Representing Uncertainties & Selecting Scenarios

download report

Transcript Representing Uncertainties & Selecting Scenarios

Can you gradually fall off a cliff?
– A glimpse at complex, self-organising systems
AIACC Training Workshop on Adaptation and Vulnerability
TWAS, Trieste
June 3-14 2002
Roger N. Jones
Atmospheric Research
The awakening of complexity
Mechanical world of
the 19th century
• Few mysteries left
for science to solve
• The universal
machine
• Mastery over
nature
• The march of
progress
• Modernism
The 20th century – transition
• Quantum physics
• Chaos and “strange
attractors”
• isolated “frame of
reference” exposed as a
scientific construct
• search for a process to
counterbalance
reductionism
• Postmodernism
Atmospheric Research
Contributions to complex systems
science
•
•
•
•
Adam Smith “The invisible hand”
Einstein, Bohr, Pauling et al. – quantum physics
Schroedinger et al. – uncertainty
Turing and Von Neumann – self-replicating automata and
game theory
• Kuhn – the scientific process is linked to social processes
• Prigogine – complex chemistry
• Lorenz, Gleick et al. – chaos
• Holland, Conway et al. – artificial life
• Bak et al. – self organising systems
• Arthur – law of increasing returns (economics)
• Capra – role of eastern philosophy
and many others
Atmospheric Research
Simple system
•
•
•
•
•
Mechanistic
Replicable
Largely linear
Can be isolated from other systems
Predictable
Atmospheric Research
Complex system
• Organic/chaotic (often described as on the
edge of chaos because both organised and
chaotic behaviour are recognised)
• Non-replicable
• Cannot be isolated from other systems
• Non-linearity and thresholds both common
• Self-organising (self-adapting)
• Bifurcations occur over time
• Uncertainty is intrinsic
Atmospheric Research
Examples
•
•
•
•
Qwertyuiop
VHS/Beta
DOS/CPM
Extinctions/radiation (evolution)
Atmospheric Research
Fractal patterns are “natural”
Atmospheric Research
Glacial cycles are driven by changes in
the Earth’s orbit
Carbon dioxide and temperature
last 420, 000 years
20
300
250
10
200
0
150
-10
400,000
CO 2 (ppm)
temperature (oC)
350
100
300,000
200,000
100,000
0
years before present
Atmospheric Research
Holocene rainfall and evaporation
– W. Victoria
Atmospheric Research
Atmospheric Research
Weather events
Atmospheric Research
CO2 emissions and concentrations
Atmospheric Research
Global warming
Atmospheric Research
Likelihood
Probability can be expressed in two
ways:
1. Return period / frequency-based
(Climate variability)
2. Single event
(Mean climate change, one-off events)
Atmospheric Research
Return period / frequency-based
probability
Recurrent or simple event
Where a continuous variable reaches a critical level, or
threshold.
Eg. Extreme temperature (max & min), Extreme rainfall,
heat stress, 1 in 100 year flood
Discrete or complex event
An event caused by a combination of variables (an
extreme weather event)
Eg. tropical cyclone/hurricane/typhoon, ENSO event
Atmospheric Research
Frequency-based probability
distributions
Atmospheric Research
Coping range under current climate
Stationary Climate
& Coping Range
Vulnerable
Coping
Range
Vulnerable
Atmospheric Research
Thresholds
A non-linear change in a measure or
system, signalling a physical or
behavioural change
Climate-related thresholds are used to
mark a level of hazard
Atmospheric Research
Single-event probability
Singular or unique event
An event likely to occur once only. Probability refers to
the chance of an event occurring, or to a particular
state of that event when it occurs.
Eg. Climate change, collapse of the West Antarctic Ice
Sheet, hell freezing over
Atmospheric Research
What is the probability of climate
change?
1. Will climate change happen?
•
IPCC (2001) suggests that climate change is occurring with
a confidence of 66% to 90%
2. What form will it take?
Uncertainties are due to:
• future rates of greenhouse gas emissions
• sensitivity of global climate to greenhouse gases
• regional variations in climate
• decadal-scale variability
• changes to short-term variability
Atmospheric Research
Range of uncertainty
M1
UNQUANTIFIABLE
UNCERTAINTY
M2
M3
M4
QUANTIFIABLE RANGE OF UNCERTAINTY
UNQUANTIFIABLE
UNCERTAINTY
TOTAL RANGE OF UNCERTAINTY
Atmospheric Research
100
100
80
50 cm
25 cm
Sea Level Rise (cm)
Sea Level Rise (cm)
75 cm
80
75 cm
60
50 cm
40
25 cm
20
75 cm
60
50 cm
40
0
0
0
100
0
Probability (%)
75 cm
60
50 cm
40
25 cm
20
0
80
75 cm
60
50 cm
40
25 cm
20
0
0
5
10
Probability (%)
100
80
75 cm
60
50 cm
40
25 cm
20
0
0
100
Probability (%)
Sea Level Rise (cm)
80
100
Probability (%)
100
Sea Level Rise (cm)
100
Sea Level Rise (cm)
Sea Level Rise (cm)
100
25 cm
20
80
75 cm
60
50 cm
40
25 cm
20
0
0
5
10
Probability (%)
0
100
Probability (%)
Non-linear climate change
• Non-linear climate events - ice ages,
Younger Dryas, collapse of the WAIS
• Climate surprises - climate events that
occur unexpectedly
• Climate surprises are likely to occur
on a regional basis under climate
change but when and where remains
unknown.
Atmospheric Research
System responses
• Resistance (e.g. seawall)
• Resilience (e.g. regrowth, rebuilding after
storm or fire)
• Adaptation (adjustments made in response to
stress)
• Transformation (old system stops, new one
starts)
• Cessation (activity stops altogether)
Atmospheric Research
Can you gradually fall off a
cliff?
Yes, if you use a model
But not in the real world
Atmospheric Research