CFAN’s forecast for the 2017 Atlantic hurricane season

by Judith Curry

Climate Forecast Applications Network (CFAN)’s  first seasonal forecast for  Atlantic hurricanes is based on a breakthrough in understanding of the impact of global climate dynamics on Atlantic hurricane activity.

Research conducted by Senior Scientist Jim Johnstone at my company Climate Forecast Applications Network (CFAN) has identified skillful new predictors for seasonal Atlantic Accumulated Cyclone Energy (ACE) and the number of U.S. landfalling hurricanes.

CFAN’s prediction for the 2017 Atlantic hurricane season:

  • ACE: 134 (average value 103 since 1982)
  • # of U.S. landfalling hurricanes: 3 (average value 1.7 since 1900)

CFAN’s research team has long-standing expertise in climate dynamics and tropical meteorology research  and in developing operational forecasts of tropical cyclones on timescales from 1-30 days. This seasonal forecast reflects the first time that this expertise has been integrated into a seasonal prediction of Atlantic hurricane activity.


North Atlantic hurricane activity varies substantially from year to year, including 5-fold variations in annual basin-scale Accumulated Cyclone Energy (ACE). The frequencies of U.S. and Florida landfalling hurricanes also changes sporadically, with recent decades witnessing multiple Category 3+ hurricanes in consecutive years of 2004 and 2005, two of the most active seasons of the past century, but also multi-year periods of quiescence, including an absence of major hurricane and Florida landfalls from 2006 to 2015.

There are numerous seasonal forecasts of Atlantic hurricane activity made by government labs, university scientists and private sector weather service providers. Most forecasts address # of named storms, # of hurricanes and # of major hurricanes. Several also predict Accumulated Cyclone Energy (ACE) and landfall probabilities. These forecasts are primarily based on ENSO forecasts, circulation regimes in the Atlantic, and statistical analysis of historical hurricane activity. Seasonal forecasts made prior to April show little skill, with most forecasts made in June or later showing skill relative to climatology, at least in hindcast mode. Notable recent forecast failures were the intensely active 2005 hurricane season and the low activity in 2013.

CFAN’s analysis of the climate dynamics of global circulation patterns has identified new predictors of Atlantic hurricane activity. CFAN’s analysis of the climate dynamics of Atlantic hurricanes includes consideration of four time scales:

  1. quasi-biennial (2-3 years), dominated by stratospheric signals
  2. interannual (3-7 years), dominated by ENSO
  3. decadal (7-16 years), dominated by Atlantic Ocean circulation patterns
  4. low frequency (15+ years)

Based on our predictability analysis, we have selected ACE and the total number of U.S. landfalls as predictands. The predictors used in this forecast are entirely empirical, based on data from January through May. Hindcasts and predictions using CFAN’s predictors for Atlantic ACE and U.S. landfalls are described below.


North Atlantic ACE is an integrative metric of tropical cyclone duration and intensity based on the square of maximum sustained surface winds during Named Storms.

CFAN’s ACE forecast model was developed from historically correlated patterns of springtime anomalies over the period 1982-2016 of sea surface temperature (SST), SST tendencies, atmospheric sea-level pressure (SLP) and lower-stratospheric temperatures. To remove redundancies in the SST and SLP predictor indices, a principal component (PC) analysis was performed, yielding a single dominant mode with significant ACE correlation (r = 0.67). This index reflects higher ACE totals with low pressure and warm surface temperatures over the North Atlantic, high SST in the SW Pacific, and high SLP and cool and falling SSTs over the eastern Pacific. The unexplained residual variability in the forecasted ACE was found to have significant negative correlations with lower stratospheric temperatures above the North Atlantic (r = -0.5). This stratospheric indicator was combined with the surface-based PC Index in a linear regression model to obtain a final ACE forecast index with improved hindcast skill (r = 0.75).

Figure 1: Comparison of modeled versus observed Atlantic ACE, for the period 1982-2016. The model forecast for 2017 Atlantic ACE is 135.

Based upon statistics since 1982, the average error for CFAN’s ACE forecast is 35, compared to an error of 52 using the climatological average.

It is instructive to compare CFAN’s ACE forecasts with the CSU forecast. The average error of 35 for CFAN’s forecast compares to an average error of 30 units for CSU’s June forecast. In comparing CFAN’s ACE forecast (Figure 1) with CSU’s ACE forecast shown in Figure 2 (Figure 1 of the CSU forecast report), it is seen that the CSU forecast has greater skill prior to 1995, whereas CFAN has greater skill since 2008.

Figure 2: Comparison of observed ACE versus the CSU ACE model.

The largest outliers in both CFAN’s and CSU’s forecast are 2004/2005. Our analysis has identified stratospheric anomalies that relate to the intense hurricane activity during 2004/2005; better understanding of these mechanisms is the subject of ongoing research.


ACE is moderately correlated with most U.S. landfall indices, however the landfall indices display remarkably weak relationships to the same surface anomalies that heavily influence ACE. Inspection of historical hurricane tracks reveals that many tropical storms never approach the U.S. coastline, instead migrating northward and eastward over the open Atlantic. Recent research highlights notable differences between basin-scale cyclone activity and the frequency of hurricanes that reach the U.S. coastline, due in part to dynamical mechanisms that tend to increase general cyclone activity, but also inhibit U.S. landfalls. It is notable that elevated ACE totals from 1995 to 2014 overlapped with a historic 2006-2014 drought of major hurricane and Florida landfalls.

Annual hurricane landfalls for 5 U.S. regions were compiled from NOAA reports of 153 total events over the 1920-2016 period.  Annual landfalls were tabulated for the Gulf Coast (TX to AL), Florida, the Mid-Atlantic (GA to MD), and New England (DE and states northward), as well as the Southeastern U.S. (excluding New England) and the Eastern U.S., including all states.

