1. 22 Nov. 2007 ESSCTC in a warmer climate? Lennart Bengtsson How might Tropical Cyclones change in a Warmer Climate? Professor Lennart Bengtsson ESSC Many thanks to Kevin Hodges and to colleagues in UK, Germany and Japan

2. 22 Nov. 2007 ESSCTC in a warmer climate? Lennart Bengtsson Tropical Cyclones in a warmer climate Societal damages are to a large extent related to extreme weather mostly in association with tropical and extra-tropical cyclones The cost for hurricane Katrina is estimated to some 100 G$ There are many examples of great loss of life due to extreme tropical cyclones. Damages are mostly related to high winds, flooding due to the high precipitation and in coastal areas to high sea-level and waves. The question whether cyclones may intensify in a future climate is consequently an issue of primary importance for society. This is further enhanced by the ongoing increased exposure to extreme weather independent of climate change.

3. 22 Nov. 2007 ESSCTC in a warmer climate? Lennart Bengtsson Tropical cyclones in a future climate what could be expected? Higher SST and higher atmospheric moisture would generally favor more intense storms ( e.g. Emanuel 1988, 1999) This is supported by modeling results by Knutson and Tuleya (2004) driving an limited area model with CMIP2+ boundary data ( 9 different models). Increasing vertical wind-shear and reduced relative humidity would counteract this tendency. Such influences occur in the tropical N. Atlantic during El Nino. Some GCM indicate reduced number of cyclones in a warmer climate How will the number of storms change? What are the general conditions controlling the number of tropical storms? What are the critical conditions in modeling tropical storms? Are results from large scale models with limited resolution credible?

4. 22 Nov. 2007 ESSCTC in a warmer climate? Lennart Bengtsson How might Tropical Cyclones change in a Warmer Climate? Present observational evidence What can we expect based on theory?What can we expect based on theory? Modeling approachModeling approach Model experimentModel experiment ResultsResults Concluding remarksConcluding remarks

5. 22 Nov. 2007 ESSCTC in a warmer climate? Lennart Bengtsson What is a tropical cyclone? It is a warm core vortex developing over tropical oceans. The maximum intensity occurs around 900hPa. Tropical cyclones are classified according to maximum intensity Tropical depression <17m/s Tropical storm 17-33 m/s Hurricane/Typhoon in categories 1 to 5 1: 33-42, 2: 43-49, 3: 50-58, 4: 59-69, 5: >69 Actual Archive of TC, see: http://weather.unisys.com/hurricane/

6. 22 Nov. 2007 ESSCTC in a warmer climate? Lennart Bengtsson Are there observational evidence that tropical cyclones are becoming more frequent, more intense? Why is it so difficult? Longer term records are needed due to internal variability There have been large changes in the observing systems making it easier to detect more tropical cyclones in later years. Some recent studies have used PDI ( time integral of max. wind cube) which is overly sensitive to observational accuracy Most researchers agree that there is no global increase in tropical cyclones, but there are different views concerning the trend in the Atlantic.

7. 22 Nov. 2007 ESSCTC in a warmer climate? Lennart Bengtsson Is the number of North Atlantic tropical cyclones significantly underestimated prior to the availability of satellite observations? by Edmund Chang and Yanjuan Guo GRL, doi:10.1029/2007GL030169, 2007 (table LB in a comment to GRL) A:1920-1965B:1966-2005Ratio B/A All3474441.28 US (landfall)1361300.96 CC (landfall)109950.87 NS(within 300km from land) 561252.23 OO (over ocean)46942.04

8. 22 Nov. 2007 ESSCTC in a warmer climate? Lennart Bengtsson There are recent claims that there is an increase in hurricane intensity ( e.g. Emanuel (2005), Webster et al. (2005) Are these findings credible? They are generally not supported by operational meteorologists According to Knutson and Tuleya (2004) any changes are probably undetectable “ for decades to come ” There are structural problems in the detection of trends Changes in observing systems Difficulties to separate a genuine change in storms from societal causes behind the huge increase in damages and damage cost

9. 22 Nov. 2007 ESSCTC in a warmer climate? Lennart Bengtsson How might Tropical Cyclones change in a Warmer Climate? Present observational evidence What can we expect based on theory? Modeling approach Model experiment Results Concluding remarks

