Attribution of - PowerPoint PPT Presentation

Loading...

PPT – Attribution of PowerPoint presentation | free to download - id: 1c0934-ZDc1Z



Loading


The Adobe Flash plugin is needed to view this content

Get the plugin now

View by Category
About This Presentation
Title:

Attribution of

Description:

Attribution of – PowerPoint PPT presentation

Number of Views:20
Avg rating:3.0/5.0
Slides: 40
Provided by: CGD95
Learn more at: http://www.bibalex.org
Category:
Tags: attribution | go1

less

Write a Comment
User Comments (0)
Transcript and Presenter's Notes

Title: Attribution of


1
Attribution of Recent Increases in Atlantic
Hurricane Activity
Kevin E Trenberth NCAR
2
Issues for detection and attribution of changes
in hurricanes
  • What has happened?
  • How good is the observational record?
  • How should hurricanes change as climate changes?
  • Are models adequate?
  • What is the role of global warming?
  • What is the role of natural variability?
  • What do models reveal?

3
Ivan 15 Sept 2004 1850 UTC
4
Katrina
5
Katrinas aftermath
Refugees in USA Aug 31 ?
6
Maria
Nate
Ophelia
7
Rita
8
Hurricane Wilma 21 October 2005
9
  • North Atlantic Hurricanes 2005
  • A record breaking year
  • Strongest Gulf hurricane month of July (Dennis)
  • Most named storms (27 normal 10)
  • Most hurricanes (15 normal 6 1970-2004)
  • First ever V, W, ?, ?, ?, ?, ?, ?
  • Strongest hurricane on record Wilma (882 mb)
  • Strongest hurricane in Gulf Rita (897 mb)
  • Most cat. 5 storms in season (4 vs 2 in
    1960,1961)
  • Deadliest hurricane in US since 1928 (Katrina)
  • Costliest natural disaster in US history
    (Katrina)
  • Highest insured losses ?40-60B vs Andrew 21B
  • Total losses ?125-200B
  • 6 of the 8 most damaging occurred Aug 04-Oct05
  • Charlie, Ivan, Francis, Katrina, Rita, Wilma
  • Hurricane Vince (October) first to hit
    Portugal/Spain

10
Atlantic Tropical Cyclone Trends
Satellites
Start of aircraft surveillance
Greg Holland
11
Atlantic Hurricane Trends
Named Cat 12 Storms Hurricanes
Cat 3-5
Greg Holland
12
Changes in hurricane frequency in the North
Atlantic Ocean

13
Issues on changing damage from hurricanes
Landfalling hurricanes are a very small fraction
of all hurricanes and the sample is small. Where
they make landfall is chance, and 10 miles (e.g.,
Andrew) can make a huge difference to
damage. The increased vulnerability of people
with increased property value building in coastal
zones, placing themselves in harms way, makes
changes in hurricane intensity even more
important. 100 years of tropical storm tracks
in Atlantic
14
  • Hurricanes
  • Depend on SSTs gt 26ºC (80ºF)
  • High water vapor content
  • Weak wind shear (or vortex comes apart)
  • Weak static stability
  • Pre-existing disturbance
  • Large variability year to year in individual
    basins.
  • El Niño means more action in Pacific, suppression
    in Atlantic
  • Large decadal variability in Atlantic

