More on Hurricanes



  1. Kenneth Fritsch
    Posted Jun 7, 2008 at 3:15 PM | Permalink

    Continuation from the thread: Hurricanes 2008

    David, if one judges the fit to a Poisson distribution to be reasonably realistic than the occurrences of intense storms would be assumed to be random in nature and reasonably unaffected by SST. If one thought that SST was a factor one could use, for a regression analysis, what is called a “generalized linear model” and include SST as an independent variable. One then calculates maximum likelihood values for the regression parameters and with it a calculation of the expected occurrence rate. From here one then performs a chi square test for goodness of fit. These calculations are handled well in R, but I do not have a good familiarity for doing them.

    An alternative method is to divide the occurrence data into 2 or 3 groups depending the value of a corresponding state variable, such as SST, by high and low and/or high, low and neutral and then determine whether the fit to a Poisson distribution improves over the fit without dividing the occurrence data. Unfortunately dividing the sparse data that we have for the SH basins since 1980 does not appear to make for a practical calculation.

    All of the foregoing is explained in the article titled “The influence of climate state variables on Atlantic Tropical Cyclone occurrence rates” by Thomas A. Sabbatelli and Michael E. Mann 2007. I think I sent you a copy of it, but if you need one I’ll email it to you. I am not a statistician, but I do not see any thing wrong with taking a quick look at a linear regression as a first approximation as a long as one understands all the limitations that assumes.

  2. Jonathan Schafer
    Posted Jun 7, 2008 at 4:05 PM | Permalink

    In response to David’s question, I’d like to add a follow-on question. Why is it assumed that higher SST’s will necessarily lead to more intense TC’s? The formation of a TC is so randomized to begin with, with so many factors that have to be in place at the right time, for a TC to even begin. Assuming it does begin, there are then continuing and new factors which come into play, which can increase and/or decrease a TC’s intensity.

    As a working hypothesis, I can understand how the higher heat content of the ocean would potentially provide more fuel for a TC during it’s lifespan, especially if the heat content is available at depth, instead of just a shallow surface increase. But translating that into actual storm intensity is a much different story. And how exactly, would you be able to measure the probability that any particular storm had an increased intensity due to a higher SST? How would you remove any and all other factors?

  3. David Smith
    Posted Jun 7, 2008 at 5:37 PM | Permalink

    Kenneth, thanks for the note and the reminder about the Sabbatelli paper – I had fogotten about it. Sometimes it’s good to revisit the basics.

    I’ve been comparing the 6-hr pressure changes of the early 1980s and recent 2000s SH cyclones. The recent SH cyclones seem to have more-rapid reported pressure changes. My conjecture is that this indicates improved (more sensitive) means of evaluating storm strength. Or, maybe it’s real and due to some change of climate. Or, maybe my review is wrong due to the short periods covered so far. I’m expanding the years analyzed to see if this observation holds water.

  4. David Smith
    Posted Jun 7, 2008 at 6:11 PM | Permalink

    Jonathan, those are good questions and many opinions abound. I hope that Judith, as a key player, could offer her summary.

    As you know, Emanuel’s potential intensity hypothesis is the key. One of the often-missed aspects of the hurricane/SST debate is that most of the “skeptics” find the potential intensity hypothesis to be plausible. I, for one, think it makes sense. But I also expect the future effects to be comparatively modest (it depends on what SST rise one assumes) and any past effects to be undetectable in the historical record, due to the relatively small SST rise so far and the nature of the data.

  5. Geoff Sherrington
    Posted Jun 7, 2008 at 6:55 PM | Permalink

    Re # 2 Jonathan Schafer

    There is some SH evidence that a SST of about 28C is needed for a tropical cyclone to form. The global SST is not important, the local SST is – I agree with the thrust of your comment. It is confusing that cyclones do not commonly form over the hottest world waters, nor head towards them.

    Excess heat in the sea is mainly in the top 100-200m. I am informed by a retired senior scientist who spent a career studying such things, that a column of atmosphere has about the same intrinsic energy as the top 3 m of sea water below it. The rate of transfer of energy depends inter alia on the SST; and latent heat conversion (evaporation, mainly)increases at about 8% per degree C. Again, the situation is more complex than that because the storm moves over the sea and derives energy from a moving scene.

    So, I am told, SST can sensitively affect storm formation, but I understand it is far less clear if it can affect intensity on present theories.

    I am not expert in these matters and rely upon the advice tendered. Please let me know if I have mentioned unreferenced material and I will endeavour to source it.

  6. Kenneth Fritsch
    Posted Jun 7, 2008 at 7:33 PM | Permalink

    Re: #2

    And how exactly, would you be able to measure the probability that any particular storm had an increased intensity due to a higher SST? How would you remove any and all other factors?

    As layperson reading the literature my general take away is that, as you indicate, the hurricane formation and intensification is a very complex process. In fact, the more I read, the more complex I understand it to be.

    First we know that there is a threshhold SST for TC formation so at some point SST has to have an effect. Pat Michaels did a paper where he actually correlated SST in storm tracks to individual hurricane intensities and as I recall he saw an SST threshhold for formation and another one for maximum intensity. His analysis would indicate that at sufficiently increased SSTs that we would not see the most intense hurricanes increasing in intensity, i.e. the maximum intensity would not increase, but one could deduce that the hurricanes with lesser intensity at lower SSTs would became more intense (up to a threshhold) at higher SSTs.

    Recently we have been made more aware of the effects of another TC variable that could reduce the number of TCs and hurricanes and perhaps even have an effect on their ability to reach higher intensity and that being wind shear. As a layperson I am impressed with effects of wind shear that I have seen correlated with TC activity from historical data.

    In recent CA posts, David Smith also brought up the observation that intense storms appear to intensify in short periods of time when apparently just the right conditions are coincident to push the storm to Cat 4 or 5. There is also much historical data that indicates that these TC and hurricane occurrences follow a Poisson distribution (that can be modified by other variables such as AMM and other cyclical climate phases) which would be consistent with their sudden intensification when the “right” conditions collide in a more or less random fashion.

    I personally view correlating SST with storm occurrence or even intensity a difficult proposition because as you noted there are many other important variables. Further complicating matters when using the historical records one must deal with changing measurement and detection capabilities over time and the cyclical nature of TC activity that can easily lead to confounding a coorelation of SST with TC activity, i.e. when they both increase over the same time period.

    Kerry Emanuel’s simple thermodynamic exposition of increasing TC intensities with increasing SSTs is probably valid as far as it goes but from my layperson’s reading of it, it is far from a comprehensive treatment that addresses all the potential factors. I have observed a disagreement between William Connelly (small to no effect)and Issac Held (not a small effect) on how much the temperature gradient from surface to troposhere would have on the TC intensity. This issue is further clouded for me by the fact that the climate models predict a warmer tropical troposphere as the surface warms(limiting storm intensity) while observations show that the tropical troposphere is cooler (tending to increase storm intensities) than predicted.

  7. Judith Curry
    Posted Jun 8, 2008 at 8:44 AM | Permalink

    Really interesting discussion, i will tune in more on friday (unfortunately horrendously busy). SST doesn’t provide much info on how much a given storm will intensify. and there can even be anticorrelations of intensity/count with ENSO. The Hoyos et al. 2006 paper showed how to assess the impact of the SST trend empirically.

    The role of SST on hurricane intensity is actually quite complex in a climatic sense. Temperature gradients are important in setting up circulations that influence wind shear. And as global sea surface temperatures warm, this warming in itself doesn’t contribute to increased intensity, but its the intensity of the atmospheric convection that seems to be increasing. Potential intensity theory is usually used to provide the link with SST, but a recent Vecchi paper showed that this is not a simple link. Personally i think potential intensity theory needs some rethinking.

  8. David Smith
    Posted Jun 8, 2008 at 9:12 AM | Permalink

    “Climate data”, Forrest Gump might say, “is like a box of chocolates – when you bite into a piece you never know what you’ll get”. That’s not surprising, because much of the data was collected for meteorological or other purposes where coarseness or gaps in knowledge were OK for their purposes. And, of course, measurement methods have changed. The important point, of course, is to bite carefully and chew well before swallowing.

    An example is the Southern Hemisphere (“SH”) tropical cyclone data. I recently took a small bite to examine its characteristics and to explore whether its quality has changed over time. The database is the one made available via Ryan and Kenneth in earlier posts and covers 1980-2007, a period when measurement technology and practices were evolving. The question is, is this 1980-2007 data good enough to allow statements to be made across that period?

    First, below is a plot of the “intense cyclone” (estimated minimum pressure below 946hPa) count since 1980. The 946hPa was a breakpoint suggested in a recent paper:

    That plot seems to show an upward trend in “intense cyclones”.

    My next step was to pick beginning and end periods and examine some of their characteristics. I chose 1980-1984 and 2001-2006, inclusive. I first looked at the pressure estimates for each 6-hr period of the storms’ durations. For example, a storm lasting 10 days would have about 40 such pressure estimates.
    (Note – some storms in the database are missing these estimates.)

    I placed the values into bins of 5hPa, for ease of display.

    The expectation is that there should be a high count near 1000hPa (the pressure of young and weak storms), decreasing towards lower pressures (the pressures of intense and older storms). Here is the plot for 2001-2006:

    I inserted a blue line to show the division point between regular and “intense” cyclones.

    The plot is pretty much what I expected. The data tails downward, relatively smoothly, towards the lower pressures. There is no unnatural-looking bumpiness.

    Next comes the plot for 1980-1984:

    The data tails downwards, as expected, but it’s an odd, uneven decline unlike 2001-2006. Importantly, there is a big bump at 946-950 hPa (just above the defined “intense” point) and then a major dropoff.

    Some might say this distribution is real and the data is OK for use but it seems like the burden would be on them to offer a physical explanation for the bumpiness, especially the one at 950 hPa.

    Does this “matter”? Well the next step was to plot the distributions of minimum pressures in the two periods. Here is 2001-2006:

    There appears to be some tendency to estimate pressures to the nearest 5 hPa but that’s reasonable, and the distribution doesn’t tail off smoothly (but that may be due to me not using 5 hPa bins for the data). All in all it looks OK.

    The red line shows the defined “intense” cyclone breakpoint.

    Now here’s the 1980-1984 plot:

    Wow. That’s a very unnatural appearance (binning would have helped, I agree). There appears to be a tendency to estimate storm intensities to the nearest 10 hPa (990, 980, 970, 960, and especially 950 hPa. To me it indicates that the meteorologists of the day did their best but simply lacked the tools/technique to more-tightly estimate minimum pressures. They tended to round the estimates, which is reasonable, but users of the data must recognize that and be careful.

    Does it matter? Suppose I do a sensitivity check and redefine “intense” as 950 hPa and below, rather than 945 hPa (there is nothing magic about 945 hPa). Using 945 hPa to define “intense” gives 19% of all 1980-1984 storms as intense, rising to 28% in 2001-2006. Using 950 hPa to define “intense” gives 44% of all 1980-1984 storms as intense, falling to 33% in 2001-2006. A reversal of fortune, so to speak.

    How about if we use 950 hPa for the entire period 1980-2006? Here goes:

    The plot becomes a very unremarkable no-trend. What a difference 5 hPa makes.

    I do not contend that there is no-trend in “intense” SH cyclones. I do not contend there is a trend. My contention is simply that the data quality, as it exists today, is simply too poor to allow much to be said with confidence. A reanalysis using satellite images may help that, but even that technology changed since 1980 and continues to improve.

    Bite with care and chew well before swallowing.

  9. Kenneth Fritsch
    Posted Jun 8, 2008 at 9:59 AM | Permalink

    Re: #8

    Does it matter? Suppose I do a sensitivity check and redefine “intense” as 950 hPa and below, rather than 945 hPa (there is nothing magic about 945 hPa). Using 945 hPa to define “intense” gives 19% of all 1980-1984 storms as intense, rising to 28% in 2001-2006. Using 950 hPa to define “intense” gives 44% of all 1980-1984 storms as intense, falling to 33% in 2001-2006. A reversal of fortune, so to speak.

    David, good job in doing my favorite test, i.e. a sensitivity check. It reveals a lot. By the way, Steve M, had a recent lead-in to a thread where he, in my mind, endorsed (or at least permitted) the use of senstivity testing by auditors.

    The next step is to determine how the pressure threshhold of less than 946 was determined. If by a prior reasoning that is fine or even by arbitrarily selecting a level (without peeking at the end results), but if it was cherry-picked that is not so fine. In fact it is, to carry the box of chocolates analogy probably a step too far, like “thumbing” the pieces of chocolate to avoid any surprises.

  10. Kenneth Fritsch
    Posted Jun 8, 2008 at 10:07 AM | Permalink

    Re: #8

    On further looking at your graphs, David, I noticed that by using 950 as a cut-off for pressure, a significantly larger number of storms can be used to look for a trend. I counted 143 storms using less than 946 and 235 using less than 950.

  11. Posted Jun 8, 2008 at 7:23 PM | Permalink

    #8 Excellent research David. As I asked a while back, what was special about 945 mb. Apparently because 950 mb had no trend.

  12. Posted Jun 8, 2008 at 8:19 PM | Permalink

    Here is a small contribution to the effort, a slightly more spatially oriented approach to show where the hundreds of points between 945 – 950 mb reside (over 950 such observations between 1980-2007) in the Southern Hemisphere.

    A link to a larger image version: Larger Figure The density map is built on 2.5×2.5 degree squares and then smoothed once using a simple box average (9 points), but those details are not critical. The Solomon Islands and Coral Sea have a huge difference for reasons that I can only speculate at the moment.