10. 22 Nov. 2007 ESSCTC in a warmer climate? Lennart Bengtsson After Emanuel The Carnot cycle concept

11. 22 Nov. 2007 ESSCTC in a warmer climate? Lennart Bengtsson What happens to the hydrological cycle in a warmer climate? See e.g. Held and Soden, 2006, J. of Clim. Observations and models show that water vapor follows temperature according to Clausius- Clapeyron expression. The increase in precipitation is much slower. This increases the residence time of water in the atmosphere. This reduces the large scale vertical mass flux This slows down the large scale circulation

12. 22 Nov. 2007 ESSCTC in a warmer climate? Lennart Bengtsson Clausius - Clapeyron expression A typical value of alfa in the lower troposphere is 0.07/K or 7% increase in the saturation water vapor for each 1K in temperature

13. 22 Nov. 2007 ESSCTC in a warmer climate? Lennart Bengtsson Globally integrated water vapor 1979-2005 From Held and Soden, 2006 DT +0.45K De s +3% Full line GFDL model Dashed line measurements

14. 22 Nov. 2007 ESSCTC in a warmer climate? Lennart Bengtsson Semenov and Bengtsson, 2002 Secular trends in daily precipitation characteristics Clim.Dyn. 123-140 +30%

15. 22 Nov. 2007 ESSCTC in a warmer climate? Lennart Bengtsson Top, IPCC models: massflux, water vapor and precipitation Below, GFDl model P = Mq ( From Held and Soden (2006))

16. 22 Nov. 2007 ESSCTC in a warmer climate? Lennart Bengtsson Energy balance for upward and downward motion Follows LqQ c = P

17. 22 Nov. 2007 ESSCTC in a warmer climate? Lennart Bengtsson How might Tropical Cyclones change in a Warmer Climate? Present observational evidence What can we expect based on theory? Modeling approach Model experiment Results Concluding remarks

18. 22 Nov. 2007 ESSCTC in a warmer climate? Lennart Bengtsson Modeling approach Direct simulation of tropical cyclones in a global GCM Using limited area models at high resolution Identifying climate predictors in a GCM (SST, vorticity, static stability, relative humidity, vertical wind-shear)

19. 22 Nov. 2007 ESSCTC in a warmer climate? Lennart Bengtsson Simulation of tropical cyclones with a GCM Typical criteria: An identifiable vortex ( often at 850 hPa) A minimum in the surface pressure Surface wind speed above a given value (> 18m/s) A warm core vortex ( reduced circulation with height)

20. 22 Nov. 2007 ESSCTC in a warmer climate? Lennart Bengtsson Direct simulation of tropical cyclones in a global GCM Disadvantage: Difficulties to resolve the intense features of a tropical storm Examples of studies: Bengtsson et al. 1995, 1996, Tellus Sugi et al. 2002, JMS, Japan McDonald et al. 2005, Clim.Dyn. Chauvin et al. 2006, Clim Dyn. Oouchi et al. 2006, JMS, Japan Yoshimori et al. 2006, JMS, Japan

21. 22 Nov. 2007 ESSCTC in a warmer climate? Lennart Bengtsson Using limited area models at high resolution Disadvantages: Generation of storms, Large scale influences difficult to handle Examples of studies: Knutson et al. 1998, Science Knutson and Tuleya, 1999, Clim. Dyn. Knutson and Tuleya, 2004, J Clim.

22. 22 Nov. 2007 ESSCTC in a warmer climate? Lennart Bengtsson Impact of CO 2 -induced warming on simulated hurricane intensity Knutson and Tuleya (2004, J of Climate) They used a high resolution limited area model driven by the SST and moisture of 9 CGCM from the CMIP 2+ project. CMIP2 uses 1%yr -1 increase over an 80-year period implying an increase by a factor of 2.2. Model calculations are undertaken in NW Pacific-, NE Pacific- and Atlantic basin Four different convective schemes are tested (no significant differences) RESULTS: Max. surface wind speed increases by 6% Min. central pressure by 14% Max. precipitation by 24% Hurricane increase by a factor of 1/2 in the Simpson-Saffir scale

23. 22 Nov. 2007 ESSCTC in a warmer climate? Lennart Bengtsson Identifying climate predictors in a GCM Disadvantage: Lack of proper understanding, ad hoc selection of predictors, overestimation of the effect of SST Examples of studies Gray, 1979 Met. over the Tropical oceans, RMetSoc Royer et al. 1998, Clim. Change Chauvin et al. 2006, Clim. Dyn.

