The Hurricane Season ( a rehash of what I have had so far)
By Joe Bastardi, WeatherBell
I see Dr Mann has a forecast out for 16 storms plus or minus 4, NOAA is out with theirs tomorrow. It is interesting to note that my friend, competitor ( he is a consultant for a rival company) and PSU 78 MASTERS in meteorology, Paul Knight and I share the same disdain for the silly overall number games. And a range of 8 for a total of 16 is a pretty big spread. Since it is a game I must play to keep people happy, please keep in mind for over 2 months now I have had 13-15 out with 6 or 7 landfalls…this year the impact will be from actually storms hitting with analog seasons of 1950,1955,1996,1999,2008. This still has not changed. It is later than the 2010 forecast for 17 which came out on Feb 14.. this year I didnt put it out till March. In any case in the race to say who said what and when, whatever happens, I have not changed yet from the 14 plus or minus 1.( 13-15)
Oh by the way, I saw what Dr Manns group is up too as he was nice enough to share it with me last year, and I have no beef with it. I just want to make sure that since it got some press, that we logged the date that forecast went public, and note the numbers and the ranges. But again, its very unlikely to have that big a range and not have the total number fall in. Last year I believe they had 23, plus or minus 5, and of course 19 is within that range. So it was there.
Technically I was wrong last year… with 17 from Feb, plus or minus 1, there were 19. So my center point missed by 2, theirs missed by 4. But in the wonderful world of academia and statistics, I was wrong, they are right. However in the world that I work in, classroom scoring is not always optimum.
But there is not much to argue about. Interestingly enough a forecast of 16 plus or minus 4 can claim a hit if there are 12, but I cant, even though I have 14. So I can be closer, yet be wrong! More wonderful statistics, eh? My personal opinion is that spreading numbers out that much is really not what I am comfortable doing. Ce le vie.. live and let live.. Que Sera Sera, whatever will be will be, but it’s not for me.
Yah I know it gets me in trouble, cause it makes me sound pompous, but I just dont like the whole number thing. My pursuit is landfall impact!!!! Still I have to give a number and it was out in March and for now, remains at 14 ( +/-1) ( 9 Hurricanes, 4 or 5 majors). But its the landfall that is the big deal to me. 07 and 10 burned me in similar ways, but people forget the non seasons forecasted in 06 and 09 for the gulf, the great forecasts in 08 and 03-05. I do believe that the error I had last year will, like 2008, lead to the much better forecast this year. There is a 10 page power point presentation I have with research and ideas that will knock your socks off, things I havent seen anywhere, and ideas I am sure will draw the ire of the crowd that are making sure anything I do wrong is shown loudly and clearly. But in the end, I am as confident going into this season as I was in 2008 for that type of year.
In hindsight, my mistake last year was simply taking the mean of what the high number should produce as far as impact, rather than digging in and seeing how close 2010
was to 2007.. from the previous nino to the nina, to the temp pattern in the heart of the hurricane season. Not this year.
When you see this foreacst, you will know, right or wrong, I did my homework. As for last year Again, give Hermine 12 more hours and a track 100 miles further east, its a big hit right into the coastal bend of Texas and Alex at 947 mb, the strongest June storm on record with Audrey, 75 miles north and earl 75 miles west and alot of the screams of too much hype would not be there. However there is a physical reason for the divergent tracks, just as when you see the DETAILS of the hurricane ideas here and how much more is put into this, you will see my physical reasoning for coming up with the year I have, as far as the enhanced threat on the US coast.
But in the terms of the number game, its 14 I came out with a couple of months ago, and that is where I stand. It’s not last year…and there are big ticket factors that will lead to the US coast being impacted quite a bit this year.
Thanks for reading, ciao for now.
lol, so, Joe was that 14 you are predicting? I just want to make sure.
Joe, will the shear relax enough for the upcoming MJO pulse to start something in the western basin?
Guesstimating the number of hurricanes in a season isn’t all that impressive… guessing their paths and strengths really would be a feat of prognosticating future doomsday events and bring some deserved attention.
If the soothsayers of long term co2 climate doomsday actually made good short term climate prognostications maybe people wouldn’t be so skeptical of all their wild claims.
It’s hard enough guesstimating ten days or one season let alone 100 years, but then Mann just can’t do without his tree ring entrails.
I applaud you Joe for having the guts to live or die (metaphorically) your less loose predictions, er, forecasts. It would be interesting to know more about the madness of your methods. [;)]
Thanks for the additional information. Kudos for not spreading out your estimate to be a full 40-50% of the entire number, like Mann and some others do.
