RealClimate logo

Technical Note: Sorry for the recent unanticipated down-time, we had to perform some necessary updates. Please let us know if you have any problems.

Hurricane Heat Kasırga AteşiChaleur d’ouragan

Filed under: — gavin @ March 1st, 2007

The big problem with much of the discussions about trends in hurricane activity is that the databases that everyone is working from are known to have significant inhomogeneities due to changes in observing practice and technology over the years. So it’s not surprising that a new re-analysis (Kossin et al, published yesterday) has been generating significant interest and controversy among the hurricane research community (see e.g. Prometheus or Chris Mooney). However, rather than this study being taken for what it is – a preliminary and useful attempt to make homogeneous a part of the data (1983 to 2005) – it is unfortunately being treated as if it was the definitive last word. We’ve often made the point that single papers are not generally the breakthroughs that are sometimes implied in press releases or commentary sites and this case is a good example of that.

Kossin et al develop an algorithm based on North Atlantic data that can be theoretically used with the coarsest data available from the earlier parts of the record and in more remote regions. While the technique works well in the North Atlantic (picking up almost all of the storms seen in the standard data), it doesn’t work as well in other basins – possibly because the characteristics of tropical cyclones are not universal, or because the coarse early remote sensing data are still not sufficient. The poorer performance in the other basins is surely a reason to anticipate that further work will be necessary to refine these estimates, and should serve as a caution to those wanting definitive conclusions.

How does this fit in with some of the previous work? Well, it confirms the large trend in the North Atlantic (seen in Emanuel, 2005), but doesn’t show significant trends in the other basins (from 1983). This isn’t directly comparable with Webster et al (2005) though, since their trends start in the 1970s, and the shortness of the new reanalysis (only 23 years) emphasizes interannual and decadal variability associated with e.g. El Nino. The Kossin et al study is therefore unlikely to shed much light on the potential global warming/hurricane intensity link.

In summary, read the papers and the comments but don’t believe the hype.

We’ll start off the discussion with a few comments we have already received on the provocative study:

1. Chip Knappenberger:

Based upon the new results of Kossin et al. (GRL, 2007), it looks like the IPCC SPM just barely covered itself in its proclamations on observed hurricanes:

There is observational evidence for an increase of intense tropical cyclone activity in the North Atlantic since about 1970, correlated with increases of tropical sea surface temperatures. There are also suggestions of increased intense tropical cyclone activity in some other regions where concerns over data quality are greater. Multi-decadal variability and the quality of the tropical cyclone records prior to routine satellite observations in about 1970 complicate the detection of long-term trends in tropical cyclone activity. There is no clear trend in the annual numbers of tropical cyclones.

From the results presented in Kossin et al. the “suggestions” of increases in intense tropical cyclone activity in regions other than the Atlantic basin are not really so well supported, at least for the last 23 years.

We’ve tried to make this point a bit more clearly here, despite the implications of the headline of Kossin’s press release.

(disclosure: I have, to some degree, been funded by the fossil fuel industry since 1992)

2. Dr. Kevin Trenberth:

1) The methodology is trained on the Atlantic. It has no parameters to allow for different structures or size of storms, and there is no good reason why it should work well on storms in other basins. Given the different land-sea configurations and the different role of ENSO in the different basins, and the fact that disturbances in other basins do not form from easterly waves from off of Africa, there is every reason to expect that storms in other basins have different characteristics. For instance, there is greater activity in the Pacific Northwest, and the tropopause is higher in the western Pacific, and this affects brightness temperatures at tops of clouds. If the size of storms differs then the fixed form of EOFs will not be able to capture that form. The analysis must be able to account for differences among basins in order to have confidence in variability or trends. It would be easy enough to test whether the storms in other basins had different characteristics by also performing an EOF analysis for each region. This basic test was not done. It should be.

2) The results are suggestive of these problems. In the SIO where the method gives 0, 1 or 2 storms vs up to 6 in the best track data, there is a serious bias. Similar large biases exist in the SPAC (up to 2 vs 5 in best track). Obviously the threshold is effectively different and it is a comparison of apples and oranges.

3) In addition, this version of the paper deals with PDI. The earlier version of the paper dealt with intensity of storms and that was abandoned because the results were not very good. In particular, the presumption is that the older results were the problem because operational methods have improved. But the Kossin et al results showed bigger and greatest discrepancies with those from best track in recent years: there is no convergence over time. This is harder to see with PDI, because the biggest storms are emphasized, but the question of why is there not good agreement in recent years is not answered.

3. Dr. Judith Curry:

The most vexing thing about the tropical cyclone data sets is the uncertainty that analyst subjectivity contributes to this. The Dvorak scheme for determining tropical cyclone intensity is notoriously subjective, see the recent BAMS article on this. The importance of what Jim Kossin has done is to take this subjectivity out of the analysis.

Kossin’s method matches well the historical data in the North Atlantic (NATL) and East Pacific (EPAC). The method was trained using North Atlantic data, and the East Pacific regime in terms of dynamical and thermodynamical conditions is very close to the North Atlantic conditions. However, Kossin’s method diverges substantially from the established data sets in the Western Pacific, South Pacific, and Indian Oceans. Does this mean that the established data sets or in error, or that Kossin’s method (trained in the Atlantic) does not translate well to the other ocean regions?

Owing to problems with dealing with historical satellite data, Kossin’s study was extended only back to 1983 (the period for which the satellite data are well calibrated), and it is almost certain that this data set cannot be extended back prior to 1977. By itself, this data set is too short to say anything about a trend in intensity. But it can in principle be used to assess uncertainties in the established data sets.

My own analysis of the discrepancies has focused on the Western Pacific (WPAC) data, where 40% of the global tropical cyclones form. During the period 1983-1987, Kossin’s data overlaps with aircraft reconnaissance data; I would expect the WPAC TC data during this period to be of comparable quality to that in the NATL and WPAC, but large discrepancies are seen. I would also expect the agreement to be better in more recent years with the advent of more sophisticated satellite systems, but the discrepancies are largest during the most recent period. Its would not be surprising for this method trained in the NATL not to work well in the other regions. The method does not allow for different structures or sizes of storms in different basins. The NATL cyclones form primarily from easterly waves, while those in other regions do not. The role of ENSO is different in the different basins. The tropopause height is higher in the WPAC. etc.

Many people in the tropical cyclone community have questioned the Emanuel and Webster et al. papers owing to uncertainties in the data sets. Other than anecdotal analyses, little has been done to quantitatively assess the uncertainties. Kossin’s paper is arguably the first important word on this subject, but it certainly won’t be the last word. To establish the credibility of Kossin’s data set outside the NATL, considerably more analysis is needed to understand discrepancies in individual basins and the nature of the discrepancies on a storm-by-storm basis.

Ingilizce’den çeviren: Figen Mekik

Kasırga olaylarındaki trendler (eğilimler) ile ilgili anlaşmazlıkların ana sebebi herkesin kullandığı veri tabanlarının arasında büyük benzeşmezliklerin oluşu. Bu uyuşmazlıkların kökünde yıllarca süren ve birbirine benzemeyen gözlemleme usulleri ve teknolojik farklılıklar var. Dolayısıyla, yeni bir analizin (Kossin ve digerleri , dün yayınlandı) kasırga arastırmacıları arasında çok büyük ilgi ve polemik yaratmış olması şaşırtıcı değil (mesela, Prometheus veya Chris Mooney). Ancak bu çalışmayı olduğu gibi kabul etmek yerine – yeni ve kullanışlı bir homojen (tektürel) veri tabanı (1983 -2005) olması – çok kişi maalesef son sözmüs gibi okuyor. Biz, pek çok kez belirttiğimiz gibi, hiç bir tek çalışmanın medya veya başka yorum sayfalarında sunulduğu gibi büyük bir gelişme muamelesi görmesi taraftarı değiliz. Bu çalışma da bunun güzel bir örneği.

Kossin ve diğerleri Kuzey Atlantik verilerine dayalı bir algoritma geliştirmişler. Bu algoritma, tortusal rekorun daha önceki kısımlarında ve ulaşılması zor bölgelerde kaba verileri kuramsal olarak kullanılabilir hale getiriyor. Bu, Kuzey Atlantik’te iyi sonuçlar veren bir yöntem olduğu halde (standart veri tabanlarında bulunan hemen hemen tüm fırtınaları açıklayabiliyor), başka havzalarda aynı derecede başarılı olamıyor. Bunun muhtemel sebebi tropik kasırgaların özelliklerinin evrensel olmayışı, veya öncül kaba verilerin yeterli olmayışı. Bu bölgelerdeki başarısızlıklar pek tabii ki algoritmalar ile yapılan hesapların ileride daha rafine olacağını umud ettiriyor, ve ivedi kesin sonuçlar arayanlara bir uyarı olmalı.

Peki, daha önceki çalışmalarla bu çalışma arasında nasıl bir uyum var? Daha önceden tespit edilen (Emanuel, 2006) Kuzey Atlantik’teki büyük ölçekli eğilimleri doğruluyor bu çalışma; ancak diğer havzalardaki büyük trendleri kestiremiyor (1983’ten bu yana). Fakat, Webster ve diğerleri (2006) ile dolaysız bir mukayese mümkün değil çünkü onların verileri 1970’de başlıyor. Ayrıca yeni analizin kısa süreli oluşu (sadece 23 yıl), El Nino’ya bağlanan senelik veya onyıllık değişkenliği vurguluyor. Yani, Kossin ve diğerlerinin çalışması olası küresel ısınma ile kasırga yoğunluğu arasındaki ilişkiyle ilgili fazla bir yeni bilgi sunamıyor.

