Earth System Modeling at NCAR

1
Earth System Modeling at NCAR

• Tim Killeen
• Director, National Center for Atmospheric
  Research
• President-Elect, American Geophysical Union
• April 3, 2006

2
National Center for Atmospheric Research

• gt 950 Staff, gt500 Scientists and Engineers
• gt 200M in total funding for FY2005
• 5 Laboratories

• Atmospheric and Earth System Sciences, including
  Climate, Weather, Atmospheric Chemistry,
  Sun-Earth Connections, Society-Environment
  Interactions and Technology Transfer/Decision
  Support
• Computational, data, and observational facilities
  for the NSF and university communities

3
Weather and Extreme Events
4
Weather Research and Forecast (WRF) Model

• Supported by NCAR as a community model for both
  academia and operational centers
• Over 3000 registered users make it the most used
  atmospheric model and provide an unprecedented
  pool of advanced research expertise
• Operational use by the National Weather Service,
  US Navy, US Army, USAF, South Korean
  Meteorological Service, Indian Meteorological
  Department
• Special forecasts are made by NCAR over the
  Antarctic in support of international operations
  there.

5
The Weather Research and Forecast (WRF) Community
Model 4 km Moving-Grid Katrina Forecast
Wind Speed (m/s) Cat 1 33-42
Cat 2 42-50 Cat 3 50-59 Cat
4 60-69 Cat 5 gt69
Initialized 27 Aug 2005 00 Z
6
Experimental Hurricane Analyses
Wind from Numerical Forecast
Likelihood of Wind Damage
Electric Grid Damage
Electric Grid Restoration
Brian Bush, 2005
7
From Scientific Analysis to End-user Information
Systems
GIS Integration
Emergency Response
Coastal Environments
8
Community Climate System Model (CCSM)
9
The CCSM Program

• Scientific Objectives
• Develop a comprehensive climate model to study
  the Earths Climate.
• Investigate seasonal and interannual variability
  in the climate.
• Explore the history of Earths climate.
• Estimate the future of the environment for policy
  formulation.
• Recent Accomplishments
• Release of a new version (CCSM3) to the climate
  community.
• Successful reproduction of ocean mixing and
  heating in the observed record.
• Simulations of last 1000 years, Holocene, and
  Last Glacial Maximum.
• Creation of largest ensemble of simulations for
  the IPCC AR4.

http//www.ccsm.ucar.edu
10
CCSM Working Groups
11
University Developers of CCSM

• Twenty-two domestic institutions currently,
  including
• Carnegie Institution
• Center for Ocean-Land-Atmosphere Studies
• Colorado State University
• Georgia Institute of Technology
• Lamont Doherty Earth Observatory/Columbia
• Massachusetts Institute of Technology
• Princeton University
• Scripps Institution of Oceanography
• State University of New York / Stony Brook
• University of Arizona
• University of California, Berkeley
• University of California, Irvine
• University of California, Los Angeles
• University of California, Santa Cruz
• University of Colorado
• University of Illinois
• University of Kansas
• University of Miami

Map of our University Development Partners
NCAR
12
CCSM PUBLICATIONS
CCSM IJHPCA Special Issue Defining software
engineering 13 papers, 2005
CCSM J. Climate Special Issue

• 26 papers accepted
• SSC (1)
• Atmosphere Model (6)
• Ocean Model (4)
• Land Model (2)
• Climate Change (2)
• Climate Variability (4)
• Polar Climate(4)
• Paleoclimate (3)

13
Contributions to IPCC Fourth Assessment

• Output 10 GB/simulated year
• Data volume for IPCC 100 TB
• Largest data volume contribution
• Eight ensemble members at T85
• Data available online
• PCMDI (IPCC archive)
• ESG (Control)
• Original history tapes NCAR/SCD
• Diagnostics on line (web)

14
CLIMATE SENSITIVITY UNCERTAINTIES
Wigley
15
CCSM-3 SEA LEVEL RISE PROJECTIONS2000-2400
16
Observed
Climate models predict heat waves Heat wave
severity defined as the mean annual 3-day warmest
nighttime minima event Model compares favorably
with present-day heat wave severity Heat waves
will become more severe in southern and western
North America, and in the western European and
Mediterranean region
Model
Future
Meehl, G.A., and C. Tebaldi, 2004
17
NCAR Paleoclimate Model
18
(No Transcript)
19
Global Annual Mean Energy Budget
Permian coupled model run for 2700 years to
new equilibrium state
Forcing of 10X increase in CO2 and
Permian paleogeography
Kiehl and Shields
CCSM3 T31X3
Global Annual Mean Surface Temperature
??Tsgt 8C
20
Inefficient mixing in Permian ocean
indicative of anoxia, consistent with large
extinction event
Kiehl and Shields (2005)
21
Greenland ice sheet