  13. David Smith
    Posted Jun 8, 2008 at 8:30 PM | Permalink

    Ken, I’ve redone the xy scatterplots using the additional storms (less than 951 hPa). The r values actually drop slightly. No significance.

    I also redid the plots of SH cyclone minimum pressures using 5 hPa bins rather than individual values. I’ll provide links to those once WordPress overcomes its current technical difficulty.

    The large majority (over 70%) of the 950 hPa storms were in the Southwest Pacific, east of 160E. Odd. I wonder which agency did those storm evaluations.

    On a different topic, I highly recommend Ryan Maue’s Powerpoint . There are some intriguing correlation locations for predicting Atlantic ACE, including the Gulf of Alaska.

  14. David Smith
    Posted Jun 8, 2008 at 8:52 PM | Permalink

    Supplemental plots:

    Minimum Pressure of SH Cyclones 2001-2006, placed in 5 hPa bins The natural pattern is more visible here than in the earlier unbinned plot.

    Minimum Pressure of SH Cyclones 1980-1984, placed in 5 hPa bins .

    SST vs Storm Count for 160E-200E, storms below 951 hPa

    SST vs Storm Count for 120E-160E, storms below 951 hPa

    SST vs Storm Count for 40E-120E, storms below 951 hPa

  15. David Smith
    Posted Jun 8, 2008 at 8:54 PM | Permalink

    Re #14

    Oops, the first plot is here .

  16. Carl Smith
    Posted Jun 8, 2008 at 8:59 PM | Permalink

    David, in the South Pacific Basin:

    W of 160E is BoM Brisbane Aust. AOR,
    E of 160E and N of 25S is FMS Nadi Fiji AOR
    E of 160E and S of 25S is NZMS Wellington NZ AOR

  17. David Smith
    Posted Jun 9, 2008 at 10:40 AM | Permalink

    Re #16 Thanks, Carl. Do Australia and the French split coverage for the Southern Indian Ocean and, if so, where does the break occur?

  18. Kenneth Fritsch
    Posted Jun 9, 2008 at 2:16 PM | Permalink

    Re: #12

    A link to a larger image version: Larger Figure The density map is built on 2.5×2.5 degree squares and then smoothed once using a simple box average (9 points), but those details are not critical. The Solomon Islands and Coral Sea have a huge difference for reasons that I can only speculate at the moment.

    Ryan, your map reminds of the NATL and the historical distributions of TCs for east and west of longitude 60w in the storm basin — except the effect would appear to be for the opposite ends of the TC intensity scale.

  19. Kenneth Fritsch
    Posted Jun 10, 2008 at 5:50 PM | Permalink

    David, in summarizing the difference between the SH intense storms using pressures less than 946 hPa and 950 hPa, I used your data below to compare the trend lines and the three SH basin for fit to Poisson distributions using the chi square p values.

    SH intense storms with pressure less than 946 hPa:

    y = 0.136x -265.4
    R^2 = 0.187

    SH intense storms with pressure less than 950 hPa:

    y = -0.024x + 56.4
    R^2 = 0.006

    Poisson fits for pressures less than 946 hPa:

    SWP: p = 0.34
    Near Aust. : p = 0.92
    SIO: p = 0.33

    Poisson fits for pressures less than 950 hPa:

    SWP: p = 0.83
    Near Aust.: p = 0.61
    SIO: p = 0.29

    While the Poisson fits (with sparse data) show a reasonable fit for both cases, the trend lines tell an entirely different story.

    I went back to the NATL storm tracks for the period 1980-2007 to look at the minimum pressures and how they related to using a less than 946 and a less than 950 hPa cut-off.

    Less than 946 hPa cut-off would have missed 6 Cat 4 hurricanes (at 947, 948, 946, 949, 950 and 946) while a less than 950 hPa cut-off would have missed 1 cat 4 storm (950). Three storms less than a Cat 4 would have been included using the less than 950 hPa cut-off, with 2 Cat 3 storms (at 934 and 946) and 1 TS at 945, while two storms less than category 4 would have been included using the less than 946 hPa cut-off with 1 Cat 3 storm (at 934) and 1 TS (at 945).

  20. Posted Jun 10, 2008 at 10:26 PM | Permalink

    (I’ve tried to post this several times over the last couple of days, but I think the spam filter keeps eating it!)

    David, in the South Indian Ocean:

    W of 90 E is the MFR Reunion Is. TCWC AOR
    E of 90E and N of 10S is the BoM Darwin Aust. AOR
    E of 90E and S of 10S is the BoM Perth Aust. AOR
    (as far as the historical database is concerned)

    The Australian region itself is a bit more complex, being divided among three Tropical Cyclone Warning Centres, being the BoM Perth TCWC, the BoM Darwin TCWC, and the BoM Brisbane TCWC – these divisions are complex to describe with coordinates, but to make it a bit easier you can see them on a map here:

    However, just to complicate matters further, responsibility for the SH area N of Perth’s AOR has been transferred from the BoM Darwin TCWC to the BMG Jakarta TCWC Indonesia within the last 12 months which you can be sure will give TC researchers of the future yet another headache to deal with! – see the BMG Jakarta TCWC website here:

    And if you thought things were complicated enough already, wait, there’s still more in the pipeline!

    Yet another change will be the upcoming handover of responsibility to PNG of their AOR at some time yet to be determined – as PNG seems like a bit of an economic and political basket case at the moment, this transfer is on an indefinite hold – however the area to the N and of the Brisbane TCWC AOR and part of the area to the NW the FMS TCWC AOR will most likely be handed over to the Port Moresby TCWC at some time – currently it is being handled from Brisbane W of 160E and FMS E of 160E.

  21. Posted Jun 10, 2008 at 10:26 PM | Permalink

    Hey … it worked! 🙂

  22. Geoff Sherrington
    Posted Jun 11, 2008 at 4:11 AM | Permalink

    Re # 19 Kenneth Fritsch

    Opinion, please – is 946 hPa a tipping point or an artefact from data collection and rounding? I don’t know, I’m interested whether true tipping points really happen in climate.

  23. Kenneth Fritsch
    Posted Jun 11, 2008 at 9:29 AM | Permalink

    Re: #22

    Opinion, please – is 946 hPa a tipping point or an artefact from data collection and rounding? I don’t know, I’m interested whether true tipping points really happen in climate.

    Geoff, I am more a data processor and a reader of TC literature, but, since I have lots of opinions, here’s mine on this observation. I suggest that that there are others posting here who are much better qualified to answer your question.

    I do not see 946 hPa as a tipping point nor do I have possession of any a prior knowledge that would indicate one should exist in this situation. I think that David Smith’s sensitivity test using 946 and 950 hPa simply demonstrated that the trend seen using 946 hPa was an artifact from the data collection.

    The reasonable fit to a Poisson distribution in both cases would indicate that much of the frequency of occurrences of these SH intense storms for the 1980-2007 period occurs mostly as a random result of just the right conditions prevailing at the same time. The differing storm count means from basisn to basin would indicate that the tendency to get the right conditions varies with basin.

    I also think that Ryan Ms maps differentiating the intensity of TCs using 946 and 950 hPa could have some bearing on explaining at least part of the difference. The maps, if I am interpreting them correctly, do remind me of what I have seen in the NATL when looking at the easier to detect TCs versus the total TCs detected and how it has changed over time.

    Using 950 hPa would seem more in line with the Cat 4 and 5 hurricanes classifications used in the NATL, but all of these classifications I suspect are rather arbitrary. Sensitivity tests like David Smith used are helpful in these cases and particularly so if a classification was suspected of being cherry picked to give a desired result – and I am not saying that David made the test because he suspected cherry picking. It does, however, go along with my motto of: trust, but verify.

  24. Bob Koss
    Posted Jun 11, 2008 at 6:52 PM | Permalink

    According to Unisys these are the Saffir-Simpson Scale pressures for the categories. 946 hPa would be just inside category 3.

    Category 1 > 980
    Category 2 965-980
    Category 3 945-965
    Category 4 920-945
    Category 5

  25. Bob Koss
    Posted Jun 11, 2008 at 6:56 PM | Permalink

    Less than sign cropped my comment.

    Category 5 less than 920

    There doesn’t seem to be a coherent relationship between pressure and wind speed in the Atlantic basin. There are Cat 5 tracks as high as 945 hPa and Cat 3 tracks as low as 923 hPa. All three categories have tracks at 930 hPa with winds ranging from 105 kts to 150 kts. If they say the central pressure is 930 without indicating a wind speed it could be any of the three categories. I think a lot of subjectivity goes into coming up with the values.

    I graphed the 1980-2007 Atlantic pressures for Cat 3-5, but for some reason I’ve been having a problem for the last couple days when trying to include a link or embed an image in my comments.

    The 948 hPa is by far the most popular and has 14 Cat 3 and 33 Cat 4 tracks.

  26. Posted Jun 11, 2008 at 7:53 PM | Permalink

    An issue with central pressure and wind speed relationships is cyclone extent.

    Small radius compact cyclones (a.k.a. ‘midget’ cyclones) have a steeper wind gradient and higher central pressure for the same maximum wind speed as larger radius cyclones. Examples of midget cyclones in the Australian region include TC TRACY 1974 and TC LARRY 2006.

    There is also the issue of the surrounding pressure field in the region around the outer closed isobar of the cyclone – if the cyclone is in an field of higher pressure it means a higher maximum wind speed for a given central pressure, and if the cyclone is in a field of lower pressure the maximum wind will be less for a given central pressure.

    The variability in both cyclone extent and surrounding field pressure confounds any investigation based on central pressure alone – but that is all we have to go on in many cases, so any investigation should be have a few caveats attached.

  27. Geoff Sherrington
    Posted Jun 11, 2008 at 10:10 PM | Permalink

    Re # 23 Kenneth Fritsch

    Thank you for the response, and from others. The # 24 categories of Bob Koss are of different size (numerically) so I guess that we can presume that past workers have used convenient groupings to envisage data more easily. Not saying this is wrong, just wondering if there is underlying physics.

  28. John A
    Posted Jun 12, 2008 at 5:17 AM | Permalink

    Piers Corbyn of WeatherAction has produced a new prediction for the first tropical storm of the Atlantic season to be around the 18-22 of June:

    “We expect it to start in the Caribbean / Gulf of Mexico region and soon develop into Hurricane strength and make landfall (75% likely) or near landfall by 21 or 22 June in Florida / Southern States or possibly Cuba or (less likley) the west side of the Gulf .

    “This Tropical storm will be associated with severe tornado events in central / SouthEast parts of the USA while New York /NorthEast USA are likely to be unusually cold” said Mr Corbyn. “After that we are 90% confident the storm will quickly die and the period 23 June to 5 July will be essentially storm free” he said.

    I reproduced the entire e-mail on my blog (minus the overformatting that Piers is so fond of)

    Time will tell.

  29. Kenneth Fritsch
    Posted Jun 12, 2008 at 11:38 AM | Permalink

    Re: #24

    According to Unisys these are the Saffir-Simpson Scale pressures for the categories. 946 hPa would be just inside category 3.

    Category 1 > 980
    Category 2 965-980
    Category 3 945-965
    Category 4 920-945
    Category 5

    To be fair these data from Bob Koss above would give a reason that the cut-off of 945 and less pressure was used for the SH intense storm classification. Cat 4 and Cat 5 storms are considered major (intense) in the NATL. That does not, however, affect the sensitivity results that David Smith obtained. I agree, as Bob Koss has implied and Carl Smith has stated, that using a minimum pressure or even wind speed as cut-offs for categories should be used with all the necessary caveats.

    Cat 4 and 5 storm classifications have been used by a number of climate scientists to make claims that hurricanes are intensifying in the NATL. Perhaps we need a another Smith sensitivity test in that storm basin. I once saw a reply to the Cat 4 and 5 claims that indicated if Cat 3.5 storms were included the claim had to be modified. I thought that replier was being facetious, but not any more.

  30. Posted Jun 17, 2008 at 10:37 AM | Permalink

    There is reason to believe that a major-El Nino could be developing, and we will know within the next 10-14 days depending on the propagation of the MJO. A major news conference is planned Major Atmospheric Shift that would basically end any chances that the Hurricane season would be close to normal. There are some inconsistencies in the press release which would imply a major cool event in the future. Perhaps word will leak out prior to the conference…

  31. David Smith
    Posted Jun 17, 2008 at 3:12 PM | Permalink

    Re #30 Ryan, you sure raised my curiosity with that press release.

    In the past 30 days, far-reaching
    changes have occurred in the atmosphere not seen in almost 70 years

    I looked back in the records and there was a long-lasting El Nino about 1940. I see nothing else in that 60-70 year window. I’ll check for major changes in the last 30 days.

  32. Jonathan Schafer
    Posted Jun 17, 2008 at 7:59 PM | Permalink


    Interesting. The last ENSO advisory said the models were indicating that an El Nino could form, or that a new La Nina could begin. I believe that was dated June 5. The indicated the current transition to ENSO neutral first.

    A majority of the recent dynamical and statistical SST forecasts for the Niño 3.4 region indicate a transition to ENSO-neutral conditions during June – August 2008 (Fig. 5). During the second half of the year, the majority of models reflect ENSO-neutral conditions (-0.5 to 0.5 in the Niño-3.4 region). However, there is considerable uncertainty during this period as some models suggest the possible development of El Niño while others show a re-development of La Niña. Based on current atmospheric and oceanic conditions and recent trends, a transition from La Niña to ENSO-neutral conditions is expected during June-July 2008.