I suspect that one ‘big ticket factor’ is the water temp in the GOM …. which may drive quite a lot of near-landfall strengthening? But something big has to happen to drive the switch from La Nina – to El Nino? And alleviate this exceptional Texas drought?
Have I got this right? I better start working on my dam I guess…. Sigh…..
Way out here in the PNW the number of storms is not critical but for those who live along the hurricane coasts (and also who underwrite the works of man) this can a big deal, particularly how many reach shore.
As for the scolds who like to disparage your forecasts Joe – all I can say is they will disparage anything and anyone who doesn’t sing their songs or dance to their tune (i.e. just about 1/2 of everyone else). Remember think of them as mosquitoes – they are annoying, but their sting is minor.
Technically I was wrong last year…
07 and 10 burned me in similar ways
In hindsight, my mistake last year
I am impressed by this. It demonstrates a dedication to the truth, humility in light of the immense complexity/uncertainty, and a reasoned assessment of forecasting skill. It is refreshing and stands in stark juxtaposition to Mann’s phony implication of certainty/accuracy, i.e. “In 2007, Mann and Thomas Sabbatelli predicted the exact number of named storms (15) for that season (see 2007 prediction).”
http://wattsupwiththat.com/2011/05/18/manns-hurricane-season-predictions/
Regardless of the actual number hurricanes this season, your forecast is much better than Mann’s…
Maybe it should be like a Price is Right thing. Closest without going over wins.
I’ll do like Mann and say 10 +/-15. With between 1 and 10 as majors and between 2 and 8 make landfall.
Oh, it doesn’t matter anyway. The earth is going to end on Saturday anyways.
For the last 25 years, I’ve been predicting 50 each year, plus or minus 49 and I’ve been right every year.
Joe,
I wouldn’t worry about there forecasts being “technically” correct when they have such a large margin for error.
Sorry,
Both throw words to the wind (sustained??)
Michael Mann and Joe Bastardi
Put me in for 18 named at $20, and 8 landfall at $10. 18 AND 8 gets me an NSF grant.
(are hurricane pools ok here??)
So if I say say 12 plus or minus 10, I can beat everyone!
This seems to be a pretty silly game, kind of like predicting how many times the ball will land in Black in twenty spins of the roulette wheel. Ten, plus or minus four would be a pretty safe bet.
>>
Ce le vie
<<
Joe, it’s “C’est la vie!” I’m not an expert in French, but I did take a couple of years in high school. In this day and age, I prefer the phrase: “C’est la guerre!” It seems more appropriate for what we do here.
Jim
Let there be one Cat 5 landfalling in NYC and all forecasts are wrong. No wait … with all that flooding, the GISS maestro will crow that he was right all along.
I would still like to see someone forecast the DATES of expected hurricane activity, the expected tracks months in advance, and the level of activity for each date for the upcoming season.
Not just some numbers pulled out of a hat. I tried last year with out having access to the background data I think one would need to get it right. For a first attempt it was not too bad, made a couple of miss assumptions that proved to be greatly in error. As a result I will not make any kind of forecast/prediction until I DO have the needed data in hand and a process for evaluating past performance.
Good to others who have access to data and history of their own track record stepping up to the plate, but I would still like to see more details forecast in advance to be happy.
Given that the 1974 tornado and Australian flooding repeated this year I am curious if Hurricane Earl last year, was a close enough repeat of the 1938 New York City hurricane, or if there is a better chance that this years storm tracks will come closer to repeating the event than last year.
Present indications are of a slow decline in global temperatures. I read this will cause an increase in temperature differential between the poles and the equator. Wind speeds will increase. Won’t this also increase the hurricane forecast? Just wondering.
I’ll be putting together a webpage that will give you plenty of (useless) climate model output from the new NCEP CFSv2 system…
I’ll take a Bastardi wild assed guess over a Mann made fact any day of the week. pg
Really, though, let’s remember that estimating uncertainty is really important. It’s especially ironic that Mann should be chided for having appropriate margin of error in his prediction. If I predict something, and the actual result falls outside my error bars, I have failed my error analysis. There’s no getting around that.
OTOH, Mann’s 15.25 -4.0 is odd too. Typically, the value is given a precision to match the margin of error, e.g. 15 – 4.
Mann’s “16 +/-4” is a model of precision compared to the UK’s very own Mystic MET, with their forecasts of “30% chance colder, 30% chance average, 40% chance warmer” prognoses. And that’s with the benefit of their shiny new £31,000,000 supercomputer (now apparently to be upgraded at a cost of a further £10,000,000. Same junk but faster.)
Here’s what Joe Bastardi and Piers Corbyn thought about that:-
http://autonomousmind.wordpress.com/2011/01/02/bastardi-and-corbyn-reply/
My prediction is 15 +/- 5 for named storms which should be about 87% accurate.