15
Better measure of tropical cyclone activity
Simplified Power Dissipation Index (Emanuel
2005)
Courtesy K. Emanuel
16
Atlantic western North Pacific
Courtesy K. Emanuel Revised
17
A large increase is seen in the number and
proportion of hurricanes reaching categories 4
and 5. The largest increase occurs in the North
Pacific, the Indian and Southwest Pacific oceans,
and smallest increase in the North Atlantic
Ocean.
From Webster et al (2005)
18
0.500.25ºC
SST in the tropics 20N-20S relative to 1961-90
mean annual means and low pass filter values to
emphasize decadal variations.
19
The Atlantic Multi-decadal Oscillation
AMO index defined by Enfield et al. (2001) as
mean SST north of equator in Atlantic then take
10 year running mean. Base period 1901-70.
20
Atlantic SSTs 10-20?N 0.92?C above 1901-70
normal. All time record. Due to weak trades and
reduced LH fluxes.
Global warming 0.45?C 2004-05 El Niño
0.2?C AMO lt0.1?C
Trenberth et al 2002 Trenberth and Shea 2006
21
Monthly SST anomalies for (A) Atlantic and (B)
Pacific tropical cyclogenesis regions Observed
(black) and 22 climate models. Model data are
smoothed 2 groups with and without volcanic
forcing (V and No-V) and end in 1999. The yellow
and grey envelopes are 1 and 2 confidence
intervals for the V averages. Santer et
al 2006
Is the variability realistic? Do the models
simulate observed?
22
Models show signal to noise of natural
variability is large trend can only arise from
increased GHGs Contribution of different
external forcings to SST changes in the Atlantic
(A) and Pacific (B) tropical cyclogenesis
regions. Results are from a 20CEN run and from
single-forcing experiments performed with the
Parallel Climate Model (PCM). Each result is the
low-pass filtered average of a four-member
ensemble. Santer et al. 2006
23
Linear regression maps of T106 ECHAM5 AGCM
simulated Atlantic TC vertical wind shear (200
-850 hPa) for regions given for 1870-2003. Color
gives statistical significance (T-test). Biggest
effect is from Pacific. Latif et al 2006 GRL
(see Aiyyer and Thorncroft 2006 JCl for obs)
24
  • What about 2006?
  • La Nina in 2005-06 winter (vs El Nino 2004-05)
  • Jan 2005 light winds, sunny
  • Jan 2006 much stronger than normal winds
  • SSTs below normal in west Atlantic earlier
    warmed midway thru season
  • Developing El Nino in Pacific
  • Unfavorable conditions for TCs in Atlantic wind
    shear etc.

Foltz and McPhaden, GRL 2006 show how the weak NE
tradewinds, anomalous latent heat fluxes and
solar radiation contributed to the record
breaking SSTs in summer 2005
25
  • In the tropics, heat from the sun goes into the
    ocean and is apt to build up Where does the heat
    go?
  • Surface heat cannot radiate to space owing to
    optically thick water vapor
  • Heat goes from the ocean into the atmosphere
    largely through evaporation that is greatly
    enhanced in tropical storms. It moistens the
    atmosphere (latent energy) and cools the ocean.

3) Heat and moisture are transported to higher
latitudes by extratropical cyclones and
anticyclones (cold and warm fronts) mainly in
winter. 4) Heat is transported upwards in
convection, especially thunderstorms, tropical
storms, hurricanes and other disturbances.
Energy and moisture from the surface is moved
upwards, typically producing rain, drying the
atmosphere, but heating it, and stabilizing the
atmosphere against further convection.
26
Tropical ocean heat balance
Hot towers convective heat transports up
Incoming radiation
Water vapor greenhouse radiation
Latent heat Rain
Evaporation Cooling
Heating
Surface radiation
Surface flux
Ocean currents
27
In the tropics, heat from the sun is apt to build
up 4) There is a competition between individual
thunderstorms and organized convection to
transport heat upwards in the general atmospheric
circulation. 5) Tropical storms are much more
effective at cooling the ocean.
28
Cold wake from Katrina and Rita in Gulf of Mexico
SST in Gulf
NASA
29
Hypothesis Hurricanes play a key role in
climate, but are not in models and are not
parameterized. Prospects are for more intense
storms, heavier rainfalls and flooding, and
coastal damage, but perhaps lower tropical ocean
temperatures?
30
Hypothesis on effects from global warming
  • Water vapor over oceans increases 7 per K SST
  • To first order, surface latent heat fluxes also
    increase by at least this amount as E
    ?CVqs(Ts)(1-RH) qs(Ts)
  • Convergence in boundary layer also should go up
    proportionately. q?, ??, vr? and vr.q? squared
  • Could also increase intensity V
  • Other feedbacks (friction, sea spray, stability
    etc)
  • Hence estimated rainfall, latent heating and
    water vapor in the storms should increased 1.072
    1.14 or 14. 7 to 21 error bars per K.
  • For observed 0.5K increase in SST this means
    increases in rainfall and latent heat release in
    storms by order 7.

31
Katrina
Katrina 28 Aug 2005 Cat 5
32
Hurricane Katrina WRF Moving grid
27 Aug 2005 00 Z
4 km WRF, 62 h forecast
Mobile Radar
33
Katrina experiments
  • Given good track forecasts of Katrina, as well as
    the diagnostics of the energy and water budgets,
    we rerun the forecast simulations with SSTs
    changed by 1?C and -1?C
  • The control run has the central pressure 892 mb
    vs observed 902 mb
  • 1?C 870 mb -22 mb
  • -1?C 910 mb 18 mb
  • Max winds 58 m/s (-1) go to 70 m/s (1)
  • Order 10 per C

34
Observed and WRF simulation
35
  • Katrina
  • Ratio of rainfall to surface LH flux is 9 inside
    100 km, and 3.9 inside 400 km. They are
    comparable at 700 km and a balance is achieved
    only by considering a radius of about 1600 km.
  • This highlights the major role of moisture
    convergence by the low level (below 1 km) inflow
    into the storm.
  • That convergence in turn is driven by the surface
    LH flux, latent heating and storm circulation.
  • Nonetheless it highlights the role of the
    large-scale environment in these storms.