    From your link

    In addition to a timely discussion of energy markets, the forum will focus
    on recent research by one of the members of the panel (Harry Van Loon) which
    suggests the potential for a significant decline in the average mean
    temperature globally beginning in the next 12 months.

    Wow, that’s saying something, and not for good. A signficant decline in the average mean temperature could be devastating to parts of the world. It will be interesting to see what they say and what actually happens.

  33. David Smith
    Posted Jun 17, 2008 at 8:43 PM | Permalink

    An interesting interview with Van Loon is here . He’s a veteran of meteorology and climatology.

  34. Jonathan Schafer
    Posted Jun 22, 2008 at 6:38 AM | Permalink

    Has anyone heard anything related to this? I think the press conference is the 26th, which is just a few days away, but if something big were brewing in the atmosphere, you’d think word would start to leak out about it.

  35. steven mosher
    Posted Jun 26, 2008 at 3:36 PM | Permalink

    any leaks on this yet?

  36. Jesper
    Posted Jun 26, 2008 at 5:05 PM | Permalink

    I wouldn’t expect anything earthshaking. This looks like a hyped-up press release for your typical quasi-statistical arm-waving climate projections. Look at the titles of the talks in the sidebars.

    Talks probably focused on the peak and decay of the recently strong La Nina, the anticipated shift from La Nina to El Nino in the next few months to 2 yrs, and statistical implications for hurricane counts…yawn….

  37. David Smith
    Posted Jun 26, 2008 at 8:39 PM | Permalink

    These are a few admittedly unsubstantive notes on a graph from a recent paper:

    World Climate Report has this article on tropical cyclone landfalls in Louisiana over the last 200 years. It shows no trend in storms except that the green dots (major hurricanes) seem to become arguably more numerous in recent decades.

    Some of the modern green dots looked surprising so I went back and looked at the record. My comments are in blue:

    When I make those (arguable) adjustments to the green dots I get this:

    The green dot pattern in this revised plot shows even less trend over the 200 years.

    Now, I have not read the paper and perhaps the author has stated reasons for inclusion/exclusion of those particular dots. I can’t imagine the reasoining, though.

    (End of unsubstantive notes.)

  38. kim
    Posted Jun 26, 2008 at 11:01 PM | Permalink

    35 (SM) Yes, anyone with knowledge please weigh in. I’m particularly interested in what Harry van Loon says.

  39. RomanM
    Posted Jun 27, 2008 at 8:43 AM | Permalink

    #37 David

    2. The 1974 green dot (Carmen) was a weak cat 2 at landfall
    3. The 1985 green dot (Elena) didn’t produce hurricane-force winds in Louisiana

    Carmen and Elena go beyond the question of the inclusion or exclusion of these two storms as green dots. Because the status of the storms would have to be determined by information of the storms’ activity outside Louisiana – information which would likely not have been available had the storms hit the coast in earlier years – the author of the graph has introduced a distinct observation bias toward undercounting major storms in the past. Given that the graph indicates an overall higher total number of hurricanes in the past, it seems that number of missing green dots in the early history would also be substantial. The criteria should remain consistent throughout the entire time period of the graph: only storms which could be judged as “major” by their status at the time of landfall deserve a green dot.

  40. Larry T
    Posted Jun 27, 2008 at 10:58 AM | Permalink

    I believe that Katrina was only a Cat 3 when it hit land with only it extra large size and the depth of its effects inland making it seem even more devastating. The rescue of so many who were in immediate danger (or roofs tops, trees etc. ) was minimized by the press treating of the New Orleans mismanagement of the evacuation as the major crisis. I am not saying that it was not a major problem or that mistakes were not made but on the majority it was successful. The cleanup has been hampered by squandering and misuse of federal funds by NO and LA.

  41. David Smith
    Posted Jul 2, 2008 at 8:46 AM | Permalink

    A couple of notes –

    1. Most of the weather models forecast that an area of disturbed weather (a seedling) in the far eastern Atlantic will become a tropical storm soon. The seedling has been moving westward, steered by the low-level environment. However, if it turns into a tropical cyclone then it will be steered by the higher-level environment. This higher-level environment, on average, tends to nudge storms slightly north of due west, veering storms away from the Caribbean and other land.

    The relevance of this is that, if the eastern Atlantic warms in an AGW world and we get storms forming farther and farther east, then there may be a tendency for those early-forming storms to veer out to sea, away from land. A warmer eastern Atlantic may result in fewer storms visiting the land areas of the western Atlantic. This current seedling may serve as an illustration of the tendency to curve. We’ll see.

    2. This week we (my wife and kids plus a thousand zillion relatives) are on the beach near Pensacola, Florida US. This beachfront community was almost totally destroyed in 2004 by Ivan and was also hit by earlier storms. In this community in 2008 the reconstructed dwellings are stronger and built more-wisely. For example, air conditioner compressor units are mostly on stilt platforms above the flood level. Insurance rates are now very high – if you want to live here, you pay. These high rates also affect the rental market, as they are passed on to renters which discourages some potential ones. This affects future construction. There is a considerable amount of unreconstructed land near the beach.

    I don’t know what portion of this changed behavior is due to government and what is due to insurers or private owners. It’s probably a combination of factors. The bottom line is that this community is adjusting things to lower its vulnerability.

  42. Kenneth Fritsch
    Posted Jul 2, 2008 at 11:16 AM | Permalink

    Re: #41

    I don’t know what portion of this changed behavior is due to government and what is due to insurers or private owners. It’s probably a combination of factors. The bottom line is that this community is adjusting things to lower its vulnerability.

    That’s all well and good, David, but are there expectations that if worse comes to worse that the state and federal governments will be there to bail people out. I believe I read recently that the federal government was going to enter into the re-insurance business with the aim of lowering some of these insurance rates. The state of Florida, I also believe, was funding a reserve of money to be used in case of future storm and perhaps other disasters. I am not sure how large that reserve is.

  43. David Smith
    Posted Jul 3, 2008 at 8:20 AM | Permalink

    The British Met Office forecast for 2008 tropical storms (July-Nov) is here . The forecast for July-Nov storms is 15. If I assume that the average ACE will be 9 then their seasonal ACE outlook is 135, in the above-average category.

    Note that the website provides a free copy of their detailed 2007 outlook, which I haven’t yet read but which should be interesting reading.

  44. David Smith
    Posted Jul 3, 2008 at 8:50 AM | Permalink

    Re #42 Point taken, Kenneth. If government provides or otherwise subsidizes low-cost insurance for risky behavior, especially if there are no accompanying restrictions on activity, then much of the current incentive is undermined.

  45. Posted Jul 7, 2008 at 8:54 AM | Permalink

    Bertha has intensified to 80 knots, which puts the ACE of the storm at just over 4 so far (thru 12Z 7/7/08). The expectation is that the storm will recurve out to sea, which surprisingly has been the prognosis from the HWRF model all along. The last hurricane in the central Atlantic was also a Bertha in 1996, which maxed out at Category 3 and hit North Carolina. Last year, no storm even attempted to traverse the same latitude waters.

    The ACE will accumulate quite fast with these Cape Verde type storms, if they continue to occur in succession. July is a little early for the season to start out there. So far, the ACE is about 5, with the Northern Hemisphere as a whole slightly above average. However, unless the Western Pacific wakes up, the 2008 calendar year will be another below average year (which is expected with the previous La Nina episode — more on this later).

  46. Posted Jul 7, 2008 at 9:56 AM | Permalink

    45 (Ryan):

    Why does recurvature surprise you?? Climatology for Cape Verde hurricanes at that latitude strongly suggests recurvature the vast majority of the time.


  47. Posted Jul 7, 2008 at 10:12 AM | Permalink

    #45, recurvature does not surprise me, HWRF’s ability to predict it correctly does.

  48. David Smith
    Posted Jul 7, 2008 at 10:31 AM | Permalink

    What impressed me about Bertha was that the GFS, and sometimes the other models, latched onto the development many days in advance. I wonder what it was about that particular seedling and environmental conditions that made the GFS peg the situation pretty well.

  49. Jonathan Schafer
    Posted Jul 7, 2008 at 11:23 AM | Permalink


    Venturing a guess, I’d say it was mostly based on very low shear and SST’s. I believe what typically inhibits these types of Cape Verde storms this early are lower SSTs and high wind shear. Bertha has been for the most part under at most 10-15 kts of shear and higher than normal SSTs for that part of the Atlantic. The only thing that surprised me was the 1.5 days she spent in less than 26C waters without really losing strength. Perhaps it was partially due to the fact that it was still a weak TC, and not as reliant on utilizing huge amounts of heat from the ocean to sustain itself.

    GFS still shows a course towards Bermuda, whereas HWRF forecasts a recurvature based on a trough moving off the east coast of the US in a couple of days. Apparently, this difference is due to GFS understating the strength of Bertha. The stronger the storm, the more likely it is to feel the upper jet stream and be affected by it.

  50. Posted Jul 7, 2008 at 11:58 AM | Permalink

    David #48, I do not give GFS much credit. Check out this animation of the past 46 forecasts by GFS: GFS Last 46 forecasts. Plotted is the maximum wind speed at 925 hPa for each grid point through 6 days, a nifty way of showing tropical cyclone tracks.
    GFS has been bullish on an African wave developing since early June, and I think got lucky for the wrong reasons. Ditto for the EPAC.

  51. Posted Jul 7, 2008 at 12:53 PM | Permalink

    49 (Jonathan):

    Another reason for a lack of Cape Verde storms this early in the season is the dry air coming off the Sahara.

    47 (Ryan): So true!!

  52. David Smith
    Posted Jul 7, 2008 at 2:37 PM | Permalink

    Re #50 Neat animation, Ryan. I’ve seen so many blown calls by the models over the years so I’m probably too easily impressed when one gets genesis (somewhat) right well in advance of the event.

  53. Larry T
    Posted Jul 7, 2008 at 2:58 PM | Permalink

    I am not a fan of the AGW argument but I would like to ask a question that may support that theory. Katrina in addition to being a Cat 5 at some point (and I believe a Cat 3 when it hit land) was of a much larger size than a normal hurricane. Is there a match between hurricane size and the supposed increase in temperature.

  54. Jonathan Schafer
    Posted Jul 7, 2008 at 3:08 PM | Permalink


    Yes, that matters too. I think this year so far, sahel dust has been average so it has not had a real impact on Bertha. The two issues with sahel dust tend to be the ingestion of dry air/dust into a TS and the reduction of SST due to greater scattering of light due to the dust particles.

  55. Jonathan Schafer
    Posted Jul 7, 2008 at 3:20 PM | Permalink



    Even if you could ascribe an increase in size to a particular storm, how would you separate what the “A” part of GW is and what the natural variability in GW is? Not all GW is “A” related.

  56. Posted Jul 7, 2008 at 7:04 PM | Permalink

    #53, I have some un-peer-reviewed research (as of yet) that shows that the hurricane power actually levels out on average for storms greater than than about 105 knots intensity in the Atlantic. This means that the size and intensity are not well correlated for the strongest of hurricanes. This in itself is not a new finding at all.

    Thus, for every Charley 2004, you get a Katrina 2005 size storm. Storm size is a function of latitude, time of year, and a bunch of other considerations like inner-core processes. No skill in forecasting storm size or intensity has been shown to my knowledge.

  57. Posted Jul 8, 2008 at 7:00 AM | Permalink

    56 (Ryan):

    I agree. I subscribe to the chaos theory when discussing storm size and intensity. It may be among the last weather mysteries man solves. My personal belief is that storm size is tied to formative environment. If there is a large envelope of atmoshere in the initial circulation, and that environment is moisture rich, and if that environment has upper air support, then maybe, just maybe, you’ll get a Katrina size storm. But not always. We should probably look at such things as odds, rather than exact numbers, since I do believe chaos theory does come into play.


  58. Posted Jul 11, 2008 at 10:56 AM | Permalink

    With Bertha swirling around aimlessly in the North Atlantic, the ACE for the Northern Hemisphere has passed 100. This is approximately 9 days ahead of climatology, which I keep diligent track of at my simple website. This somewhat above climo activity is expected due to the spring La Nina conditions which favor active Western Pacific and Indian Ocean activity. The opposite is generally true after an El Nino winter.

    Bertha has accounted for over 15 units of ACE, which is nearly 25% of last year’s activity in the North Atlantic.

    Yet, unless the Eastern Pacific and Western Pacific basins begin to show signs of activity again, climatology will quickly pass up 2008, and quickly plunge to below normal activity around the first of August.

  59. bender
    Posted Jul 11, 2008 at 5:08 PM | Permalink

    Ryan and David,
    Recall my prediction of a below normal season and also my reasoning, based on significant 5 and 10-yr autocorrelations in hurricane frequency. Perhaps the MJO is the source of these autocorrelations? Regardless, I’m sticking with my numbers for 2008, 2009, 2010.

  60. Jonathan Schafer
    Posted Jul 11, 2008 at 8:39 PM | Permalink

    A question about Bertha…She has been categorized as anywhere from CAT 1 to CAT 3 over her lifespan, but until today, had not been visited by the hurricane hunter aircraft at all. Categorization was entirely based on satellite estimates. How realistic is it that Bertha was actually a CAT 3? Is the satellite measurement accurate enough to make that kind of determination?

  61. Posted Jul 11, 2008 at 9:15 PM | Permalink

    #60, the Dvorak scheme has been subjectively applied by skilled forecasters as well as by automated objective techniques. There is considerable analysis that goes into an intensity determination in the Atlantic. Bertha is not extraordinary in terms of its structure, just its longevity at this point.

