Summaries of Recent Storm Research

Science tightens the link between hurricanes and climate change

Posted: 08-Nov-2006; Updated: 28-Dec-2006

Climate change factored in the brutal 2005 hurricane season, according to a recent research paper. (Photo: Courtesy NOAA)

Global warming was a major factor in the record-breaking 2005 hurricane season, according to a paper in the June 2006 issue of Geophysical Research Letters. From a policy perspective, this study by Kevin Trenberth and Dennis Shea (of the National Center for Atmospheric Research) is essential because it says "yes" to the question "Is global warming causing more destructive hurricanes?"

Two other research threads have addressed similarly important scientific questions. First, how much do various environmental factors contribute to hurricane intensity? And second, when did the effect of global warming on hurricanes first appear? The answer to the first question is that sea surface temperature, which is increasing because of global warming, plays a huge role in hurricane intensity. And while scientists continue to uncover the details, all the evidence so far indicates that global warming started intensifying hurricanes about 30 years ago.

Here we provide summaries of (and links to) the main papers that have contributed to this lively and essential discussion, in the order in which they were published. We will update this site as new research appears.

The paper summaries

Emanuel, K. 2005. Increasing destructiveness of tropical cyclones over the past 30 years. Nature 436: 686-688.
Abstract | Full text [PDF]

Contribution: Presents data showing that the intensity (or destructive potential) of hurricanes has doubled over the past 30 years and is linked to rising sea surface temperatures (caused by global warming).

Details: Kerry Emanuel (from the Massachusetts Institute of Technology) examined 55 years of data from the North Atlantic and North Pacific and found a correlation between sea surface temperatures and the destructive potential of hurricanes. His analysis showed that the destructive potential of hurricanes—defined by a storm's wind speed and duration—has approximately doubled over the past 30 years. In both ocean regions, there is a close relationship between water temperatures and hurricane strength. In other words, when sea surface temperatures were cooler, hurricanes had less destructive potential; when sea surface temperatures were warmer, hurricanes had greater destructive potential. Starting in about 1975, sea surface temperatures in the North Atlantic and North Pacific began to increase dramatically, and the destructive potential of hurricanes followed suit.

Webster, PJ, GJ Holland, JA Curry & H-R Chang. 2005. Changes in tropical cyclone number, duration, and intensity in a warming environment. Science 309: 1844-1846.
Abstract | Full text [PDF]

Contribution: Expands analyses to include all of the world's hurricane basins; their results strengthen the link between global warming and hurricane intensity.

Details: Peter Webster and his colleagues (from the Georgia Institute of Technology) assembled a 35-year record of hurricane intensities and sea surface temperatures in all six hurricane regions of the world (North Atlantic, the western North Pacific, the eastern North Pacific, the South Indian, the North Indian, and the Southwest Pacific). Their analysis showed that, worldwide, the total number of hurricanes per year has not changed. However, there has been a large global increase in the percentage of storms that reached Category 4 or 5. In the early 1970s, 20% of the world’s hurricanes reached Category 4 or 5, but by the 1990s, about 35% of all storms reached these levels. This increase in storm intensity accompanied a steady global rise in sea surface temperatures.

It has been argued that the only reason the U.S. has seen a recent spate of really bad hurricanes is that the North Atlantic is reaching the peak of a normal regional cycle of hurricane intensity. However, this study shows a worldwide increase in storm intensity; the number and percentage of Category 4 and 5 hurricanes has increased in all of the ocean basins around the world. (In fact, of all these regions, the North Atlantic had the smallest increase in intense hurricanes.) This suggests that the trend is global in nature and not part of a multi-decadal natural cycle.

Hoyos, CD, PA Agudelo, PJ Webster & JA Curry. 2006. Deconvolution of the factors contributing to the increase in global hurricane intensity. Science 312: 94-97.
Abstract | Full text [PDF]

Contribution: Brings rigorous statistical analysis to global datasets; their results indicate that rising sea surface temperature is the only environmental factor that explains rising hurricane intensities around the globe.