36
Peak azimuthally-averaged surface tangential
winds increase 47.0 to 50.5 to 56.0 m s-1, or 4.5
m s-1 per K (8.9 of control). Minimum pressure,
average 920.8 hPa, fell 11.5 hPa per K increase
in SST.
37
  • Precipitation is dominated by moisture
    convergence
  • Surface flux of moisture is essential amounts to
    gt1500 Wm-2.
  • Substantial increases with increasing SSTs rain
    increased by 19/K inside 400 km.

38
Katrina 49-54 hours Area averages inside radius,
except for winds. Mean vorticity/divergence
inside R is 2Vt/R and 2Vr/R. R 100 km
Units R 400 km Control Change Control
Change /K SST /K
SST Wv Transport 19.8 11.6 mm/h
2.7 22.4 Rainfall 19.1 9.4 mm/h
4.3 18.8 LH flux 2.1 20.2 mm/h
1.1 25.5 Vt 31.2 4.2 m s-1
12.1 19.0 Vr -12.4 2.5 m s-1
-4.3 14.9 SST 304.2 K 304.2 SLP 950.0 7
.4 hPa 982.9 4.2 hPa
Ratio Rainfall to LH flux 100 km 9.0 400 km 3.9
39
Changes in Hurricane 0-400 km
Rain enhanced 19 per K
Surface LH flux enhanced 25 per K
Winds enhanced 10-18 per K well outside eye
Inflow enhanced 17 per K beyond 280 km
Eyewall peak winds enhanced 9 per K
40
Best track data
Model
WRF Katrina results of surface fluxes as
function of maximum wind at any grid point.
30N-30S Best track data
41
For 1990-2005 over 0-400 km radius (51011 m2),
ocean cooling is 0.52, 0.58, 1.84?1022 J/yr, or
0.16, 0.185, 0.58 PW.
42
TC flux climatology
Surface LH (total enthalpy) fluxes 0-400 km
radius Cat 1 2 3 4 5 .
LH 548 623 682 766 865 SH 80 85 101 129
154 Katrina model Enthalpy 628 708 783 895 101
9 W m-2 Precip 3.2 3.0 4.3 4.7 5.1
mm/h Freq all 22.7 11.1 7.6 8.0 1.6 Best
track data 30N-30S 18.8 10.1 7.3 8.0 1.5
gives evaporative, total enthalpy, precip ocean
cooling of 0.16, 0.185, 0.58 PW over a year.
43
TC flux climatology
The results suggest an evaporative, total
enthalpy, precipitation ocean cooling of 0.16,
0.185, 0.58 PW over a year. Over the tropical
ocean 20?N to 20?S the LH is equivalent to 1.5 W
m-2 , or 1.1 ?C/year over a 10 m
layer. Globally this is 0.36 and 1.13 W m-2 vs
CO2 radiative forcing 1.5 W m-2. It matters!
And it is not included in climate models.
44
Implications for climate models
1) In models, the thunderstorms and convection
are not resolved and are dealt with by sub-grid
scale parameterization. 2) However, most (all?)
climate models have premature onset of
convection, as seen in the diurnal cycle over
land, and feature convection too often and with
insufficient intensity. (cf Lin et al. 2006 J
Cl) 3) This characteristic likely means that
sub-grid scale convection is overdone at the
expense of organized convection (MJO, tropical
storms, etc see Lin et al. 2006, JC). 4) Hence
models likely under-predict changes in
hurricanes. 5) Hurricanes are missing in models
SSTs may get too warm increased TCs keep SSTs
cooler.
45
Research questions for detection and attribution
of changes in hurricanes
  • Need to reprocess the satellite record.
  • Need measures of activity size, duration,
    intensity, rainfall, track, ACE, PDI etc
  • How is TC environment changing and why?
  • Models must improve in simulation of natural
    variability ENSO, AMO, PDO
  • Need to improve climate models Resolution
    precipitation (frequency, intensity, amount),
    atmospheric stability, convection (sub-grid
    scales), tropical transients (storms, MJO,
    easterly waves)
  • Coupled problem must have ocean model
  • How to parameterize effects of hurricanes?
About PowerShow.com