  62. Jim Arndt
    Posted Jul 11, 2008 at 10:09 PM | Permalink

    Ryan # 61

    I know that size is not speed but if they didn’t visit then how can the sats. know the wind speed except by clouds. Don’t you need DOPLER or direct measurements to know wind speed? Maybe they used a GISS algore-ithm. / sarcasm off

  63. Jim Arndt
    Posted Jul 11, 2008 at 10:17 PM | Permalink

    Ryan #61

    I almost forgot in my sarcasm. If this year is the same as 1936 then we should look at the storms that occurred in 1936, including their path in the Atlantic. We have similar heat waves and tornadoes along with a La Nina. Look at South America, The Argentine glacier is about to have a breakout same as the LAST TIME in 1936. All these events must be related to weather patterns and 1936 is a good fit. All we need is a dust bowl but agriculture is way ahead of that game.

  64. Posted Jul 12, 2008 at 12:08 AM | Permalink

    #63, Jim, I think rice patties would fare well this summer in the Midwest just like in 1993. Analogue techniques have shown some promise for large-scale expected conditions but show limited if any skill. Each year is unique.

    The Dvorak technique is basically a look-up table which typically relies on geostationary satellite imagery at a minimum. We have plenty of recon flights over the past several decades with concurrent satellite coverage to fairly accurately analyze intensity.

    Bertha is setting “records” which are based upon its longevity as a hurricane so early in the season. It is currently surrounded by dry air over rather average ocean-heat content waters. Hurricane Jeanne in 2004 looked like a blown out tire as it performed a loop near the Bahamas. It eventually completed an overhaul of its eyewall structure and sailed westward to Florida.

  65. Kenneth Fritsch
    Posted Jul 12, 2008 at 10:27 AM | Permalink

    Because of my participation on some of these TC threads, I have been paying more attention to individual TCs such as Bertha. Bertha, so far, has been spending a lot of time out at sea and thus is building ACE index points that are rather benign.

    When a storm such as Bertha avoids land it obviously can live longer and this brings me to my point: When one might want to do a straight forward correlation of ACE index with a potential for personal and property damage, a storm like Bertha, so far, would seem to complicate that relationship.

    Obviously since Bertha will probably be a record breaker for duration and thus is not a typical TC, we would need to look at other storms to analyze a general phenomenon that might be called benign ACE. Has this been done previously?

    Another general question I have about Bertha is will her long live make an understanding of what can sustain a hurricane any more complete?

    Another thought on long lived hurricanes is that the longer they live the more likely they are to eventually hit land. Is that true?

  66. David Smith
    Posted Jul 13, 2008 at 7:52 AM | Permalink

    Re #65 Kenneth, here are my two cents –

    Storms like Bertha indeed chalk up many ACE points and in a sense distort a season. Perhaps Bertha is a reminder that there is no single measure of a hurricane season which tells the full, accurate story.

    I’ve never seen an ACE study which removed these high-ACE fish-killer storms. My sense is that they are randomly distributed with no trend, so any study would probably reveal little new info. But, maybe that’s wrong and you’ve raised my curiosity so I’ll look at the records later today.

    Bertha formed in the far eastern Atlantic above a region of anomalously warm water near Africa. Once the seedling became a cyclone it came under the influence of mid-level winds which typically veer systems a bit north of west and out into the open Atlantic. That’s what happened. Had the seedling delayed its development it would have chugged due westerly, closer to inhabited areas. Thus, a warmer eastern Atlantic might actually reduce the number of systems reaching the populated western regions. An unadvertised potential benefit of AGW, I suppose.

    My sense is that the conditions which lead to long-lived storms are fairly well understood and that Bertha won’t add to that knowledge.

    On duration and landfall, most of the longer-lived storms in the record are those which formed far at sea and then got stuck in the weak steering winds of the mid-Atlantic (25N-35N). They eventually get unstuck with most moving northward into the cooler mid-latitude waters. There they get absorbed by mid-latitude weather. My bet is that there is little relationship between duration and landfall, after about 5 days duration.

  67. Bob Koss
    Posted Jul 13, 2008 at 6:00 PM | Permalink

    Here is the storm and track distribution since 1900 for landfalling and non-landfalling storms. I defined landfall as having come within 30nm. of any land.
    Storms no landfall
    Storm Peak Wind:
    Mode Median Mean S. Dev.
    45 65 70 25
    SS TS Cat1 Cat2 Cat3 Cat4 Cat5 Total
    14 114 52 42 23 16 2 263
    Track Wind:
    Mode Median Mean S.Dev.
    35 45 53 23
    ST TS Cat1 Cat2 Cat3 Cat4 Cat5 Total
    1491 3702 1436 640 227 114 4 7614
    Storms that made landfall
    Storm Peak Wind:
    Mode Median Mean S. Dev.
    45 70 76 30
    SS TS Cat1 Cat2 Cat3 Cat4 Cat5 Total
    10 306 158 85 93 71 29 752
    Track Wind:
    Mode Median Mean S.Dev.
    35 45 54 27
    ST TS Cat1 Cat2 Cat3 Cat4 Cat5 Total
    5093 10605 3608 1608 1078 708 117 22817

  68. Jim Arndt
    Posted Jul 13, 2008 at 6:19 PM | Permalink

    Ryan #64,

    Thank you, I see what you mean that the year maybe a good analog but don’t use it to predict that year but see the patterns and see the fit. I think rice will do very well in the next 10 or so years in the midwest. California is going to suck wind though.

  69. jeez
    Posted Jul 13, 2008 at 6:57 PM | Permalink

    ryan, it’s paddies.

  70. Kenneth Fritsch
    Posted Jul 14, 2008 at 4:04 PM | Permalink

    Bob Koss, in post #67 you have values for Track Wind. What do those values represent?

  71. Bob Koss
    Posted Jul 14, 2008 at 8:10 PM | Permalink

    Kenneth Fritsch,

    Those are tallies by category of all tracks associated with the qualifying storms. It does not mean all the tracks hit or missed landfall.

    e. g. the last line of my comment #67 shows 117 Cat5 tracks out of the 22817 total tracks associated with the 752 land falling storms.

    The mode, median, mean, S.Dev are wind speeds in knots for the dataset. I automatically generate those figures though I haven’t found that information to be of any use.

  72. Kenneth Fritsch
    Posted Jul 15, 2008 at 9:25 AM | Permalink

    Bob Koss per Post #71:

    I guess I am not able to understand that your data show multiple tracks for an individual storm. My simple-minded understanding sees only one track for a storm — unless, I guess, it reforms. What am missing here?

  73. Bob Koss
    Posted Jul 15, 2008 at 10:17 AM | Permalink

    Kenneth Fritsch,

    My fault.

    I have a bad habit of not explaining myself well. Track wind is the categorized tally of individual plots or observations for the storms.

  74. David Smith
    Posted Jul 16, 2008 at 7:26 AM | Permalink

    The Pew Center has a press release relating the length of the hurricane season to global warming. Normally I put Pew studies in the same category as Limbaugh, etc but this one does quote Holland and Webster so it may be worth a few minutes of review. Unfortunately I haven’t found a copy of the study so my comments are pretty limited at this time.

    I did take a look at early season (January thru July) storm count since 1900 and divided the count into our Easy Detect (either hit land (including islands) or came close enough to land to be detected from shore) and at-sea storms.

    The Easy Detect group, which covers 82% of the early-season count, has this pattern:

    It’s the classical no-trend plot.

    I also plotted the “at-sea” category which I’ll post when I return home. It, too, is our classical pattern of few storms prior to the satellite era then ramping up strongly in recent decades. We’ll see how well that correlates with SST.

    One of the basic arguments (warmer SST have lead to a longer hurricane season) is plausible so I really don’t have a basic argument with the premise. My interests are in whether the historical data show a weak relationship or a strong relationship (recall that wind shear also plays a major role in the start and end of a season) and whether there’s any indication of a strength or landfall trend in all of this.

  75. Posted Jul 16, 2008 at 11:21 AM | Permalink

    #75, this story appeared out of the ether in an effort to provide a climate change connection to long-lived cutoff-storm Bertha. Webster and Holland were simply interviewed and provided opinions or really just anecdotes as filler for the article. The story references no peer-reviewed study. Simple back-of-the-envelope calculations show no glaring or statistically significant trends in storm season length. I did a dumb calculation on storms from 1960-2008 for the average date when a particular letter appears to illustrate the difficulty in making such claims, which are plainly loose with the facts and climate science fraud.

    A, B, C, D, E, F, G, H, I, J, K, L, M
    July 9,Aug 2,Aug 16,Aug 27,Sep 4,Sep 14,Sep 17,Sep 19,Sep 23,Oct 2,Oct 13,Oct 12,Oct 13

    Obviously as the letters go on, the sample size decreases as 2/3 of the past 50 years have not seen an L or an M storm. Sadly, on average, we see the L named storm a day earlier than the K named storm.

  76. Posted Jul 16, 2008 at 11:54 AM | Permalink

    75 (Ryan):

    Sadly, that passes for correct math these days.

  77. Posted Jul 16, 2008 at 3:19 PM | Permalink

    You may have heard that Bertha is now the longest serving member of the hurricane database in July. This storm developed near the Cape Verde Islands and tracked quickly at first to the WNW and then blew a tire in the Bermuda triangle.

    Historically Bertha is a below average Cape Verde hurricane by the following metric: ACE per advisory.

    Since 1950, I found roughly 60 hurricanes (>64 kts) that formed south of 25N and east of 35W in the SE Atlantic.

    The average number of advisories written (at TS strength+) is 42 (stdv 13) and the average ACE is 27 (stdev 16). The average ACE per advisory is 0.63 (stdv 0.27).

    Ivan had 59 advisories and an ACE of 70 which equates to 1.2 ACE/advisory.

    Bertha will have an ACE of about 25 spread over 60 advisories, which is about 0.42 ACE/advisory, which is in the bottom fifth of the sample.

    So, if we are going to speculate on future storm activity based upon one data point named Bertha, then I conclude that these Cape Verde storms that occur earlier will be much weaker.

    Also, since 1950, there is no trend in ACE per storm advisory.

  78. David Smith
    Posted Jul 16, 2008 at 8:43 PM | Permalink

    The early-season, entirely-at-sea time series is below:

    The entirely-at-sea storms tend to be in the eastern Atlantic. Few appear in the record prior to the start of satellite and recon coverage. The frequency picked up circa 1995, as did total-season frequency.

    Perhaps the eastern Atlantic SST was cool 1900-1960 then warmed starting in the 60s, creating more storms. Well, the monthly SST anomaly for the MDR is here

    There was a peak period circa 1940 which generally declined towards the 70s and then slowly rose towards 2000, at which time the anomaly reached 1940s levels. It then peaked in the mid-2000s.

    R for the SST and storm count correlation ranges from 0.15 to 0.30, depending on the periods used.

    Not much of a relationship.

  79. Posted Jul 23, 2008 at 5:24 AM | Permalink

    Here is the latest hurricane path predictions from the webcomic

  80. Kenneth Fritsch
    Posted Jul 23, 2008 at 12:20 PM | Permalink

    Re: #78

    David, I am not certain how many of these observations such as the one you made in this post are independent, but they have added to significant numbers and point to the changing detection capabilities over time that could readily confound a correlation between SST and TC counts and activity in the NATL. I think, in most or all cases, where we have looked for alternative explanations from changes in detection, and correct me here if I have misspoken, we have not found any that appear reasonable.

  81. Posted Jul 29, 2008 at 11:29 PM | Permalink

    The tropics have been relatively active during July 2008 with a NH ACE value of 80, with 31 from the EPAC, 12 from the WPAC, and 37 from the NATL. This is in comparison to 2007 which managed half of that. Most of the meat of the season is still ahead since the 25% milepost is typically reached on August 4, with the halfway point on September 7.

    I plotted up a simple bar chart to show the monthly ACE for the NH since April 2004. NH ACE Monthly Bar Chart .

    2008 is so far consistent with a waning La Nina with below to near-average Pacific activity and a slightly above average Atlantic.

    With global warming supposedly accelerating and affecting all facets of society, politics, the environment, and people’s psychiatric health, tropical cyclone activity must be quite the lagging indicator.

  82. Posted Jul 30, 2008 at 5:02 AM | Permalink

    78 (David Smith):

    Interesting graph on the “Atlantic MDR SST Anomoly”. I’d swear that it corresponds very well (eyeballing) to the PDO switching which occurred roughly from 1910-1940 (+PDO), 1940-1975 (-PDO), and 1975-2005 (+PDO). If my eyes are not deceiving me, then I do not enderstand the correlation between the PDO and the Atlantic SST anomoly, which, while it is also mutlidecadel in length, follows its own schedule offstep from the PDO. If you know why, please explain??

    Thanks, Tom

  83. David Smith
    Posted Jul 30, 2008 at 5:03 AM | Permalink

    The first two months of the official Atlantic season have been busy, with four tropical cyclones. Does that mean we’re headed for a busy year?

    Well, here’s a plot of the seasons 1948-2007, with the number of storms in June-July plotted against the number of storms in the rest of the season:

    My read is that one cannot use early-season activity (as measured by storm count) as an indicator of the rest of the season. There may be other measures (maybe ACE, SST, sea-level pressure or something else) which would work, but not storm count.

  84. David Smith
    Posted Jul 30, 2008 at 5:06 AM | Permalink

    Re #82 Tucker, I haven’t looked at that but will do so. Thanks for the observation.