Details: The strong relationship between hurricane intensity and sea surface temperature suggests a link between global warming and hurricanes. But is sea surface temperature really the cause, or is the relationship a coincidence? Carlos Hoyos and colleagues (at the Georgia Institute of Technology) compiled 35 years of detailed tropical storm and hurricane data from all six hurricane regions of the world to determine the statistical linkage between the occurrence of the most intense hurricanes (Category 4 or 5) and the relevant environmental factors that are thought to influence hurricane intensity; i.e., sea surface temperature, humidity, wind shear, and zonal stretching deformation. Their results showed that sea surface temperature is the only factor that can explain the global increase in the number of Category 4 and 5 hurricanes over the past 35 years. The other factors contributed to short-term (but not long-term) and regional (but not global) patterns of hurricane intensity. Since increasing sea surface temperatures are driven by global warming, this study adds even more evidence that global warming is causing stronger hurricanes.

Michaels, PJ, PC Knappenberger & RE Davis. 2006. Sea-surface temperatures and tropical cyclones in the Atlantic basin. Geophysical Research Letters 33: L09708.
Abstract | Full text [PDF]

Contribution: Despite statements to the contrary, their study of hurricanes in the Atlantic confirms earlier results—rising sea surface temperatures cause more intense hurricanes.

Details: Patrick Michaels and his colleagues (at the University of Virginia) focused on the Atlantic basin and used 24 years of data to examine the complex relationship between sea surface temperatures and hurricane wind speeds. They found that warmer waters had hurricanes with greater maximum wind speeds, although the maximum winds speeds did not seem to increase any further at the highest temperatures observed. They concluded that "rising sea surface temperature will act to increase the percentage of major hurricanes but not the ultimate intensity of these storms." Their analyses also suggested that there may be a temperature tipping point for major hurricane development—that is, water temperatures must be above 83 degrees Fahrenheit for hurricanes to jump to Category 3 or higher. If true, this result adds further evidence that global warming is likely to affect hurricane intensity.

Although the authors state that their results "weaken the notion of a simple cause-and-effect relationship between rising sea surface temperatures and stronger Atlantic hurricanes," their study in fact concludes that there is a direct linkage between the two and, thus, that global warming acts to increase the intensity of future hurricanes. Furthermore, while Michaels et al. argue that the impact on hurricane intensity from the current amount of global warming should be "too small to reliably measure," their approach, based on an analysis of individual storms (instead of annual averages), is not well-suited to identifying long-term trends.

Klotzbach, PJ. 2006. Trends in global tropical cyclone activity over the past twenty years (1986-2005). Geophysical Research Letters 33: L10805.
Abstract | Full text [PDF]

Contribution: Despite statements to the contrary, results are largely consistent with those of Webster et al.—the number and percentage of the most intense storms have increased.

Details: Phillip Klotzbach (Colorado State University) revisited the question of whether hurricanes are becoming more intense. He restricted his study to a 20-year period—divided into two 10-year blocks (1986-1995 and 1996-2005)—and then compared hurricane activity between blocks. Even with these limitations, his result—that the total number of Category 4 and 5 hurricanes increased 10% between time periods—is not inconsistent with the results of previous studies (e.g., Webster, et al.). When Webster et al.’s more detailed data are grouped in the same way (i.e., lumped into two 10-year blocks), the resulting increase in hurricane intensity is the same as in Klotzbach’s study. The percent increase is smaller using the Klotzbach approach because the coarse 10-year resolution obscures the large increase in hurricane intensity that occurred in the late 1980s and early 1990s. This and the fact that Klotzbach chose to omit 15 years of data (from 1970-1985) obscure the size of the long-term trend. Klotzbach’s other result, that there has been no change in the total number of storms, is consistent with earlier studies.