Özet olarak, makaleleri ve yorumları okuyun, ama her yazılana inanmayın.

O grande problema com a maior parte das discussões sobre tendências na atividade dos furacões é que os conjuntos de dados que todos estão utilizando são conhecidos por não serem homogêneos devido a mudanças nos procedimentos de observação e na tecnologia ao longo dos anos. Assim, não é surpreendente que uma nova re-análise (Kossin et al, publicado em 28/02/2007) gerou um significativo interesse e controvérsia no seio da comunidade de pesquisadores de furacões (veja, por exemplo, Prometheus ou Chris Mooney). No entanto, ao invés de tomar este estudo por aquilo que ele é – uma tentativa preliminar e útil de tornar homogênea uma parte dos dados (1983 a 2005) – ele está sendo tratado como se fosse a última e definitiva palavra sobre o assunto. Nós freqüentemente afirmamos que trabalhos isolados não são em geral os grandes avanços (“breakthroughs”) sugeridos pela imprensa ou pelos sítios de comentários, e este caso é um bom exemplo.

Kossin et al desenvolvem um algoritmo baseado em dados do Atlântico Norte que podem teoricamente ser utilizados com os dados de menor resolução do início das séries históricas e de regiões mais remotas. Enquanto esta técnica funciona bem no Atlântico Norte (detectando quase todas as tempestades observadas no dado padrão), ela não funciona tão bem em outras regiões – possivelmente porque as características dos ciclones tropicais não são universais, ou porque a resolução dos dados iniciais de sensoriamento remoto ainda é insuficiente. O pior desempenho em outras regiões é certamente uma razão para antecipar que mais trabalho será necessário para refinar estas estimativas, e deveria servir como um sinal de alerta para aqueles buscando conclusões definitivas.

Como esta pesquisa se encaixa com alguns dos trabalhos precedentes? Bem, ela confirma a forte tendência de aumento no Atlântico Norte (visto em Emanuel, 2005), mas não mostra tendências significativas em outras regiões (desde 1983). Este resultado não pode, entretanto, ser diretamente comparado com os de Webster et al (2005), uma vez que suas tendências começam nos anos 1970, e a brevidade da nova re-análise (somente 23 anos) enfatiza a variabilidade interanual e decadal associada ao El Niño, por exemplo. Então, é improvável que o estudo de Kossin et al venha a esclarecer muito sobre a ligação potencial entre aquecimento global e a intensidade dos furacões.

Em resumo, leia os trabalhos e comentários mas não acredite no oba-oba.

traduzido por Fernando M. Ramos e Ivan B. T. Lima

Traduit par Etienne Pesnelle

Le gros problème dans la plupart des débats concernant les évolutions de l’activité cyclonique, c’est que les bases de données avec lesquelles tout le monde travaille ont notoirement des non-homogénéités significatives, dues aux changements dans les pratiques et les techniques d’observations survenus au fil des ans. Il n’est donc pas surprenant qu’une nouvelle ré-analyse (Kossin et al, publié hier) ait provoqué un intérêt marqué et une controverse au sein de la communauté de recherche cyclonique (voir ainsi Promotheus ou Chris Mooney). Cependant, plutôt que de prendre cette étude pour ce qu’elle est – une tentative préalable et utile d’homogénéiser une partie des données (1983-2005) – on la traite malheureusement comme s’il s’agissait du
mot de la fin. Nous avons souvent fait remarquer que les articles isolés ne sont pas les percées que laissent parfois supposer les communiqués de presses ou les sites de discussion, et ce cas en est un bon exemple.

Kossin et al ont développé un algorithme, fondé sur les données nord-atlantiques, qui peut être théoriquement appliqué aux données les plus grossières disponibles dans le passé le plus lointain et concernant des régions plus éloignées. Alors que la technique marche bien sur l’Atlantique du Nord (recueillant pratiquement toutes les tempêtes trouvées dans les données standards), elle ne fonctionne pas aussi bien sur d’autres bassins – peut-être parce que les caractéristiques des cyclones tropicaux ne sont pas universels, ou parce qu’il n’y a pas assez de mesures grossières de télédétection disponibles dans le passé. La moindre performance de l’algorithme dans les autres bassins est certainement une bonne raison pour penser que des travaux complémentaires seront nécessaires pour affiner ces estimations, et devrait servir d’avertissement à ceux qui veulent des
conclusions définitives.

Comment cela s’articule-t-il avec une partie des travaux précédents ? Hé bien, cela confirme l’évolution de tendance importante en Atlantique Nord (vue par Emanuel en 2005), mais ne montre pas d’évolution de tendance significative dans les autres bassins (depuis 1983). Cela n’est toutefois pas directement comparable à Webster et al (2005) puisque leurs évolutions de tendance commencent dans les années 1970, et la courte durée (23 ans seulement) de leur ré-analyse souligne la variabilité inter-annuelle et décennale qui est associée à El Niño par exemple. Il est donc assez improbable que l’étude de Kossin et al puisse éclairer quelque peu le lien potentiel entre réchauffement climatique et intensité des ouragans.

En résumé, lisez l’article et les commentaires mais ne prenez pas le battage médiatique pour argent comptant.

120 Responses to “Hurricane Heat Kasırga AteşiChaleur d’ouragan

  1. 1
    Nils Simon says:

    There is a freely accessible pdf version of this article available at Kossin’s website. See

    [Response:Thanks. We linked it in. -gavin]

  2. 2
    S. Molnar says:

    On a slightly tangential topic, what is the state of the “global warming will cause a permanent el nino” theory

    [Response: While I wouldn't quite characterize this in terms of a "permanent El Nino", a reasonably up-to-date discussion of possible climate change influences on ENSO is provided in our previous article El Nino and Global Warming --mike]

    and what are the implications for Atlantic hurricanes?

    [Response: There is also some discussion of this point in the RC article linked above. --mike]

    On a very tangential topic, I wonder [edited] [about] this comparison of apples and oranges.

  3. 3
    Craig Allen says:

    Is it likely that a useful proxy will ever be found for hurricanes?

    [Response: It's definitely a tricky problem, but there is some progress on local landfalling hurricanes - overwash in coastal ponds in Massachusetts, tree ring isotopes in Georgia etc. Open ocean proxies don't have the resolution, and so a complete record will never happen - the hope is that if landfalling data can be significantly extended it might give reasonable long term statistics - but right now my understanding is that this is all very preliminary. Any comments to contrary welcome! - gavin]

  4. 4

    Hi Guys-

    Klotzbach 2006 uses the same dataset as Webster et al. 2005 (note typo in your post, not 2006) and finds no trends in numbers of Cat 4-5 or ACE other than in the Atlantic.

    So shouldn’t Kossin et al. actually reinforce Klotzbach 1986 and suggest that any global increase in storm intensity (or numbers of intense storms) necessarily occurred before 1986 (or 1983)?

    Whether that increase is due to data or climate is a separate issue, but isn’t there multiple lines of evidence now that there are no global trends since the early- to mid- 1980s, whereas in the Atlantc there has been?

    Klotzbach paper:


    [Response: Thanks - fixed the typo. -gavin]

  5. 5
    Thom says:

    Call me confused….But since when are Chris Mooney and Roger Pielke Jr. part of the “hurricane research community?” Neither of these two are scientists.


    [Response: The reference was mostly to the comments sections where bona fide scientists are mixing it up. -gavin]

  6. 6
    Ike Solem says:

    At least the argument that it’s simply the North Atlantic Oscillation that’s responsible for the increased intensity trend in the Atlantic has apparently been discarded; at least it isn’t mentioned at Prometheus:

    “The paper does find the Atlantic to be more active over the past 23 years. No one in the world has ever questioned whether or not the Atlantic has been more active over the past 3 decades. Any assertion that the Atlantic has become more active is hardly “fresh evidence.” This is not news.

    That assertion is false for a number of reasons; for example see this article in science 2001:

    “The Recent Increase in Atlantic Hurricane Activity: Causes and Implications

    Stanley B. Goldenberg, Christopher W. Landsea, Alberto M. Mestas-Nunez, William M. Gray

    suggested that the hurricane data was too spotty to support the trend:

    One may ask whether the increase in activity since 1995 is due to anthropogenic global warming. The historical multidecadal-scale variability in Atlantic hurricane activity is much greater than what would be “expected” from a gradual temperature increase attributed to global warming. There have been various studies investigating the potential effect of long-term global warming on the number and strength of Atlantic-basin hurricanes. The results are inconclusive. Some studies document an increase in activity while others suggest a decrease. Tropical North Atlantic SST has exhibited a warming trend of ~0.3°C over the last 100 years; whereas Atlantic hurricane activity has not exhibited trendlike variability, but rather distinct multidecadal cycles as documented here and elsewhere. The extreme activity in 1995 has been attributed in part to the record-warm temperatures in the North Atlantic. The possibility exists that the unprecedented activity since 1995 is the result of a combination of the multidecadal-scale changes in Atlantic SSTs (and vertical shear) along with the additional increase in SSTs resulting from the long-term warming trend. It is, however, equally possible that the current active period (1995-2000) only appears more active than the previous active period (1926-1970) due to the better observational network now in place. During the previous active period, only 1966-1970 had continual satellite coverage. Further study is essential to separate any actual increase from an apparent one due to more complete observations.