  85. RomanM
    Posted Jul 30, 2008 at 5:36 AM | Permalink

    #83 David

    Could I suggest that switching the X and Y axes would make the graph a bit more informative? We are used to seeing the predictor on the horizontal and the response on the vertical axis. You could just replace the original graph file with a new one.

  86. David Smith
    Posted Jul 30, 2008 at 7:58 AM | Permalink

    Re #85 Certainly, Roman, if that helps. It’ll be later today.

  87. Kenneth Fritsch
    Posted Aug 1, 2008 at 9:39 AM | Permalink

    Re: #85

    Could I suggest that switching the X and Y axes would make the graph a bit more informative? We are used to seeing the predictor on the horizontal and the response on the vertical axis. You could just replace the original graph file with a new one.

    RomanM, your comments remind me that we might assume that we would use a previous period to predict a later period, but that is not always the case in climate science. Please note what Sabbatelli and Mann did in the article linked below: they related seasonal TC counts to an ENSO index value for the next year. Below are the details along with comments that Steve M made on this relationship.

    The influence of climate state variables on Atlantic Tropical Cyclone occurrence rates by Thomas A. Sabbatelli and Michael E. Mann from the link:

    Various alternative indices of the El Nino/Southern Oscillation (ENSO) are available. We employed the boreal 135 winter (DJF) Nin˜o3.4 index (SST averaged over the region 5_S-5_N, 1200-1700W) favored by many investigators [e.g., Trenberth, 1997]. Use of alternative (e.g., Nin˜o3) ENSO 138 indices yielded similar conclusions. The Nin˜o3.4 index was taken from the Kaplan et al. [1998] data set and updated with subsequent values available through NCEP. The boreal winter (DJFM) NAO index was taken from Jones et al. [1997], updated with more recent values from the University of East Anglia/CRU. For simplicity, the ‘year’ was 143 defined to apply to the preceding storm season for both 144 indices (e.g., the 1997/1998 El Nino and winter 1997/1998 145 NAO value were assigned the year 1997).

    Then from Steve M’s linked thread he comments as follows:

    You have to watch a bit carefully here, because it turns out that Mann regresses the storm count against the following Nino and NAO indices not the predecessor. Mann justifies this with the tricky phrase:

    However, we do find a statistically significant lagged correlation relating the Nino3.4 index to the MDR SST series for the following year’s storm season, consistent with the observation elsewhere [Trenberth and Shea, 2006] that ENSO events influence tropical Atlantic SST in the following summer.

    If you’re not watching carefully, you’d assume that preceding winter’s climate data would be used to predict tropical cyclone counts rather than the following winter’s data. But you’d be wrong, as I’ll show shortly. Now for the purposes of estimating past tropical cyclone levels, as Mann and Sabattli try to do, this is not necessarily the end of the world. If there’s a relationship between 2006 cyclone counts and 2006-2007 Nino 3.4 and 2006-2007 NAO, one can utilize this for past estimates. But the causality relationship is certainly not what one expects and this surely warrants a little discussion.

  88. RomanM
    Posted Aug 1, 2008 at 12:13 PM | Permalink

    #87 Kenneth

    Despite the issues that you raise, David’s graph in #83 was interpreted by him as:

    My read is that one cannot use early-season activity (as measured by storm count) as an indicator of the rest of the season.

    If that is indeed the point he wishes to make, then we have a clear choice for a predictor and a response. Because the axes are reversed from the way that we are used to viewing such a predictive relationship, the graph can be easily misinterpreted – a violation of good statistical practice. It also suggests that the line superimposed on the graph might also have been created using the number of August-September storms as the independent variable in a regression which is not the appropriate approach to support his interpretation of the data.

  89. bender
    Posted Aug 1, 2008 at 1:15 PM | Permalink

    An interesting set of issues is being raised here in terms of regression vs. inverse regression.

    Regression produces biased estimates of slopes (overly flat). You may reverse David’s axes, get a different slope and significance level, and propose that one as “correct” [as RomanM is suggesting]. However the fact is that events in the earlier time-period do not cause events in the later time period. They may be correlated, but through a common cause spanning both periods. In which case it is fair to debate which variable should act as the regressor.

    In such cases an unbiased regression method may be preferable.

    [In this case, however, I expect the answer is the same: no relation.]

    Steve: bender, stay tuned for the Brown 1987 discussions which will put a lot of the inverse regression issues into context.

  90. Posted Aug 1, 2008 at 1:50 PM | Permalink

    The most recent GFS forecast has no tropical cyclone activity over the globe during the next 7 days. This would be a most unusual occurrence, at least unprecedented since 1974. In fact, there has been an Atlantic storm every year since 1999 during July 29 – Aug 7, and at least one storm somewhere on the planet since 1974. However, it looks like we are in the middle of another TC inactivity period continuing last year’s rotten numbers.

    #87, I just submitted a paper to the GRL that relates the SST’s in the North Pacific during the Spring with Northern Hemisphere tropical cyclone activity. During the past 30 years, the relationship between April-June SST in the Gulf of Alaska and Northern Hemispheric ACE is about R = 0.95. The SSTs there are forced by winter-time stochastic atmospheric forcing, basically the movement of the Aleutian Low largely measured by the North Pacific Oscillation (NPO). The following spring’s SSTs then force anomalous trade winds that can impact the tropical SSTs again, finishing the loop begun in the previous fall. I concluded that TCs are tracers and are signatures of the prevailing climate pattern. Should not be a shocking result to anyone, except those that push the global warming explanation.

  91. RomanM
    Posted Aug 1, 2008 at 2:04 PM | Permalink

    #89 bender

    I still stand behind my comment that the graph here has a directed focus (trying to “predict” the number of later storms given the number of earlier ones) that determines which variables go more appropriately on which axis. I would much rather see the pattern of points in the graph graph look like:

    The line would be quite different. There is also a hidden element in the graph (only 39 visible points of 60 altogether) so the picture is somewhat incomplete although R-squared would likely not change much. In fact, I find the low R-squared surprising, since one would expect that conditions which influence the formation (or lack thereof) of hurricanes in the earlier part of the season might be expected to still be similarly influential in the latter part of the season as well.

  92. bender
    Posted Aug 1, 2008 at 2:17 PM | Permalink

    The line would be quite different.

    Careful what you say. First, yes, David’s line would be “different” were it reflected in this way. But my point is that it would no longer be the appropriate regression line, as you are now regressing Y on X, rather than X on Y. And I bet the new line is no different from David’s: slope ~zero, p value non-significant.

    Second, I also told you why the fact the new line would be “different” would be trivial. The correct test of the hypothesis is not biased regression because the issue where is not how does X cause Y, but how are X and Y related to some other common variable, Z, external to the analysis. I repeat: X may come before Y, but the arrow of time is not the arrow of causality.

    The hidden points can be revealed by adding some random jitter in the plotting. Their number will weight the regression and should be shown in the data plot.

  93. Kenneth Fritsch
    Posted Aug 1, 2008 at 2:45 PM | Permalink

    RomanM @ #88:

    Despite the issues that you raise, David’s graph in #83 was interpreted by him as:

    Roman, I was not defending David’s graph, I was simply using it as an excuse to bring up the Sabbatelli and Mann relationship again. I am certain that David is as appreciative as I am of your statistical advice and prodding (and that he will fix his graph per your recommendations).

    I sometimes plot the data in reverse order in using Excel depending on the order of the data that I highlight initially, but have gotten pretty good at reversing the axes to make the y axis the dependent variable even if only to satisfy myself that I am doing it right..

    That Sabbatelli and Mann relationship has a further repercussion in that if one looks at the relationship of TC count versus the Nino index using the winter months directly preceding the TC season it breaks down badly. I think the point of Steve M (and Bender) is that one can find correlations in this manner that will be suspected of being spurious unless one can show some other linkage with another variable. In this case, as Steve M noted, Sabbatelli and Mann point to a lagged correlation of SST to the Nino index, but it would appear to be in the wrong direction.

  94. bender
    Posted Aug 1, 2008 at 2:50 PM | Permalink

    If I may speculate … I think RomanM may be overweighting that one data point (JJ=7,AD=20) in his mind’s eye. Regression will take care of that little outlier by weighting all points equally – something inexperienced humans are typically bad at. The positve slope I presume he’s hoping for will not emerge as significant.

  95. RomanM
    Posted Aug 1, 2008 at 3:06 PM | Permalink

    #92,93, Bender and Kenneth, it was not my intention to go to the mat over this. My suggestion was just that, a suggestion, that the data would be easier for most folks to understand if it was presented in the way we are used to seeing it: predictor on the horizontal and predictee on the vertical. There was no suggestion that the mathematical result would somehow necessarily be different if looked at the other way, nor was the issue one of which was better here, regression or inverse regression (the latter which I have been looking at in the context of calibration for a couple of months), for actually trying to make a prediction.

    David does good stuff and he doesn’t need help. This just distracts from the points he has raised. After 40 years of teaching this stuff, one month of retirement isn’t long enough for me to change my habit of offering unsolicited advice on matters such as this. 😉

  96. bender
    Posted Aug 1, 2008 at 3:10 PM | Permalink

    Your advice is good advice. I’m going further, telling you what I think the result of following your advice will be.

  97. Kenneth Fritsch
    Posted Aug 1, 2008 at 4:42 PM | Permalink

    David does good stuff and he doesn’t need help. This just distracts from the points he has raised. After 40 years of teaching this stuff, one month of retirement isn’t long enough for me to change my habit of offering unsolicited advice on matters such as this.

    Roman, I can, and should, only speak for myself, but your help, unsolicited or solicited, is much appreciated — as some posters must be aware per your analysis of the USHCN CRN ratings using (the correct) multiple regressions in another thread.

    In David’s case, you are right to point out that the dependent variable should be the on the y axis. R^2 will be the same either way, but the “Mannian” explanation needs to be different.

    I am too old to worry about being wrong as I am more in hurry to get things right.

  98. David Smith
    Posted Aug 4, 2008 at 8:34 PM | Permalink

    If Edourard behaves as forecast then the storm center will pass over my house about 18:00Z August 5, eighteen hours from now. I’ll offer eyewitness (eye witless?) news reports for CA readers tomorrow about conditions near the eye.

  99. David Smith
    Posted Aug 5, 2008 at 8:33 AM | Permalink

    Edouard did not behave as expected and reached land east of my house. I’m currently about 20 miles from the center and winds are maybe 20 MPH, which is not even a good thunderstorm. The coastline east of here had winds of 50MPH for a few hours. I think the official storm windspeed is 65 MPH, near-hurricane, but it’s hard to find much evidence of it outside of a small area.

    The point is, weak systems like this could easily have been missed or ignored in the times before the modern weather service and especially in sparsely-settled areas.

  100. Jonathan Schafer
    Posted Aug 5, 2008 at 11:37 AM | Permalink


    Hey, at least you will get some rain out of it. Here, just NW of D/FW, we won’t likely get anything. And we need something to break up this teapot we’ve been sitting in for about 3 weeks now. It hasn’t been this hot since 1998. Thank God we had a bit of rain this spring or we’d be looking at a whole lot more days over 100.

  101. bernie
    Posted Aug 5, 2008 at 11:42 AM | Permalink


    Thank God we had a bit of rain this spring or we’d be looking at a whole lot more days over 100.

    Interesting climate model!!
    I hope the heat wave ends soon.

  102. Jonathan Schafer
    Posted Aug 5, 2008 at 1:52 PM | Permalink


    I think 20 out of last 21 days have been at or above 100, broken only by one day at 99. Frankly, up until around July 4th, it had actually been a good year. But we’ve been under this persistant high pressure that’s been very stubborn this year. Normally, it will retrograde, then move back, then move forward, etc., so it’s not always parked directly on top of us. But this year, since early/mid July, it’s just mostly been parked right on top of us. The few times it has retrograded, we’ve been able to sneak a bit of rain, but not much. Actually, we got a pretty good rain from the shear axis left over from TS Dolly after she moved out of NM and across TX. But Sun/Mon, highs were 105 – 107, and that’s pretty doggone hot. Not as bad as the 112 I sat through in 98, but still hot.

    Last year as bad as this was 2000, which was also pretty bad. That was a La Nina year like this one. But, as I said, the spring rains put enough moisture in the ground to keep the temps down for awhile. Once that big H parked over us though, and baked the moisture out the ground, temps have been increasing.

    Looks like we’ll get some reprieve from Edouard, even though we won’t likely see any rain.

    Hope David is doing well down in S TX and staying safe.

  103. M. Jeff
    Posted Aug 5, 2008 at 1:53 PM | Permalink

    Jonathan Schafer, August 5th, 2008, at 11:37 am says:

    And we need something to break up this teapot we’ve been sitting in for about 3 weeks now.

    Yep, so far this year DFW has had 25 days of 100F or above, compared to 5 last year. Normal is 16.

  104. Jonathan Schafer
    Posted Aug 5, 2008 at 2:03 PM | Permalink

    Ran across this today…

    Tropical Storm Risk August 2008

    Basically says we are in for a very active season.

    only 3% likelihood it will be near normal (defined as an ACE Index Value in the middle tercile historically (70-115)) and NO CHANCE it will be blow normal (defined as an ACE Index Value in the lower tercile historically (

    * Bold my emphasis

    Well, we’ll see. Most of the ACE so far has come from a single storm that wandered the Atlantic for seemingly forever, but never struck land.

    I also liked the NOAA Atlantic 2008 ACE Forecase, with values ranging from 88 – 184. Can you say hedging your bets?