Sriver, R & M Huber. 2006. Low frequency variability in globally integrated tropical cyclone power dissipation. Geophysical Research Letters 33: L11705, doi:10.1029/2006GL026167.
Abstract | Article (subscription required)

Contribution: Using a different dataset to examine global hurricane intensity, the researchers confirm earlier findings that storms are stronger.

Details: Ryan Sriver and Matthew Huber of Purdue University examined over 40 years of storm wind speed data. (This information came from the European Centre for Medium-Range Weather Forecasts Reanalysis Project, not the Joint Typhoon Warning Center and the National Hurricane Center, which Kerry Emanuel, Peter Webster and colleagues used for their earlier studies.) Sriver and Huber used the wind speeds to calculate the power dissipation (strength) of storms in all of the world’s hurricane basins (not just the North Atlantic and North Pacific as in Emanuel’s study). Even though the data for the studies came from different sources, the results were essentially the same: hurricane intensity has increased over the past few decades, globally as well as in the Atlantic and Pacific basins. In fact, Sriver and Huber's work suggests that Emanuel may have underestimated the size of the increase.

Mann, ME & KA Emanuel. 2006. Atlantic hurricane trends linked to climate change. EOS 87 (24): 233-244.
Full text [PDF]

Contribution: Finds that global warming and aerosol pollution, not natural regional cycles, have been responsible for the long-term sea surface temperature variations in the Atlantic hurricane region.

Details: Rising sea surface temperatures are increasing hurricane intensity. How much of this effect is due to global warming and how much of it is due to natural variability? Mann and Emanuel (from Penn State and MIT, respectively) tackled this question with a statistical analysis of sea surface temperature data from the past 130 years, focusing on the Atlantic hurricane region. They first compared sea surface temperatures in the Atlantic to the global average sea surface temperature and found that Atlantic temperatures closely followed the global trend. The match wasn’t perfect, however, meaning that global warming had most, but not all, of the influence on rising sea surface temperatures in the Atlantic. The remaining influence—particularly a slight Atlantic cooling between about 1950 and 1980—appeared initially to belong to the Atlantic Multi-decadal Oscillation (AMO), a long-term sea surface temperature cycle that has been blamed for recent increases in Atlantic hurricane intensity. However, when Mann and Emanuel added to their calculations the known cooling effect of aerosol pollution during the late 20th century, the effect of the AMO virtually disappeared. Mann and Emanuel concluded that "there is no evidence that a natural climate oscillation such as the AMO contributes to long-term tropical North Atlantic sea surface temperature variations."

Landsea, CW, BA Harper, K Hoarau & JA Knaff. 2006. Can we detect trends in extreme tropical cyclones? Science 313: 452-454.
Abstract | Full text [PDF]

Contribution: Questions the validity of the work of Emanuel and Webster et al. by arguing that the pre-1990 hurricane database is not sufficiently reliable to infer multi-decadal trends in hurricane intensity.

Details: In a Science Perspectives piece, Christopher Landsea and his co-authors reviewed the methodologies and protocols used to monitor and document hurricane characteristics over each of the major ocean basins since the early 1970’s. They argue that because of uncertainties in the methodologies used in the early part of the record as well as operational changes on the methods used, the hurricane database is not reliable enough to accurately discern a trend in hurricane intensity over this period. Regions primarily dependent upon satellite imagery for their hurricane data are likely, they argue, to have an artificial upward trend in hurricane intensity. The authors suggest that efforts to reanalyze the data from earlier measurements may remove some artifacts and make for more reliable trend analysis. (Read Environmental Defense commentary on this paper.)

Trenberth, KE & DJ Shea. 2006. Atlantic hurricanes and natural variability in 2005. Geophysical Research Letters 33: L12704.
Abstract | Full text [PDF]

Contribution: Shows that half of the unusual warmth in sea surface temperatures during the record 2005 Atlantic hurricane season was due to global warming.

