hurricanes, but if you take out even one of the two Cat 5's you end up with a significantly lower than normal ACE number, as 15 of the season's 19 storms were fairly weak and short-lived--Barry and Pablo, while padding the hurricane count a bit, were at hurricane intensity for a combined 15 hours.

And (in a comparison I've often made) the storms this year that have failed to reach hurricane intensity have been quite a bit higher in ACE value than last year.

2013/14 are another good instance--2013 had double the number of storms but produced half the ACE as 2014.

[Nuno]

It's a good in-depth look at a season beyond simply the number of storms, hurricanes, and majors, and is key for historical cataloging and comparison when those numbers alone don't tell the whole story. 1964 for example had six hurricanes which in itself doesn't seem like a lot, many of those had high ACE counts and contrasting with 2019, a more active season similarly with six hurricanes, but if you take out even one of the two Cat 5's you end up with a significantly lower than normal ACE number, as 15 of the season's 19 storms were fairly weak and short-lived--Barry and Pablo, while padding the hurricane count a bit, were at hurricane intensity for a combined 15 hours.

And (in a comparison I've often made) the storms this year that have failed to reach hurricane intensity have been quite a bit higher in ACE value than last year.

2013/14 are another good instance--2013 had double the number of storms but produced half the ACE as 2014.

So hammy, if we reach Alpha this year which is almost a lock, but haven't reached a sufficient hyperactive ACE number, it still wouldn't count as a hyperactive season?

[toad strangler]

It's a good in-depth look at a season beyond simply the number of storms, hurricanes, and majors, and is key for historical cataloging and comparison when those numbers alone don't tell the whole story. 1964 for example had six hurricanes which in itself doesn't seem like a lot, many of those had high ACE counts and contrasting with 2019, a more active season similarly with six hurricanes, but if you take out even one of the two Cat 5's you end up with a significantly lower than normal ACE number, as 15 of the season's 19 storms were fairly weak and short-lived--Barry and Pablo, while padding the hurricane count a bit, were at hurricane intensity for a combined 15 hours.

And (in a comparison I've often made) the storms this year that have failed to reach hurricane intensity have been quite a bit higher in ACE value than last year.

2013/14 are another good instance--2013 had double the number of storms but produced half the ACE as 2014.

So hammy, if we reach Alpha this year which is almost a lock, but haven't reached a sufficient hyperactive ACE number, it still wouldn't count as a hyperactive season?

Using ACE alone to determine hyperactivity is nonsense.

[Nuno]

It's a good in-depth look at a season beyond simply the number of storms, hurricanes, and majors, and is key for historical cataloging and comparison when those numbers alone don't tell the whole story. 1964 for example had six hurricanes which in itself doesn't seem like a lot, many of those had high ACE counts and contrasting with 2019, a more active season similarly with six hurricanes, but if you take out even one of the two Cat 5's you end up with a significantly lower than normal ACE number, as 15 of the season's 19 storms were fairly weak and short-lived--Barry and Pablo, while padding the hurricane count a bit, were at hurricane intensity for a combined 15 hours.

And (in a comparison I've often made) the storms this year that have failed to reach hurricane intensity have been quite a bit higher in ACE value than last year.

2013/14 are another good instance--2013 had double the number of storms but produced half the ACE as 2014.

So hammy, if we reach Alpha this year which is almost a lock, but haven't reached a sufficient hyperactive ACE number, it still wouldn't count as a hyperactive season?

Using ACE alone to determine hyperactivity is nonsense.

Thats my whole point. I don't think you can link hyperactivity to ACE, as it would require very specific storm intensities and so duration. Which neither is required for a damaging cyclone anyways. ACE is fun as a broad metric to compare things to, but any season that reaches Alpha regardless of ACE is hyperactive imo

Everything is spinning up in a way we haven't seen in the Atlantic before. Even 2005 never featured this crowded of a basin

[tolakram]

There are so many metrics, some objective and some subjective.

ACE is objective, that's a good thing in my opinion.

Number of storms is a little less objective IMO because sometimes naming a storm can be subjective, but still decent.

Landfall is a crap shoot, pure luck, and strength at landfall is also pure luck. If a hurricane is strengthening at landfall but never made it to major status is that really a reflection of seasonal activity? Hanna comes to mind, and now Sally. Sally had some great conditions for a day or so but steering fell apart and it stopped moving. Is this an indicator of seasonal quality?

Landfall cost. Cost can be seen as an objective measure, but if I build a house on the coast in 2000 and it gets destroyed by a hurricane in 2001 is the cost to replace a true measure of seasonal strength? What if the hurricane hit in 1999?

I think ACE is a good metric to indicate the overall conditions and strength of a season. In my opinion there is no difference between a 150 ACE season with 10 storms and a 150 ACE season with 25 storms. 'Quality' is an amateur metric.