However, 15 +/- 8 should get you to over 90%. And 17 +/- 13 should be darn close to 100%. People get paid for this?
Now! Actual numbers from an actual meteorologist, instead of crap from the arch-liar Mann.
Got to quibble, though, Joe. I call 4-6 landfalls(on CONUS shoreline) and only 3 major storms. I think last year’s pattern is going to repeat as far as storm intensity goes. Then again, I is not a meteorological science thingie guy, so what do I know.
I forgot one thing as far as the live and let live, ce le vie, etc.
that the most important thing of all may be that no matter what the weather,
Chaka Kahn loves us all in a NATURAL way, not an AGW way.
No matter how hard I try to drive the age of disco away… it still is a fact no matter
what people do with climate… or hurricanes
Is that 14 named storms OR 14 hurricanes? If it is 14 named storms, you are on the low end. NOAA is now over-zealous in naming tropical cyclones. Storms that would not have had a name 10 years ago get one today. I believe that NOAA is quick to name storms because they know perception is reality. Who knows about the ACE? But they do know about the alphabet. So using more letters makes people think global warming is causing more storms. So, is the prediction for 14 named storms or 14 hurricanes?
Predicting total numbers just turn this into a game. Tell me the probability of my 100 miles of coastline being struck this year so I can appropriately react to the risks.
That is a great post Joe and I appreciate your courage to make a prediction that has meaning. My question is, when can we see the “10 page power point presentation I have with research and ideas that will knock your socks off?”
The total number doesn’t bother me. But I saw an interview with you a few weeks ago where you predicted that we would see more U.S. landfalls this year. Yuck.
So, how many media predictions are we going to have that a hurricane will hit New York this year?
Joe! Welcome back! Will you be doing a European/British weather forecast as you did before? Do hope so…
I predict 20 +/-20 big swirling cloudy windy rainy dangerous possibly tornado spawning weather systems in the Atlantic basin. 20 +/-20 of them will blow leaves off of trees. And my prediction is good for next year also. There, beat that. Now, let’s keep careful score and see if I am right.
I predict (if we have a football season) that the Denver Broncos will win 8 games, plus or minus 4.
I love Joe Bastardi…
With this kind of game, it is right enough and tight enough to be wrong 50% of the time.
With a wide spread it is too easy to be right all the time, but a spread of +/- 1 shows that you care about being tight as well.
Joe,
How about a link to the PowerPoint presentation you mentioned. I’m sure many beside myself would be interested.
w^3
Is there a derby for this, like sea ice?
Mann and Kozar 18.86 +/- 4.3
NOAA 15 +/-3
Bastardi 14 +/-1
When scientists give a +/- range it usually means that based on some calculations, they are 95% confident that the true answer will be in that range. They could easily give a more narrow range if they wanted to be only 90% or 80% or 70% confident, there are formulas to do that.
When non scientists like Bastardi give a +/- range, I do not know what it means to them. The wideness or narrowness of a range is no information by itself without saying what it means.
So Joe failed to be 4X as accurate as NOAA?
Overall, I’d say the value of the forecasts is proportional to the inverse square of the standard deviation. Which makes his 16X theirs.
Credibility is foremost to any forecaster. This includes credibility in the quality of forecasts (we want to be accurate) as well as credibility when evaluating the quality of forecasts (we must acknowledge both successes and failures).
I know Joe and consider him a friend. But I have to comment on the excerpt below, taken from his 2011 seasonal hurricane forecast, looking back on past seasons:
“But its the landfall that is the big deal to me. 07 and 10 burned me in similar ways, but people forget the non seasons forecasted in 06 and 09 for the gulf, the great forecasts in 08 and 03-05”
After reading this you would think the 06 landfall forecast was pretty good, but we just don’t remember “the non-season forecasted in 06 for the gulf.” Perhaps we don’t remember because that forecast was forgettable, arguably one of the worst ever issued. Here’s that 2006 hurricane landfall forecast in graphical form:
http://vortex.accuweather.com/adc2004/pub/images/promos/hurr2006/hurrisk.jpg
According to Joe, “the big deal” is the forecast of 5 hurricane landfalls, 3 of them major. Actual: 0 and 0. End of story. By nearly any reasonable metric, a huge bust.
The interpretation of the 06 forecast presented here by Joe is not an isolated oversight. The same spin on the 06 forecast was used in the write-up for the 2007 Accuweather seasonal hurricane forecast (see http://wx-man.com/blog/?p=608).
Don’t get me wrong here. I’m not criticizing the quality of the forecast or suggesting that I could do better. But putting the 06 forecast in this light reeks of disingenuousness and undermines credibility.