Details: Kevin Trenberth and Dennis Shea (from the National Center for Atmospheric Research) rendered the debate on whether hurricane intensity has increased or not largely moot from a policy perspective by showing that global warming played a major role in fomenting the record-breaking hurricane season of 2005. They did this by identifying the importance of the various factors that could have contributed to the Atlantic’s extraordinary warmth in 2005, that in turn led to the worst Atlantic hurricane season on record. They took 130 years of sea surface temperature data (both North Atlantic and global average temperatures) and subtracted the global trend from the Atlantic temperature record to reveal the Atlantic Multi-decadal Oscillation (AMO). The amount of warming in 2005 that was due to the AMO alone was very small—10% or less—whereas global warming contributed approximately half of the extra heat. Trenberth and Shea also used previously described relationships between Pacific and Atlantic sea surface temperatures during El Niño years to determine that 25% of the warming in 2005 was due to heat associated with the 2004-2005 El Niño. (The remainder of the warming is associated with random year-to-year variability.) In their conclusion, the authors point out that global warming "is guaranteed to continue" and that, even though not every season will be as extreme as 2005, global warming "provides a new background level that increases the risk of future enhanced [hurricane] activity."

Elsner, JB. 2006. Evidence in support of the climate change-Atlantic hurricane hypothesis. Geophysical Research Letters 33: L16705.
Full text [PDF]

Contribution: Shows that global air temperatures predict Atlantic sea surface temperatures after a lag time of one to nine years.

Details: James Elsner, from Florida State University, applied a statistical causality analysis to a data set consisting of global mean air temperatures, Atlantic sea surface temperatures and hurricane positions and winds from 135 consecutive hurricane seasons from 1871-2005. He wanted to test two competing hypotheses:
Rising air temperatures due to global warming cause increased Atlantic sea surface temperatures, which in turn contribute to more intense hurricanes; or
Both global air temperatures and increasing hurricane intensity are caused by natural, cyclical fluctuations in Atlantic sea surface temperatures related to the Atlantic Multidecadal Oscillation.
Elsner’s analysis indicated that the first hypothesis held true. Specifically, he found that global warming-caused increases in global air temperatures have played a direct role in increasing Atlantic sea surface temperatures (with a lag time of one to nine years) and that increases in Atlantic sea surface temperatures play a direct role in increasing hurricane intensity.

Curry, JA, PJ Webster & GJ Holland. 2006. Mixing politics and science in testing the hypothesis that greenhouse warming is causing a global increase in hurricane intensity. BAMS (August 2006): 1025-1037.
Full text [PDF]

Contribution: Judith Curry and her colleagues at Georgia Tech and the National Center for Atmospheric Research review the current scientific debate about the link between hurricanes and global warming. They discuss the appropriate procedures for testing the complex "causal chain" linking global warming to hurricane intensity. They also highlight the valid scientific questions that have been raised concerning recent research on this topic. Finally, Curry et al. identify the criticisms that are scientifically and/or logically flawed and thus should be excluded from the on-going debate.

Kafatos, M, D Sun, R Gautam, Z Boybeyi, R Yang & G Cervone. 2006. Role of anomalous warm gulf waters in the intensification of Hurricane Katrina. Geophysical Research Letters 33: L1780.
Abstract | Article (subscription required)

Contribution: Finds that the intensity of Hurricane Katrina was correlated to the difference between sea surface and air temperatures; because of its short-term, small-scale focus, this paper is likely to be of more interest to hurricane forecasters than to climate scientists.

Details: Although hurricane forecasters are very good at predicting storm tracks, they are less skilled at predicting a hurricane’s intensity along that track. In this study, Menas Kafatos and his colleagues at George Mason University focused on Hurricane Katrina to better understand the factors that were involved in this storm's rapid intensification over the Gulf of Mexico. They compiled sea surface temperature, air temperature, wind speed and dew point data from a NASA satellite and NOAA buoys in the Gulf of Mexico. They found that Katrina reached peak intensity when the difference between sea surface temperatures and air temperatures was greatest. Their data also suggested that Katrina may have been particularly sensitive to sea surface temperatures in the northeast quadrant of the storm's track.

