|Above: Hurricane Sandy’s wind field on October 28, 2012, when Sandy was a Category 1 hurricane with top winds of 75 mph. This surface wind data is from a radar scatterometer on the Indian Space Research Organization’s (ISRO) Oceansat-2 satellite. Wind speeds above 65 kilometers (40 miles) per hour are yellow; above 80 kph (50 mph) are orange; and above 95 kph (60 mph) are dark red. Sandy’s winds in excess of 40 mph spanned a region of ocean 900 miles in diameter, making Sandy the largest Atlantic hurricane on record. Image credit: NASA.|
As the planet warms, the extra heat and moisture available to hurricanes are expected to make them more destructive. Computer modeling work consistently indicates that greenhouse warming will increase the average intensity of global tropical cyclones by 2 - 11% by 2100, with an increase in the number of Category 4 and 5 storms (though the majority of the models predict that the number of weaker tropical cyclones will decrease). Global warming is also expected to make tropical cyclones rainier, with about 20% more precipitation falling within 100 km of the storm center. But how might the areal size of these great storms change in our warming climate? Storm size is a crucial factor in determining how large a hurricane’s storm surge will be and how much damage it will cause. With hurricane storm surges progressively riding inland on top of ever-increasing sea levels, any increase in storm size could greatly increase storm surge damage. We are just beginning to study how storm size may change with global warming, but preliminary model results I heard last week at the American Meteorological Society’s 33rd Conference on Hurricanes and Tropical Meteorology showed that we may well have to reckon with an increase in storm size causing additional destruction by Atlantic hurricanes as the climate warms. The uncertainties are high, though, and there is no evidence that storms have been getting bigger since 1981.
The case of Hurricane Sandy
Despite being only a Category 1 hurricane just before making landfall in the U.S. in October 2012, Hurricane Sandy caused a catastrophic $55 billion in damage (other damage estimates are as high as $72 billion). At peak size, twenty hours before landfall, Sandy had tropical storm-force winds that covered an area nearly one-fifth the area of the contiguous United States. Since detailed records of hurricane size began in 1988, only one tropical storm (Olga of 2001)—and no hurricane—has had a larger area of tropical storm-force winds. Sandy's area of ocean with twelve-foot high seas peaked at 1.4 million square miles--nearly one-half the area of the contiguous United States, or 1% of Earth's total ocean area. Most incredibly, ten hours before landfall (9:30 am EDT October 30), the total energy of Sandy's winds of tropical storm-force and higher peaked at 329 terajoules--the highest value for any Atlantic hurricane since at least 1969. This is 2.7 times higher than Katrina's peak energy, and is equivalent to five Hiroshima-sized atomic bombs. At landfall, Sandy's tropical storm-force winds spanned 943 miles of the U.S. coast. No hurricane on record has been wider; the previous record holder was Hurricane Igor of 2010, which was 863 miles in diameter. Sandy's huge size prompted high wind warnings to be posted from Chicago to Eastern Maine, and from Michigan's Upper Peninsula to Florida's Lake Okeechobee--an area home to 120 million people. Sandy's winds simultaneously caused damage to buildings on the shores of Lake Michigan at Indiana Dunes National Lakeshore, and toppled power lines in Nova Scotia, Canada--locations 1200 miles apart!
A 2014 study by Alice Zhai and Jonathan Jiang, Dependence of US hurricane economic losses on wind speed and storm size, found that Sandy’s diameter was about three times greater than the average hurricane. Had Sandy been an average-sized hurricane, they estimated that damage from the storm would have been about 20 times less (with an uncertainty range of $1 billion - $5 billion in damage, instead of the actual $55 billion that occurred). In general, for a given maximum hurricane wind speed, the storm surge can vary by 30% depending upon how large the hurricane is (see Irish et al., 2008, The Influence of Storm Size on Hurricane Surge.)
|Figure 1. Map of the spatial distribution of tropical cyclone size from 1999 - 2009 using wind measurements from the QuikSCAT satellite. Size values are shown in kilometers for the radius of 12 m/s (27 mph) winds, plotted on a logarithmic scale. Gray lines denote associated storm tracks from best-track database. Dashed lines denote boundaries of basins. The western Pacific basin contains most of the largest storms in the dataset, though it contains some very small storms as well, indicative of both the very large storms that form from the monsoon gyre (Lander 1994) as well as the “midget” typhoons that occasionally form along the gyre periphery or else adjacent to the subtropical high. The eastern Pacific basin contains almost exclusively smaller storms. The Atlantic basin displays primarily smaller storms evident at lower latitudes in the central and eastern Atlantic, a mix of sizes at intermediate latitudes, and large sizes at high latitudes, where storms are likely transitioning to extratropical systems (Hart and Evans 2001). The north Indian Ocean basin is characterized by small and medium-sized storms, though this may simply reflect the confining coastline geometry of South Asia. In the Southern Hemisphere, both the south Indian Ocean and the South Pacific do not appear to exhibit any characteristic size. Image credit: Chavas et al., 2014, Observed Tropical Cyclone Size Revisited, Journal of Climate (American Meteorological Society).|
New modeling results for the Atlantic
At last week’s American Meteorological Society’s 33rd Conference on Hurricanes and Tropical Meteorology, Ben Schenkel of the University of Oklahoma presented modeling results showing that we may need to be concerned about Atlantic hurricanes growing larger in size as the climate warms. His team ran three separate models that studied how tropical cyclones might change in size by the end of the century under a moderate global warming scenario (RCP4.5). Two of their models showed an 8 – 9% increase in size of Atlantic tropical cyclones by 2100, as measured by the radius of 8 m/s (18 mph) or greater winds. Their other model showed no significant change in tropical cyclone size. The two models that showed the increase in size were the relatively coarse 25-km resolution GFDL HiFLOR model (which was connected to a model of the ocean), and the 2012 version of the GFDL hurricane model running at high resolution (6 km grid) with data taken from the coarser-resolution HIRAM model (using a technique called “downscaling”). The model that showed little change in future tropical cyclone size was the 2006 version of the GFDL hurricane model, running at 9 km resolution with downscaled data from the coarser-resolution ZETAC model. The models that showed an increase in storm size predicted that the size increases would not occur at genesis time, but later in the storm’s life cycle. The authors are in the process of extending their analysis using a higher-resolution model (4.5 km horizontal grid spacing) to look at Hurricanes Irma, Maria, and Katrina in both the current and late 21st century climate to better understand the uncertainty in their results.
Previous modeling studies show an increase or little change globally in tropical cyclone size
Knutson et al. (2015) also studied how tropical cyclones might change in size by the end of the century under the same moderate global warming scenario (RCP4.5). Using the 2012 version of the GFDL hurricane model, running at 6 km resolution with downscaled data from the 50-km resolution HiRam model, they found that globally, median storm size stayed nearly constant (+1%), but was not uniform across ocean basins. A decrease in storm size of 8% was found in the northwest Pacific; there was no statistically significant change in the north Indian basin, and small increases were predicted in all other basins (+7% to +15%). The differences in sample size allowed the northwest Pacific signal to balance those of the other basins. Notably, the two largest increases were found in the northeast Pacific (+15%) and North Atlantic basins (+11%).
A more recent 2017 study was done by Yohei Yamada of the Japan Agency for Marine-Earth Science and Technology: Response of Tropical Cyclone Activity and Structure to Global Warming in a High-Resolution Global Nonhydrostatic Model. His 14-km resolution model, which assumed a more extreme business-as-usual A1B global warming scenario, found that tropical cyclone size increased by about 12% globally, with increases in the Atlantic, Northwest Pacific, and Southern Hemisphere. However, there were decreases in the Eastern Pacific and North Indian Oceans. He argued that the reason his results differed from Knutson et al. (2015) for the Northwest Pacific was differences in locations of the storms. Yamada had more storms at higher latitudes, and these storms tend to be larger, as explained in the next section.
An earlier study (Kim et al. 2014), using a global warming scenario with a doubling of CO2 levels, found a marginal 3% increase globally in tropical cyclone size. This study used a coarser model with 50 km resolution.
Storms at higher latitudes tend to be larger
As hurricanes move towards Earth's poles, they acquire more spin, since Earth's rotation puts more vertical spin into the atmosphere closer to the poles; this extra spin helps storms grow larger, and we commonly see hurricanes grow in size as they move poleward. In addition, storms that are about to undergo the transition to a cold-cored extratropical storm as they move towards higher latitudes typically expand greatly in size during this process. As a result of these two influences, the largest storms in the Atlantic Extended Best Track database tend to reach their maximum size at a far northerly location, close to or at the time of the final advisory issued by the National Hurricane Center, when the tropical cyclone is transitioning to an extratropical storm. Here are the largest Atlantic tropical cyclones from 1988 – 2016, as measured by area covered by tropical storm-force winds. Also shown is the latitude where the storm first achieved its largest size (note that the area of Texas is approximately 270,000 square miles, so all of these storms were roughly 2 – 3 times the size of Texas):
Tropical Storm Olga, 2001: 780,000 square miles (latitude 30.8°N)
Hurricane Sandy, 2012: 560,000 square miles (latitude 39.5°N)
Hurricane Lili, 1996: 550,000 square miles (latitude 34.0°N)
Hurricane Igor, 2010: 550,000 square miles (latitude 48.5°N)
Hurricane Nicole, 2016: 520,000 square miles (latitude 39.3°N)
Hurricane Karl, 2004: 430,000 square miles (latitude 45.5°N)
|Figure 2. Time series of global hurricane size inferred from 3-hourly infrared geostationary satellite imagery, in degrees of latitude (DDLAT), using the radius of winds of at least 5 knots (R5), from 1981 – 2011. The maximum size for each hurricane is shown along with a trend line (solid line). The regression equation and percent of variance explained (R2) are also listed at the bottom. Image credit: Knaff (2014), An Objective Satellite-Based Tropical Cyclone Climatology, Journal of Climate (American Meteorological Society).|
Tropical cyclone size does not appear to have changed significantly over the past 35 years
It is difficult to determine if the size of tropical cyclones is already changing due to global warming, since we have a high-quality storm size database utilizing data from the Hurricane Hunters for only one ocean basin, the Atlantic. This Extended Best Track database only extends back to 1988. Climate is defined on time scales of 30+ years, so we are only now at the point where we might have enough statistics to do preliminary work on how storm size may be changing. There has been one study of storm size globally using lower-quality infrared satellite data back to 1981, though, and this data set does not show a significant change in tropical cyclone size in any ocean basin for the period 1981 – 2011 (Figure 2).
What controls tropical cyclone size is poorly understood
What controls tropical cyclone size is poorly understood, as detailed in Chavas and Emanuel (2014). They explained that storm size is correlated only weakly with latitude and intensity, since the outer- and inner-core regions appear to evolve nearly independently. Storm size is likely controlled by the structure of the initial disturbance, and the environment in which the storm is embedded; storms that are in an area of high relative humidity at mid-levels of the atmosphere are likely to be larger.
|Figure 3. Relationships between storm size and relative sea surface temperature (SST). Relative SST is the SST within the local environment of the storm minus the tropical mean SST. Mean values calculated in integer bins of 1°C, with the final bin including all data above 3°C. The bottom curves are for the radius of 12 m/s (27 mph) winds; this radius increases by approximately 48 km per 1°C increase in relative SST. The top curves are for the radius of calm winds surrounding the storm. Image credit: Chavas et al., 2014, Observed Tropical Cyclone Size Revisited, Journal of Climate (American Meteorological Society).|
Better models and theory are needed
Until we have a better theoretical understanding of what controls tropical cyclone size, predictions of how this quantity might change in the future climate should be viewed as uncertain. Furthermore, the models used so far to simulate storm size are inadequate for reliable predictions. Observations of storm size taken by the QuikSCAT instrument from 1999 – 2009 (Figure 3) showed that storm size varies depending upon how warm sea surface temperatures (SSTs) in the storm’s local environment are in comparison to SSTs in the world-wide tropics. This “relative SST” is a difficult quantity to get right in a climate model. Furthermore, work by Chavas and Emanuel (2014) found that models with a resolution of at least 4 km are needed to reliably predict the maximum winds of a hurricane; the maximum winds are connected to how large a tropical cyclone gets. None of the model results published so far have had a resolution that good.
Encouragingly, satellite estimates of tropical cyclone size going back to 1981 do not show any trends. However, preliminary modeling results described here show that we may well have to reckon with larger tropical cyclones causing additional destruction in some ocean basins in coming decades as the climate warms.