• Volume 2.8 million km3
• (7 m sea level equivalent)
• Area 1.7 million km2
• Mean thickness 1.6 km
• Accumulation 500 km3/yr
• Surface runoff 300 km3/yr
• Iceberg calving 200 km3/yr

Annual accumulation (Bales et al., 2001)
Bill Lipscomb, LANL
22
(No Transcript)
23
CCSM Average Aug Sea Ice
a) Present
b) Last Interglacial
Sea level at last interglacial was several
meters above present
Otto-Bliesner et al., Science, 2006
24
Need for High Resolution
25
Data visualization and exploration in ArcGlobe
GIS Climate Change Scenarios
100 year (2000-2099) change in mean July surface
temperature
26
Towards an Earth System Model
27
(No Transcript)
28
Ongoing Activities

• Include biogeochemistry and ocean ecosystem model
  for the carbon/nitrogen cycle
• Include dynamic vegetation and land use changes
  in the land component
• Include both the direct and indirect effects of
  aerosols and land ice sheet model
• Atmospheric chemistry component has been added
  include the effects of tropospheric ozone
• Different dynamical cores, nested regional
  climate models

29
Nested Regional Climate Models
30
Science Driven Demand for Supercomputing
31

3/1/03 Over One Petabyte
11/1/04 Over Two Petabytes
32
Sustained GFLOPs at NCAR mid-2007
33
A Vision for the Geosciences

• Establish a Petascale Collaboratory for the
  Geosciences with the mission to provide
  leadership class computational resources that
  will make it possible to address, and minimize
  the time to solution of, the most challenging
  problems facing the geosciences.

• Science requirements have been detailed
• Geosciences is ready for transformative
  capabilities, via access to Petascale computing
  not just for atmospheric sciences but for all of
  Geosciences

34
Birds Eye View
35
Petascale Geosciences Collaboratory
Other National Labs and International Centers
NSF Geosciences Research Community ATM, OCE,
EAR
EAR Geoscience Resource Center
NSF Geosciences Collaboratory Center
OCE Geoscience Resource Center
ATM Geoscience Resource Center
NCAR Facility Partner Somewhere near Boulder
OCE Geoscience Resource Center
EAR Geoscience Resource Center
ATM Geoscience Resource Center
36
The Earth System Grid (ESG)
http//www.earthsystemgrid.org
37
Earth System Modeling Framework Rollout
DAO fvCAM
14 major Earth system codes Each is coupled to
a code never coupled to before - 3 by July
03 - all 14 by July 04
DAO analysis
GFDL FMS B-grid atmosphere
NASA
NASA
NSIPP atmosphere
NSF
NASA
GFDL MOM4 ocean
NSF
NSIPP ocean
Earth System Modeling Framework Development
NASA
GFDL POM or MOM4
NSF
LANL CICE
DOE
NCAR CAM
NSF
MITgcm ocean
MIT
NCEP/NCAR WRF
NOAA NSF
Early adopters of the ESMF
CLM
?
NCEP atmosphere
NOAA
8/04
GSFC Global LIS
NOAA
Early
NCEP analysis
NASA
NOAA
NCEP model
UCLA AGCM
UCLA
All JMC validation codes compliant - April
04
11/04
Adoption
LANL HYPOP
DOE
1/05
- Broad use - Enhancement - Maintenance
Unprecedented software sharing ease among the
nations major Earth system models
One of GISS, COLA, IRI, JPL, LLNL, Colorado
State, U.Illinois, Scripps, U.Miami, NOAA FSL,
Florida State, Rutgers, ORNL, Air Force Weather
Agency, U.Washington
ESTO Earth Science Technology Office
Jim Fisher
38
1st generation Earth System Model
Coupler
Atmosphere
Ocean
Sea Ice
Land
C/N Cycle
Dyn. Veg.
Ecosystem BGC
Land Use
Gas chem.
Prognostic Aerosols
Upper Atm.
39
Closing Comments

• Earth System Sciences is coming of age
  throughboth intellectual and infrastructural
  advances. It is clear that not that many
  high-capability Earth System Modeling centers can
  be afforded worldwide, with all the needed
  features
• Community earth system model suites
• Grid computing
• Software frameworks
• Design for Education
• Data systems
• Visualization, data assimilation
• Open source approaches
• A World-scale Collaboratory
• It may be time for significant international
  collaboration in building and analyzing these
  complex earth system models