[ElectricStorm]

Maybe we should start classifying a season based on activity and intensity. If a season is like 2020 so far, it would be considered hyper active but not hyper intense due to low ACE, but if a season only has 10 names storms, but like 160 ace, it wouldn't be all that active, but it would be hyper intense

[tolakram]

Maybe we should start classifying a season based on activity and intensity. If a season is like 2020 so far, it would be considered hyper active but not hyper intense due to low ACE, but if a season only has 10 names storms, but like 160 ace, it wouldn't be all that active, but it would be hyper intense

But overall what does this tell us about the season. Number of opportunities for development is also rather random. Great conditions but few waves might lead to a low count high ACE season. Good conditions and a lot of waves might be what 2020 is. If 2020 ends with lower ACE, regardless of the number of storms, I think that's a decent indicator. That said it also runs into the issue of running out of space. I forgot about number of storm days, which is yet another decent indicator. 2005 was so incredibly intense and at the same time individual storms were rather short lived.

I have no good conclusion here, other than no single metric works perfectly, but that doesn't indicate it's a bad metric.

[ElectricStorm]

Maybe we should start classifying a season based on activity and intensity. If a season is like 2020 so far, it would be considered hyper active but not hyper intense due to low ACE, but if a season only has 10 names storms, but like 160 ace, it wouldn't be all that active, but it would be hyper intense

But overall what does this tell us about the season. Number of opportunities for development is also rather random. Great conditions but few waves might lead to a low count high ACE season. Good conditions and a lot of waves might be what 2020 is. If 2020 ends with lower ACE, regardless of the number of storms, I think that's a decent indicator. That said it also runs into the issue of running out of space. I forgot about number of storm days, which is yet another decent indicator. 2005 was so incredibly intense and at the same time individual storms were rather short lived.

I have no good conclusion here, other than no single metric works perfectly, but that doesn't indicate it's a bad metric.

Yeah. I just think there should be another way to determine a "hyperactive" season other than just ACE

[Nuno]

Maybe we should start classifying a season based on activity and intensity. If a season is like 2020 so far, it would be considered hyper active but not hyper intense due to low ACE, but if a season only has 10 names storms, but like 160 ace, it wouldn't be all that active, but it would be hyper intense

I think intensity and activity have to be separate. Hyperactivity is based to me, on total activity, regardless of intensity or landfall. Also, who determines the parameters for hyperactivity?

I also did a curious search, as I dont recall ACE being mentioned much in 2005 on the communities I frequented and a search of S2K shows that ACE wasn't brought up in any of the 2005 storm threads. As a matter of fact, the earliest references to ACE on s2k isn't until 2007. So a despite accumulating a lot of cyclonic energy by the technical and objective definition, 2005 seems to have been judged more by its sheer number of storms. As tolakram put it, there shouldn't be a single metric that acts as end all-be all, which it seems discussions about seasonal ACE quickly devolve into.

I have no good conclusion here, other than no single metric works perfectly, but that doesn't indicate it's a bad metric.

Hmm, I don't think ACE is a bad metric, until people try to use it as a singular descriptor of activity beyond generalized seasonal comparisons.

[Hammy]

Maybe we should start classifying a season based on activity and intensity. If a season is like 2020 so far, it would be considered hyper active but not hyper intense due to low ACE, but if a season only has 10 names storms, but like 160 ace, it wouldn't be all that active, but it would be hyper intense

I think intensity and activity have to be separate. Hyperactivity is based to me, on total activity, regardless of intensity or landfall. Also, who determines the parameters for hyperactivity?

I also did a curious search, as I dont recall ACE being mentioned much in 2005 on the communities I frequented and a search of S2K shows that ACE wasn't brought up in any of the 2005 storm threads. As a matter of fact, the earliest references to ACE on s2k isn't until 2007. So a despite accumulating a lot of cyclonic energy by the technical and objective definition, 2005 seems to have been judged more by its sheer number of storms. As tolakram put it, there shouldn't be a single metric that acts as end all-be all, which it seems discussions about seasonal ACE quickly devolve into.

I have no good conclusion here, other than no single metric works perfectly, but that doesn't indicate it's a bad metric.

[b]Hmm, I don't think ACE is a bad metric, until people try to use it as a singular descriptor of activity beyond generalized seasonal comparisons.[/b]

I'll throw my two cents in on this as I was part of those discussions. Hyperactivity is, in regard to the forecasts and whether or not they'll bust, based on a specified ACE number--similar to rapid and explosive deepening being based on a set change within a given amount of time.

The point I was trying to make in those was that the hyperactive forecasts will indeed bust if the season fails to reach the statistical level on which those forecasts are based.

[Nuno]

I'll throw my two cents in on this as I was part of those discussions. Hyperactivity is, in regard to the forecasts and whether or not they'll bust, based on a specified ACE number--similar to rapid and explosive deepening being based on a set change within a given amount of time.

The point I was trying to make in those was that the hyperactive forecasts will indeed bust if the season fails to reach the statistical level on which those forecasts are based.

