Climate modelling uncertainties

1
CLIMATE MODELLING STRATEGIES FOR REDUCING THE
UNCERTAINTIES A. Sorteberg
2
A NATIONAL CENTER OF EXCELLENCE
Institute of Marine Research

• University of Bergen
• Geophysical Institute
• Dept. of Earth Science
• Dept. of Geography
• Dept. of Botany

Nansen Environmental and Remote Sensing Center
3

• EXPERISE
• CLIMATE MODELLING
  PALEOCLIMATOLOGY
• PHYSICAL OCEANOGRAPHY CHEMICAL OCEANOGRAPHY
  

4

MAIN SCIENTIFIC QUESTIONS

• WHAT CAUSES CLIMATE VARIABILITY IN THE NORTH
  ATLANTIC AND ARCTIC REGIONS? TO WHAT EXTENT ARE
  THEY PREDICTABLE ?
• WHICH PROCESSES DRIVE PAST, PRESENT AND FUTURE
  CLIMATIC CHANGES ? HOW CAN WE DISTINGUISH BETWEEN
  NATURAL VARIABILITY AND IMPACT OF HUMAN
  ACTIVITIES ?
• WHAT CAUSES RAPID CLIMATE CHANGES AND WHAT ROLE
  DOES THE OCEAN CIRCULATION PLAY IN THESE CHANGES
  

5

5 SCIENTIFIC GROUPS

• Coupled ocean-ice-atmosphere modelling and future
  climate
• Bergen Climate Model (BCM)
• Climate processes and identification of ongoing
  climate changes
• Process modelling Atmosphere, ocean, ice
• Ocean observations
• The fate of the oceanic carbon sink/biogeochemical
  cycles
• Ocean observations
• Modelling

6

5 SCIENTIFIC GROUPS

• Multi-decadal to century scale climate
  variability
• Paleoclimatic reconstructions
• Abrupt climate change
• Paleoclimatic modelling

7
UNCERTAINTIES RELATED TO CLIMATE MODELLING
TEMPERATURE LAST 1000 YEARS
8
A hierarchy of uncertainties
MAGNITUDE OF EXTERNAL FORCING
SOLAR VARIABILITY/VOLCANOES SOCIO-ECONOMIC MODELS
MAGNITUDE OF CLIMATE CHANGE
CLIMATE MODELS
MAGNITUDE OF ENVIRONMENTAL CHANGE AND ITS
IMPACT
IMPACT MODELS
9
CLIMATE MODEL UNCERTAINTIES

• GLOBAL SYSTEM UNPREDICTABILITY
• UNCERTAINTIES RELATED THE EXTERNAL FORCINGS
• MODEL DEFICIENCIES
• SHARED DEFICIENCIES DUE TO CURRENT LEVEL
• OF SCIENTIFIC KNOWLEDGE
• MODEL FORMULATIONS
• RESOLUTION
• LEVEL OF COMPLEXITY
• CLIMATE SYSTEM UNPREDICTABILITY
• UNCERTAINTIES RELATED TO INTERNAL CLIMATE
• VARIABILITY

10
CLIMATE SYSTEM UNPREDICTABILITY

• THE CLIMATE SYSTEM BEARS THE CHARACTERISTICS OF A
  CHAOTIC SYSTEM.
• SIMULATIONS WITH SLIGHTLY DIFFERENT INITIAL
  CONDITIONS WOULD EXHIBIT DIFFERENT
  CHARACTERISTICS IN EACH EVOLUTION WHICH MAY BE
  DESCRIBED IN TERMS OF A FREQUENCY DISTRIBUTION OF
  DIFFERENT OUTCOMES
• THE MODELS DIVERGENCE FROM A SINGLE SOLUTION CAN
  THEREFORE BE SEEN AS A MANIFESTATION OF BOTH
• REAL INTERMODEL DIFFERENCES
• THE FACT THAT THE MODEL SPREAD ARE REPRESENTING
  THE FREQUENCY DISTRIBUTION OF THE CHAOTIC
  BEHAVIOUR OF THE CLIMATE SYSTEM

11
INTERMODEL DIFFERENCES
UNCERTAINTIES RELATED TO MODEL SENSITIVITY TO
CO2 CHANGE
1 INCREASE IN CO2 per YEAR
?T 2CO2 1Cº
?T 1.5CO2 0.5Cº
IPCC, 2001
12
INTERMODEL DIFFERENCES
UNCERTAINTIES RELATED TO MODEL SENSITIVITY TO
CO2 CHANGE
BLUE LINE RANGE ?T 2CO2
MODEL DEFICIENCIES ? CLIMATE SYSTEM
UNPREDICTABILITY ?
IPCC, 2001
13
STRATEGIES FOR DEALING WITH CLIMATE SYSTEM
UNPREDICTABILITY
THE USE OF ENSEMBLE SIMULATIONS

• MULTIMODEL ENSEMBLES
• REPRESENT A PROBABILITY SPACE OF CLIMATE CHANGE
  OUTCOMES
• ENSEMBLES USING THE SAME MODEL
• REPRESENTS THE SPREAD IN CLIMATE CHANGE OUTCOMES
  DUE TO THE CHAOTIC NATURE OF THE SYSTEM

• COMPUTATIONALLY EXPENSIVE !