    So, progress has indeed been made. Now the question is whether the hypothesis that this is all due to the North Atlantic Oscillation can be discarded or not – would that be yes, or is the jury still out?

    Note that even more recently, the following perspective in science was published:
    “Science 28 July 2006

    Can We Detect Trends in Extreme Tropical Cyclones?
    Christopher W. Landsea, Bruce A. Harper, Karl Hoarau, John A. Knaff.

    Subjective measurements and variable procedures make existing tropical cyclone databases insufficiently reliable to detect trends in the frequency of extreme cyclones.”

    It’s rather strange that Prometheus is just sweeping this whole issue under the rug. Does that indicate acceptance of a global-warming related trend in hurricane intensity in the North Atlantic?

  7. 7


    The scientific debate on what has occurred in the Atlantic is alive and well. Some scientists assert that the 1995-2006 period is more active in the Atlantic than any time in the past 50-100 years (e.g., Webster, Holland, Emanuel), and others argue that it is not (Goldenberg, Landsea, Gray). The current state of the peer-reviewed literature provides support for both perspectives (see WMO consensus statement). E.g., see the exchange between Emanuel and Landsea in Nature December 2005:

    I am sure that you will hear more about the AMO, AMM, AGW etc. and the Atlantic in forthcoming papers. However, conclusive results widely accepted are not likely to appear anytime soon …

    Gavin- I was expecting a bit more of a substantive response on #4! How about it?;-)

  8. 8
    Jim Kossin says:

    I thank Dr. Trenberth for his comments. I’ll respond here, and I’ll apologize in advance for the length of my response. It may be longer than the article itself…

    The algorithm does in fact “allow for different structures or size of storms” and it is entirely justifiable to expect it to perform reasonably well outside of the Atlantic. We state that

    “The EOFs contain information about hurricane eye temperature (when an eye is present), the height of the convective eyewall clouds, and the average radial structure of cloudiness around the storm [cf. Kossin et al., 2007], and these factors are correlated with hurricane intensity. For example, warmer eye temperature and higher eyewall clouds (indicated by colder cloudtop Tb) are strongly related to greater intensity. This is the foundation of the Dvorak Enhanced Infrared (EIR) technique [Dvorak, 1984], which is utilized by all tropical forecast offices in every ocean basin to estimate hurricane intensity with geostationary infrared imagery. In the Dvorak EIR technique, eye Tb and cloud-top Tb are directly related to intensity with a look-up table. The EOFs represent these temperatures but also contain additional information about radial structure such as eye size and radial extent of the cold cloud-tops above the eyewall. This additional information is also related to intensity [Kossin et al., 2007].”

    We also state that

    “The relationship between Tb (represented by the PCs) and intensity is analogous to the basis of the Dvorak technique, and latitude has been shown to modify this relationship [e.g., Kossin and Velden, 2004] as well as affect radial structure and size [Kossin et al., 2007].”

    The relationship between these EOFs and storm size and structure is also very clearly documented in Mueller, et al. (2006) and Kossin et al. (2007). Both papers are in WAF. These are recent papers and Dr. Trenberth is probably unaware of them. Yes, the EOFs are fixed by construction, but their weighted linear combinations do describe size, and the PCs then describe size variability. This is a fundamental concept of course. It is also been documented that storm size is actually not that well correlated with intensity (compare Hurricanes Charlie and Katrina for example; same intensity, vastly different sizes).

    Furthermore, the Dvorak technique (used in all basins) actually makes little to no distinction regarding storm size. For example, when an eye is present, the raw Dvorak estimates depend only on two temperature measurements: eye and eyewall. Size makes no difference whatsoever in these cases.

    As to the basin-dependence, it seems that Dr. Trenberth is unfamiliar with the Dvorak technique. The intensity in the best track, as deduced by the Dvorak technique, is estimated in an analogous manner in all basins. We very clearly state in the paper that the relationship between the Tb field and intensity is fundamentally the same (within a fixed bias that does not affect the normalized values or the trend). We state:

    “The fundamental relationships between infrared imagery (as well as the other predictors) and intensity do not change among the various ocean basins, and the Dvorak technique is applied in much the same way everywhere. There are differences in the method for converting raw Dvorak technique output into intensity estimates, but these differences can only create a temporally constant bias and this will not affect the trend analyses shown in the next section.”

    This is a statement of fact that any hurricane expert would be aware of and it would not be debated. Dr. Trenberth can consult with any hurricane forecaster (i.e. the creators of the best track) to confirm this. The Dvorak technique also does not depend on ENSO or basin size, etc. and nor should our algorithm. The goal was to create homogeneous Dvorak-type (i.e. IR-based) intensity estimates to test against trends in the existing heterogeneous Dvorak-based records. The increased activity in the Northwest Pacific, tropopause height, land-sea configurations, basin size, ENSO, and the role of easterly waves that Dr. Trenberth elaborates on are not at all relevant in this context. Ultimately, it seems that Dr. Trenberth is questioning the tenets of the Dvorak technique itself, but it’s not our goal to convince anyone of the validity of a method that has been applied globally for 30 years by every forecast office on every continent (see Velden et al. 2006). It would also be inappropriate for us to give a history of the Dvorak technique in our paper, and we have a reasonable expectation that an interested reader will consult the references we offer. I strongly encourage Dr. Trenberth to do this.

    Dr. Trenberth’s remark about the EOFs in the Western North Pacific being different than the EOFs from the Atlantic is a good one. This was looked at, and should have been mentioned. The EOFs are essentially identical. I’ve put the figure on my website.
    As I mentioned above, the different linear combinations of the EOFs are responsive to size, and the PCs, which are the predictors in the algorithm, control these combinations. A large storm in the Atlantic, like Gilbert (1988) or Floyd (1999) or Katrina (2005) is very similar to a large storm in the W. Pacific. On average, the storms in the W. Pacific are larger, but that doesn’t mean that the EOFs could not be combined appropriately to represent this. The Atlantic has large storms too.

    As a final point about our algorithm performing reasonably well outside of the Atlantic, note that the algorithm performed remarkably well in the Eastern Pacific.

    As to point 3), this is simply false and I don’t know what Dr. Trenberth is basing this somewhat inflammatory comment on. The first draft of our paper considered mean kinetic energy (MKE) instead of PDI. The results were the same for both MKE and PDI (or ACE for that matter). In fact, most presentations I’ve given lately have used the MKE results because I could make all the same points using my old slides. We originally looked at MKE because we thought that an intensive variable based on intensity would be more useful than an extensive variable like PDI, which is strongly dependent on the number and duration of storms. Ultimately we had a lot of comments that there was no need to introduce a new variable into the mix and if we could show the same thing with PDI, which also helps to more directly address Emanuel (2005), then that might be the better route. Again, the results don’t change so I don’t see this as a big deal.

    As to the differences between the start versus the end of the time series, I think Dr. Trenberth is missing a key point to this analysis. There may be biases, and in fact we expect biases when we apply the algorithm to other basins. Just as raw Dvorak “T-numbers” are converted to winds differently in the West Pacific versus the Atlantic, our algorithm is expected to have a systematic bias. The point is that the bias is not time-dependent. We are testing for trends, not making contemporaneous absolute comparisons. This type of confusion is what motivated us to standardize all the variables. A systematic bias affecting the absolute values is ok. The consistency over time is the important goal.

    Dr. Trenberth is incorrect to say that the estimates should be better in the latter parts of the records than the beginning. First, all basins from the JTWC best track are measured almost entirely with the Dvorak technique. There is no ground truth available to “recalibrate” the forecasters. The Dvorak technique is a subjective technique and it has in fact been hypothesized that forecasters have been overestimating intensity in recent years because of the higher quality/resolution satellite data (Harper and Callaghan 2005). So this is quite opposite from Dr. Trenberth’s claim that accuracy should improve with time. Furthermore, Dr. Trenberth is also apparently unaware (although it was stated clearly in the paper) that the W. Pacific had aircraft reconnaissance until 1987 at which point they were forced to rely entirely on Dvorak satellite estimates of intensity. So we would expect the accuracy to be greatest in the earliest part of the record (1983-1987). We state:

    “Since aircraft reconnaissance into storms was routine in the Northwest Pacific during the earlier period of our record 1983-1987, the best track intensities during this period are likely to be more accurate than the later period 1988-2004 after the termination of reconnaissance. This indicates a systematic over-estimation of intensity in the later period of the JTWC record when compared to the UW/NCDC data.”

    Dr. Trenberth, I am puzzled by how forcefully you are dismissing our results while you apparently have virtually no knowledge of how intensity estimates are formed or how best track records are constructed. You seem to know very little about many of the things that a hurricane researcher would consider very fundamental. There’s no crime in this of course, but you’re taking an adversarial tack on these things and I think that’s unreasonable.

