Transcript Holland-Prediction Across Scales CCSM 0606

Predicting the Earth System
Across Scales: Both Ways
Greg Holland
NCAR
Summary:
Rationale
Approach and Current Focus
Improved Simulation of Tropical Modes
Requirements for Tropical Cyclone Response to Climate
Change
Holland, CCSM Workshop 0606
ESSL
Rationale: Regional Climate
• Downscaling
• Upscaling
Holland, CCSM Workshop 0606
ESSL
Downscaling and Upscaling: Atlantic
Tropical Cyclone Trends
North Atlantic Tropical Cyclones 1905-2005
30
Satellites
Tropical Cyclone Number
Individual Year
25
20
15
10
5
9-Year Running Mean
0
1900
1920
1940
1960
1995
1980
2000
2020
Year
Holland, CCSM Workshop 0606
ESSL
Upscaling: Mean Rainfall
Holland, CCSM Workshop 0606
ESSL
Upscaling and Downscaling:
Tropical Modes
Observed
CCSM
(Lin et al 2006)
All tropical modes are poorly handled by current climate
models. This impacts everything from tropical cyclones to
ENSO and interactions with the extratropics.
Holland, CCSM Workshop 0606
ESSL
The Opportunity
CCSM/CAM
Community Models
WRF Community Model
Holland, CCSM Workshop 0606
ESSL
Approach and Current Focus
Approach: Nesting the NCAR Weather Research
and Forecasting Model into CAM as a 2-way Nested
Regional Climate Model (NRCM);
– Stage 1: Downscaling over North America (done);
– Stage 2: 2-way atmospheric scale interactions in the
tropics (current);
– Stage 3: NRCM in CAM/CCSM, with coupled ocean model;
as a community facility (next).
Current Focus: Tropical scale interactions:
– Importance of mesoscale organization of convection
and its related surface exchanges and radiative
influences for forcing tropical modes;
– Tropical mode forcing of tropical cyclone
development and intensification.
Holland, CCSM Workshop 0606
ESSL
Model Setup
Tropical Channel, 36 km, N/S boundaries 1-way nested into
NCEP Reanalysis with specified SST, Kain-Fritsch Cu
Parameterization, CAM radiation and YSU boundary layer.
Precipitable Water (mm)
Wind Speed (m/s)
4 km nested domain inside 12 km
and 36 km domains, fully 2-way
interactive, Dudia cloud physics,
CAM radiation and YSU boundary
layer.
Holland, CCSM Workshop 0606
ESSL
NRCM Experiments
• CAM at T170 Resolution: Jan 1, 1966 to Jan 1 2001
• 36km Channel Model: Jan 1, 1996 to Jan 1, 2001
• 36km Channel + high resolution SST's: Jan 1, 1999 to Jan
1, 2000
• 36km Channel + fluxes: Jan 1, 1996 to Jan 1 1998
• 36km Channel + 12km Maritime domain: Jan 1, 1996 to Feb
12, 1998
• 36km Channel + 12/4km Maritime domains: Jan 1, 1997 to
Jul 1, 1997
Holland, CCSM Workshop 0606
ESSL
Precipitable Water: 1997
Holland, CCSM Workshop 0606
ESSL
Power Spectra
NCEP Analysis
NRCM Simulation
Inertia-Gravity
Modes present, but
too small
Improved Rossby Wave, Kelvin and MJO Modes
Holland, CCSM Workshop 0606
ESSL
MJO and Easterly Wave Simulation
Holland, CCSM Workshop 0606
ESSL
July 22nd over
eastern Indonesia
MJO wet phase
28 days
August 19th in
eastern North
Pacific
39 days
October 8th
westerly wind
burst
Holland, CCSM Workshop 0606
ESSL
Easterly Wave Interactions with MJO
0722:00
0711:12
0701:12
0715:12
0705:12
0718:12
0708:12
0722:12
Holland, CCSM Workshop 0606
ESSL
NRCM Tropical Cyclone Simulation: Seasonal
Distribution 1996-1998
5.0
6.0
NIO NRCM
NIO OBS
Number of TCs
Number of TCs
6.0
4.0
3.0
2.0
1.0
-
5.0
Simulated
Observed
ENP NRCM
ENP OBS
4.0
3.0
2.0
1.0
-
1
2
3
4
5
6
7
8
9 10 11 12
1
2
3
4
5
Month
6.0
5.0
SIO OBS
4.0
3.0
2.0
1.0
5.0
6.0
SWP NRCM
SWP OBS
4.0
3.0
2.0
1.0
-
1
2
3
4
5
6
7
Month
8
9
10 11 12
Holland,
1 2CCSM
3 4 5 Workshop
6 7 8 9 10 11 0606
12
Month
Number of TCs
SIO NRCM
Number of TCs
Number of TCs
6.0
6
7
8
9 10 11 12
8
9 10 11 12
Month
5.0
NAT NRCM
NAT OBS
4.0
3.0
2.0
1.0
1
2
3
4
5
6
7
Month
ESSL
Annual Tropical Cyclone Statistics
35
20
15
SWIO WAUS
SWP
NWP
ENP
NAT
35
25
20
15
10
SWIO WAUS
SWP
NIO
NWP
ENP
OBS
NRCM
OBS
NRCM
OBS
NRCM
OBS
NRCM
OBS
NRCM
OBS
NRCM
OBS
5
0
SWP
NWP
NAT OBS
ENP OBS
ENP
NAT NRCM
ENP NRCM
WNP OBS
WNP NRCM
NIO OBS
NIO
NAT
Yellow are Observed; Blue are Simulated.