40
Thank you!
41
Dynamic, Intelligent Forecast System(DICast)
Combines multiple models and observational data
to generate a tuned site forecast for each
parameter at each forecast lead time.
42
DICast Performance
Developed originally for The Weather Channel in
1999 and improved since, DICast routinely
outperforms NWS guidance forecasts
Climatology
NWS Guidance
DICast
DICAST
RMS Error oC
18 U.S. Cities Max Temp January 2002
Forecast lead time (days)
43
Natural forcings do not fully explain observed
late 20th century warming

• NCAR Climate model with only natural forcings
  (volcanic and solar) does not reproduce observed
  late 20th century warming
• When increases in anthropogenic greenhouse gases
  and sulfate aerosols are included, the model is
  able to reproduce observed late 20th century
  warming

Meehl, G.A., W.M. Washington, C. Ammann, J.M.
Arblaster, T.M.L. Wigley, and C. Tebaldi, 2004
Combinations of natural and anthropogenic
forcings and 20th century climate. J. Climate,
17, 3721-3727.
44
Intergovernmental Panel on Climate Change (IPCC)
Fourth Assessment Report

• NCAR Community Climate System Model (CCSM-3).
• Open Source
• 8-member ensembles
• 11,000 model years simulated
• T85 - high resolution

45
Changes in frost days in the late 20th century
show biggest decreases over the western and
southwestern U.S. in observations and the model
46
WRF 36 km res global channel Precipitation in
1996 CAM3 at T170 res
47
NINO3 spectra Black obs, blue control, red
new
48
Multi-Century Coupled Carbon/Climate Simulations
2.0
14.1
13.6
-2.0
Surface Temp.
Net CO2 Flux (Pg C/yr)
1000
0
0
1000
year
year

• Fully prognostic land/ocn BGC and
  carbon/radiation
• Atm-Land CO2 flux 70 PgC/yr ?? Atm-Ocean CO2
  flux 90 PgC/yr ??
• Net Landocean CO2 flux 0?1 PgC/yr
• Stable carbon cycle and climate over 1000 years

Doney, Lindsay, Fung and John Accepted by J
Climate
49
Natural Disaster Prevention and Mitigation

• During 1990s natural disasters (floods,
  droughts, earthquakes, storms, strong winds,
  torrential rains, and mudslides) hit the world
  500-800 times a year and cost more than 600
  billion (more than previous four decades
  combined)
• 2 billion people affected gt400,000 killed

50
Why?

• More people
• Ecosystems are frayed (deforestation and wetland
  destruction, soil erosion)
• People migrate to disaster-prone regions
• More intense hydrological cycle with global
  climate change
• More than 50 live in cities (many new city
  dwellers live on vulnerable hillsides and
  floodplains)

51
By the year 2050

• Perhaps 9 billion people
• gt6 billion tons per year of greenhouse gases
• gt60 million tons per year of urban pollutants
• Withdrawing 30 of available fresh water
• Converting 65 of frontier forests
• 80 of people in cities
• 25 near earthquake faults
• 2 within 1meter of mean sea level
• So, what is the role of the geosciences?

52
Private Sector and Industry too?YES - Acquired
knowledge must be transitioned to support modern
society
53
Towards a Global Weather Research Forecast Model
(WRF)
54
Ensemble Forecasts

• Every six hours these forecasts will be shared
  and grouped to form an ensemble of 200 forecasts.

GWS2025-Anthes, 1999
55
Projections for Global Surface Temperature
290
288
286
Meehl et al, 2005
56
Changes in Sea Ice Coverage
Meehl et al, 2005
57
NCAR/Ohio State Operational Antarctic Forecasts
The forecast was provided in time for BBV to get
out and avoid damage to the aircraft, for which
we are most grateful!
We thank you for providing support and
weather forecasts to our remote tent camp of 6
persons during the 2005-06 season. These
forecasts increased our preparedness for severe
weather and did much to improve our overall
safety in the field.
Christine Siddoway PI, G-088
Nov 15, 2005, Fosdick Mountains - last chance
to leave
58
(No Transcript)
59
Horizontal Discretization of Equations
T31
T42
T85
T170
Strand
60
Observing and Modeling the Earth System
The GLOBAL C CYCLE
The GLOBAL N CYCLE
61
(No Transcript)
62
SimulatingtheGlobal Earth System
Atmosphere
Hydrosphere
Cryosphere
Biosphere