But since when has hyperactivity been defined primarily by ACE? This is something I cannot seem to find. I want to find wording from NOAA or NHC that specifically define hyperactivity strictly via ACE and I can't find it, or even a scientific study. It all seems a bit subjective. S2K in the middle of the 2nd highest ACE season (2005) never once mentioned ACE in a single storm thread according to the archives. So clearly the community here, both pro and amateur, weren't using ACE as a primary factor for defining hyperactivity. I found a 2007 outlook (the same season where ACE is finally discussed more widely on S2K) that states the following:

For the 2007 Atlantic hurricane season, the ACE index is expected to be in the range of 125% to 210% of the median. The upper portion of this range is above the 175% baseline that Goldenberg et al. (2001) used to define a hyperactive season.

https://www.cpc.ncep.noaa.gov/products/ ... cane.shtml

So if 2020 doesn't reach 175% of median ACE, despite, reaching the Greeks, it isn't considered hyperactive?

[toad strangler]

I have no good conclusion here, other than no single metric works perfectly, but that doesn't indicate it's a bad metric.

Agree. I have made a bunch of statements against the idea of ACE being the sole determination of activity but I also never said it was a bad metric. It just doesn't make any sense for it to hold a ton more weight over other factors such as landfalls, number of named storms, and total days between 6/1 and 11/30 with a named storms. This is just a start.

This might be a stupid question but does the ACE meter shut off after a landfall or does it continue until it dissipates?

[Hammy]

I'll throw my two cents in on this as I was part of those discussions. Hyperactivity is, in regard to the forecasts and whether or not they'll bust, based on a specified ACE number--similar to rapid and explosive deepening being based on a set change within a given amount of time.

The point I was trying to make in those was that the hyperactive forecasts will indeed bust if the season fails to reach the statistical level on which those forecasts are based.

But since when has hyperactivity been defined primarily by ACE? This is something I cannot seem to find. I want to find wording from NOAA or NHC that specifically define hyperactivity strictly via ACE and I can't find it, or even a scientific study. It all seems a bit subjective. S2K in the middle of the 2nd highest ACE season (2005) never once mentioned ACE in a single storm thread according to the archives. So clearly the community here, both pro and amateur, weren't using ACE as a primary factor for defining hyperactivity. I found a 2007 outlook (the same season where ACE is finally discussed more widely on S2K) that states the following:

For the 2007 Atlantic hurricane season, the ACE index is expected to be in the range of 125% to 210% of the median. The upper portion of this range is above the 175% baseline that Goldenberg et al. (2001) used to define a hyperactive season.

https://www.cpc.ncep.noaa.gov/products/ ... cane.shtml

So if 2020 doesn't reach 175% of median ACE, despite, reaching the Greeks, it isn't considered hyperactive?

Not in the sense NOAA uses it in official context--as far as I know ACE is their metric for the term. Not to say the season shouldn't be considered such, but that's the metric used for their seasonal forecasts and usage. I'll add that I'm not sure when ACE came into the public sphere, in fact I don't recall hearing about it much prior to 2013.

That said we have a fair chance of reaching that if the GFS run plays out with Paulette.

[ClarCari]

Has anyone else though size and winfield should be took into consideration?
This is a genuine point bc you have small Cat.5s like Camille and Andrew producing more ACE than a very large Cat.3 like Katrina and Rita were at landfall.
I don’t buy at all the a small Cat.5 produces than same ACE and energy and momentum physically speaking as a large Cat.5.
I strongly believe size needs to be taken into account.

[dexterlabio]

My amateur opinion is that ACE says a lot about the conditions that tropical cyclones experience in one particular season. I get that it should not be taken as the sole determinant of a hyperactive season, but it is a very important feature when it comes to scientific analysis. When we have more than 24 tropical storms that formed in a year but only had a total of 120 ACE, of course we are going to ask why. If the season had only around 12 TS but managed to top 200 ACE, we are going to be curious as well.

If we have anomalously high number of tropical cyclones in one season but the ACE is around average, a quick hypothesis (at least for me) is that conditions was ripe for TC genesis but something in the atmosphere or SST prevents these storms from getting really intense...then the next step here is figuring out why..

[TheStormExpert]

How about we add impact to the scale of ACE and change the name?

[al78]

I have no good conclusion here, other than no single metric works perfectly, but that doesn't indicate it's a bad metric.

Agree. I have made a bunch of statements against the idea of ACE being the sole determination of activity but I also never said it was a bad metric. It just doesn't make any sense for it to hold a ton more weight over other factors such as landfalls, number of named storms, and total days between 6/1 and 11/30 with a named storms. This is just a start.

This might be a stupid question but does the ACE meter shut off after a landfall or does it continue until it dissipates?

ACE accumulates as long as the storm is tropical or subtropical and is at least 34 kts intensity. Landfalling is irrelevant.