    As a general parting note, I’d like to respond to the comment in the introductory section of this website:

    “However, rather than this study being taken for what it is – a preliminary and useful attempt to make homogeneous a part of the data (1983 to 2005) – it is unfortunately being treated as if it was the definitive last word. We’ve often made the point that single papers are not generally the breakthroughs that are sometimes implied in press releases or commentary sites and this case is a good example of that.”

    I’ve been greatly concerned about this for quite some time, as public statements made by other scientists suggested that our work would really get to the bottom of a number of questions. These were generous and supportive statements, but I knew that people were overstating our work. In addition to using the phrase “as a first step” in a number of places, my concern led me to insert the following into our paper:

    “…the method applied here, while far from a panacea, is a reasonable first cut at the problem. Future analyses will hopefully expand on our work.”

    I for one have no illusions regarding the importance of this work, but I think it is indeed a reasonable first cut at the problem and I think we’ve presented the results objectively and transparently.
    –Jim Kossin

  9. 9
    Jim Kossin says:


    Thanks for your comments. I think you have hit a number of things on the mark. I just want to offer a couple of responses (don’t worry, this will be much shorter than my last message)

    Some of your comments regarding basin differences, ENSO, etc were addressed in my comment to Dr. Trenberth. Those differences should not matter when considering best track intensity estimates. Also the comparisons of how close the series are at the beginning versus the end is not relevant since you can translate either curve up or down as you like. We only care about the trends here and not absolute contemporaneous comparisons.

    I will have to disagree with your claim that “the East Pacific regime in terms of dynamical and thermodynamical conditions is very close to the North Atlantic conditions.” I’d like to put this forth on the Tropical Storm listserv to get more feedback, but I would say that Eastern Pacific storms are very different from Atlantic storms. The SST structure is completely different, as is the shear. Storms intensify and weaken very differently there. There is lots of very cold water and stratus decks in the area that kill storms within hours if they stray too close. There is a pronounced ITCZ in the neighborhood. I would guess that the static stability is different there too. Basically all the things storms care about are different between the basins. So with all of this being said, the fact the our algorithm performs well in the E. Pacific as well as the Atlantic can’t be chaulked up to the storms and environment being the same.

    Let me suggest a different possibility for why the algorithm finds bigger disparities everywhere except the ATL and EPAC: The Atlantic and E. Pacific best tracks are constructed by NHC forecasters who are the best trained hurricane forecasters in the world, and they stay at their jobs for decades. Their interest in best track quality is exceedingly high. The best track from the other basins are constructed at JTWC by temporarily assigned military personnel, many of whom have no experience whatsoever with hurricanes. They are cycled through every couple years and never are allowed to get very good at hurricane forecasting. It’s not their career. This is why the JTWC best track comes with such a strong written warning regarding the quality of the intensity estimates. The creation of the best track at JTWC is also a very low priority item (ask anyone who works there). They do it when they can and they don’t really have the manpower. That’s why the 2005 JTWC best track was only made available about a month ago, while the 2006 NHC best track has already been available for months. Yes, this is all anecdotal, but it does offer an alternative reason for why we found differences in the JTWC basins and not in the NHC basins.

  10. 10

    Is there any connection between global climate change and tornadoes? And are there more tornadoes now? And are they stronger or in new places or new seasons?

  11. 11
    Chris Mooney says:

    Susan K,
    The IPCC says there is insufficient evidence to determine whether there is any trend in tornadoes. We had a significant go at it on the global warming tornado question, with some scientists participating, here

  12. 12
    Judith Curry says:


    If you look at the diagram (Fig 4) in Webster et al. plotting NCAT45 vs time, the appropriate diagram to consider is NCAT45/N (the fractional number of NCAT45). This is because in the latter part of the time series there is a decrease in the total number of tropical cyclones, largely owing to a large multidecadal cyclone in the WPAC (which comprises 40% of the global tropical cyclones), see Fig 3 in Webster et al. The NCAT45/N curve in Fig 4 does show more of a linear trend (albeit with some zigs and zags).

    The main issue is that the 35 year period used by Webster is marginally on the short side for doing a global trend. While averaging among all the basins does average out much of the basin-specific multidecadal cycles, owing to the large number of WPAC tropical cyclones, we still see a signal from the Pacific multidecadal cycle in the global average. Going to shorter periods like 20 years really doesn’t make much sense (physically or statistically) in trying to decipher a trend.

  13. 13
    Judith Curry says:


    We all appreciate the extra effort you made to comment here, it is such a help when the authors spend some time here to answer questions.

    I still suspect that the higher SST values and much stronger and larger tropical cyclones particularly in the Western Pacific are in a different region of “parameter space” that wasn’t included in your training data set. While size isn’t included in the Dvorak scheme, it does reflect some differences in dynamics that may influence intensity in more subtle ways that may not be captured by your objective analysis scheme. Do you have any thoughts on how you might convince us “skeptics” on this in terms of a further analysis that might be done?

    With regards to the Dvorak scheme, I was astonished by the Wu article in the December 06 EOS. I view the differences in these 3 analyses of the Western Pacific tropical cyclone intensity to be “error bars” on the s Dvorak scheme, largely associated with subjectivity of the analyst. Pretty large error bars indeed.

    I have one other question. I understand why you only went back to 1983, but based on your expert judgment and the analysis you have done, how far back in time would you expect the NATL and EPAC data to be reliable? 1977? 1973? Certainly prior to 1970, we the the kerfuffle between landsea and emanuel regarding Landsea’s 1993 intensity correction to sort out.

    By the way, I view your confirmation of the intensity data since 1983 in the NATL and EPAC to be a huge help in trying to sort this whole thing out.

  14. 14

    What most people seem to forget here, is that on one hand there are fundamental thermodynamical arguments which demand that hurricane intensity increase over time, while on the other hand, there are these two big chunks of ice sitting in both polar regions, which will counteract the warming process in their own special way. These two different effects will compete for many centuries to come.

  15. 15
    Ray Ladbury says:

    Re 14: Greenhouse heating would be a secondary or even tertiary source of the energy of tropical–regardless of where they occur. Personally, I think that the level of effort being directed at teasing our such a signature or disproving it is overblown. Increased tropical storm intensity will likely be way down on the list of serious impacts of climate change. Much higher on the list will be the melting of those two big chunks of ice you refer to, and the subsequent release of more greenhouse gases trapped in permafrost. The signature of anthropogenic climate change is strongest there, just where the models predict we’ll see it. Looking for signatures in tropical storms, while an interesting diversion, is nonetheless a diversion from strengthening the case for anthropogenic causation and the need to do something about it. I do not intend by this to denigrate the very good work by Jim Kossin or others. It is important research. It just isn’t central to the argument for anthropogenic causation of climate change.

  16. 16
    Judith Curry says:

    Ray, the uncertainties in the tropical cyclone data preclude the cylones from being used as a “smoking gun” for gobal warming. The importance of the hurricane-global warming link in the overall global warming debate is more psychological/emotional. For the first time (after Katrina and the Emanuel/Webster papers), the public sensed that one degree warming could potentially have some serious consequences, if one degree would make future Katrinas more likely. So in terms of near term impacts of global warming, the risk of increased hurricane activity is pretty high on the list of things that the public is worried about.

  17. 17
    Lynn Vincentnathan says:

    What gets me is how contrarians use the hurricane issue (like “scientists cannot say Katrina was caused by GW,” or “there’s no increased trend in the Pacific”) to imply that GW has been disproved. Increased hurricanes are only ONE possible effect of GW, and even if (for the sake of argument) they were found even in the future not at all to be linked to GW, that would in no way disprove GW.

    We reg folks only have time for news headlines (not sci articles), so the man-in-the-street reads, “Katrina not caused by GW,” & thinks “GW has been disproved & I can rest easy,” while his wife reads in the Drs office, “GW increasing Hurricane intensity” (a reporter read “warming oceans increasing hurricane intensity” and got confused, and I don’t fault him for it, bec he/she’s very much like that person-in-the-street). And the couple is so busy fighting about GW that they fail to buy those CF or LED bulbs, or even turn off lights not in use.

    I’m trying to think like a denialist: since GW is not happening, then all its supposed effects have to be shown to be bogus as well, or due to other factors. These effects are serious problems for them. It’s tiring just to watch contrarians jump through all their inelegant hoops and jump at very thin threads (what next, after cosmic rays?).

    OTOH, for GW believers, if there were lack of evidence for increased hurricanes it would be a total NON-PROBLEM. Since there are 2 types of energy, kinetic and heat, we might expect that some of the increasing GW heat would translate into increased hurricanes & storms…or maybe it’ll be doled out in bits and pieces that no one hardly notices…or maybe it’ll just mainly stay in a heat form and not get kinetic on us. Weather is a lot more chaotic than climate, and there are many many factors impacting it, including Sally in Cincinnati singing, “Rain, rain go away.”

    We have no trouble at all understanding how internal COMBUSTION engines in our cars translate into driving over to the beach, so this heat-to-motion (GW to storms) should be a no-brainer. Eventually, if not right now, we would expect to see increase in storm intensity and perhaps frequency in a globally warming world, all things being equal…which they are not, since even weathermen can’t well predict next week’s weather due to some butterfly flapping it’s wings in Japan gumming up the wind system.