24. 22 Nov. 2007 ESSCTC in a warmer climate? Lennart Bengtsson How might Tropical Cyclones change in a Warmer Climate? Present observational evidence What can we expect based on theory? Modeling approach Model experiment Results Concluding remarks

25. 22 Nov. 2007 ESSCTC in a warmer climate? Lennart Bengtsson Tracking methods and vortex identification Tracks are followed from its generation (6x10 -6 s -1 ) until it disappears as an extra-tropical cyclones north of 60N We calculate the total life-time of the TCs We are able to identify the transition from a tropical to an extra-tropical vortex Alternatively we use the wind speed at 925 hPa as a selection criteria for intense storms We have also calculated the potential dissipation index, PDI. See Emanuel, Nature, 2005

26. 22 Nov. 2007 ESSCTC in a warmer climate? Lennart Bengtsson Objectives of the modeling study Tropical cyclones (TC) in ECHAM5 in the Northern Hemisphere (Using high global resolution) How do the TCs respond to anthropogenic climate change and how does this depend on resolution? What changes occur in intensity, life time and power dissipation index? What possible mechanisms control the change in TC? What are the dominant factors?

27. 22 Nov. 2007 ESSCTC in a warmer climate? Lennart Bengtsson Hurricane Katrina August 2005 ECMWF operational analyses, 850 hPa vorticity

28. 22 Nov. 2007 ESSCTC in a warmer climate? Lennart Bengtsson Katrina vorticity at different levels

29. 22 Nov. 2007 ESSCTC in a warmer climate? Lennart Bengtsson Selection of TC indicators We use criteria for minimum vorticity at 850 hPa (V), minimum vertical vorticity gradient (G) between 850 and 250 hPa, and number of time steps of 6 hrs (T) when these conditions are fulfilled. (V, G, T) = (6, 6, 4) vorticity at 850 hPa = 6x10 -5 s -1 vorticity 850- 250 hPa = 6x10 -5 s -1 conditions fulfilled at least 24 hrs (6, 6, 4) is defined as a TC

30. 22 Nov. 2007 ESSCTC in a warmer climate? Lennart Bengtsson Selection of criteria for selecting tropical storms (TC) (using ECMWF operational analyses ) All Tropical Storms Hurricanes, Typhoons, Cyclones >33ms -1 (6, 6, 4) (10, 6, 4) (12, 6, 4) 20037533714839 20047236795241 20058038836248

31. 22 Nov. 2007 ESSCTC in a warmer climate? Lennart Bengtsson Objectives of the present study Tropical cyclones in ECHAM5 We have used scenario A1B and studied the periods 1861-1890, 1961-1990 and 2071-2100 We have explored the coupled T63 run (3 runs) for all periods and T213 time - slice 1961-1990 and 2071-2100 T319 time - slice 1971-1990, and 2080-2100 We have also used three AMIP2 runs (20 years) with T63 and T159 as a validation study

32. 22 Nov. 2007 ESSCTC in a warmer climate? Lennart Bengtsson How might Tropical Cyclones change in a Warmer Climate? Present observational evidence What can we expect based on theory? Modeling approach Model experiment Results Concluding remarks

33. 22 Nov. 2007 ESSCTC in a warmer climate? Lennart Bengtsson Structure of modeled tropical cyclones This shows the averaged structure of the 100 most intense storms at the time when the reach their maximum intensity

34. 22 Nov. 2007 ESSCTC in a warmer climate? Lennart Bengtsson Tangential (left) and Radial winds (right) for the T63 resolution. Negative values inflow. Average of 100 tropical cyclones. At the time of maximum wind. Radius 5 degrees. South

35. 22 Nov. 2007 ESSCTC in a warmer climate? Lennart Bengtsson Tangential (left) and Radial winds (right) for the T213 resolution. Negative values inflow. Average of 100 tropical cyclones. Radius 5 degrees. The flow is predominantly inward to the rear and left of the storm and outward to the front and right (Frank 1977 MWR) Observations:

36. 22 Nov. 2007 ESSCTC in a warmer climate? Lennart Bengtsson Tangential (left) and Radial winds (right) for the T319 resolution. Negative values inflow. Average of 100 tropical cyclones. Radius 5 degrees.