We identified several precursors in the ocean, troposphere and stratosphere that reflect circulation patterns that apparently influence hurricane tracks and U.S. landfalls. The clearest signal of this circulation regime is seen in the Arctic and the global stratospheric circulations. Using observations from the Arctic during spring, a new Arctic Index has been developed that provides the basis for CFAN’s prediction of the number of U.S. landfalling hurricanes.

The predictors for landfalls vary with the low-frequency Atlantic circulation regime. For the current regime (warm phase of the AMO), we considered the period 1995-2016. Figure 3 compares the modeled number of U.S. hurricane landfalls with observations.

Figure 3: Comparison of modeled and observed annual U.S. hurricane landfalls, for the period 1995-2016.

For 2017, CFAN’s forecast for the number of U.S. landfalls is 2.7-3.4 (average value 3), with a mean absolute error of 1. For reference, the average number of hurricane landfalls since 1900 is 1.7. During the first half of the period, the model landfalls are generally lower than observed, whereas the modeled landfalls are slightly higher than observed since 2009. Because of the short training period (owing to the Atlantic regime change in 1995), the confidence in landfall forecast is lower than for the ACE forecast.

Figure 4 provides scatter plots of the Arctic Index versus the number of U.S. landfalls. For values of the Arctic Index > 0, there are no years since 1995 with zero U.S. landfalls. Identifying the regional location of landfalls is statistically less robust, but with the strongest signal for Florida.

Figure 4. Scatter plot of the springtime Arctic Index versus the number of hurricane landfalls, for the period 1995-2016.


How do CFAN’s forecasts for 2017 Atlantic hurricane activity compare with other forecasts?

  • NOAA Climate Prediction Center forecasts a 45% chance of an above normal hurricane activity, with a predicted ACE range of 75%-155% of the median.
  • The Colorado State University team under the leadership of Phil Klotzbach forecasts normal hurricane activity, with an ACE forecast of 100 (with a mean absolute error of 30) and a near average probability of U.S. landfalling hurricanes.
  • Tropical Storm Risk forecasts average hurricane activity with an ACE forecast of 98 (with a mean absolute error of 48), with a 40% probability that U.S. landfalls will be above average
  • CFAN forecasts an ACE value of 134, with a mean absolute error of 35, and above average U.S. landfalls (3).

CFAN’s forecast is unique in making a specific prediction of the number of U.S. landfalling hurricanes. Further, CFAN’s forecast uses new predictors that potentially provide greater prediction skill for years when there is no clear signal from ENSO (El Nino/La Nina). To date, the 2017 Atlantic hurricane seasonal forecasts have been regarded by forecast providers as highly uncertain owing to continued ambiguity regarding the forecasts for ENSO, although the latest ENSO forecast look to be flatly neutral for the rest of 2017, with slightly positive values.


We are continuing to conduct research on the climate dynamics of Atlantic hurricanes:

  • Investigation the dynamics of stratospheric connections to Atlantic hurricane activity
  • Improve understanding of the causes for the anomalous 2004/2005 hurricane activity
  • Early season predictability and prediction (December and April)
  • Multi-year forecasts (2-5 years)
  • Predicting the next regime shift in the Atlantic
  • Investigation of landfall dynamics during 1926-1971

In future, we plan to integrate the ECMWF seasonal forecasts more thoroughly into the forecast, along the lines of the paper by Kim and Webster (2010). 

Further information about CFAN’s tropical forecast products can be obtained [here].

JC reflections

Apart from the intrinsic interest in the Atlantic seasonal hurricane forecast,  CFAN’s forecast is an interesting example of the sociology of private sector research and how it differs from academic research.

If you are paying close attention, you will see that I do not provide sufficient information for this forecast to be reproduced.  While reproducibility is the mantra (if not the norm) in academic research, in the private sector there are big counter incentives  to giving away your ‘trade secrets’.

Underlying this forecast model is some very significant research into climate dynamics.  Will it slow down academic research progress not to make the details of this research public?  Maybe, but I’m not too worried about it since academic research is focused on other things.

Developing this forecast cost CFAN about $40K in salaries and overhead.  CFAN has one client that is partially supporting this research and forecast product.  In general, CFAN’s research is funded by the occasional government grant, client contracts, and overhead.  More funding is needed for our hurricane climate dynamics research to continue and for regular seasonal hurricane forecasts to be sustainable.  Sponsoring subscribers to the hurricane forecasts will receive full technical reports.  We shall see how this funding model works.

So this is a very different model for climate research.  With President Trump’s funding priorities and cuts, not to mention the endemic group think in academic climate research, this may turn out to be a good path to follow. Time will tell.

104 responses to “CFAN’s forecast for the 2017 Atlantic hurricane season

  1. Judith, what does it matters, or am I mistaken – that you do think there is substantial risk of extreme weather events driven by climate change? If the risks are not greater than historically then why expend the effort to forecast more accurately? And then is it not better for Trump to save taxpayer money?

    • Seasonal hurricane forecast is a big deal for the reinsurance sector and emergency managers. There is large year-to-year variability in hurricane activity, driven by the 4 time scales discussed in the post (not to mention weather roulette).

      Given these large variations, its difficult to separate out any AGW impact (if anything, possible small AGW impact on hurricane intensity).


        There does seem to be some correlation with long term SST but there are factors which cause warmest years to have a slump in ACE. It then depends on how much of the warming of the last 300y you are willing to attribute to AGW.

        Clearly it is not a one to one linear dependency, which will be a great disappointment to all those who want to draw a straight line through any and all data they can find and arbitrarily attribute a non zero trend in either direction as being caused by human emissions of CO2.

        Climatology of the last three decades has started with the ASSUMPTION that the linear trend is ‘obviously’ AGW and they then try to do some science to explain the remaining wiggles.