    The important point here is that GW is going to entail many types of serious harms, one of which will likely be increased storms…if not now, then most likley later. And we reduce our GHGs like maniacs now, not that we think it’ll reduce Africa’s current drought, but that it will help in the future.

    An important point in the article, I felt, was: “the very real and dangerous increases in recent Atlantic hurricane activity will no doubt continue to provide a heightened sense of purpose to research addressing how hurricane behavior might change in our changing climate…”

  18. 18
    yartrebo says:

    Re #14:

    The cold stored in polar ice is small compared to the solar forcings related to global warming. It melts without having much cooling effect, and in short order there is net warming because of the reduced albedo of wet snow vs. dry snow and bare rock vs. snow cover.

  19. 19
    Ray Ladbury says:

    Re 16. Thanks for the response, Judith. I agree that hurricanes seem to be what the public responds to. However, we know that the public in general does a poor job of accurately perceiving risk. What is more, I think that the denialist camp has exploited the lack of a clear answer on tropical storms and the public’s seeming obsession with it to obfuscate the very clear evidence that does exist for climate change and the anthropogenic causation thereof. How many times have we heard denialists cite the mild Atlantic hurricane season in 2006 as “proof” that we’re all just a bunch of alarmists? This is just turnabout for the irresponsible chatter that blamed Katrina and every other hurricane on climate change in 2005.
    The fact that the general public is ignorant of the nature of scientific evidence and scientific consensus means that we have to be very careful in what we link to climate change.

  20. 20
    Ray Ladbury says:

    Re: 17. Lynn, you may be giving too much credit to the denialist camp. Some are contrarians–bent on showing how much smarter they are than all the smart people. Some are sheep, merely spouting the party line without any understanding of the science. And some really don’t care, figuring, probably correctly, that if they can make enough money, they can insulate themselves against any adverse effects of climate change. I saw this type when I lived in Africa a lot. They were more than happy to leave their country in squalid conditions as long as they had money to withdraw from Swiss bank accounts that they could use to insulate themselves from the squalor.

  21. 21
    SecularAnimist says:

    Judith Curry wrote: “So in terms of near term impacts of global warming, the risk of increased hurricane activity is pretty high on the list of things that the public is worried about.”

    I think that worry is appropriate, given the massive damage inflicted by a single hurricane — Katrina — which even the world’s wealthiest nation has had, and continues to have, great difficulty dealing with. In the US alone, multiple major cities in addition to New Orleans are extremely vulnerable to such storms. If things had gone slightly differently, hurricane Rita’s impact on Houston could have caused a disaster even worse than Katrina’s impact on New Orleans.

    I would suggest that another near term impact that the public ought to be worried about is drought. An extreme, prolonged drought affecting North America’s prime agricultural regions would be devastating to the entire world, and unlike sea level rise which is so often discussed as a threat from global warming, a “mega-drought”, like a mega-hurricane, could occur suddenly in any given year, without warning.

  22. 22
    Jim Kossin says:


    I absolutely agree with you that storms in W. Pacific are different than Atlantic storms. And it’s not just SST. The monsoon trough there affects storms in a variety of ways unique to the region. (as an aside, Emanuel 2000 showed that WPAC storms and ATL storms evolve almost identically when intensity is normalized by local potential intensity). But this is not necessarily relevant to what we did. The Dvorak technique is applied the same way there and all that’s changed between ATL and WPAC is the pressure-wind relationship (by the way, the Dvorak technique was also trained in the WPAC). When an eye exists, which is certainly the case for the mature storms that Webster et al(2005) and Emanuel (2005) were most concerned with, the intensity is based on two temperatures; warmest pixel in the eye and temperature of coldest ring around the eye. The “Dvorak rules” that are used to convert raw T-numbers to a CI-number (Dvorak’s final output) have nothing to do with what basin it’s being applied in.

    I’ve just revisited Dvorak (1984) and can’t find any aspect to the Dvorak technique that reflects any basin-dependent differences in dynamics. Can you tell me specifically what you’re referring to? Different wind-pressure relationships aren’t relevant to our work, since this wouldn’t introduce a trend, and again, trends were all we were after.

    It seems that we are converging on the need to create a better Dvorak technique, and I think this is a worthy topic, but our goal was to compare with the existing Dvorak estimates, warts and all. That’s what the best track used by Webster etal, Emanuel, and almost everyone else, is based on, so I’m not sure why anyone would expect our algorithm to take all these other local environmental factors into account. The cross-validation in our paper also suggests that the Dvorak estimates are more accurate than ours, but not that much. I was really surprised by this considering the simplicity of our algorithm.

    To answer your question about convincing skeptics, our algorithm could be tested against intensity estimates from early recon data in the WPAC. It would require finding the data, but I haven’t made such an attempt. Do you have a grad student who enjoys a good hunt? But again, if the Dvorak technique is applied in the same way in both basins, and strong storms are measured simply by IR brightness temperatures, (and all the previous work used these Dvorak estimates to measure trends), then I don’t see why it is so hard to believe that our algorithm could be applied to both basins.

    As we mentioned in the article, the folks at NCDC are working to extend the data back to about 1977 or so. The first geostationary satellite SMS-1 was launched in 1971 (if memory serves). I think that the GOES series started in 1973. There were polar orbiters in the 1960′s. My colleagues at NCDC have described how difficult it is to collect this historical data, but they are working on it. It’s a very unfortunate constraint and even in the best of circumstances, I don’t see how this type of satellite reanalysis will really get to the bottom of the pre-1970 Atlantic correction factor debate. What our analysis does do is confirm that the explicitly stated data issues that come with the JTWC best track do have a temporal dependence in them, and as it happens, this bias introduces an upward trend in metrics based on intensity.

  23. 23
    Ray Ladbury says:

    Re 21: Something we do agree on. I think drought is a much more likely, demonstrable and potentially much more devastating impact than increased hurricane intensity. So to, are the more intense rainfall events likely to result from climate change–and note that these can occur in the same places as droughts. The impacts I am most concerned about are those that strike at the very infrastructure of civilization–particularly our agriculture–in a world with over 9 billion people in it.

  24. 24
    Charles Muller says:

    It seems there are two debates here. One concerns the science, Jim Kossin et al. paper. Other the “hype” (RC article) or the “mixing of science, politics and the media” (Judith Curry et al 2006, BAMS).

    Concerning science, I recall the first main sub-hypothese in Curry 2006 : “the frequency of the most intense hurricanes is increasing globally”. In the light of Kossin 2007 (or previous critics on data accuracy like Landsea 2006), we can just say that there’s no consensus for the moment and that the statement still needs to be confirmed, before any assumption on tropical SST increase / intensity and GW / tropical SST increase. As Jim puts it nicely and modestly (many thanks to him for participating), “future analyses will hopefully expand on our work.” Longer trends will be useful for progress in our undestanding. I’m interested by Jim’s answer to Judith’s question in #13: “how far back in time would you expect the NATL and EPAC data to be reliable?”

    Concerning the hype, hurricane activity seems a classical example of strenghts and weakenesses of popular sensibilization to GW, and its potental counterproductive effects. As Judith observed in #16, the emotional impact of extreme events like Katrina has been sometimes perceived as a good occasion for incriminating GHGs and changing our mind about them. But if a more rational exam concludes that we don’t know the real culprit (maybe GHGs, maybe not), it could well lead to the opposite conclusions among laymen (science is definitely full of uncertainties, medias always exaggerate, all that hype around GW is politically biased, etc.). An that’s true IMO for many points unsufficiently documented, but highly mediatized because of their psychological impact (rate of sea-level rise, effects on biodiversity, tipping points in carbon cyles, global frequency of extreme events, etc.)

  25. 25
    Hank Roberts says:

    > how difficult it is to collect this historical data … a very unfortunate constraint
    Aside: a topic (elsewhere, sometime) on this would be welcome. Got data on paper? digitizing and distributed, volunteer proofreading are working for Project Gutenberg. If data’s on old tape or 12″ CP/M floppy disks …

  26. 26
    Andrew Sipocz says:

    I think the folks working on this question (changes in tropical storm frequency and intensity) are not seeing the forest for the trees.

    Most TS losses occur from the storm surge, the water pushed over the land, or inland flooding caused by huge amounts of rainfall often generated by lesser tropical storms hung up over coastlines. Neither of these appear to be particulary linked to storm intensity or storm frequency.

    Jennifer Irish (one of Andrew Dessler’s collegues though a similar entry onto his blog generated no discussion) at Texas A&M University recently gave a talk on analyses that showed a weak correlation between storm intensity and surge height or area. For example, Katrina had the highest recorded surge height but was barely a category 3 when it hit, versus both Camille (a 5) and Andrew (a 5). Other category 3 storms have had a third the surge height as Katrina. Rather she found storm size was a more important factor. Katrina, Wilma, Rita, all three were large storms that owed (according to what came out of the tropical storm prediction center) their rapid increase in size to high sea surface temps (Gulf loop currents).

    It seems to me that determining the effect AGW has on the generation of large storms (versus total number or intensity) would be a more fruitful line of research.

    Also, I’d think modeling storm size would be easier than storm intensity for the same reasons predicting average global temperature is easier to predict than next week’s weather. Storm intensity appears to be subject to rapid changes caused by chaotic preturbations.

