Transcript jones_regclimmod - START - SysTem for Analysis Research and

PRECIS
Providing REgional Climates for Impact Studies
The Hadley Centre regional climate modelling system
A presentation to the Tyndall Centre AIACC training workshop
by
Richard Jones and Simon Wilson
23 April 2002
Hadley Centre for Climate Prediction and Research
Outline

Background: Motivation and PRECIS overview

PRECIS components and how to operate them

Future climate scenarios available from PRECIS

Applying PRECIS in impacts assessment

Limitations of PRECIS and future developments
Hadley Centre for Climate Prediction and Research
Motivation

UNFCCC requirement to assess national vulnerability and plans for
adaptation

UNFCCC requirement to submit National Communications

Both need estimate of impacts

Impacts need scenarios of future climate
Hadley Centre for Climate Prediction and Research
Predicting Impacts
Hadley Centre for Climate Prediction and Research
Advantages of a Regional Model

General circulation models
(GCMs)

Regional climate models (RCMs)
– local detail
– large-scale patterns
– runs on a PC
– large computer
– realistic extreme events
– no skill in extremes
– represent smaller islands
Hadley Centre for Climate Prediction and Research
GCM versus RCM
Mediterranean temperature changes in summer
GCM
RCM
3
4
5
6
Hadley Centre for Climate Prediction and Research
7 8
Temperature °C
Regional Climate Model
• 50km resolution
•19 vertical levels
• limited area
Hadley Centre for Climate Prediction and Research
What is PRECIS?

Providing REgional Climates for Impact Studies

Regional climate modelling (RCM) system that can be applied to any area of
the globe

Generate detailed climate predictions
Hadley Centre for Climate Prediction and Research
Why develop PRECIS?

Need of all countries for detailed climate change scenarios to
assess vulnerability.

A PC version of the RCM will enable developing countries to
generate their own national scenarios

Capacity building

Technology transfer
Hadley Centre for Climate Prediction and Research
Who funds PRECIS?
DEFRA
DFID
UNDP
Hadley
Cent re' s
RCM
PC version
of t he RCM
Training
m at erial
Hadley Centre for Climate Prediction and Research
PRECIS components
PRECIS
USER Interface
Regional Model
Domain
Scenario
Period
Diagnostics
Run RCM
Hadley Centre for Climate Prediction and Research
Graphic package
How will PRECIS be supplied?
On DVD containing

RCM source code and executables

Software to:
– set up and monitor
– archive data
– process generated data
– display processed data
– extract data

Guidance material
Hadley Centre for Climate Prediction and Research
Boundary conditions

Requests through
– PRECIS website
– helpdesk
Hadley Centre for Climate Prediction and Research
Support and follow-up

Helpdesk
– voice
– web-based

Updates

Exchange of data and methods
Hadley Centre for Climate Prediction and Research
PC requirements to run PRECIS
Processor Operating Memory Disk- Mass Fortran
system
space Store compiler
Minimum
1GHz
Desirable
Fastest
possible
Suse Linux 512Mb
Red Hat
Linux
Hadley Centre for Climate Prediction and Research
1Gb
60Gb
DAT
tape
100Gb DLT Processor
tape optimised
PRECIS components and their use

User-interface to set up RCM experiments

The latest Hadley Centre RCM

Data-processing and graphics software

Boundary conditions from latest Hadley Centre GCM

Training course and materials
Hadley Centre for Climate Prediction and Research
PRECIS: THE USER INTERFACE
Hadley Centre for Climate Prediction and Research
PRECIS: SETTING UP THE DOMAIN
Hadley Centre for Climate Prediction and Research
The PRECIS RCM

Climate model: comprehensive representation of physical processes in
atmosphere and on land
– dynamics - atmospheric circulations, cyclones, fronts
– radiation - effects of greenhouse gases and aerosols
– clouds - radiative effects, sulphate aerosol effects
– precipitation - convection, large-scale condensation
– land-surface - soil hydrology (4 levels), vegetation

Horizontal grid - 50km standard, set by user

19 vertical levels - variable thickness, 1st is 50m

Limited domain - typically 5000 x 5000km

Driven by GCM or observed boundary contitions
Hadley Centre for Climate Prediction and Research
A TYPICAL DOMAIN OF AN RCM
Hadley Centre for Climate Prediction and Research
Schematic representation
of the PRECIS RCM applied
over SE Asia
Hadley Centre for Climate Prediction and Research
Data-processing and graphics
(CDAT- available freely from PCMDI)

Data display system:
– to monitor ongoing experiments
– to display (processed) results of experiments

