Climate Scenarios in Vulnerability, Impact and Adaptation Assessments: an overview

1
Climate Scenarios in Vulnerability, Impact and
Adaptation Assessments an overview AIACC
Scenarios Training Course Norwich, 16-25 April
2002 Dr Mike Hulme
2
Norwich
What we might ideally like . daily weather,
for a place, for now and for a future year
3
Why does creating climate scenarios give us so
many problems?

• Problem 1. Models are not accurate .
• so we cannot use data from climate models
  directly in environmental or social simulation
  models

4
Why does creating climate scenarios give us so
many problems?

• Problem 2. Different climate models give
  different results
• so we have difficulty knowing which climate
  model(s) to use

5
Why does creating climate scenarios give us so
many problems?

• Problem 3. It is expensive to run many
  (global/regional) climate model experiments for
  many future emissions .
• . so we often have to make choices about which
  emissions scenarios from which we build our
  climate scenarios

6
Why does creating climate scenarios give us so
many problems?

• Problem 4. Climate models give us results at the
  wrong spatial scale
• so we have to develop and apply one or more
  downscaling methods.

7
Our problems would be much easier if .

• Climate models were fully accurate
• Different climate models gave the same results
• One could run a GCM experiment over 200
  simulated years in one day on a PC
• Climate models had a resolution of 1km

But they dont!
8
Be clear about what you need ..

•  How many scenarios do you want? Which
  uncertainties are you going to explore?
•  What non-climate information do you need in
  your scenario(s)?
• Do you need local data for case studies/sites,
  or national/regional coverage?
•  What spatial resolution do you really need
  300k, 100k, 50k, 10k, 1k? Can you justify this
  choice?
•  Do you need changes in average climate, or in
  variability?
•  Do you need changes in daily weather, or just
  monthly totals?
•  What climate variables are essential for your
  study?

9
A framework for conducting integrated assessment
of climate change for policy applications
NB. this has a UK interpretation
10
Historical climate data . necessary as a
baseline and also to explore historical/current
vulnerability
11
The four IPCCSRES storylines
a major international effort to construct an
overarching framework for thinking about the
future with regard to emissions of greenhouse
gases .. global and continental rather than
national and local.
12
Downscale SRES socio-economic scenarios
13
or create socio-economic scenarios
bottom-up for local communities or regions
14
Incremental Scenarios for Sensitivity Analysis
Advantages easy to construct and apply, allows
sensitivity of sectors/models to be explored
Disadvantages arbitrary (and unrealistic)
changes, not related to wider scenario frameworks
15
Scenarios from Global Climate Model Experiments
Advantages easily accessible, numerous model
runs, global in scale, numerous variables
Disadvantages coarse resolution (300km), daily
extremes poorly represented
16
WINTER PRECIPITATION FROM 9 CLIMATE MODELS

17
Overcoming problem 2 (model differences) -
vintage - validation - credibility -
resolution - accessibility - politics!
18
Problem 4 Spatial scale downscaling options
19
Scenarios from Regional Climate Model Experiments
Advantages higher resolution (50km), local
geography well represented, daily weather
extremes more realistic
Disadvantages few runs available, can be
time-consuming to run, not good for representing
uncertainties in risk assessment
20
Spatial scale may still be a problem (problem
4)
21
UKCIP 1998 GLOBAL MODEL
UKCIP 2002 REGIONAL MODEL
50km grid
300km grid
22
The model will still not be accurate (problem 1)
23
WINTER PRECIPITATION OVER BRITAIN
24
Overcoming problem 1 (accuracy)

• Use raw GCM/RCM outputs (rarely done)

• Add model-derived changes to an observed baseline
• mean changes (common)
• inter-annual changes (less common)

• Calibrate and perturb a weather generator Rob
  Wilby

• Applying an empirical downscaling method Bruce
  Hewitson

25
- Simple interpolation of model changes (300k,
50k, 14k) onto an observed climatology (1k,
timeseries) - Weather generators to places
or catchments or grids - Statistical
downscaling
Overcoming problem 4 (scale) (may have to be
tackled whether using GCMs or RCMs)
26
Simple interpolation combining observed data
with modelled changes
27
Scenarios from Weather Generators
Advantages site or locality specific scenarios,
long and multiple daily weather sequences produced
Disadvantages requires a lot of historic data
to calibrate, based on empirical relationships
which may change, climate model data not always
available
28
Representing uncertainties (especially emissions
uncertainties problem 3) remains an issue
29
Scenarios from Climate Scenario Generators
Advantages easily to explore uncertainties,
multiple integrated scenarios, accessible
Disadvantages coarse resolution (300km), no
daily data, not readily updated
30
MAGICC/SCENGEN framework
31
Probabilistic Scenarios for Risk Assessment
Advantages makes (some) uncertainties explicit,
good for risk assessment, can be applied at
different scales
Disadvantages not yet a well developed
methodology, requires a lot of model data to
develop, expert assumptions still needed
32
UKCIP02 scenario strategy
33

• How will you link climate and non-climate
  scenarios?

Decided to link SRES futures with UK climate
scenarios We chose A1FI, B2, A2 and A1 1-to-1
mapping of climate and non-climate scenarios
34

• What non-climate information is needed in your
  project?

Non-climatic socio-economic indicators for
the UK have been produced by the UK Climate
Impacts Programme for each of the four SRES
storylines
35

• What types of uncertainties are critical to your
  project?

Both emissions and modelling uncertainties are
important Our strategy was to explicitly quantify
the emissions uncertainty (4 different emissions
spanning the IPCC range), but only to provide
general guidance about the relative importance of
modelling uncertainty
36

• What climatic variables are required for I, A V
  assessments in your project?

A wide range - T, P, SL, CO2, RH, snow, cloud,
etc., surface rather than upper air, however We
aimed to produce generic climate scenarios for
many different applications (UKCIP, government
policy, academic research, public awareness,
etc.)
37

• At what spatial and temporal scales are these
  variables required?

We decided that we must have information at 50km
resolution We needed to analyse both monthly and
daily data
38

• What baseline climate data are you planning to
  use?

A 5km gridded, dataset for UK 26 surface climate
variables Monthly series for 1961-2000
39

• Which project(s) in your region you envisage you
  will be able to collaborate with to develop
  climate scenarios?

We worked with the government (Ministry of
Environment), climate modellers (Hadley Centre)
and users (through the UK Climate Impacts
Programme)
40
(No Transcript)
41
(No Transcript)
42
(No Transcript)
43
Designing climate scenarios is largely an
exercise in handling uncertainties
Source Hadley Centre
44
Purpose(s) of (climate) scenarios .. To make
predictions of the future wrong

• To provide data for impact/adaptation/assessment
  studies
• To act as an awareness-raising device
• To aid strategic planning and/or policy formation
  
• To scope the range of plausible futures
• To structure our knowledge (or ignorance) of the
  future
• To explore the implications of decisions
• To function as learning-machines, bridging
  analyses and participation

45
Temperature and precipitation effects of natural
variability
46
(No Transcript)