37. 22 Nov. 2007 ESSCTC in a warmer climate? Lennart Bengtsson Hurricane Mitch, Oct -Nov. 1998 Kepert 2006 JAS

38. 22 Nov. 2007 ESSCTC in a warmer climate? Lennart Bengtsson Hurricane Mitch: Tangential winds (left) and Radial winds (right) at 500 m. Units: 5 m/s Kepert 2006 JAS outflow West

39. 22 Nov. 2007 ESSCTC in a warmer climate? Lennart Bengtsson Tangential wind cross section Temperature anomaly T63 resolution

40. 22 Nov. 2007 ESSCTC in a warmer climate? Lennart Bengtsson Tangential wind cross section Temperature anomaly T213 resolution

41. 22 Nov. 2007 ESSCTC in a warmer climate? Lennart Bengtsson Comparison with observations from the Tropical Warning Centers and with ERA-40 re-analyses

42. 22 Nov. 2007 ESSCTC in a warmer climate? Lennart Bengtsson Super Typhoon 21 (1991) in ERA-40 (left) and selected similar storm in ECHAM5 (right) Intensity (vorticity at 850 hPa)

43. 22 Nov. 2007 ESSCTC in a warmer climate? Lennart Bengtsson Hurricane genesis (a) observed, (b) ERA-40 and (c ) ECHAM5 T159

44. 22 Nov. 2007 ESSCTC in a warmer climate? Lennart Bengtsson Hurricane track density (a) observed, (b) ERA-40 and (c ) ECHAM5 T159

45. 22 Nov. 2007 ESSCTC in a warmer climate? Lennart Bengtsson Hurricane track density (Atlantic) (d) observed, (e) ERA-40 and (f ) ECHAM5 T159

46. 22 Nov. 2007 ESSCTC in a warmer climate? Lennart Bengtsson What may happen in a warmer climate? We have used the AIB scenario And the coupled MPI model at T63 resolution used in the IPCC 4th assessment Higher resolution experiments use the transient SST from T63 (time - window) We study C20 (1961- 1990) C21 ( 2071-2100)

47. 22 Nov. 2007 ESSCTC in a warmer climate? Lennart Bengtsson What is A1B? Middle of the line scenario Carbon emission peaking in the 2050s (16 Gt/year) CO 2 reaching 450 ppm. in 2030 CO 2 reaching 700 ppm. in 2100 SO 2 peaking in 2020 then coming done to 20% thereof in 2100

48. 22 Nov. 2007 ESSCTC in a warmer climate? Lennart Bengtsson SST difference (C 21-C 20)

49. 22 Nov. 2007 ESSCTC in a warmer climate? Lennart Bengtsson TCs at T63 resolution C19 (black), C20 (red) and C21(blue)

50. 22 Nov. 2007 ESSCTC in a warmer climate? Lennart Bengtsson Hurricane genesis T213 From top C21-C20, C21 and C20

51. 22 Nov. 2007 ESSCTC in a warmer climate? Lennart Bengtsson Hurricane track density, T213 From top C21-C20, C21 and C20

52. 22 Nov. 2007 ESSCTC in a warmer climate? Lennart Bengtsson Number of TCs/year (T213) for C 20 and C21 for wind speed and vorticity T213All (6, 6, 4)>2x10 -4 s -1 >5x10 -4 s -1 >1x10 -3 s -1 20C (1961-1990)10497406.0 21C (2071-2100)9490499.8 T213>18ms -1 >33ms -1 >50ms -1 20C (1961-1990)100333.7 21C (2071-2100)92364.9