        Predicting ACE in actually independent of the cause of changes in SST. It just takes SST as one of the inputs.

      • Do reinsurers who use hurricane forecasts make more money than those who don’t, that is who use historical averages.

        It might be hard to tell, what with market-clearing prices. It would come out as fewer bad years for the users of forecasts, or not, in that they’d hope to avoid average bids with an unfavorable forecast and make below average bids with favorable forecasts.

    • John Carpenter

      There is a substantial risk of extreme weather events regardless of climate change. Even if the risks are not greater than historic (due to climate change), historic events are enough to tell us we have not been adequately prepared for them in the past and represent substantial risk for the future. More accurate forecasts could only help our ability to prepare. There is no down side to better accuracy…., if you can truly get it.

      • David Wojick

        Indeed, the big downside is believing forecasts that turn out to be wrong. Believing a false forecast is often worse than believing none.

      • dougbadgero

        “Believing a false forecast is often worse than believing none.”

        Amen brother

      • Which is why the only forecast I put any weight on is looking out of the window in the morning. I no longer even have any interest in whether weather forecasts got it right or not.

        “What is the forecast for today?” Dunno, I don’t bother looking.

    • Uncle Robot ==> I have lived in Hurricane Country for the last 15 years, much of it on a vulnerable sailboat in the Caribbean. Hurricane forecasts are part and parcel of daily life in the season. How many, when, where, how strong….? Predicted hurricane paths determined where we moved the boat and when.
      In the Northeast US, state, county, and municipal officials are concerned with winter snow forecasts — how many snow days, how much snow — how many snow plows do they need, how much road salt, how many employees to keep on thru the winter.
      In the corn belt (and other rain-dependent agricultural areas) the farmers need to know when, where and how much rain their will be — it regulates plowing, planting, and harvests in some cases.
      Regional and near-term (weeks/months) forecasting is the most important part of weather forecasting — and our economy and well-being is oddly dependent on it.

  2. Roger Knights

    TYPO: “instead migrating northward and westward eastward over the open Atlantic.”

  3. Pingback: CFAN’s forecast for the 2017 Atlantic hurricane season – Enjeux énergies et environnement

  4. 75% chance Florida takes a hit this year? Yikes!

    Do you think the major hurricane drought for a US landfall is over also Judith?

    • We haven’t looked at major hurricanes, but the statistics for this are even skimpier than for total hurricanes

  5. We shall see how you do. With all due respect it seems like reading tea leaves and playing the odds to me.

  6. You wrote: “The frequencies of U.S. and Florida landfalling hurricanes also changes sporadically…” I thought Florida was part of the U.S.

  7. A gutsy move, Dr. Curry, and I hope it pays off for you.

    • PhySciTech

      Well, that’s how science is supposed to work.

      • Don Monfort

        Explain why predicting a specific number of landfalls is science. How meaningful for science will it be, in the unlikely event that that there are exactly (3) landfalls in 2017? Are the other expert organizations predictions not as scientific, because they don’t go out on a limb and predict an exact number of landfalls? How about a probability for landfalls being exactly (3)?

        I don’t know much about the science, but this seems to be more about marketing than science and potentially sets up a big flop. What do you say if it’s (0) landfalls, or (6)? I guess you could fall back on the fact this stuff is unpredictable…

      • Don

        Our met office has recently taken to give weather forecasts with a range of temperatures. For example ‘there will be heavy showers in the east and more continual range in the west with temperatures ins range from 9 to 15 degrees centigrade.’

        That is not helpful. What do you wear? 9 is very chilly whilst 15 is very mild.

        This could be the approach Judith needs to take: ‘there will be between none and 12 hurricanes this year’

        That should cover all eventualities and Shecould always say her method is endorsed by the Met office. :)

      • Tony B, surely that range is for a large area with different conditions. For example, I live at 1800 ft above sea level. Less than 15 miles away it is 900 ft and often 5 degrees F warmer. I look at ridge tops that are 1200 ft above me and cooler still. But the local forecast knows these differences and often gets them right.

      • Don Monfort

        Is your reply supposed to be amusing in its entirety, Tony? If not, what has it got to do with explaining why predicting a specific number of landfalls is science? Do you think a prediction of precisely (3) landfalls is going to be more helpful than the predictions of the other experts? How likely is (3) to be accurate? If three landfalls turn up in the first month of the season, do the folks near the sea assume that the coast clear, until next year?

      • Don, I would say that predicting that number is engineering. But discovering a new way to predict it is science. We have both in this instance. Interestingly, the greater possibility of being wrong the more scientific it is, provided the success or failure feeds back into the underlying hypotheses.

      • David

        This is when they give a national forecast then summarise it at the end instead of region by region

      • Don Monfort

        I don’t get it, David? Discovering a new way to predict something is not necessarily science. What if the method of prediction ain’t useful? I will take any even money bets on (3).

      • Don, discovering a new way to predict something is the core definition of science, broadly speaking. Unless the case is chaotic, in which case science means discovering why it is not predictable. When it comes to climate we need more of the latter. For example, I doubt that 20 years of statistics is sufficient to know that there will be at least one hurricane landfall in the Atlantic US.

      • Don Monfort

        What if the new way doesn’t actually predict, David? Suppose it’s (0) landfalls this year? They try again next year and predict (0) landfalls. It turns out to be (3). Are we still talking science? My point is that predicting a specific number is not necessarily science. Particularly, if you don’t give the freaking probability of it being exactly (3). What happened to the uncertainty monster? This looks to me like a marketing ploy designed to differentiate their services from the competition. Not a good idea, in my opinion.

      • Don Monfort

        I just took the time to read the rest of the comments.
        “Modeled range is 2.7-3.4, with MAE of 1. These numbers are in the blog post. Figure 4 shows that since 1995, there are no years with high positive Arctic Index that have zero landfalls. So i have high confidence in 1 or more landfalls

        Because of the transient nature of the skillful predictors (e.g. varying with phase of AMO), the period over which the model is developed is not long enough for high confidence in the actual number of landfalls”

        I would have gone with “1 or more landfalls”, instead of (3).

      • Ask your reinsurer friends about the average cost of landfalls, Don Don. Here’s something to cheer you up:

        Betting two landfalls under is a neat way to lose enough money to ever buy in these predictions forever again.

      • Does it help the science of insurance to consider those who say the rising costs in damages is do to the rising number of coast-dwellers and not CAGW?

      • Thank you for your concerns, Hark, but we’re not talking about science. Science would be a bit more open than that about its process.

        It’s not science but it’s important. That it’s not science may not hinder scientific progress. On the contrary. Think about it.

  8. Reblogged this on Climate Collections.

  9. Ben Franklin made some money with long-range weather forecasting. I wonder what models he used?

  10. Applied research by private companies fulfills an important role in knowledge progress. Hurricane, cyclone and monsoon intensity prediction are very, very important as they are among the most destructive manifestations of climate/weather variability. In the end the generated knowledge ends in the general pool of scientific knowledge without being paid from taxes, while generating salaries and improving economical performance, so it is a win-win situation.

    I wish you a great success in your endeavor. I’ll try hard not to wish more hurricanes than average for this coming season ;-)

  11. Inquiring minds want to know: may UK’s elderly again be forced to burn books this winter to stay warm during?

  12. Judith Curry: ACE: 134 (average value 103 since 1982)
    # of U.S. landfalling hurricanes: 3 (average value 1.7 since 1900)

    Can you supply a range estimate, here or to your clients?

    thank you for the essay.

    • Modeled range is 2.7-3.4, with MAE of 1. These numbers are in the blog post. Figure 4 shows that since 1995, there are no years with high positive Arctic Index that have zero landfalls. So i have high confidence in 1 or more landfalls

      Because of the transient nature of the skillful predictors (e.g. varying with phase of AMO), the period over which the model is developed is not long enough for high confidence in the actual number of landfalls

  13. Steven will not approve. 40k sounds a lot cheaper than a grant for that work. Maybe there is a deal with the government in the cards…

  14. If a single hurricane makes landfall, the news will be screaming how it’s Trump’s fault. I know that isn’t exactly relevant to the forecast, but I can make it relevant by phrasing it this way.

    There’s a 75% chance newspapers will scream that Trump has destroyed all those buildings in Florida.

    • I am not sure I understand this American fixation on hurricanes. They cause less damage and deaths and injuries than either thunderstorms or tornadoes.


      • Tonyb.
        Fixation is a time honored American pastime.
        Probably at the heart of our successes and failures.
        I also think this ‘climate change’ thing is best described as a fixation rather than rational observation of facts.
        My opinion on the second has been highly influenced by you. :)

      • tonyb
        New Orleans in 2005 Katrina was a major disaster.!!

        Gulf coast US is subsiding into below sea levels with earth berms likely to be overtopped in Hurricane. Disaster with massive damage and heavy loss of lives. Building sea walls may help.

        Tornadoes are more random and little can be done to stop them.

        Missisippi river dumps channelized water with sediment in the deep ocean instead of helping rebuild land forms. New Orleans may have to be really built up to have storm surge protections for billions of $s.

        Much more likely to bring disaster than 2*C in 100 years.

      • rebel and scott

        I amass an eclectic selection of books, preferably printed some time ago, which are useful at giving snapshots on previous weather.

        I have a book on British gardening from the 1930’s which talks of the gradually warming climate and details the list of plants that could then be grown. After losing the ability to grow many of them again in the 1960’s and 70’s they became possible again in ensuing decades, but some are now tricky again, like runner beans and tomatoes.

        Our climate/temperature in the UK is on a slight downward decline over the last 15/20 years or so, which mirrors this finding, which does not stop us having periodically very warm weather like the current spring (although I had to put on our central heating a dozen times during May.)

        My point on hurricanes was that as it is essentially affecting those living on the land/sea axis at certain latitudes, it affects a relatively small amount of people, notwithstanding the biggies such as New Orleans which seemed to have as much to do with inadequate infrastructure as anything else.

        If we didn’t have the Thames Barrier London would be regularly flooded and similarly there are many cities around the world such as New Orleans that are known to be vulnerable. They can be readily identified as they have had problems in the past (Galveston) and surely funds should be set aside to deal with these known problems>

        Back to my main point: I have a 1945 book called ‘the elements rage’ in which hurricanes (amongst other types of weather) are discussed. It seems to me that hurricanes are these days less frequent and much less costly in terms of life now, than they were in the first half of the last century.

        conversely thunderstorms appear to have increased markedly. they redistribute air in an extraordinary manner and with some 100,000 in America annually I do wonder if they are more worthy of putting under the microscope than hurricanes, causing more damage, loss of lives and, as I say, the potential to export cold/warm air in a manner that may be climatically meaningful


      • You send the new guy to stand in the wind near the shore dressed for the weather. He doesn’t have too much information to tell us. The time frame is longer. Predictions can be made. Grocery stores can be bought out. The national guard may show up. Roads will flood. Waves will crash. And my favorite, surfters may show up.

      • Tony b, it is the concentration of deaths and destruction that makes it interesting. Individual plane crashes are more interesting than the much larger sum of car crashes.

        “See the bubble headed bleach blonde, she comes on at five. She can tell you bout the plane crash, with a gleam in her eye.” From “Dirty Laundry” (from memory, may be off some). A personal favorite, being a sometime journalist. “Kick’em when the’re up. Kick’em when the’re down.”