  27. 27
    Hank Roberts says:

    > tornado
    I’ve seen it said these can’t be used to conclude much about warming, but — we know the stratosphere is cooling; we know the heat engine works; we know tornado-forming conditions reach high enough to divide the jet stream — isn’t that going to increase the temperature difference? Or is the stratosphere not warming at the level of the jet stream yet? Mostly curious whether hurricanes also poke up enough to split the jet stream wind.

    “… nobody imagined that the Lowden tornado would be the first of thirty-nine ….
    At 2pm, the storm experts at SELS noticed that the jet stream was splitting into two branches, like a river flowing around an island. That island was located above and along the Illinois/Wisconsin border. ….…/wxpics/tornado/Huffman.gif…/palmsunday1965/index.php

  28. 28
    Stephen Berg says:

    This is probably one of the best threads ever on a science blog. It is very encouraging, educational, and enlightening for non-experts to have esteemed scientists like Drs. Curry, Kossin, and Trenberth discussing their observations and hypotheses about this important topic for all to see, rather than simply leaving it to a conference setting where scientists are the only ones present.

    Congratulations all for your efforts and thank you for sharing this with the rest of us!

  29. 29
    Steve Sadlov says:

    RE: #9 – As a confessed East Pacific TC geek, I was also quite shocked by Dr. Curry’s assertion regarding basin similarity. Another interesting thing is the exceedingly low occurrence of recurvature for storms that do not transition to typhoons and continue into WPAC. Again, sheer, stratus and cold water are key factors. And by way of comparison, we really do have cold water out here in a way that you will not encounter in say, Iberia or NW Africa litorals. The California Current is quite strong, and, our upwelling is fierce, clear down to Los Cabos. You have to pass Los Cabos into the Sea of Cortez to get into the truly warm water, or of course, go further south, to where the current truly leaves the shore.

  30. 30
    Dan Allan says:

    It seems to me that it is a mistake to overemphasize the metrics regarding hurricane intensity in the past 30 years.

    It was pointed out previously at RC that the case for AGW does not rest with the hockey-stick, but with the theoretical argument that (a) CO2 is increasing due to burning of fossil fuels, and (b) more CO2 in atmosphere means warming temperatures.

    Moreover, as each hurricane season is relatively short, it is especially susceptible to the caprices of “weather”, i.e., a season may be strong or weak due to a particular weather pattern that is just a few weeks long and that does not reflect any long-term trend whatever. Moreover, the number of data points (seasons) is still relatively small. And, as increase in ocean temperature lags increase in atmospheric temp., even if no AGW signal is yet visible in the hurricane data, this may not mean very much in terms of the future impact of AGW on hurricane intensity.

    So are there good theoretical reasons to expect stronger hurricanes due to higher SSTs? I would certainly think that there are, given that high SSTs are routinely used (by AGW-denier Bill Gray, et al) to predict the intensity of individual seasons. But maybe there are other factors I am not aware of, as a layperson.

  31. 31
    Lynn Vincentnathan says:

    RE “The IPCC says there is insufficient evidence to determine whether there is any trend in tornadoes.” That may be from a scientific view. From a concerned-person-view, I tend to consider nearly every weird weather event (logically expected as something GW might cause) as impacted by GW, until proven otherwise. Back in 1992 during Hurricane Andrew, I thought, yep, global warming….when will they ever learn, when will they every learn. And those tornadoes this week…

    We live in a globally warming world. The burden of proof should be on the skeptics to prove GW is NOT impacting such&such. I know this is not science or nonsense to scientists, but it makes sense to me.

    So if anyone has proof at, say, .05, that GW is not presently increasing storm intensity, I might even consider believing them. (And that might be hard to do, since we might be in a trend right now that doesn’t show up clearly in the data for another 10 or 20 years. So in 2039, the definitive study will come out: “We’ve been in a whopper of a GW-caused increased storm trend for four decades, but, of course, back in 2007 they couldn’t have known that.)

    In any case, I won’t stop reducing my GHGs.

  32. 32

    The cold stored in polar ice is small

    I wasn’t aware that cold was ‘stored’. Anyways, clearly I am referring to the ability of meltwater to interfere with ocean currents, in their own special way, which is clearly yet to be determined. The ice sheets and the oceans are the sleeping giants here, and how their interactions play out as the ice sheets melt and the oceans overturn will determine the course of evolution of the warming which we will inevitably experience. We have just embarked on the great see saw ride in the last 30 years or so, clearly we have a lot to learn. In addition to an international polar project, we need an international ocean project, the continents already clearly being overrun with humans.

  33. 33
    Ray Ladbury says:

    Re: 31. Lynn, your post reminds me of the quip that stated that economists had managed to predict 10 of the last 4 recessions.
    Ultimately, the thing is that we cannot know with certainty whether any single weather event was influenced by climate change. And the variability of weather is such that you could find “evidence” for any trend you wanted if you looked at a short enough sequence.
    Ultimately, climate change poses risks–that is a certain probability of a certain adverse event. We know how to deal with risks–you allocate effort (read $$) in proportion to the risk. So we should already be shelling out a significant wad of cash to try to mitigate risk, and as our certainty grows, so should the funds allocated. The thing is that the public’s perception of risk is distorted, and they tend to over-react to catastrophic but improbable risks that are imminent while discounting very real risks in the “distant” (read >5 years) future. It is probably our evolutionary heritage that makes us look for the rare poisonous snake while getting lost on the trail and starving to death.

  34. 34
    Dan Vimont says:

    Hello all,

    Thanks for everyone’s interest in our work. As a co-author, I am somewhat disturbed by overall response of the media, and to a certain extent of some of the scientific community to our work. In this respect, I agree, to a certain extent, with Gavin’s assessment of our work (of course, I think it’s a little more convincing that Gavin suggests, but I’m a co-author, what do you expect?).

    An essential component of scientific method is the continual testing and verification of hypotheses put forth within our respective communities. This ensures that the result of a given test is not contaminated by the particular circumstances (in our case, a particular data set) under which that test is performed. Hence, we conduct the same tests as Emanuel (2005) and Webster (2005), only using a data set that is more appropriate for analyzing climatic signals in hurricane activity. We find that a major result of these papers – that hurricane activity is increasing in most ocean basins – is not reproducible, due probably to the quality of the data that was originally used to establish this claim. This result is independent of the fact that we’re looking at a short time period – the two data sets clearly diverge. Interestingly, we find very good agreement in the Atlantic and East Pacific, as explained by Kossin, above.

    Are our results a definitive statement that global warming is “honing in on the Atlantic” (as depicted in the media)? Of course not! We’re looking at trends in a 23 year time series – we’re not delusional. Do the results shoot down the hypothesis that hurricane intensity may increase in the future? Again, of course not. BUT, the results do cast serious doubt on the quality of data used to establish this claim (the “best track” data). Hence, I feel our work is an important contribution to the ever-evolving and growing body of evidence that ultimately either disproves a scientific hypothesis, or contributes to its acceptance as a theory. We are very careful to cast our results in that light.

    On a contrary note, I worry that Gavin may have missed a key point of our research when stating (in the introduction above) that our technique “doesn’t work as well in the other basins”. The implication is that the method should not be trusted outside the Atlantic and East Pacific basins due to a discrepancy between the best track data and the UW / NCDC analysis. Another, more likely interpretation is that the inhomogeneities in the methods by which the “best track” data are constructed lead to serious errors in long term climatic signals in that data set in the first place. Rather than casting doubt on the UW/ NCDC results, the results indicate that we should seriously question the validity of climatic signals in the best track data outside of the Atlantic or Eastern Pacific. Thus, our paper sheds a lot of light on previous claims of a strong global warming / hurricane link that use the best track data. And, it highlights a very interesting scientific question – why is it that hurricane activity is increasing in the Atlantic but nowhere else, even though SST is increasing everywhere?

    As a parting note, I’ll again thank everyone for their interest, and for the ongoing debate. Regardless of our particular qualms with specific research results, we’re all in this for one purpose – to make sure we get this right!

    - Dan Vimont

    [Response: Dan, thanks for the comment. This is a little outside my field and so I was not trying to pass judgment on your study. In general though, I've learnt that where there are discrepencies, there are often interesting problems to be looked into (on all sides), and that it's best not to prejudge what future research will find. I recall the 'satellite/surface' discrepancy or the ice age tropical temperature issue - much was said that proved rather premature. I'm trying and avoid that! -gavin]

  35. 35
    Tim Jones says:

    Speaking of papers…
    “…the Senate Republican Policy Committee that (Sen. Kay Bailey Hutchison) chairs issued a 10-page document titled, “Global Warming: The Settled Versus the Unsettled Science.” Beyond three general areas of scientific agreement, the Republican document explains “there is considerable uncertainty” about man-made climate change.”