53. 22 Nov. 2007 ESSCTC in a warmer climate? Lennart Bengtsson Accumulated precipitation ( in mm and for an area with a radius of 5 degrees) along the track of the TCs for C20 and C21(T213) TCs reaching >33ms -1 C21 C20 Total increase 30%

54. 22 Nov. 2007 ESSCTC in a warmer climate? Lennart Bengtsson How might Tropical Cyclones change in a Warmer Climate? Present observational evidence What can we expect based on theory? Modeling approach Model experiment Results Concluding remarks

55. 22 Nov. 2007 ESSCTC in a warmer climate? Lennart Bengtsson Climate change and the water cycle Atmosphere appears to conserve relative humidity. This means that atmospheric water vapor follows Clausius- Clapeyron relation. (Held and Soden, 2006, J. Clim.) We see an increase of 27% in atmospheric water vapor at C21 compared to C20 Precipitation must be balanced by evaporation. Evaporation is driven by the surface energy balance which increases slower than atmospheric water vapor. In fact it can even diminish at a high aerosol concentration (ECHAM4). Precipitation increases by 6% both globally and in the tropics This means that the residence time of water in the atmosphere increases from 8.7 to 10.3 days or by 16%.

56. 22 Nov. 2007 ESSCTC in a warmer climate? Lennart Bengtsson Result from present study: Reduced number of hurricanes There is s reduction in the number of tropical cyclones in agreement with most previous studies. We suggest this is due to a weakening of the tropical circulation. This can best be seen as a slowing down of the hydrological cycle by some 16% The radiative cooling of the tropics increases due to more water vapor in the upper troposphere and dynamical cooling due to increased static stability. This processes can compensate warming from release of latent heat (6%) with a less active tropical circulation.

57. 22 Nov. 2007 ESSCTC in a warmer climate? Lennart Bengtsson Result from present study: More intense hurricanes Given favorable atmospheric condition we suggest the ideas put forward by Emanuel and Holland comparing a hurricane with a Carnot cycle can be applied. This will provide an energy input broadly proportional to the specific humidity at a higher temperature (following SST) The intensification of the tropical cyclones depends on sufficient model resolution to accurately describe the convergence of momentum which generate the very high wind speeds at the core of the vortex. For TCs reaching 33ms -1 PDI (power dissipation index) increases by 16%

58. 22 Nov. 2007 ESSCTC in a warmer climate? Lennart Bengtsson Why do we see a reduced number of tropical cyclones in a warmer climate? We suggest the following mechanism There is a reduction in the large-scale tropical circulation at climate warming due to increased static stability and a rapid increase in the amount of water vapor in the atmosphere. This generally leads to less favorable conditions to generate organized onset vortices which are seen as key atmospheric conditions for the generation of tropical storms.

59. 22 Nov. 2007 ESSCTC in a warmer climate? Lennart Bengtsson Why do we see an intensification of the tropical storms in a warmer climate? We suggest the following process Due to chance favorable conditions are likely to evolve In such situations the mechanisms proposed by Emanuel may be applied, that means that the maximum wind speed is proportional to the sqr. of the available latent energy provided to the storm To demonstrate this with a GCM requires sufficient resolution ( say 50-100 km or less)

60. 22 Nov. 2007 ESSCTC in a warmer climate? Lennart Bengtsson What is the effect of even higher resolution? 20-year simulation at T319 resolution

61. 22 Nov. 2007 ESSCTC in a warmer climate? Lennart Bengtsson T319 Intensity changes 20C to 21C

62. 22 Nov. 2007 ESSCTC in a warmer climate? Lennart Bengtsson Tropical cyclones in different regions, T319 resolution

63. 22 Nov. 2007 ESSCTC in a warmer climate? Lennart Bengtsson Tropical cyclones in different regions, T213 resolution

64. 22 Nov. 2007 ESSCTC in a warmer climate? Lennart Bengtsson Conclusions from the T319 resolution (40 km) Results confirm the finding from the T213 experiment The number of tropical cylones are reduced by some 10% at the end of the 21st century The stronger storms are becoming more intense (Cyclones with wind speeds >50m/s increase from 12 to 17 per year compared to the end of the 20th century) Intensification occurs in all tropical storm regions

65. 22 Nov. 2007 ESSCTC in a warmer climate? Lennart Bengtsson END