In the past 16 months, two exceptionally intense tropical cyclones, Haiyan and Pam, have struck the western Pacific with devastating effect. Haiyan may have had the highest wind speeds of any tropical cyclone on record, but we will never know for sure because we do a poor job estimating the intensity of storms that are not surveyed by aircraft. (Currently, only North Atlantic tropical cyclones are routinely reconnoitered by aircraft, and only if they threaten populated regions within a few days.) Pam’s analyzed intensity puts it within 10 knots of the most intense storms on record in the South Pacific, but here again this is within the error bars of satellite-derived intensity estimates.
Pam’s high intensity and terrible impact on Vanuatu have invariably raised the question of the possible effect of global warming on its characteristics. For example, Vanuatu’s President Baldwin Lonsdale blamed the disaster partly on climate change. Just as predictable is the backlash to the effect that no single event can be attributed to climate variations of any kind. What can we say about the effects of climate change on South Pacific tropical cyclones?
We can begin by looking at the record of tropical cyclones in that region. Unfortunately, for the reasons discussed above, these records are poor and those that exist only go back to about 1980, though there are longer records of storms making landfall in Australia. Perhaps the best existing analysis of South Pacific tropical cyclones is that of Kossin et al. (2013), who homogenized the satellite data record from 1982 to 2009 to create a temporally consistent record, and compared that to the problematic historical data base of storms over the world. While the historical data in the South Pacific region show a highly significant upward trend in the incidence of high intensity events, the satellite-based record shows a less prominent and significant trend of 2.5 m/s per decade with a p value of 0.09. Thus there is some evidence of a trend toward higher intensity of high category tropical cyclones in the South Pacific over the period 1982-2009, but it is not conclusive and in any event spans a limited time interval.
We can also look at trends in important environmental factors that are known to influence tropical cyclones. The usual suspect among these is sea surface temperature (SST) and there has been much talk about the elevated SST’s in the region where Pam developed. But SST by itself is not the main factor in the existing theory for the upper bound on tropical cyclone intensity, known as the potential intensity; instead, the potential intensity depends more nearly in the difference between SST and a measure of the bulk temperature of the troposphere as well as the temperature of the tropopause. An expression of the potential intensity, measured in maximum possible wind speed, is
where Vp is the potential maximum wind speed, Ts is the surface temperature, Tt is the tropopause temperature, hs* is the saturation moist static energy of the sea surface, and h* is the saturation moist static energy of the free troposphere, which is nearly uniform with height if the lapse rate is moist adiabatic. In the deep tropics, temperature is nearly uniform on pressure surfaces because there is not enough Coriolis acceleration to balance strong pressure gradients, thus h*, which is just a function of pressure and temperature, is horizontally as well as vertically uniform in the free troposphere.
Therefore, the potential intensity depends mostly on variations of SST (which controls hs*) for climate variations that do not affect the mean temperature of the troposphere. But global warming very definitely does affect the temperature of the tropical free troposphere, so it is not possible to conclude, as alas many have, that increasing SST per se means increasing tropical cyclone intensity (though it usually does signify more TC-related rain).
It is not difficult to calculate the actual potential intensity from SST and atmospheric soundings, and this can be done as well for reanalysis and global model data sets. The map below shows the potential intensity at 12 GMT on 7 March, 2015, calculated from the NCEP operational analysis at that time.
The track of PAM is superimposed in blue.
Pam traversed a region of potential intensities around 75 m/s, consistent with the storm’s estimated peak intensity. These values are not unusual in this region, but for the past few decades they have been increasing, if the reanalyses are to be believed. I calculated potential intensity trends over the period 1980-2012 using three different reanalysis products: NASA’s MERRA, the European Center’s ERA Interim, and the NCAR/NCEP reanalysis. The MERRA and ERA Interim reanalyses show upward trends of around 2 m/s per decade in the region where Pam formed, while NCEP’s trend is closer to 3 m/s per decade; all these trends have p values less than 0.1. These trends are broadly consistent with the Kossin et al. (2013) trend of 2.5 m/s per decade in the observed intensity of high category tropical cyclones in this region.
Thus the weight of evidence points to increasing potential intensity in the region where Pam developed, and consistent with this, increasing intensity of the highest category storms based on satellite-derived measurements. But to what do we attribute such increases? The roughly 30-year period over which we have reliable reanalyses and satellite measurements is too short to rule out the influence of natural climate variability, such as the Pacific Decadal Oscillation. We can at least check to see what kinds of trends climate models produce. I looked at eight CMIP 5 models whose output I had ready access to and calculated linear trends of potential intensity over the period 2006-2100 under the RCP 8.5 emissions pathway. The eight models were the NCAR CCSM4, a super-parameterized version of the same, referred to as SP-CAM, The GFDL CM3, the UKMO HADGEM-ES, the IPSL CM5A-LR, the Max Planck MPI-ESM-MR, the CCSR/NIES/JAMEST MIROC5, and the Meteorological Research Institute’s MRI-CGCM3. Of these, two models showed insignificant trends in the region in which Pam developed, and the rest showed positive trends averaging around 0.5 m/s per decade, considerably less than the observed trend over the last 30 years. The largest trend was produced by the GFDL model, whose global trend distribution is shown below. (White areas represent p values less than 0.1., and the units are m/s per decade.)