    “Global Warming: The Settled Versus the Unsettled Science.”
    “Trends in Hurricane Numbers and Intensity”

    “The IPCC Summary notes that there is evidence for increased hurricane intensity in the North Atlantic, which is correlated with higher sea surface temperatures. It adds this substantial cautionary note, however: that “multi-decadal variability and the quality of the tropical cyclone records prior to routine satellite observations in about 1970 complicate the detection of long-term trends in tropical cyclone activity.” It also states that even though there is some evidence of increased intensity, “There is no clear trend in
    the annual numbers of tropical cyclones.”28

    In November 2006, the World Meteorological Organization (WMO) held a two week International Workshop on Tropical Cyclones (or hurricanes) where leading hurricane researchers wrote a Statement on Tropical Cyclones and Climate Change. The Statement says that recent research has found “evidence for a substantial increase in the power of tropical cyclones in the West Pacific and Atlantic basins during the last 50 years.” It notes, however, that some scientists argue that the trend may not be real, and instead is due to improved monitoring of strong tropical cyclones in recent years.

    With regard to changes in hurricane activity, the statement says that the Atlantic basin has the most reliable historical hurricane records. The causes of multi-decadal variability in major hurricane activity in this basin is still being debated. Some scientists argue that Atlantic hurricane activity fluctuates naturally. The Statement notes one study that disputes this, arguing that variability is due to a combination of changes in solar activity, volcanic and manmade aerosols, and greenhouse gases.

    Does Human Activity Contribute to Changes in Hurricane Intensity?

    The IPCC Summary assesses the likelihood of increases in intense hurricanes as “likely [better than 66 percent chance] in some regions since 1970,” and states that the likelihood of a human contribution is “more likely than not,” though in the footnote it also notes that the “magnitude of anthropogenic [human] contributions [was] not assessed.”

    The WMO Statement notes that, “Though there is evidence both for and against the existence of a detectable anthropogenic signal [i.e., human component] in the tropical cyclone climate record to date, no firm conclusion can be made on this point.” Part of the difficulty in identifying the human component, if any is, as the IPCC Summary itself notes, that tropical cyclone activity varies naturally over the decades. “This variability makes detecting any long-term trends in tropical cyclone activity difficult,” according to the WMO statement. The statement also notes that, “The recent increase in societal impact from tropical cyclones has largely been caused by rising concentrations of population and infrastructure in coastal regions.”

    Finally, the Statement notes that, “The scientific debate…is not as to whether global warming can cause a trend in tropical cyclone intensities. The more relevant question is how large a change: a relatively small one several decades into the future or large changes occurring today? Currently, published theory and numerical modeling results suggests the former.”29


    Climate change continues to be a controversial scientific issue. The IPCC process is a major effort to provide a comprehensive overview of the state of climate science, though as is evident even from the Summary – which, due to the necessity for brevity, tends to downplay uncertainties – there are still many difficult obstacles and uncertainties to overcome in determining the influence of human activities on the climate.”

    There are three areas upon which there is general agreement. The average temperature has risen, atmospheric concentrations of greenhouse gas have risen, and the latter has probably contributed to the former. The important question is whether changes will be relatively benign or catastrophic. This is still an open question. It would be prudent for policymakers to continue to closely monitor the science.”

    This is it? All we have to do is monitor the science?

    I’m so relieved. Oink.

  36. 36
    Charles Muller says:

    #34 And, it highlights a very interesting scientific question – why is it that hurricane activity is increasing in the Atlantic but nowhere else, even though SST is increasing everywhere?

    Please to read you, Dan. Do you have an hypothesis about your own question?

  37. 37
    Vince says:

    What gets me is how alarmists use the hurricane issue (like “scientists say Katrina was caused by GW,” or “there’s increased trend in the Atlantic”) to imply that GW has been proved. Increased hurricanes are only ONE possible effect of GW, and even if (for the sake of argument) they were found even in the future to be linked to warming, that would in no way prove GW.

    We reg folks only have time for news headlines (not sci articles), so the man-in-the-street reads, “Katrina not caused by GW,” & thinks “GW has been disproved & I can rest easy,” while his wife reads in the Drs office, “GW increasing Hurricane intensity” (a reporter read “warming oceans increasing hurricane intensity” and got confused, and I don’t fault him for it, bec he/she’s very much like that person-in-the-street). And the couple is so busy fighting about GW that they fail to buy those CF or LED bulbs, or even turn off lights not in use.

    I’m trying to think like an alarmist: since GW is happening, then all its supposed effects have to be shown as well, or not due to other factors. These absent effects are serious problems for them. It’s tiring just to watch alarmists jump through all their inelegant hoops and jump at very thin threads (what next, after polar ice caps aren’t melting?).

    OTOH, for GW skeptics, if there were evidence for increased hurricanes it would be a total NON-PROBLEM. Since there are 2 types of energy, kinetic and heat, we might expect that some of the increasing solar warming heat would translate into increased hurricanes & storms…or maybe it’ll be doled out in bits and pieces that no one hardly notices…or maybe it’ll just mainly stay in a heat form and not get kinetic on us. Weather is a lot more chaotic than climate, and there are many many factors impacting it, including Sally in Cincinnati singing, “Rain, rain go away.”

    We have no trouble at all understanding how internal COMBUSTION engines in our cars translate into driving over to the beach, so this heat-to-motion (solar warming to storms) should be a no-brainer. Eventually, if not right now, we would expect to see increase in storm intensity and perhaps frequency in a globally warming world, all things being equal…which they are not, since even weathermen can’t well predict next week’s weather due to some butterfly flapping it’s wings in Japan gumming up the wind system.

    The important point here is that solar warming is going to entail many types of serious harms, one of which will likely be increased storms…if not now, then most likley later. And we reduce our GHGs like maniacs now, not that we think it’ll reduce Africa’s current drought, but that it will help in the future.

    An important point in the article, I felt, was: “the very real and dangerous increases in recent Atlantic hurricane activity will no doubt continue to provide a heightened sense of purpose to research addressing how hurricane behavior might change in our changing climate…” so “give us more money!!!”.

  38. 38
    Mark A. York says:

    RE#36 I do. And it comes from Emanuel I believe, which is to say the Pacific and Indian Oceans are already warmer, thus this is an opening in the natural system that needs to catch up given the rising global mean temperature. If my nonclimate background is off, I’m confident someone with more expertise will adjust this accordingly.

  39. 39
    Ike Solem says:

    Given that this issue is still generating much disagreement, are that aspects that everyone can agree on? In particular, would anyone question that the observed trends in sea surface temperature in all basins can be attributed to anthropogenic global warming? Hurricanes are not the best place to look for a global warming signal, but SSTs and surface ocean heat content are.
    (Ref: )

    It seems that the hurricane issue is of greater interest than SSTs for the reasons stated in the comments. However, just looking at the isolated case of Katrina, if Katrina hadn’t encountered a deep warm pool of water in the Gulf of Mexico, it wouldn’t have intensified over the Gulf (historically rare behavior for hurricanes to do that). This brings in another variable, the depth of the warm layer, as well as the SST issue. In the absence of counteracting factors, it’s hard to argue that hurricanes won’t have a higher probability of intensifying along the southeastern United States. One could also argue that there is a chance that ‘permanent El Nino conditions’ could shift wind patterns in the Atlantic basin making in less probable that hurricanes will develop vertical structure – but that’s a bit too convenient to be reassuring. It is the kind of argument that contrarians will sieze upon, however.

    The practical issue is what matters to people who live in these areas, however – and I think most climate scientists would recommend reinforcing levees and not building in flood plains, and it seems that the insurance industry is taking the same approach. Whether or not a trend in hurricane intensity has been detected is still apparently open to question – but the trend in SSTs seems unquestionable.

  40. 40
    Pascal says:

    Dan Vimont

    Hi Dan

    when I look at fig.3 “percent of 2sigma hurricanes” I find an increasing tendancy of 1.8 per cent by decade.

    What do you think of it?

  41. 41
    wahoo44 says:

    As a member of the non-scientific “skeptic” camp (but not a denialist), I find this particular topic and most other topics introduced on this site to be extremely valuable in forming my thoughts on the entire GW/human impact debate. Thanks for providing this resource (I even can use it at times to help with my insomnia while trying to follow some of the more technical aspects of the comments..again, a great public service).

    All joking aside, this thread highlights a trend I see in much of the RC blogs……drifting over into commentary/politics/hype versus sticking to science. Why is Katrina used as an example of hurricane intensity, when it was documented as only a Category 3 hurricane (albeit with a large storm surge)? If you take New Orleans out of the equation, Katrina would only be a footnote other than to those directly affected along the coastline of MS/LA. The larger question is the extreme risk (both in human and economic terms) that the exponential growth of the human footprint along the US Southern Coastline has created. Nobody cares if a Category 5 hurricane hits undeveloped doesn’t get ratings! What research is being done on the growing human risk along the various coastlines of the world, both from Hurricanes and (God forbid) rising sea-levels?

    Also, one final dig. Unless Global Warming has only been a problem since January 2001 (date of a certain inauguration), please keep the political commentary out of postings…it’s beneath this fine effort.

  42. 42
    Ray Ladbury says:

    Re 38. Wahoo, Katrina is significant because prior to landfall it was a really big storm. It was a very large storm and it had (at the time) the 4th deepest low pressure ever measured. Its near coincidence with Wilma and Rita, which were even deeper raised eyebrows. So for climate folks, yes, we do care about a Cat 5 in an uninhabited area, because it is another piece in a trend.
    As to the political sniping, I agree that it is unprofitable. However, I think you have to understand the level of frustration in the scientific community and in the climate community in particular with the tendency in the current administration to “manage information”. To some extent, there is an inevitable conflict between politicians, who thrive on being able to manage and spin information, and scientists, who require free exchange of information to do their jobs. However, by any measure, this administration has been heavy handed in “editing” scietific reports. While I think that this tendency goes all the way to the top, I think the worst offenders have been ham-handed political appointees, who have made things much worse than they had to be. However, I will take your admonition to heart and make an effort to avoid political bashing. The debate is not between liberal and conservative, but between science and anti-science.