Santer, BD, TML Wigley, PJ Gleckler, C Bonfils, MF Wehner, K AchutaRao, TP Barnett, JS Boyle, W Brüggemann, M Fiorino, N Gillett, JE Hansen, PD Jones, SA Klein, GA Meehl, SCB Raper, RW Reynolds, KE Taylor, and WM Washington. 2006. Forced and unforced ocean temperature changes in Atlantic and Pacific tropical cyclogenesis regions. PNAS (Published online before print September 12, 2006. 10.1073/pnas.0602861103).
Abstract | Article (subscription required)

Contribution: Finds that human-caused global warming is primarily responsible for the recent increase in sea surface temperatures in the Atlantic and Pacific hurricane regions.

Details: Scientists know that hurricane intensity is closely related to sea surface temperatures: the warmer the water, the stronger the storm. Scientists also know that sea surface temperatures in the world’s hurricane basins (where storms first develop) have been increasing and hurricanes have been getting stronger. What is causing the seas to warm, and therefore the storms to intensify? A group of scientists led by Ben Santer at Lawrence Livermore National Laboratory used 22 climate models to determine what factors can explain the observed warming in the Atlantic and Pacific hurricane basins. They ran the models over the period 1906-2005. One set of simulations included just natural factors—essentially, what the climate would have been like had the Industrial Revolution never occurred. A second set of simulations included both natural factors and human-caused factors like increases in greenhouse gases. The models did a good job reproducing observed temperature changes only when they included increases in greenhouse gases. Santer and his colleagues calculated that there is an 84% chance that 67% of the recent ocean warming is due to rising levels of greenhouse gases.

Slott, JM, AB Murray, AD Ashton & TJ Crowley. 2006. Coastline responses to changing storm patterns. Geophysical Research Letters 33: L18404.
Abstract | Article [PDF] (subscription required)

Contribution: Jordan Slott and his colleagues at Duke University and Woods Hole Oceanographic Institute modeled how coastlines in the Carolinas could change if tropical storms become stronger. Their results indicate that even small increases in storm wind speeds are likely to have large impacts. Particularly along coastlines with complex shapes, some shoreline segments are likely to erode much more quickly than would be expected from sea level rise alone, while other areas could grow as sediments build up. The research reveals shortcomings in current coastal management plans that assume the effects of climate change—primarily sea level rise— will affect coasts uniformly.

Evan, AT, J Dunion, JA Foley, AK Heidinger & CS Velden. 2006. New evidence for a relationship between Atlantic tropical cyclone activity and African dust outbreaks. Geophysical Research Letters 33: L19813.
Abstract | Article [PDF] (subscription required)

Contribution: Reports that the presence of dry, dusty air over the tropical Atlantic tends to suppress Atlantic hurricane activity.

Details: The link between hurricanes and global warming has been the subject of several recent scientific studies, but scientists are also studying the effects of other environmental variables on storm development. One of these is the Saharan Air Layer (SAL), a warm, dry and dusty air mass that can form above northern Africa and travel across the tropical Atlantic Ocean. Hurricane researchers have hypothesized that the SAL could disrupt storm formation. To test this theory, Amato Evan and his colleagues at the University of Wisconsin-Madison used satellite data to look at the correlation between dust cover (a proxy for the SAL) and hurricane activity, which they defined as the number of days in which at least one named tropical storm was in a particular study region.
The research covered a southerly area of the Atlantic Ocean (between northwest Africa and the Carribbean) and a six-week period (August 20-September 30) each year between 1982 and 2004. They found that hurricane activity was generally lower in years with more dust and that the correlation was strongest when El Niño and La Niña years were excluded from the analysis. Evan and his colleagues found that dust cover could account for about 25% of the variance in hurricane activity.
(Read Environmental Defense commentary on this paper.)