  105. Jonathan Schafer
    Posted Aug 5, 2008 at 2:04 PM | Permalink

    Oops, forgot about the stupid less than sign. Should be

  106. Jonathan Schafer
    Posted Aug 5, 2008 at 2:04 PM | Permalink

    Third try

    < 70

  107. M. Jeff
    Posted Aug 5, 2008 at 2:13 PM | Permalink

    Jonathan Schafer is one useful source for DFW climate info

  108. David Smith
    Posted Aug 5, 2008 at 2:52 PM | Permalink

    Re #100 Jonathan my backyard received 3.7 inches of rain from Edouard. Things are well-watered now.

  109. David Smith
    Posted Aug 5, 2008 at 3:10 PM | Permalink

    Hurricane researchers Tuesday increased the number of storms they predicted for the 2008 Atlantic hurricane season, calling for a “very active” season with an above average likelihood of a major hurricane making landfall in the U.S.
    Forecasters at both Colorado State University in Fort Collins and at Tropical Storm Risk in Britain both issued updated forecasts which raised the number of storms from predictions earlier in the year. The National Oceanic and Atmospheric Administration (NOAA) was scheduled to release its latest forecast on Thursday morning.
    The predictions came on the same day that Tropical Storm Edouard struck the Texas Gulf Coast with strong winds and heavy rain. It was the fifth named storm of the Atlantic hurricane season, which began June 1 and ends Nov. 30. Two of those storms, Bertha and Dolly, became hurricanes.
    Researchers William Gray and Philip Klotzbach of Colorado State University upped their predictions, calling for 17 tropical storms, nine of which would become hurricanes. Of those hurricanes, they said five would be “major” or “intense” Category 3 or above with winds in excess of 111 mph. In April and June, they had predicted 15 named storms, eight of which would develop into hurricanes, four of them intense.
    Tropical Storm Risk called for 18 named storms, with 10 of them becoming hurricanes and nearly half of those being intense storms. Its earlier reports, in April, June and July, closely mirrored the predictions of the CSU researchers.
    An average season calls for 11 named storms, six of which become hurricanes.
    “Based on current and projected climate signals, Atlantic basin and U.S. landfalling tropical cyclone activity are forecast to be about 90 percent above the 1950-2007 norm in 2008,” TSR said. “There is a high — about 95 percent — certainty that activity will be in the top one-third historically.”
    TSR said it based its predictions on July-September trade wind speed over the Caribbean and tropical North Atlantic and the forecast August-September sea surface temperature in the tropical North Atlantic.
    “TSR anticipates the trade wind predictor having a strong enhancing effect on activity and the sea surface temperature having a small enhancing effect,” it said.
    Gray and Klotzbach — whose predictions in 2006 and 2007 were wrong — said they raised their predictions “due to a combination of a very active early tropical cyclone season in the deep tropics and more favorable hurricane-enhancing sea surface temperature and sea level pressure patterns in the tropical Atlantic.
    “The primary concern with our current very active seasonal forecast numbers is the continued ocean surface warming in the eastern and central tropical Pacific,” they said. “Although it seems unlikely at this point, there is a possibility that an El Niño could develop this fall.”
    But Dennis Feltgen, a spokesman for the National Hurricane Center, said there was nothing unusual about seeing so many storms so early in the season.
    “The hurricane season goes from June through November,” he noted. “It’s not unusual at all to have one or two storms in June and a couple of storms in July. If we’re in an above average season, it’s not unusual at all.
    “It’s a misconception to think that we’re going to have an above average season but we’re still not going to have them [storms] until the middle of August,” he said. “Once hurricane season begins, we can have a storm just about any time.”
    Tropical Storm Risk dismissed global warming as a factor for previous hurricanes, including Dennis, Katrina, Rita and Wilma in 2005.
    “The global warming arguments have been given much attention by many media references to recent papers claiming to show such a linkage,” they said. “Despite the global warming of the sea surface that has taken place over the last three decades, the global numbers of hurricanes and their intensity have not shown increases in recent years except for the Atlantic.
    “Although global surface temperatures have increased over the last century and over the last 30 years, there is no reliable data available to indicate increased hurricane frequency or intensity in any of the globe’s other tropical cyclone basins besides the Atlantic,” they said.

    Posted Aug 5, 2008 at 9:16 PM | Permalink

    #108 David … Not my backyard, but polar low, curly swanlike, not
    TH-like, “Britta” , [named by, not to be confused with
    “Sturmtief Britta” N Germany Oct 31 2006…The namegiving of severe
    storms is a fairly recent phenomen over here while we await a tropical or at least a subtropical climate
    over most of Europe …or???] There is some interesting reading on
    the net on polar lows and my old “living room hurricane” “Vince”
    of E Atlantic 2005. Also PL connection to flooding in N Algeria November 2001, Algiers
    Port 110 mm … ANYHOW our little “Britney” dumped 146 mm that would
    be 5.75 INCHES, I guess in Kollebäckstorp, also a private station N slope of Hallandsås,
    [the horst? that they never? come through to build a new railway
    tunnel…] At least 2 people killed, one russian captain in S Baltic,
    fell overboard…Here in the Stockholm area we got some 2 inches, very
    much needed after the summer “drought”…
    [Meanwhile, severe frost in N Lappland,
    frost for 5th consecutive night in August.] This storm had winds
    up to 33m/s , cat 1, in a couple of places, both W coast and S Baltic!!

    Posted Aug 5, 2008 at 10:29 PM | Permalink

    #110 Addendum … GUST winds were up to 33 m/s, GWS?
    SMHI has a neat animation of them spanning 18h! [“Sammanfattning
    av ovädret”] AND please check “European Polar Low Working Group”
    Click “NH gallery” 1987 Feb 27 0410 UTC, top left, can a system
    be more symmetrical…?? That’s why PL’s sometimes are called
    “arctic hurricanes”… PS “Britta’s” lowest pressure was close
    to 982 mb… “Edouard” 996 mb I think??

  112. Kenneth Fritsch
    Posted Aug 8, 2008 at 11:10 AM | Permalink

    I was surfing the TV the other night and come upon a cable show discussing hurricanes with the usual consensus view on historical AGW trends and correlations with more intense TCs, at least, in the NATL. After this blurb, they interviewed Emanuel Kerry during which he had what I thought were some interesting and perhaps new things to say about TCs and climate. I had heard parts of Emanuel’s thoughts on these matters before but had not heard it all in one place.

    Briefly, he commented that:

    Hurricane modeling initially did not give an upper limit for the hurricane wind velocities that are seen in nature until they put in parameter(s) that dealt with the negative feedback due to the friction that is created when the hurricane pulls up spray from the oceans surface. He noted that this phenomenon is too complex for the models to resolve directly and that even determining the location of the ocean surface under this condition becomes difficult because these droplets smear out a transition layer.

    He then noted that while the hurricanes can pull some small amount of heat out the oceans into the air, the overwhelmingly large effect on cooling the ocean surface is caused by the mixing of the oceans warmer surface with the cooler depths below.

    Emanuel went on to make what I thought were far reaching statements about this process potentially having a profound effect on transporting warm tropic waters to the poles for cooling. He went so far to say, or at least as I interpreted it, that while the climate can affect TCs, that TCs can in turn have a significant effect on the climate. He said that to understand how the climate works we have to understand the TC/climate interaction. He seemed to me to be indirectly saying also that TCs could provide a negative feedback on increasing SSTs. He even noted that when some periods in the earth’s past when the poles where much warmer than present and the tropics not much warmer than present that TCs could have been part of mechanism for this phenomenon of cooling the tropics and warming the poles.

    Is anyone here sufficiently familiar with these thoughts (or theories?) of Emanuel’s to let me know if I have, at least, in a general sense, interpreted them correctly or what it was that he was saying?

  113. Kenneth Fritsch
    Posted Aug 8, 2008 at 11:18 AM | Permalink


    That of couse should be Kerry Emanuel and not Emanuel Kerry.

  114. David Smith
    Posted Aug 8, 2008 at 1:28 PM | Permalink

    Re #112 Kenneth, I think you’ve captured the gist of his thoughts, at least the portion I’ve read from time to time. His views are probably towards the extreme while most others think that tropical cyclones, while powerful, are small players in global heat redistribution. Personally, I’m with the majority on this one.

    There’s one sea spray story that I find interesting. An aircraft investigating a hurricane encountered salt water droplets many thousands of feet above the ocean surface. The plane was flying through a no-rain part of the hurricane so that the salt water droplets, when they touched the plane surfaces, were not washed away by rain. The result was a crust of salt which entered the engines and caused three of the four to shut down. The plane was headed for the ocean when it flew into a rain squall which washed off the salt and allowed the stalled engines to be restarted.

  115. Posted Aug 8, 2008 at 2:00 PM | Permalink

    #112, I am of the mindset that tropical cyclones are at the mercy of the prevailing atmosphere-ocean conditions and form opportunistically when the environmental conditions permit. These favorable conditions are modulated by the large scale climate variability modes such as ENSO, PDO, NAO, etc. which are both spatially expansive but also temporally important. Just as tree rings can be “climate integrators” of previous ENSO episodes, yearly TC activity also has a “climate fingerprint”.

    Emanuel’s comments also suggest to me at least one null hypothesis: When there are fewer TC’s, some other atmosphere-ocean dynamical feature is “picking up the slack” or perhaps the TC’s are “unnecessary” for the climate “balance” to be maintained.

  116. Kenneth Fritsch
    Posted Aug 8, 2008 at 3:44 PM | Permalink

    Re: #115

    With my layperson take on the TC and climate data, I would tend to agree with Ryan M’s view, but at the same time admit to being interested in at least attempting to understand some of the background for these theories that are being bandied about. I would assume that Emanuel must have had some back of the envelop calculations on the amount of heat that TCs could potentially move around the globe.

    In this interview, which was not the first that I have seen with Emanuel, he comes across as a rather unassuming type of guy without any signs of arrogance. He also commented that before teaching a course on the physics of TCs he thought that he knew what essentially drove TCs but during his teaching he said he realized that what he was offering did not make sense to him anymore. I am not sure from the interview that he thinks he has it right yet.

    Of course, climate moderated by TCs might be considered more in line with explanations of climate being moderated by PDO and other like processes/relationships than the more narrow consensus view of AGW.

  117. bender
    Posted Aug 8, 2008 at 4:13 PM | Permalink

    he realized that what he was offering did not make sense to him anymore


    Climate is complicated. In my experience, few who say they “understand” complex emergent phenomena in 3D dynamical systems actually do. The mathematics are not simple – they just aren’t – and mathematics is the only way to demonstrate and communicate true “understanding”. If Emanuel’s “understanding” was based on heuristics, not mathematics, this could explain why his “understanding” was subject to change. It was not true understanding; it was superficial heuristics.

    I would pay to see Kerry Emanuel take lectures from Gerald Browning. I would pay.

  118. Scott-in-WA
    Posted Aug 8, 2008 at 8:22 PM | Permalink

    bender: I would pay to see Kerry Emanuel take lectures from Gerald Browning. I would pay.

    It could be that I’ve simply missed it between here and the CA Message Board, but is there a discussion topic (or a post) somewhere on CA concerning CCSP’s July 2008 report concerning the strenghs and limitations of climate models.

    It would be most interesting to hear what Gerald Brown has to say about this report. It is probably around here on CA somewhere, and I just haven’t found it.

    Climate Models: An Assessment of Strengths and Limitations, Final Report, Synthesis and Assessment Product 3.1, July 2008

  119. Scott-in-WA
    Posted Aug 9, 2008 at 7:32 AM | Permalink

    More concerning “Climate Models: An Assessment of Strengths and Limitations.” Here is the press release from DOE’s web site at:

    July 31, 2008

    Climate Change Science Program Issues Report on Climate Models

    WASHINGTON, DC – The U.S. Climate Change Science Program (CCSP) today announced the release of the report “Climate Models: An Assessment of Strengths and Limitations,” the 10th in a series of 21 Synthesis and Assessment Products (SAPs) managed by U.S. federal agencies. Developed under the leadership of the U.S. Department of Energy (DOE), this report, SAP 3.1, describes computer models of the Earth’s climate and their ability to simulate current climate change.

    “Complex climate models are tools that provide insights and knowledge into how future climate may evolve. To assure that future climate projections are used appropriately, it is crucial to understand what current models can simulate well, and where models need improvements,” said David Bader, with DOE’s Lawrence Livermore National Laboratory and the coordinating lead author for this SAP report. “This report makes an important contribution in helping to describe and explain the current state of high-end climate modeling for the non-specialist.”

    The SAP 3.1 report describes complex mathematical models used to simulate the Earth’s climate on some of the most powerful supercomputers, and assesses their ability to reproduce observed climate features, and their sensitivity to changes in conditions such as atmospheric concentrations of carbon dioxide. The report notes that “the science of climate modeling has matured through finer spatial resolution, the inclusion of a greater number of physical processes, and through comparison to a rapidly expanding array of observations.” The authors find that the “models have important strengths and limitations.” The report assesses how well models simulate the recent observational period; it does not deal with climate change predictions.

    The report organizes the discussion of these strengths and limitations around a series of questions, including: What are the major components and processes of the climate system that are included in present state-of-the-art climate models? How uncertain are climate model results? How well do climate models simulate natural variability? How well do climate models simulate regional climate variability and change?

    The report documents the improvement in climate model fidelity over the past decade. As emphasized by the Intergovernmental Panel on Climate Change (IPCC), modern models faithfully simulate continental to global scale temperature patterns and trends observed during the 20th century. Despite this progress, a number of systematic biases across the set of climate models remain, particularly in the simulation of regional precipitation. On smaller geographic scales, when compared against the current climate, the simulated climate varies substantially from model to model. The report notes that “an average over the set of models clearly provides climate simulation superior to any individual model,” and concludes that “no current model is superior to others in all respects, but rather different models have differing strengths and weaknesses.”