      • Tonyb
        Human experience is biased and probably influenced by the fossil fuel industry,
        Computer models are pure and without sin.
        The farther into the future the projection, the more pure, as the sinners will either forget or move on to the afterlife.

      • That’s because places that get hit by hurricanes are hardened against them and the population knows 24-hours or more in advance if they are going to be hit by them. If you knew a tornado was going to hit at 5 p.m. tomorrow on Elm Street, it wouldn’t kill anyone.
        What’s more interesting is the fact that damage and injury reduction depends on people hearing, believing and acting on the forecasts. For the last decade we’ve been given forecasts of average to above average hurricane seasons, which means the entire coastal population of the U.S. has been conditioned to think a “normal” or even above normal hurricane season is “no hurricanes.”
        Which means that there will be a whole lot of folks entirely unprepared for the next significant landfall.

  15. CFAN differs from the other 3 forecasts. It will be interesting in late October to see how this pans out. On the one hand, rooting for CFAN. On the other hand, living directly on the Atlantic in Fort Lauderdale, not so much.

  16. ristvan
    I thought you lived in Wisconsin?

    • That is my dairy farm with its associated 4br/2bath farmhouse– hunting, fishing, snowmobiles. Used to keep riding horses. I also have a town home in the Chicago suburbs on a golf course. Moved permanent residence to Fort Lauderdale in a career change to join a startup. Stayed because no state income tax and homesteaded property taxes. A largely virtual existence anyway as global senior consultant and exec.

      • What kind of horses? I have Walkers and Rockies. Both largely indifferent to extreme events, except tornados of course. The only thing that drives them inside is flies, which have yet to appear.

      • For more see my Exploring instinct as expert knowledge.

      • DW, we had three quarter horses for the kids (one a former national champion stallion, retired and gelded) and two Arabian mares for long distance riding. My Gazada was 16.5 hands, and a handful. We had a two horse Merhow trailer, and took the younger Arabian Bathsheba plus the best quarter horse Sonny down to a Chicagoland stable about a half hour from Winnetka where my wife and daughter could train indoors all winter long. Made sense, because I would load up the F250 HD farmmtruck tower with three face cords of firewood from the farm. Never had to buy any for the fireplace at the Chicago manse. (When we used the farm actively in winter up until the kids went to college, I would put each year 8-10 facecords into the cellar. All hand split. Never needed a gym membership.

      • DW, looked at your horse cognition blog. Much truth.
        We actually had as my horse handler teacher (I knew nothing at the outset of my decades long farm horse adventure about handling, saddling, ridng, farrowing–hoofs have frogs?!?) a famous local SW Wisconsin horse whisperer, John Bonine. He learned in Montana from the original. I do not claim horse whisperer skills, only that I was a student of a real second generation whisperer.
        The basics are that horses speak to their environment, their herd, and us, using instinctive body language. Learn the body language, then talk back. Easier said than done. I ended up somewhere better than a deaf mute horse but definitely a learning disability challenged low horse IQ.
        BTW, the 1998 Robert Redford movie Horsewhisperer has a hokey plot, but is full of actual horsewhisperer truth even if he is not one. I have been there and experienced that.

      • Yr blog is fascinating, David.
        Led me to this… talking with horses.
        I talk ter birds. At least not ter trees, (yet.)

      • Per Frank James, my ancestors, who raised fine horses and mules, used to sell getaway horses to the James Gang… Jesse and Frank, and the Younger Gang.

    • Thanks.
      I enjoy Wisconsin from the rural parts. Mother in law from there.

      Also like SF and northern California in Trinity Mountains and San Diego.

      I wish there were more time to explore more places.
      Interesting writeups from your articles.

  17. Pingback: CFAN’s forecast for the 2017 Atlantic hurricane season – NZ Conservative Coalition

  18. Judith

    Is there solid rationale for ACE using named storms as a baseline metric? Can’t total energy be the same if there are a higher number of large storms just below the threshold?

    • Rob, not Judith but know a little bit about ACE since wrote it up. Energy is proportional to V^2 times time at V. So unnamed storms wont contribute a lot of cumulative energy. The metric seems good enough. H/t Ryan Maue.

  19. ENSO forecasts are now resolutely neutral. I have an idea that this reflects an assumed persistence but the system seems instead to be in a dynamic flux.

    Both polar annular modes are trending less positive in recent times – and this is modulating both Atlantic and Pacific states as both north and south Pacific gyres spin up. From March –

    To the most recent image –

    I frequently review these images to monitor the evolution of SST in key regions. What we had in March was a La Nina Modoki which looks to be evolving into a dispersal of warm surface water in the equatorial eastern Pacific, more equatorial upwelling, cooler north eastern Pacific temperature, a cooler Southern Ocean and cooler Northern Atlantic and Pacific conditions. It all seems symptomatic of a system in flux.

    I am looking for a La Nina to emerge in the Austral spring – which would conventionally be expected to lead to less wind shear in the Atlantic and a lower than average cyclone season. Even more speculatively – I am looking for extreme ENSO events that mark a transition between multi-decadal Pacific states.

  20. Ristvan
    Given your comments above relating to writing ACE.

    From what I can determine, ACE is the gross energy created by a cyclone event. It takes zero account of the wind velocity, volume and angle of entry contributing to and maintaining the cyclone. No wind no cyclone ?

    If this is the case there is no clear measurement of the net efficiency derived from the cyclone’s other inputs such as sea surface temperature, vertical pressure column etc.

    What influence is the wind, and what influence contributing factors on the gross output ?.

    A comparison using the Americas cup yachts currently racing in Bermuda is appropiate, where by they achieve a boat speed considerably faster than the wind. Efficiecy.The same could be said for cyclones ?

  21. Judith, very impressive stats, but seeing the performance over the next 5 or 10 years will be more telling. I looked at the 06z and 18z GFS runs today (June 8) and both runs show a tropical storm moving northward across the Gulf of Mexico around June 18-23, which is still pretty far out to have much confidence. The 06z run indicated it might reach minimal hurricane intensity and hit Louisiana, but the 18z run shows a weaker system hitting southeast Texas around June 23. Oddly, the 18z run forecast for June 23 also shows a very early hurricane hitting Barbados at 360 hours, which is highly unlikely, but interesting that the model is indicating unusually favorable conditions.

    • The GFS has been trying to spin up a hurricane around Central America for the past month. The ECMWF shows a very low probability of anything. The ECMWF is very likely correct in its prediction over the next two weeks, and not the GFS

  22. Living at the Jersey shore we always looked forward to Dr William Gray’s hurricane predictions. Guess I am showing my age but when he spoke…we listened

    • Carl ==> Phil Klotzbach apprenticed under Dr. Grey for many years and produces a forecast here dedicated to him.

  23. Peter Lang


    Well done. You seem to be doing really well so far. I hope your business succeeds. I also hope you will be able to find enough time to keep CE going and to keep contributing to the public debate – especially testimonies to Congress.

  24. ENSO to remain flat?
    No el Nino?
    Probably right, given the strength of anchovy catches in Peru:

    which spells strong upwelling and thus brisk trade winds.


    Outline for MEI webpage (updated on June 8th, 2017)

    …The next update for the MEI will hopefully take place before July 8. ENSO-neutral conditions appear to have been replaced by a fledgling El Niño, at least in the MEI-sense. Meanwhile, the PDO continues in positive territory, thus not interfering with El Niño-typical impacts. Daily updates of the ENSO status can be found at the TAO/TRITON website, confirming the recent disappearance of equatorial cold anomalies, and the replacement of easterly wind anomalies by weak westerly anomalies near the dateline.

    • “The El Niño–Southern Oscillation (ENSO) remains neutral. The Bureau’s ENSO Outlook remains at El Niño WATCH, meaning there is around a 50% chance of El Niño developing in 2017—double the normal likelihood. However several indicators have shown little or no increase for several weeks, suggesting El Niño development has stalled for now.

      Sea surface temperatures across the tropical Pacific remain warmer than average, though cooling has occurred in some areas over recent weeks in response to stronger than average trade winds. The Southern Oscillation Index has also eased to near zero values. All other ENSO indicators also remain neutral.
      Four of eight international climate models suggest tropical Pacific Ocean temperatures may exceed El Niño thresholds during the second half of 2017, down from seven of eight models that were forecasting a possible event in April. Virtually all models have reduced the extent of predicted ocean warming compared to earlier in the year, indicating that if El Niño forms, it is likely to be weak.”

      As I said – the indicators are neutral – as are the model forecasts. The La Nina Modoki introduces into the system a different set of variables – with a double Walker Cell and water level increases on both sides of the Pacific. The usual ENSO indicators need to be understood in the light of conditions that are neither La Nina or El Nino.

      The La Nina Modoki has faded – as it does in April/May. Sea surface anomalies show that elevated heights have dissipated on both sides of the Pacific. The potential energy that creates the conditions for El Nino is not in evidence.

      Cooling in the eastern Pacific and stronger than average trade winds in recent weeks suggest to me that the coupled ocean and atmospheric dynamics are with the evolution of a La Nina normal in the Austral spring.

      • Go argue with Wolter, or reality:

      • The reality is a weak El Nino – but more probably increased eastern Pacific upwelling. IMHO.

        The near term evolving ENSO dynamic is something that Claus Wolter doesn’t address in his monthly update. The cold anomaly – the La Nina Modoki – in the central Pacific has disappeared. The latest MEI seems to have ticked into El Nino territory – although this doesn’t define El Nino conditions. The northern Pacific has cooled in the past month – and is not covered by JISAO yet. Trade winds are strengthening in the east. I don’t need to argue with Claus Wolter about anything. Current ENSO indicators are neutral.

        What can be seen in this is cold anomalies pushing north from the Southern Ocean into the Peruvian current – as the Southern Annular Mode drifts into negative territory. We are feeling the chill as far north as Central Queensland. It results in the initiation of upwelling in the eastern equatorial Pacific – which seems to be happening over the past few weeks.

        Upwelling sets up wind and current feedbacks that drive the system to a La Nina normal state. As I said – the ocean and atmospheric dynamics seem to be favouring La Nina in the Austral spring.

        Unlike JCH I don’t deal in trendology and I prefer not to bother with the misleading certainties he pretends to on almost no evidence.

      • TAO/TRITON data are pretty much as I said – and the mean of international ENSO models is neutral. JCH has been predicting an El Nino for months – despite my cautions about the spring predictability barrier.

        The prognosis is for a weak El Nino at most – but I think that the dynamics favour increased flow in the Peruvian and Californian currents and more cold and nutrient rich upwelling.,-17.78,307/loc=-119.226,0.295

      • . JCH has been predicting an El Nino for months – despite my cautions about the spring predictability barrier. …

        This is a lie. What I have said, very early on, was that 2017 could easily be either a record warmest year or a near record warmest year… either 1st or 2nd. I have stated what is factual… that ENSO forecasts were for El Niño. Fact… they were predicting El Niño.

        You are a despicable little pos.

      • June 8 IRI ENSO update:

      • “Historical accuracy of models is lowest in late (Austral) autumn, but begins to improve for outlooks generated in June.”

        Here’s the POAMA model.

        And here’s JCH whining about how the models have changed and being angry and insulting. Pathetic.

      • Good to see JCH and Robert going toe to toe. JCH is very proud at predicting doom and gloom so I hope the Cheifio wins.
        For what it is worth it is a random walk process reverting to the mean. Being positive it is closer to El Niño at the moment so a small deviation up for a few months give El Niño.
        Obviously it has a higher chance of doing so from where it is than when it is negative. On the other hand the overall odds of it going down become higher the more it moves away from zero. It is all pur guesswork.
        Extra effects though give an idea of which way the trend will move short term. If the sea surface temps are getting hotter the trend will be upwards for a few weeks. Robert gives some of these indicators with a well thought out explanation. It is still subject to the vagaries of the currents. JCH sticks with the trend whenever it is up going and disappears when it is going down, like a good stockbroker.
        Makes for interesting watching.
        Are you still predicting an El Niño and a hottest ever year for 2017, JCH?

      • JCH | June 10, 2017 at 10:04 am |
        . JCH has been predicting an El Nino for months – despite my cautions about the spring predictability barrier. …
        This is a lie. What I have said, very early on, was that 2017 could easily be either a record warmest year or a near record warmest year… either 1st or 2nd. I have stated what is factual… that ENSO forecasts were for El Niño. Fact… they were predicting El Niño.

        JCH | April 2, 2017 at 12:22 pm |
        “the combination of external forcing and internal variability should lead to accelerated global warming. This accelerated warming appears to be underway, with record high GMST in 2014, 2015, and 2016.
        Call me names. Call me a cheerleader. Don’t care… I called this.
        The stadium is waving a HEATWAVE… for how long, nobody knows. But for now, long enough to make 2017 a record warmest year, and maybe even 2018”
        JCH | April 4, 2017 at 4:06 pm |
        El Niño is forecast,

      • Now this is a prediction months ago.
        Apology, anyone?
        ” JCHJanuary 25, 2017 at 12:32 PM
        The current post El Niño SST map looks sort of like those in JAN 2004 and JAN 2006. Both were followed by the resumption of El Niño later in those years. Not going to be surprised if there is an El Niño in 2017, and I don’t think a 4th warmest year in a row is impossible.”

  26. This is OT, but EPA’s Pruitt has endorsed the Red Team proposal:

    Dr. Curry and this blog (!) are quoted. Of course the alarmists are against it, claiming the science is settled. Such a debate could set the stage for reversing the EPA endangerment finding, on the grounds of uncertainty. EPA just delayed the horrendous ozone rule, arguing uncertainty, so Pruitt knows this strategy.

    • What is most needed is to have unbiased, objective, rational, analyses of the costs and benefits of global warming. I expect, if this was done without being controlled by the climate alarmists it would clearly show global warming will be beneficial for this century and long after.

      Go Trump! Get it done, asap!

  27. Also OT, but closely related, my Climate Change Debate Education project has posted its Strategic Plan:

    A Red Team vs Blue Team debate might provide useful educational content. Almost everything that is presently available at the K-12 level is one-sided, on the alarmist side. It is not a matter of the students getting into the climate change debate, which can be very technical. It is a matter of seeing that the debate is real. It is like going to a laboratory and seeing that the science is real.

    • A Red Team vs Blue Team debate might provide useful educational content. Almost everything that is presently available at the K-12 level is one-sided, on the alarmist side.

      The most important thing the students need to understand is that there is not threat and global warming is beneficial, not damaging, if the Social Cost of Carbon is negative (i.e. a benefit). What is most needed is for a substantial part of the climate research expenditure and effort to be diverted from trying to understand and project climate change, to estimating the real, total economic impact of global warming – as Tol and other’s have been doing but have had to rely on sparse to negligible objective, unbiased, competent studies of impacts.

      Remind the students, in the end politically sustainable policy decisions always boil down to “It’s the Economics, Stupid

  28. Geoff Sherrington

    Dr Curry,
    One can think of less speculative income bases, but you seem to have high demand. I do not know how much subjective input you have put into CFAN, so it would not be correct for me to be worried about wishing you well in a gambling venture. But, in my reading, you are doing very well over the years at being right. So I can happily say that I wish you well.

  29. Pingback: | thePOOG

  30. I have a very simple question. Why are there no hurricanes (tropical cyclones or whatever you want to call them) in the SE Pacific and the South Atlantic oceans? (I would post a graphic, but I don’t know how.)

    In lieu of a graphic see:


  31. Predictions or averages?
    This appears to be an average of the likely conditions for this year given the (semi-secret) input you choose to use.
    A prediction would be more like
    “We expect three land touching hurricanes to develop this year, one in mid July and a further 2 at the start and end of August. Conditions will be such that the expected ACE will be 145 this year.”
    Averages are what the BOM use 3 monthly in Australia but call predictions. Anyone wanting a laugh could check there latest two tries. Predicting a warm dry 3 months MJJ the weather became exactly the opposite, it poured down. Anyone with access to the weather patterns must have known that 2 wet weeks were due at the start of that forecast but they did not change it. GIGO.
    While I understand the insurance concept the whole point of insurance is to cope with the unknown.
    The best one can say is after such a long run of outs it is time for a few hurricanes to make landfall.
    Not statistically correct technique but commonsense. Even if we have 10 years with no land touching coming up we are still closer to it happening than last year.
    I appreciate that conditions now imply a higher likelihood of hurricanes this year, but it is no guarantee of anything actually happening.