It is interesting that the largest increases are at the polar peripheries of the tropics, indicating a general expansion of the regions that are thermodynamically favorable for tropical cyclone development; this general feature is present in most of the model potential intensity trends as well as the reanalysis trends over the past 30 years, and may be behind the poleward migration of the latitudes at which observed tropical cyclones reach their peak intensity, as documented by Kossin et al. (2014).
The disparity between the reanalysis potential intensity trends over the past 30 years and the projected trends over this century suggests either that most of the observed increase in potential intensity (and actual intensity of high category storms) is due to natural variability, that decreasing anthropogenic aerosol loading over that period may have played a role, or that the model projections are too conservative. Yet the projected increase is not insignificant, amounting to about 5 m/s over 100 years. Note from the figure above that there are somewhat larger increases elsewhere, particularly in the northern hemisphere.
All of this is consistent with the strengthening consensus that the frequency of high category tropical cyclones should increase as the planet warms (Knutson et al., 2010). Basic theory and a variety of numerical simulations support this, as well as the projection that tropical cyclones should produce substantially more rain, owing to the increased moisture content of the tropical atmosphere. This is important because most destruction and loss of life are caused by high category storms and their attendant storm surges, and by freshwater flooding from torrential rains. Most of the disagreement in the literature on tropical cyclone projections concerns the incidence of weak storms, but these are usually far less consequential in spite of being more numerous.
Besides the oft-discussed issues of TC frequency and intensity, changes in genesis locations and tracks are potentially very important, as are the diameters of TCs, which affect the area covered by strong winds and which greatly affect the magnitudes of storm surges. In the present climate, the diameter of storms, as measured by the radius at which their circular wind component becomes indistinguishable from environmental winds, appears to be log normally distributed, with a mean around 420 km (Chavas and Emanuel, 2010). There is some indication from modeling studies that the size of the dangerous inner core scales with the potential intensity divided by the Coriolis parameter (Khairoutdinov and Emanuel, 2014). If this turns out to be true in nature, then storm inner core dimensions should increase over time.
While Pam and Haiyan, as well as other recent tropical cyclone disasters, cannot be uniquely pinned on global warming, they have no doubt been influenced by natural and anthropogenic climate change, and they do remind us of our continuing vulnerability to such storms. Destructive TCs in any one place tend to be generational…enough time for people to forget them and go back to risky behavior, including over-development of coastal regions. We adapt more successfully to the more frequent events which are always in the back of our minds (and often in the front). But this human adaptation time scale may be longer than the time over which climate change affects storms, so that comparatively small changes in the frequency of generational events can have large social consequences. When a 100-year event becomes a 50-year event, it may take a few destructive hits before we adapt to the new reality. This is of particular concern with tropical cyclones, where the application of existing damage models to projected changes in tropical cyclone activity predict large increases in damage, as documented, for example, in the recent Risky Business report commissioned by Michael Bloomberg, Hank Paulson, and Thomas Steyer*.
Now if only we could better measure tropical cyclones to record how they may change in coming years.
_________
* [Update, 1pm ET] It has been pointed out that in my reference to the Risky Business report, I might have mentioned that I contributed synthetic hurricane event sets that were used by Risk Management Solutions, Inc., to estimate damages from tropical cyclones.
* [Update, 1pm ET] It has been pointed out that in my reference to the Risky Business report, I might have mentioned that I contributed synthetic hurricane event sets that were used by Risk Management Solutions, Inc., to estimate damages from tropical cyclones.
References
Chavas, D. R., and K. A. Emanuel, 2010: A QuickSCAT climatology of tropical cyclone size.Geophys. Res. Lett., 37. 10.1029/2010GL044558.
Khairoutdinov, M. F., and K. Emanuel, 2014: Rotating radiative-convective equilibrium simulated by a cloud-resolving model. J. Adv. Model. Earth Sys., 5, In press.
Knutson, T. R., and Coauthors, 2010: Tropical cyclones and climate change. Nature Geosci., 3,157-163.
Kossin, J. P., T. L. Olander, and K. R. Knapp, 2013: trend analysis with a new global record of tropical cyclone intensity. J. Climate, 26, 9960-9976.
Kossin, J. P., K. A. Emanuel, and G. A. Vecchi, 2014: The poleward migration of the location of tropical cyclone maximum intensity. Nature, 509, 349-352.
Posts
No comments:
Post a Comment