    The report also describes “downscaling,” which is the use of methodologies to generate higher resolution information from global models results for applications on the regional and local scales. Several downscaling examples such as applications focusing on water resources and surface climate change are illustrated to demonstrate how model results can be applied to a diverse set of problems.

    To develop the SAP 3.1, DOE chartered a Federal Advisory Committee comprised of 29 members drawn from academia, government scientists, non-profit and for-profit organizations that drafted and oversaw the review of the report in accordance with the CCSP guidelines. The lead authors include David Bader (coordinating lead author) and Curt Covey, Lawrence Livermore National Laboratory; William J. Gutowski Jr., Iowa State University; Isaac Held, NOAA Geophysical Fluid Dynamics Laboratory; Kenneth Kunkel, Illinois State Water Survey; Ronald Miller, NASA Goddard Institute for Space Studies; Robin Tokmakian, Naval Postgraduate School; and Minghua Zhang, State University of New York, Stony Brook. SAP 3.1 is the third and final SAP that DOE coordinated for the CCSP.

    The SAP 3.1 report and additional information about the CCSP are available from the Climate Change Science Program.

    Information about DOE’s climate change research is available from the Office of Biological & Environmental Research.

    Media contact(s):
    Jeff Sherwood, (202) 586-5806

  120. Posted Aug 14, 2008 at 12:23 PM | Permalink

    Cat 5 Fay? Not the most encouraging set of forecasts coming from HWRF today:

    Pay attention to 92L as it continues to show promise… Ryan’s Weather Maps from FSU

  121. SteveSadlov
    Posted Aug 14, 2008 at 4:31 PM | Permalink

    Ryan, be that as it may, this season’s another dud thus far in NATL. And now, late Septembereqsue synoptics seem to be in place. Things may actually start to wind down soon.

  122. Jonathan Schafer
    Posted Aug 14, 2008 at 9:55 PM | Permalink


    Dr. Masters would disagree with your assessment I believe. Today’s blog post talks about a change to the jetstream, which would allow the Bermuda H to expand further west. This would keeps storms further S, farther away from the dry air/sahel dust. It would also keep them from recurving in the central Atlantic, and allow them to head into the Carribbean instead. This is forecast to take place in about 7 days, which really, is when hurricane season really ramps up.

    Of course, if it weren’t for Bertha hanging around forever, the existing ACE would be almost non-existent to date. One thing’s for sure though. An active early season doesn’t necessarily correlate to an active middle/late season and an inactive early season doesn’t necessarily correlate to an inactive middle/late season.

    That said, I still hold to my prediction of lower than normal ACE, despite the upticks predicted by Klotzbach/Gray and TSR noted in 104 above.

  123. SteveSadlov
    Posted Aug 15, 2008 at 10:18 AM | Permalink

    The Jet Stream is reaching Alabama.

  124. Jonathan Schafer
    Posted Aug 15, 2008 at 11:10 AM | Permalink


    I refer to this post Major Steering Current Shift Coming

    As I discussed in last week’s blog on steering currents, the hurricane steering pattern for all of July and the first two weeks of August over the North Atlantic has predominantly acted to recurve hurricanes out to sea. The jet stream has been “stuck” in a standing wave pattern, where it dips southward over the East Coast of the U.S., creating a trough of low pressure capable of recurving tropical storms once they get north of the Caribbean Sea (20° latitude). This pattern is in contrast to the steering pattern that set up in 2004 and 2005, when a ridge of high pressure set got stuck over the Eastern U.S. A ridge in this location does not allow hurricanes to recurve, and the U.S. took a terrific battering those years.

    This year’s steering pattern is about to make a major shift towards the steering pattern observed in 2004 and 2005. According to recent 500 millibar (mb) upper-air forecasts from the GFS model. and ECMWF model, the trough of low pressure over the U.S. East Coast will be replaced by a ridge of high pressure 7-10 days from now. As a result, the surface Bermuda High will extend far to the west over the Eastern U.S. This pattern will mean that fewer hurricanes will be recurving beginning a week from now, and the threat to the U.S. Gulf Coast will increase. Conversely, the threat to Bermuda and the Northeast U.S. will diminish.

    There is no way of telling how long this new steering pattern might stay in place. It could last only a few days, or remain in place for several months.

    So, it would appear as though the trough will be replaced by H pressure with more of a zonal flow. At least in the short term.

  125. Posted Aug 15, 2008 at 11:48 AM | Permalink

    This is mid-August, the “tropics” are in the middle of a pattern that comes around this time of year, every year, called summer. Since hurricanes can develop anytime from June – November, I will discount the random musings or ramblings about the jet stream over Alabama. I am also someone who doesn’t want anymore ACE to accumulate this year 🙂 I am in that camp that hopes 92L does not develop any further and is torn apart by the mountains of Hispaniola. On the other hand, a weak Cat 1 hurricane would provide welcome rain to the peninsula of Florida.

    Hurricanes are beneficial in moderation, as with many other liquids.

  126. David Smith
    Posted Aug 15, 2008 at 11:59 AM | Permalink

    Hurricanes are beneficial in moderation, as with many other liquids

    🙂 Ryan, I plan to steal that line.

  127. Posted Aug 20, 2008 at 4:09 PM | Permalink

    If you look back at #120, and the very boisterous HWRF model forecast for Fay, which back on August 14th was still the glimmer known as Invest92L, then you see that the “track” forecast was very good!

    Example of a model arriving at the correct track location but obviously for the wrong reason and via the wrong route. Thankfully, even though Fay is undergoing a complete overhaul and starting to get cooking over the past hour (6 pm Eastern), it has not reached hurricane status (yet). Updated model diagnostics are on my weather site at FSU.

    Looking for some suggestions to start a new hurricane thread for the upcoming week. Little new research has been published in eons and with the exception of Fay and Bertha, the hurricanes and global warming bandwagon clearly needs an infusion of new ideas or observations.

  128. Posted Aug 20, 2008 at 5:40 PM | Permalink

    127 Ryan,

    I as a member of the Gator Nation, would like a more neutral thread. One based more on the natural variability of tropical cyclones and multidecadal trends. One less intense on poking at AGW and more dedicated to stimulating improvement in models. While the models for Fay did a good job to the Keys, they totally lost it after Fay hit land (as in the Keys and Florida).

    Since I live in the Keys I am often forced to make big decisions on whither to stay or go. In the last 8 years I only left for was Wilma. I slept through many in my motor home and a few I spent in a local watering hole. If you can remember back to the Keys forecast for Ivan in 2004 I believe, I stayed put because the models made no sense. For that one I moved to a better RV lot and finally got free cable.

  129. Posted Aug 21, 2008 at 8:30 AM | Permalink

    captdallas #128,

    While the models for Fay did a good job to the Keys, they totally lost it after Fay hit land

    Not really. When Fay was over the Keys, ECMWF had already forecasted her path towards the Atlantic waters and then back to the Florida panhandle or Georgia-Alabama. Go to the ECMWF site to check past forecasts.

    Recently ECMWF has predicted, almost always 1, 2 days in advance with respect to the others models (global or maybe not), the correct route of tropical cyclone in the NA or WP.

  130. Posted Aug 22, 2008 at 10:33 PM | Permalink

    Papers in Press on the Geophysical Research Letters site that will be of interest to the CA community. They are available for download if you have an institution that has subscribed to the journal.

    First up:

    United States and Caribbean tropical cyclone activity related to the solar cycle, by Elsner and Jagger (2008)

    Abstract: The authors report on a finding that annual U.S hurricane counts are significantly related to solar activity. The relationship results from fewer intense tropical cyclones over the Caribbean and Gulf of Mexico when sunspot numbers are high. The finding is in accord with the heat-engine theory of hurricanes that predicts a reduction in the maximum potential intensity with a warming in the layer near the top of the hurricane. An active sun warms the lower stratosphere and upper troposphere through ozone absorption of additional ultraviolet (UV) radiation. Since the dissipation of the hurricane’s energy occurs through ocean mixing and atmospheric transport, tropical cyclones can act to amplify the effect of relatively small changes in the sun’s output thereby appreciably altering the climate. Results have implications for life and property throughout the Caribbean, Mexico, and portions of the United States.


    Smoothing of Climate Time Series Revisited, by Michael E. Mann (2008)

    Abstract: We present an easily implemented method for smoothing climate time series, generalizing upon an approach previously described by Mann [2004]. The method adaptively weights the three lowest order time series boundary constraints to optimize the fit with the raw time series. We apply the method to the instrumental global mean temperature series from 1850-2007 and to various surrogate global mean temperature series from 1850-2100 derived from the CMIP3 multimodel intercomparison project. These applications demonstrate that the adaptive method systematically out-performs certain widely used default smoothing methods, and is more likely to yield accurate assessments of long-term warming trends.

    Last one:

    False causality between Atlantic hurricane activity fluctuations and seasonal lower atmospheric wind anomalies, by Kyle L. Swanson (2008)

    Abstract: Statistical studies suggest a link between anomalies in seasonally averaged lower atmospheric dynamical fields and Atlantic hurricane activity. Here we show that lower atmospheric seasonal wind anomalies result primarily from the presence of the hurricanes themselves. This is done by assuming a hypothetical vortex structure whose radial structure is constrained by observations derived from aircraft probing of tropical cyclones and whose vorticity magnitude is scaled to time varying, best track intensities. Seasonal vorticity anomalies associated with Atlantic hurricane activity are accumulated by summing these idealized vorticies along observed tropical cyclone tracks. Winds associated with these seasonal vorticity anomalies explain the bulk of observed hurricane activity-related fluctuations in the seasonally averaged lower tropospheric wind. Hence, seasonal wind anomalies appear to have little causal information relevant to understanding why hurricane activity in the Atlantic has fluctuated in the past, and may be of limited value in projecting future hurricane activity

  131. Kenneth Fritsch
    Posted Aug 23, 2008 at 1:56 PM | Permalink

    Ryan @ #130

    Ryan, the paper titled “False causality between Atlantic hurricane activity fluctuations and seasonal lower atmospheric wind anomalies”, by Kyle L. Swanson (2008) in your list sounds fascinating to me – if I understand what the abstract is referencing.

    I am aware of recent papers that have shown excellent correlations of vertical wind shear 200-850 hPa and wind direction at a lower atmospheric level (920 hPa??) with TC activity in the NATL, but am not sure whether the paper you listed is referring to the lower atmospheric level only. It would appear that the author is saying that in contrast to previous results that indicated that “vertical wind shear (and direction) is a key environmental variable that controls tropical cyclone development” TCs are controlling wind shear and/or direction.

    This would also seem to be in the same vein as Kerry Emanuel theorizing about TCs having a significant influence on climate.

  132. David Smith
    Posted Aug 23, 2008 at 3:23 PM | Permalink

    Ryan I’m interested in the Elsner paper and will take a look.

    The abstract reminded me of an old plot which I never understood. The plot is linked below (can’t post it with the WP transition underway). I never developed even weak conjecture about the visual appearance of cyclical behavior:

    Different topic – you folks in the Florida Panhandle must be tired of Tropical Storm Fay’s rains. I saw a report of nearly twenty inches of rain near you over the last two days.

  133. Posted Aug 24, 2008 at 6:53 PM | Permalink

    #131, Ken, rather than keep you in the dark, I have a directory that has recent papers (usually in pre-press format, which avoids copyright issues). The papers I described are stuck in there as well as a new Swanson 2008 prize winner, in my opinion in the new journal called G-Cubed.

    Nonlocality of Atlantic tropical cyclone intensities, by Kyle L. Swanson

    Abstract: The assumption that tropical cyclones respond primarily to sea surface temperatures (SSTs) local to their main development regions underlies much of the concern regarding the possible impacts of anthropogenic greenhouse warming on tropical cyclone statistics. Here the observed relationship between changes in sea surface temperature and tropical cyclone intensities in the Atlantic basin is explored. Atlantic tropical cyclone intensity fluctuations and storm numbers are shown to depend not only upon SST anomalies local to the Atlantic main development region, but also in a negative sense upon the tropical mean SST. This behavior is shown in part to be consistent with changes in the tropical cyclone potential intensity that provides an upper bound on storm intensity. However, Atlantic tropical cyclone intensity fluctuations are more nonlocal than the potential intensity itself and specifically vary along with Atlantic main development region SST anomalies relative to the tropical mean SST. This suggests that there is no straightforward link between warmer SSTs in the main development region and more intense tropical cyclones.

    Fay has been a bitch here in Tallahassee. 20+ inches of rain east of the city, with around 10 near downtown and the campus of FSU. Major roads are flooded, trees knocked down, and sewage issues in many locations.

    And, to get excited about a “model” forecast, the GFDL for Invest 94L is predicting Category 6 Gustav, nearing the Gulf of Mexico in 5-days.

  134. Kenneth Fritsch
    Posted Aug 25, 2008 at 11:28 AM | Permalink

    Re: #133

    Thanks, Ryan, the links are much appreciated.


  135. Posted Aug 25, 2008 at 6:11 PM | Permalink

    Category 6 Gustav has shown its ugly head in the HWRF mesoscale hurricane model: worth keeping track of, especially when the 5 day forecast takes a 910 mb storm in the SE Gulf of Mexico. HWRF+GFDL maps on my webpage

  136. David Smith
    Posted Aug 25, 2008 at 8:30 PM | Permalink

    Ryan, does HWRF do a decent job on its intensity forecasts?