  32. “An extremely simple univariate statistical model called ‘IndOzy’ was developed to predict El Nino-Southern Oscillation (ENSO) events. The model uses five delayed-time inputs of the Nino 3.4 sea surface temperature anomaly (SSTA) index to predict up to 12 months in advance. The prediction skill of the model was assessed using both short- and long-term indices and compared with other operational dynamical and statistical models. Using ENSO-CLIPER(climatology and persistence) as benchmark, only a few statistical models including IndOzy are considered skillful for short-range
    prediction. All models, however, do not differ significantly from the benchmark model at seasonal Lead-3–6. None of the models show any skill, even against a no-skill random forecast, for seasonal Lead-7. When using the Nino 3.4 SSTA index from 1856 to 2005, the ultra simple IndOzy shows a useful prediction up to 4 months lead, and is slightly less skillful than the best dynamical model LDEO5. That such a simple model such as IndOzy, which can be run in a few seconds on a standard office computer, can perform comparably with respect to the far more complicated models raises some philosophical questions about modelling extremely complicated systems such as ENSO. It seems evident that much of the complexity of many models does little to improve the accuracy of prediction. If larger and more complex models do not perform significantly better than an almost trivially simple model, then perhaps future models that use even larger data sets, and much greater computer power may not lead to significant improvements in both dynamical and statistical models. Investigating why simple models perform so well may help to point the way to improved models. For example, analysing dynamical models by successively stripping away their complexity can focus in on the most important parameters for a good prediction.” file:///C:/Users/rober/Documents/enso.pdf

    “The forecast skill of individual models and the MME depends strongly on season, ENSO phase, and ENSO intensity. A stronger El Niño is better predicted. The growth phases of both the warm and cold events are better predicted than the corresponding decaying phases. ENSO-neutral periods are far worse predicted than warm or cold events. The skill of forecasts that start in February or May drops faster than that of forecasts that start in August or November. This behavior, often termed the spring predictability barrier, is in part because predictions starting from February or May contain more events in the decaying phase of ENSO.”

    The predictability barrier was hinted at by Judith in terms of how ‘skillful’ predictions are. Not at all beyond some 3-4 months – and less at certain times and ENSO phases. There is an object lesson here.

    • The predictability barrier for ENSO refers specifically spring. A reorganization in the Pacific takes place March-May, tied to the seasonal cycle. Hence ENSO predictions before May show little to no skill, whereas by June (certainly July), the signal is in place and there is prediction skill often as long as 6 months if there is a strong signal.

      We are investigating some global signals that might serve as precursors across the spring predictability barrier, to produce longer range forecasts. Work in progress.

      • Yes – I referred to it as the Austral autumn above. In general – predictability drops off rapidly and while there may be a prediction the reliability is a priori unknown.

        ENSO is a difficult one because of the rapid transitions caused by feedbacks in wind and currents. In La Niña, the Pacific trade winds intensify causing sun warmed surface water to pile up against Australia and Indonesia. Cool subsurface water rises in the east. In an El Niño, the trade winds falter and warm water spreads out eastwards across the Pacific Ocean. ENSO has a influence on global surface temperatures, Australian, American, Indian and African rainfall and Atlantic cyclones. ENSO varies between La Niña and El Niño states over 3 to 7 years but also over periods of decades to centuries. One mode of ENSO variation involves changes in both the frequency and intensity of La Niña and El Niño over at least a few decades.

        There is still a ‘debate’ about whether ENSO is a purely resonant phenomenon or whether it is stochastically forced. It seems clear to me – for a number of reasons – that the trigger is the polar annular modes. But how predictable are those? The suggestion is that they are modulated by solar UV/ozone chemistry in the stratosphere. But how predictable is solar UV and how predictable is the influence on the annular modes? Even if it were predictable – it is merely a highly variable trigger in a chaotic system. Atlas supports the world but it is turtles all the way down.

      • Oh and there are a number of problematic conditions for predictability – as detailed above. Not just the spring barrier.

  33. @judith
    ‘in consecutive years of 2004 and 2005, two of the most active seasons of the past century’

    Sorry to bother you with this, but 2004 and 2005 belong to THIS century, not the past one. 😁

  34. Judith Curry, I admire your courage in attempting to predict the results of the hurricane season. Using historical (in-sample) data to construct a predictive model is fraught with subtle problems and traps. You have to account for changing variables and the effects of those variables on hurricane frequency and intensity over the historical period. Some of this problem can be alleviated by using appropriate cross validation techniques with k-fold cross-validation coming to mind. I would assume with your background and expertise in this area of climate science you would be able to control any over-zealous model builder who might want to use fits to variables and parameters that do not make good scientific sense.

    My questions to which I do not necessarily expect answers are:

    1. A client would need in the end information concerning the damage a hurricane could do to a business entity or to an insurance enterprise. Predicting a regional ACE with frequency and intensity in itself would not necessarily provide that information unless the location were rather precisely known. Therefore my question would be if the ACE and landfall parameters are merely used to give a measure of the chance of damage to a particular location or alternatively if the entity requiring the predictions has risks in many potential areas of the damage how much would the location prediction – even in a general way- play in these probabilities of damage.

    2. Obviously if you are allowed to update your predictions with real time data your prediction capability would increase many fold just as is the case of those modelers who do this for hurricane tracking and intensity. Is that type of prediction of value to your clients? And if so would not those predictions depend on entirely different models?

  35. Pingback: Weekly Climate and Energy News Roundup #273 | Watts Up With That?

  36. Hello, Judith,
    I’m cheering for CFAN’s success. Meanwhile, can you point me to any studies showing, for some or all of the other hurricane activity forecasters you mention, how well their forecasts have matched future reality? As a consumer, I tend to discount “hindcast” claims heavily, since a model can be “tuned” to give results that match past reality to any degree.


    “Only two previous tropical storms in recorded history have developed this far east before July (in the region of the deep tropics between the Lesser Antilles and Africa): One formed on June 24, 1933, and the other on June 19, 1979. The two seasons during which these June storms formed were quite active and included some of the most infamous storms in history.”

  38. Will tropical storm Cindy count as one of the three landfalls?

    • technically, no. Our scheme predicts hurricane landfalls. However, the early easterly wave that developed (Bret) is a harbinger of an active season. Stay tuned.