Data processing system:
– to calculate climate statistics
– to calculate changes in climate
– to form climate scenarios
Hadley Centre for Climate Prediction and Research
PRECIS: MONITORING THE RCM
Hadley Centre for Climate Prediction and Research
SIMULATION OF A TROPICAL CYCLONE
in the PRECIS regional climate model
Global climate model
Regional climate model
Hadley Centre for Climate Prediction and Research
WINTER DAILY RAINFALL OVER THE ALPS
Hadley Centre for Climate Prediction and Research
CHANGES IN WINTER PRECIPITATION
between present day and the 2080s due to A2 emissions.
Global climate model
Regional climate model
Hadley Centre for Climate Prediction and Research
SOUTHERN AFRICA SUMMER RAINFALL Change by the 2080s
relative to the present-day
PRECIS
scenario:
percentage
reduction
in future
rainfall
Hadley Centre for Climate Prediction and Research
Training course and supporting material

Training will be provided to maximise the usefulness of PRECIS and will focus on:
– Interpretation of PRECIS results by regional experts
– Construction of regional climate change scenarios
– Building capacity in countries/regions using PRECIS

PRECIS will be supplied with:
– a workbook covering the background science, system description and use as presented in the training
– a technical manual explaining technical details about the system and how to use it, again as in the
training
Hadley Centre for Climate Prediction and Research
Future climate scenarios available from PRECIS

PRECIS can provide:
– climate scenarios for any region
– an estimate of uncertainty due to different emissions
– an estimate of uncertainty due to climate variability

Data available from PRECIS
– comprehensive for atmosphere and land-surface
– grid-scale box average quantities
– maximum time resolution 5 minutes
Hadley Centre for Climate Prediction and Research
CO2 EMISSIONS PROFILES
under IPCC SRES scenarios
Source:
IPCC
Hadley Centre for Climate Prediction and Research
GLOBAL TEMPERATURE RISE
due to four SRES emissions scenarios
Hadley Centre for Climate Prediction and Research
Applying PRECIS in impacts assessments

Obtain current state from impacts model via:
a) observed data e.g. as used in model validation
b) RCM control data
1) direct output, e.g. temperature, precip., radiation
2) statistically downscaled for point locations e.g. gusts

Obtain future state from impacts model using:
– RCM climate changes (in means, variability, extremes) applies to observed data
(compare with a)
– RCM future climate data directly (compare with b1)
– statistically downscaled as b2 (compare with b2)
Hadley Centre for Climate Prediction and Research
BAY OF BENGAL CYCLONE
simulated by the PRECIS regional climate model
Resulting storm surge simulated using
the Proudman Oceanographic Laboratory model
Hadley Centre for Climate Prediction and Research
CHANGE IN ANNUAL SURFACE RUNOFF
between today and the 2080s, for A2 emissions
Southampton University hydrological model
Hadley Centre for Climate Prediction and Research
CURRENT LIMITATIONS OF PRECIS

10 year simulations take over a month

It is a model hence imperfect, i.e. uncertainty due to use of only one
RCM is not captured

Uncertainty due to imperfect driving data (i.e. from only one GCM)
is not captured

Resolution of 50km insufficient for some impacts applications

Coastal or marine impacts have to be inferred (e.g. storm surges) or
data is not available
Hadley Centre for Climate Prediction and Research
UNCERTAINTIES IN DETAILED CLIMATE CHANGE
SCENARIOS

Uncertainty in future emissions
– PRECIS uses 4 SRES emissions scenarios

Natural variability
– PRECIS can run in ensemble mode

Uncertainty in the response of the climate system
– PRECIS uses boundary conditsion from 1 GCM

Uncertainty in the detailed climate change
– PRECIS uses only one RCM
Hadley Centre for Climate Prediction and Research
CHANGE IN S ASIA SUMMER RAINFALL with PREDICTED BY NINE
GCMs (A2 emissions)
A2 emissions scenario
Hadley Centre for Climate Prediction and Research
% CHANGE IN
WINTER
PRECIPITATION
Hadley Centre for Climate Prediction and Research
REPRESENTATION OF THE PHILIPPINES
WITH DIFFERENT MODEL RESOLUTIONS
25km RCM resolution
50km RCM resolution GCM 300km resolution
Hadley Centre for Climate Prediction and Research
FUTURE DEVELOPMENTS

Continuously upgraded to new processors

Ability to run PRECIS RCM from different GCMs
– at least two to be added this year
– relevant cooperation from GCM centres being sought

Intercomparisons of all RCMs being organised
– No plans currently to incorporate other RCMs

Resolution of model being improved to 25km

10km RCM with ocean component being built
Hadley Centre for Climate Prediction and Research
Final remarks

You can contact us via:
– Email: [email protected]
– Tel: +44 1344 854938
– Fax: +44 1344 854898

Web-site for information on PRECIS
– www.hadleycentre.com/models/PRECIS.html
Hadley Centre for Climate Prediction and Research