14
STRATEGIES FOR DEALING WITH CLIMATE SYSTEM
UNPREDICTABILITY
THE IMPORTANCE OF ACCOUNTING FOR NATURAL
VARIABILITY

• CONTROL INTEGRATIONS OVER SEVERAL CENTURIES TO
  MAP INTERNAL SYSTEM VARIABILITY
• ENSEMBLES INCREASES THE SIGNAL TO NOISE RATIO
  WITH vn (nNUMBER OF SIMULATIONS)
• IF THE ERROR FOR DIFFERENT MODELS IS RANDOM WITH
  ZERO MEAN, THE ENSEMBLE MEAN WILL PROVIDE THE
  BEST ESTIMATE OF THE SIGNAL

15
STRATEGIES FOR DEALING WITH CLIMATE SYSTEM
UNPREDICTABILITY
THE IMPORTANCE OF ACCOUNTING FOR NATURAL
VARIABILITY
CLIMATE CHANGE ANTROPOGHENIC NATURAL

SIGNAL NOISE
CHANGE MEAN OVER YEAR 31-60
BCM ENSEMBLE MEMBERS
CMIP2 MODELS
16
STRATEGIES FOR DEALING WITH CLIMATE SYSTEM
UNPREDICTABILITY
THE IMPORTANCE OF ACCOUNTING FOR NATURAL
VARIABILITY
ZONAL MEAN TEMPERATURE TREND (ºC/DECADE) OVER
YEAR 31-60
ZONAL MEAN TEMPERATURE TREND (º C/DECADE) OVER
YEAR 1-80
17
CLIMATE SYSTEM UNPREDICTABILITY AN EXAMPLE
THE FATE OF THE THERMOHALINE CIRCULATION UNDER
CLOBAL WARMING AND ITS IMPACT ON THE SIMULATION
OF CLIMATE CHANGE
ANNUAL TEMPERATURE DEVIATION FROM ZONAL MEAN
THE THERMOHALINE CIRCULATION
18
CLIMATE SYSTEM UNPREDICTABILITY AN EXAMPLE
THE FATE OF THE THERMOHALINE CIRCULATION UNDER
CLOBAL WARMING AND ITS IMPACT ON THE PROCJECTED
CLIMATE CHANGE
0.5 increase in CO2 per year up to 750 ppm
1 increase in CO2 per year up to 750 ppm
Stocker and Schmittner, 1997
19
CLIMATE SYSTEM UNPREDICTABILITY AN EXAMPLE
CHANGE IN ANNUAL TEMPERATURES 20-30 YEARS AFTER
THC COLAPSE
-3 to -8ºC
M. Vellinga, Hadley Center, 2001
20
MODEL DEFICIENCIES
LEVEL OF COMPLEXITYFROM ATMOSPHERE TO EARTH
SYSTEM MODELS
Mid-1970s Mid 1980s Early 1990s Mid
1990s Mid 2000
21
MODEL DEFICIENCIES
MODEL RESOLUTION
THE INCREASE IN COMPUTER POWER
MULTI CENTURY SIMULATIONS
Early 1990s Mid 1990s Late 1990s
HORIZONTAL RESOLUTION 4.5º x 7.5º 4º
x 5º 2.5º x 2.5º VERTICAL
RESOLUTION 9
15 30

• GRIDSQUARE AREA REDUCED BY A FACTOR OF 4-5
• VERTICAL RESOLUTION INCREASED BY A FACTOR OF 3

22
SUMMARY

• UNCERTAINTIES RELATED TO CLIMATE CHANGE
  PROJECTIONS ARE RELATED TO
• OUR SCIENTIFIC KNOWLEDGE
• MODEL COMPLEXITY
• THE INHERENT CHAOTIC BEHAVIOUR OF BOTH THE
  CLIMATE AND THE GLOBAL SYSTEM
• THE CLIMATE CHANGE PROJECTIONS SHOULD THEREFORE
  BE DESCRIBED IN TERMS OF A FREQUENCY DISTRIBUTION
  OF DIFFERENT OUTCOMES
• MAIN PATHWAYS TO REDUCE THE MODELLING
  UNCERTAINTIES ARE
• INCREASED SCIENTIFIC UNDERSTANDING
• COMPLEXITY OF THE MODELS
• MULTI MODEL ENSEMBLE SIMULATIONS
• SINGLE MODEL ENSEMBLE SIMULATIONS
• MULTI SCENARIO SIMULATIONS

23
(No Transcript)