1998
NRCM
Annual Number
30
SWIO WAUS
SWP OBS
OBS
NRCM
OBS
NRCM
OBS
OBS
NIO
NRCM
OBS
OBS
NRCM
0
NRCM
0
NRCM
5
OBS
5
NRCM
10
NIO NRCM
10
SWP NRCM
15
1997
Waus OBS
20
25
Waus NRCM
25
SWIO OBS
Annual Number
30
SWIO NRCM
1996
30
Annual Number
35
SWIO: Southwest Indian Ocean;
WAUS: Western Australia
SWP: Southwest Pacific
NIO: North Indian Ocean;
NWP: Northwest Pacific;
ENP: Eastern North Pacific;
NAT: North Atlantic.
NAT
Holland, CCSM Workshop 0606
ESSL
Tropical Cyclones and Climate Change
1900
1950
2000
30
EA SST Anomaly (C)
25
1
20
0.5
15
10
0
5
Number of Tropical Cyclones
1850
1.5
-0.5
1850
1900
1950
2000
0
Year
Holland, CCSM Workshop 0606
ESSL
AEW Development
Total Tropical Cycones vs East Atlantic SST
Year
1920
1940
1960
1980
2000
2020
8
26.8
7
26.6
6
26.4
5
26.2
4
Est Atlantic SST
TC's Forming <20 Lat
26
R=0.65
9-y Running Mean
25.8
1900
1920
1940
3
Number of Tropical Cyclones
SST (C)
1900
27
1960
1980
2000
2
2020
Year
Holland, CCSM Workshop 0606
ESSL
Importance of AEW Development
6
Equatorial
90%
Category
5
4
92%
3
81%
2
55%
1
36%
0
42%
0
10
20
30
Genesis Latitude
40
50
(Data 1945-2004)
>85% of all Major Hurricanes develop from Easterly Waves!!!
Holland, CCSM Workshop 0606
ESSL
East/West Atlantic and Gulf SSTs
0.60
EA 9-Y
0.40
WA 9-Y
GM 9-Y
0.20
(0.20)
(0.40)
(0.60)
1850
1900
1950
Holland, CCSM Workshop 0606
2000
ESSL
Global Surface Temperature Variability
Volcano
1970
Solar
Sulfate
0.60
EA 9-Y
0.40
SST Anomaly
EPAC
9-Y
There is 0.20
no known
natural
forcing mechanism that
can explain the surface temperature increases since
1960 (Meehl et al 2004, 2006)
Ozone
(0.20)
(0.40)
Notice the
warm 1940-50 period in the “natural cycle”
(0.60)
1850
The NAO
1900
1950
2000
Holland, CCSM Workshop 0606
Year
ESSL
So what is happening?
Compare the record 2005 season with a mean of
1991-1993:
• 2005: 27 Storms, 12 hurricanes, 5 cat 4-5;
• 1991-1993: 22 storms (total), 12 hurricanes, 2 cat 4-5
• 2995 30% of all AEWs became named storms (normally
10%);
• AEWs produced 10 of the 14 hurricanes in 2005, all
category 3-5 hurricanes, all tropical cyclones in July and
August, and 8 of the 11 tropical cyclones in September and
October. Two AEWs also generated two tropical cyclones
each, a rare event that last occurred in 1988;
• For 1991-1993 only one hurricane developed from AEWs,
there were no July storms and only one in October.
Holland, CCSM Workshop 0606
ESSL
2005 vs 1991-1993
Holland, CCSM Workshop 0606
ESSL
NAT SST-Hurricane Relationships
7000 km
Min OLR
SS
S
5
4S1
5 3
54 S
S
Max SST
21S
3
Tropical cyclone formation locations
North Atlantic oceanic and atmospheric response to
global warming.
Holland, CCSM Workshop 0606
ESSL
Vorticity Changes; du/dx<0
July 2005
Holland, CCSM Workshop 0606
ESSL
Assessing Potential Climate Change
Impact on Tropical Cyclones:
Requirements
• Resolve hurricane structure
• Simulate tropical atmospheric response to
oceanic heat changes
• Simulate tropical modes accurately
• Simulate feedback between hurricanes and
ocean/atmospheric circulations
Coupling CCSM/CAM with WRF and a GOM/ROM
enables this
Holland, CCSM Workshop 0606
ESSL
Summary
• The weather and climate scales are part of a
continuum of action across scales;
• We have the tools in CAM/CCSM and WRF to
open up new frontiers in understanding and
predicting these 2-way interactions;
• The Petascale Computing Facility provides the
capacity;
So Lets Do It
Thank You
Holland, CCSM Workshop 0606
ESSL