  43. 43
    Ray Ladbury says:

    Re: 37. Vince, if you wish to “think like an alarmist” might I suggest looking at the evidence. The question of whether climate change is influencing hurricane strength is still open and debated within the scientific community. The question of whether humans are changing climate is settled. Do not confuse the two issues.

  44. 44
    Charles Muller says:

    #39 A better link would be the Santer et al. 2006 PNAS study on XXth century tropical SST trends, as simulated in forced and unforced situations (thereafter, full text, pdf):

  45. 45
    Lynn Vincentnathan says:

    Dear RC folks, you’ve got to do an article on the recent Senate (Repubican) document, “Global Warming: The Settled Versus the Unsettled Science” at:

    What is your take on it?

  46. 46
    Lynn Vincentnathan says:

    #33, thanks, Ray. I get your point. I think I was trying to argue the prudence perspective r/t risk. So, while “we cannot know with certainty whether any single weather event was influenced by climate change,” we cannot know if it wasn’t. And prudence (one of those Godly virtues, like mortification) would require us to mitigate GW, even without high certainty that it is happening, or that it is impacting severe weather events. Of course, in prudent ways, like all the measures that do not harm us (some even helping us).

    And policy-makers, you’d think, would not be able to argue against that position…unless they are solely focused on helping those few fossil fuel companies that refuse to diversity into alternative energy. And even those companies could benefit by becoming more energy/resource efficient/conservative.

  47. 47
    Kevin Trenberth says:

    Some of my comments were posted perhaps a bit out of context. I welcome the response from Jim Kossin and I wish to emphasize that I was not dismissing their results at all. On the contrary I welcome them as a useful, but incomplete, contribution to the science of tropical cyclones and climate change. The development of the new dataset is a very good first step also but much more can and must be done. It takes resources and commitments from funding agencies, and I have been working to that end for some time (in paricular through my role as Chair of the WCRP Observations and Assimilation Panel) by calling for extra efforts in reprocessing satellite observations and reanalyses.

    Jim Kossin may be right, in part, that my criticism of their work applies also to the Dvorak technique, but it is not at all obvious to me that the latter applied to the new data set would give the same results as in their recent study. It is very bothersome that there is no convergence of results over time, and maybe this calls current practices into question, but those aspects should be resolved.

    Some of the big picture things are being missed here. Basically with global warming there is an expectation of increases in tropical storm activity. Please note that the IPCC SPM statement very carefully focussed on activity and not just intensity. Such activity may extend to individual thunderstorms, but it is certainly expected also for tropical cyclones owing to their relationship with the ocean. Only the organized TCs take significant heat out of the ocean ameliorating global warming of the oceans. Then the increased activity can be manifested in several ways. These could be increases in numbers, intensity, size, and duration. We have some data on numbers and duration but not the other quantities. Currently intensity is a spot value and not integrated over the size and duration of the storm. Size is not kept as part of best track data. There are published listings of many possible metrics that can be retrieved and kept and the need is to go back and reprocess the data to recover a much more comprehensive set.

    The physical expectation is that TCs are apt to become more intense and bigger with warming. But because one big storm takes much more heat out of the ocean than two smaller storms, the expectation is that numbers may decline. Duration could vary: a more intense storm may peter out more quickly by having used up the available energy in the immediate vicinity (from convergence in the atmosphere as air spirals in, to surface moisture from evaporation in the strong winds) unless it moves into a new environment. The track of the storms seems to be largely a crap shoot depending mostly on the weather situation at the time. I know of no arguments about landfalling vs nonlandfalling storm trends that make sense.

    It seems that many agree that the environment in each TC basin is different. The tropopause differs substantially (relates directly to potential intensity) and the sea surface temperature patterns and land configurations vary greatly. The sooner we can come to grips with the issues and set about compiling the requisite datasets and analyzing them, the better off we will be. The Kossin et al studies are good first steps, but just first steps.

    Kevin Trenberth

    [Response: Kevin is right to point out that we may not have provided the full context for interpreting the various quotes that we provided, including those attributed to him. That's our fault, and we're thankful to him for taking the extra time in his busy schedule to come in and provide a fuller context for the remarks. -mike]

  48. 48
    Rod B. says:

    Ray says in #19 (as have others) …”How many times have we heard denialists cite the mild Atlantic hurricane season in 2006 as “proof” that we’re all just a bunch of alarmists? This is just turnabout for the irresponsible chatter that blamed Katrina and every other hurricane on climate change in 2005…”

    As a skeptic (I’m certain the Holocaust took place) you’re absolutely right, the nil hurricane season of 2006 does not disprove AGW. But it’s logical many would claim that in response to all the cheering for AGW proof from the 2005 hurricanes, which seemed like a solid good point at the time but now has to be called irresponsible as part of the back-peddling.

    [Response: 2006 was an average year, not 'nil', and if you look back at the pages of RealClimate on the issue you will not see any 'cheering'. Indeed, you will see plenty of statements pointing out the fallacy of taking one event or season as proof of anything. It is easy (and a little lazy) to only rail against the stupidest or most extreme position on the 'other side' of an issue. Serious discussion only occurs when people address the more considered statements. - gavin]

    [Response: In fact, 2006 was slightly above average w/ 10 total named storms, and substantially above average for an El Nino year (for which the typical number of named storms is closer to 7). Part of the explanation is that SSTs over the main development region were anomalously warm, favoring development even in the face of an unfavorable shear environment (associated with the incipient El Nino during Fall '06). - mike]

  49. 49
    Ike Solem says:

    It’s worth noting how the scientific discussion of the hurricane issue has differed from the media discussion. Right during and after the 2005 hurricane season the most common media explanation was that hurricane activity had a 20- to 30- year cyclical behavior that was dependent on the North Atlantic Oscillation. (see C. Landsea on this at ) The data to support that theory thus seems scantier than the data that supports an increasing trend in hurricane intensity, yet the NAO lnk was widely repeated on CNN withou much questioning. Similarly, there was an immediate media rush to blame this winter’s anomalous record warmth on a mild El Nino event, though that explanation is surely only partial… and there was also record warmth in Moscow. It would be nice to see reporters covering the scientific discussion more carefully, in other words.

    These two topics are related in that the current ENSO forecast ( ) shows a higher chance of La Nina conditions this fall. Note to reporters: a scientist’s willingness to make predictions of the future is an indication of the current level of understanding of the science; for example Hansen et al predicted that Pinatubo’s eruption in 1991 would produce a significant aerosol cooling effect, and they were right; but would anyone be willing to predict that La Nina (assuming conditions set in) will result in a record hurricane season this fall?

  50. 50
    Jim Kossin says:

    Re: Gavin’s reply to #34.

    Gavin, I appreciate most of what you’re saying and as I mentioned in my earlier comments I certainly don’t consider our findings to be the last word, and I encourage others to expand on our work. But you are in fact passing judgment by saying:

    “While the technique works well in the North Atlantic (picking up almost all of the storms seen in the standard data), it doesn’t work as well in other basins – possibly because the characteristics of tropical cyclones are not universal, or because the coarse early remote sensing data are still not sufficient.”

    This statement is not actually justified. First, the technique works exceedingly well in the Eastern Pacific as clearly documented in our paper (also see my comment #9 above). Second, as Dan Vimont correctly points out, it is the fact that the agreement is not as good in the other basins that is the key finding here, it’s not a suggestion that our results are somehow questionable. There is no reason to believe that our method would not capture the salient features in basins besides the Atlantic, and this can be confirmed by any hurricane specialist including those who operationally apply the Dvorak technique and/or participate in constructing the best track. An alternate hypothesis for the basin differences is offered in my comment #9.

    It seems that most of the people who have commented formally here on our paper are not actually in a particularly good position to comment on whether our algorithm, or any IR-based method such as the Dvorak technique itself, is applicable in all ocean basins. While you and the others who have commented (somewhat negatively) on the validity of our results are certainly experts in your fields, I doubt that any of this group would claim any expertise in the procedures and applications of the Dvorak technique or the operational protocols of the forecast offices that construct the best track data that most of the recent studies have been based on. Yet, this is where the all of the main objections seem to be coming from. This is a clear disconnect that in all fairness should be considered before stating that our algorithm “doesn’t work well in other basins”. Unfortunately, these words are the first thing people see when they visit your blog and they are in fact not well-founded.

    Could our results be proven wrong by subsequent studies? Absolutely. Am I sure that they are giving us the right answer? No. Are there warts on our algorithm? Yes. Are there warts on the algorithms used to estimate the intensities that form the best track records. Yes. We’re all estimating intensity from IR imagery. Not exactly uber precise. Would hurricane specialists question the validity of our algorithm outside of the Atlantic? No. Would climate scientists question the validity of our algorithm outside of the Atlantic? Apparently yes (said with tongue planted firmly in cheek).
    –Jim Kossin

Switch to our mobile site