  137. Jonathan Schafer
    Posted Aug 25, 2008 at 8:48 PM | Permalink


    Can’t answer for HWRF but SHIPS intensity models have been reasonably accurate from what I’ve seen. No stats on it though. As far as track goes, this will be interesting. NOGAPS has it recurving out to the Alantic, after being picked up by a trough exiting the east coast, but NOGAPS has consistently performed the worst of all the models. The rest of the well-known models basically take it into the GOM or Yucatan peninsula. The question is at what strength. How much will the mountains of Hispanola take out of it. And then, if it hits Cuba as well before turning west, how much will that take out as well. Right now, GFDL is calling for strong CAT1 at Haiti, then drop 25mph at least, then possibly another 10 – 20 if it hits Cuba. It could then be anywhere from weak TS to a strong TS as it enters the GOM. At that point, it will depend on wind shear and exact track.

    I’m sure the Floridians don’t want to see any sign of Gustav.

  138. Posted Aug 25, 2008 at 9:30 PM | Permalink

    Intensity forecasts on average show no skill, but can of course do so in fits and starts. HWRF and GFDL were terrible for Fay’s intensity forecast while it was near Haiti mainly because the track was so far off. The human element is still critical to intensity forecasting and will likely remain so for at least the next decade.

    The mesoscale models require good initialization of the storms to achieve any semblance of accurate track and intensity. Otherwise, they are criticized for being right for the wrong reason. If climate models suffered the same scrutiny, heads would roll.

  139. David Smith
    Posted Aug 26, 2008 at 5:14 AM | Permalink

    Re #138 Thanks, that’s what I thought. I see the intensity forecasts as curiosities, worth noting but not worth taking seriously. My personal practice is to pay attention to the 200mb and 300mb winds to see if they support a major storm (assuming that land proximity, SST, etc support major status).

  140. Jonathan Schafer
    Posted Aug 26, 2008 at 6:20 AM | Permalink

    Gustav is now a CAT 1 Hurricane with 90MPH winds. That didn’t take long to go from invest to TD to TC. The big question right now is what will the track be. Will it hit Hispanola and get knocked down or will it miss it to the south. GFDL and UKMET both show tracks to the south of Hispanola. If that happens and there is no loss of strength, we could be looking at a CAT4 or higher in the GOM in a few days. Then, the big question is where does Gustav go from there.

  141. Michael Jankowski
    Posted Aug 26, 2008 at 7:24 AM | Permalink

    RE #128 and #129, I saw a Tampa weatherman predict Fay would go up the state, hit the Atlantic, and veer across the state to the Gulf. I think this was a several hours after it had hit Key West (5pm news, or maybe it was 10pm). Of course, there were about two seconds where they had a “spaghetti map” of model runs on his map. While I tend to see somewhere from 6-12 models in the typical spaghetti maps, this one seemed to have over 50 – seemingly covering the entire Gulf and southeastern US with possibilities. That was the first prediction I’d see that Fay would act as she did. The projected paths I saw elsewhere seemed to project Fay up into the heart of the US and arguing over which Great Lake she’d end up over in several days.

    If I hear another person remark about “historic 4 landfalls,” I am going to throw up. Counting Key West as a separate landfall from south FL is like claiming my shoelaces are an article of clothing during strip poker.

  142. Posted Aug 26, 2008 at 1:54 PM | Permalink

    Michael #141,
    probably that weatherman is not used to look at the best global model available… 🙂

  143. Posted Aug 26, 2008 at 11:58 PM | Permalink

    Gustav forecast to be a Category 6 hurricane by the mesoscale models, HWRF and GFDL, and slam into New Orleans just in time for the 3 year Katrina anniversary. These are just two models and part of the suite that is available for NHC guidance. However, the official forecast is nearly the same track (not timing, however) as the HWRF and GFDL. Couple links to watch: HWRF GUSTAV 00Z You can get to GFDL by navigating to the bottom of that mouseover page.

  144. Michael Jankowski
    Posted Aug 27, 2008 at 6:49 AM | Permalink

    I tend to use the NHC for updates to tracking and path projections. I noticed the projected path hasn’t changed much from yesterday (other than extending), but none of the projections from 8am thru 5pm yesterday suggested Gustav would be reduced to tropical storm strength at this time. The 8am update today predicts Gustav won’t reach hurricane strength again until sometime Thursday. So maybe this current weakening will help limit the strength Gustav grows to or is a sign that Gustav won’t become quite the monster he appeared he might be.

    But looking at all that open gulf water and a projected path towards the New Orleans area is quite disheartening.

  145. Posted Aug 28, 2008 at 7:26 AM | Permalink

    The Jog SW to Jamaica is a bit unexpected and pleasant.

  146. David Smith
    Posted Aug 28, 2008 at 8:35 AM | Permalink

    Re #145 Captain, one man’s pleasant surprise is another man’s dread! Those of us to the west are going to have several days of anxiety and someone around here will get the short straw. HWRF continues to show cat 4 and 5 strength, ouch.

  147. Posted Aug 28, 2008 at 10:15 AM | Permalink

    david, Without a doubt. I have been through enough to not wish it on anyone. Hopefully, Jamaica will tear it down enough it will have trouble reforming to cat 3 of greater.

  148. Posted Aug 28, 2008 at 10:20 AM | Permalink

    Rats! now Hanna has pick up and there is another are north of PR/VI forming.

  149. David Smith
    Posted Aug 28, 2008 at 12:43 PM | Permalink

    By the way, CaptainDallas, I like your websites. If I’m ever in Florida looking for a fishing trip I’ll give you a call.

  150. Posted Aug 28, 2008 at 5:49 PM | Permalink

    Why thank you David. I need to quit cruising CA and write more. BTW, I am impressed by your graphs. Pondering the tropical weather and remembering back to Tsonis, I was wondering if you have plotted the Decadal Atlantic hurricane cycle versus temperature. There is a whole lots of convection going on right now as there was in 2005. The hurricane cycle has to be related to various oscillations and has to have a significant influence temperatures.

    Timing of the hurricane cycle change led the recent flat/declining temperature trend by a few years. It would be interesting see how global temperature responds to more ‘canes.

  151. Posted Aug 29, 2008 at 2:47 PM | Permalink

    After just being reminded by the weather channel that hurricanes AND weather ridges and troughs are currently AGW, related my mind wandered the Accumulated Cyclone Energy (ACE). I think it would be interesting to have a Global Convection index (GCI). An estimate of energy released monthly/annually due to convection based on satellite observation. I know that this would be a PITA but it could be a good project for a grad student. Stupid idea probably but I would be interested in seeing how it changes with global temperatures and various oscillation.

    Anyone know if such a product exists or is being attempted?

  152. Jonathan Schafer
    Posted Aug 30, 2008 at 10:07 AM | Permalink

    Gustav is up to CAT 3 now. The model tracks are generally converging on a central/western LA landfall. There’s a lot going on over the US though, that will impact the track and speed it ultimately takes. A trough approaching the west coast will lead to stronger ridging across the central US. That could continue to pull Gustav to the west. Several model runs even have Gustav heading back to the GOM after a LA landfall. Most though are pusing it up through central/western LA into E TX. That’s about as far out as they go right now.

    David, I hope you are battening down the hatches. I heard someone mention a Corpus landfall, so you could see some action in Houston if that happens. Gustav is gaining size as well, so a lot more people could be affected.

    I’m less worried about Hanna. If she does hit anywhere, my guess it will be as a strong TC/minimal hurricane. Some models show her recurving to the north by the Bahamas and missing the US altogether. It will depend on timing of an exiting trough and whether it will be enough to keep her out to sea.

  153. Jonathan Schafer
    Posted Aug 30, 2008 at 1:57 PM | Permalink


    Well, living in TX for 11 years doesn’t mean I know my Texas cities all that well. I was confusing Corpus with Beaumont/Pt Arthur area. If it hits Corpus David, you’ll be fine.

  154. Posted Aug 30, 2008 at 2:23 PM | Permalink

    Gus is a cat 4 now and the center is about 195 nm from my house. I am a good 0.3 miles outside of the TS wind advisory. Winds are near 30 knots with gusts to around 33 here. Anyone living on the Gulf Coast of TX, LA or MS may want to visit relatives in Ks for the weekend. BTW, mandatory evacuations are as much of a pain as the storm, two miles per hour for three or more hours requires Valium. Either leave early or take the road less traveled.

    Good Luck

  155. BarryW
    Posted Aug 30, 2008 at 3:13 PM | Permalink

    If it hits west of NO it could be bad, very bad for them. They need to just give up on that location (or maybe hire the Dutch to build a proper dike system). At least they’re taking it seriously, and I don’t think Jindal is going to let Nagin’s incompetance endanger anyone this time.

    I heard a projection on the radio about it possibly reaching Cat 5. There go the gas prices again, just when they were coming down.

  156. Posted Aug 30, 2008 at 3:34 PM | Permalink

    Re #155 (BarryW): Probably on the gas prices, but I don’t think anyone can build a proper dike system for high cat fives. I am thinking a little more westerly track which is bad news for Galveston. They need to declare barrier islands national parks and be done with it or make them be self insured. The keys are at least above sea level rocks with a barrier reef, Hispaniola and Cuba which helps protect things. If we could just teach vehicles to swim better the Keys would not have had much damage from Wilma.

  157. David Smith
    Posted Aug 30, 2008 at 4:09 PM | Permalink

    Cpatain, I think there’s satellite data available to support that but I don’t know of any indices which track it. Indeed, it’d be interesting to see how atmospheric oscillations (PDO, NAO, etc) other than ENSO affect global precipitation.

    Jonathan, I just placed a large dollar bet that Gustav won’t have a destructive impact on Houston.

  158. Jonathan Schafer
    Posted Aug 30, 2008 at 4:56 PM | Permalink


    That’s a man’s bet David. The EWMCF run has Gustav headed right towards you guys. Hopefully it will make landfall in less populated areas of the coast but even at 72 hours out, there’s still significant uncertainty as to the track.

    #155, It could reach CAT 5 but it won’t be CAT 5 when it makes landfall. Typically, hurricanes in the GOM lose strength as they approach land, at least for the US. A combination of factors really. I have seen a few that hit MX that hard though and of course, it’s different for the FL peninsula on the Atlantic side as well.

  159. Posted Aug 30, 2008 at 5:02 PM | Permalink

    Re #158 (Jonathan Schafer): Ryan had a typo where he called Gustov a cat 6. That may have been a Freudian slip. This does have a shot at breaking records.

    David, if it wasn’t for Hanna I would say come down for some fishing. Be careful.

  160. Posted Aug 30, 2008 at 5:08 PM | Permalink

    I love the new paste link thing! John A and MrPete are my new heros.

  161. BarryW
    Posted Aug 30, 2008 at 5:52 PM | Permalink

    Re 158

    Your right iit will probably weaken before landfall, but that may not help the oil rigs.
    Even though it may be less intense wind wise, what would be the affect relative to storm surge if it intensified just prior to landfall even though land fall was say at cat 3? Katrina was a cat 3 and I seem to remember some comment that the surge was worse because it had been stronger right before landfall building up a larger surge than a cat 3 would. That might be faulty remembering on my part though.

  162. MrPete
    Posted Aug 30, 2008 at 7:58 PM | Permalink

    FWIW, I have a plane ticket to visit N’Orleans Sep 5-9… and need to get south of N.O. across the river. I was planning to help my dad clean up a “normal” mess… guess we may have a slightly bigger mess on our hands… if I can even get there. If I can, perhaps we’ll have some on-site photos. No weather reports though so probably too boring for CA.

  163. kim
    Posted Aug 30, 2008 at 8:17 PM | Permalink

    Brendan Loy is blogging as ‘Weather Nerd’ at Pajamas Media. Some may remember him from his Katrina coverage. Where did that high, dry, air come from?

  164. Posted Aug 30, 2008 at 8:28 PM | Permalink

    Re #161 (BarryW):Thus far the GOM offshore oil wells have handled big storms very well. The problem is with land based refineries, mainly with access after storms. Evacuating and re-crewing offshore platforms causes a few days lost production, but refinery damage and limited access can cause several months in production delays. Building more refineries in less susceptible locations would level the spikes due to hurricanes. Building new refineries to use less desirable types of sour crude would also reduce fluctuations in distillate supply and increase our global stature as a distillate player.

    In the US we are mainly light sweet only refining and farm out the hard stuff, sour Saudi and Canadian oil sands etc. to the ROW.

  165. BarryW
    Posted Aug 30, 2008 at 8:57 PM | Permalink

    Re 164(captdallas2)

    I think there was some minor damage after Katrina on the rigs so they did take some time to get back on line, but your right the refineries are the major problem. NIMBY is the first thing screamed whenever anyone mentions building a refinery. What’s it been, 20 years?

  166. Posted Aug 30, 2008 at 9:19 PM | Permalink

    Re #165 (BarryW): Easy 20 years. NIMBY is getting played. If we aren’t going to drill don’t let the Row drill. The most asinine policies I have every seen. We should lead the world not farm it out to the world. Clean coal is a technology we should export.

    BTW, Gus has a projected cat 5 path, (intensity projections are not all that great) a touch East of the last line I predicted. That will tend to drop its intensity. If it curves West it could maintain more steam.

  167. Posted Aug 30, 2008 at 9:37 PM | Permalink

    Okay, going to bed. We had winds to 40 knots but they are dropping. A few tornado alerts and a few inches of rain. Now we wait for Hanna. Good luck Gulf coast.

%d bloggers like this: