Andrew Price and Andrew Yool

1
Andrew Price and Andrew Yool Southampton
e-Science Centre Southampton Oceanography Centre
2
In order to predict the future
The central scientific goal of the GENIE project
is to study the forcing and feedbacks driving the
glacial-interglacial cycles that dominated the
Earths climate over the last 1 million years By
better understanding the processes (physical and
biogeochemical) which regulated these cycles in
the past, we can be more confident about the
predictions climate models make for the future
3

• Orbital parameters affect incident radiation and
  climate
• Biological and geological processes interact
  with, and feedback upon, the climate (via, for
  instance, CO2)

4
How GENIE fits in
GENIE intends to study these climate cycles by
building a new Earth system Model of Intermediate
Complexity (an EMIC) A key component for the
success of GENIE will be its harnessing of new
e-Science techniques For example making use of
Grid resources for large ensemble simulations
data management and analysis adopting new
programming techniques to facilitate model
construction and Grid-based execution
5
The (final) GENIE model
6
The current GENIE model
Simpler, but fast - 1000 y per 1 h CPU time on
a PC
7
c-GOLDSTEIN grid and bathymetry
8
Challenges
How to ?
integrate state-of-the-art Earth system modules
repeatably and flexibly
Collaborative grid-based component programming
improve model composition tools
Modern development environment
Management of distributed compute and data
resources
integrate large-scale hardware systems in a
flexible way
Data archives visualisation of simulation runs
share, post-process, archive, re-use modeling
results
Inter-comparison of alternative modules/ models
test hypotheses about Earth System Modeling
9
The underlying technology

• Wrapping of component models
• XML schema, Java, .NET, Web Services technology
• Scripting environment
• E.g. Matlab, Python (Jython)
• Portal
• Web-based
• Repositories for components and data
• Database system
• Computational Grid infrastructure
• Condor pools, Beowulf clusters, linked by
  middleware
• Meta-scheduler
• Monitors the Grid, runs model on best platform/s

10
Data Management
Client
Grid
Geodise Database Toolbox
Java Client Code
Jython Functions
SOAP
Apache Axis
Matlab Functions
Metadata Database
CoG
GridFTP
XML Schema
11
Computational Toolbox
12
Grid Computation
National Grid Service (GT2)
Oxford
Leeds
WS Client
Java CoG
RAL
Manchester
SOAP
GRAM GridFTP
Flocked Condor Pools
WS
13
Scripting a Tuning Study
MATLAB
function RMS_Error cgoldstein(params)
optimum fminsearch( _at_cgoldstein, params, )
GENIE Database
gd_query(results)
Grid Resource
CG binary
gd_putfile(CG binary)
config file
gd_putfile(config file)
gd_jobsubmit(RSL)
results file
gd_getfile(results file)
gd_archive(results)
return RMS_Error
14
Matlab Optimisation Toolbox
Specify a starting
point parameters
0.5 Perform the
minimisation optimum
fminsearch( _at_cgoldstein_1D, parameters,
optimisation_parameters )
Specify a starting point
parameters 420
5000000 Perform
the minimisation optimu
m fminsearch( _at_cgoldstein_2D, parameters,
optimisation_parameters )
15
OptionsMatlab

• Matlab interface to the Options design
  exploration system
• http//www.soton.ac.uk/ajk/options/welcome.html
• State of the art design search and optimisation
  algorithms
• Design of Experiment methods
• Response Surface Modelling
• Over 30 search methods including
• Adaptive Random Search (ADRANS), Powell's Direct
  Search (PDS),
• Simplex Method (SIMP), Genetic Algorithm (GA),
• Simulated Annealing (SA), Evolutionary
  Programming (EP)

16
OptionsMatlab
Available Optimisation Methods 1.1 for OPTIVAR
routine ADRANS 1.2 for OPTIVAR routine DAVID
1.3 for OPTIVAR routine FLETCH 1.4 for OPTIVAR
routine JO 1.5 for OPTIVAR routine PDS 1.6 for
OPTIVAR routine SEEK 1.7 for OPTIVAR routine
SIMPLX 1.8 for OPTIVAR routine APPROX 1.9 for
OPTIVAR routine RANDOM 2.1 for user specified
routine OPTUM1 2.2 for user specified routine
OPTUM2 2.3 for NAG routine E04UCF 2.4 for bit
climbing 2.5 for dynamic hill climbing 2.6 for
population based incremental learning 2.7 for
numerical recipes routines 2.8 for design of
experiment based routines 3.11 for Schwefel
library Fibonacci search 3.12 for Schwefel
library Golden section search 3.13 for Schwefel
library Lagrange interval search 3.2 for
Schwefel library Hooke and Jeeves search 3.3 for
Schwefel library Rosenbrock search 3.41 for
Schwefel library DSCG search 3.42 for Schwefel
library DSCP search 3.5 for Schwefel library
Powell search 3.6 for Schwefel library DFPS
search 3.7 for Schwefel library Simplexsearch
3.8 for Schwefel library Complexsearch 3.91 for
Schwefel library twomembered evolution strategy
3.92 for Schwefel library multimembered
evolution strategy 4 for genetic algorithm
search 5 for simulated annealing 6 for
evolutionary programming 7 for evolution
strategy
gtgt OptionsInput createOptionsStructure(4.0)
DNULL -777 OLEVEL 2 MAXJOBS 100
NVRS 12 VNAM 'SCLTAU' 'INVDRAG'
'OCNHORZDF' ... LVARS 1.3000 2.0000
2500 ... UVARS 2.1000 4.8000 5700 ...
VARS 1.7000 3.4000 4100 ...
ONAM 'RMSERROR' OMETHD 4.0000 DIRCTN
-1 NITERS 1000 OPTFUN
'cgoldstein_12D' OPTJOB 'optjobparallel'
GA_NPOP 100 gtgt OptionsOutput
OptionsMatlab(OptionsInput)
17
Twin-Test Experiment
Attempt to recover a known state of the model
using a Genetic Algorithm.
Performed 10 generations of a 100 member
population. Then applied a local Simplex search
of the best candidate.
Population too small to find optimal solution
suitable for finding local minima
18
Tuning using Observational Data
Apply the same method but calculate the RMS error
statistic by comparing the model state with NCEP
observational data.
19
e-Science Summary
Environmental Scientist
Application
Middleware
Grid
20
Fusing science and e-science
One successful use of e-science so far has been a
large-scale study of the bistability of the
oceans thermohaline circulation (THC) The THC
is responsible for large-scale distribution of
heat and salt throughout the ocean, and has
climatic consequences such as the warming of
western Europe (via the Gulf Stream) Changes
to the budgets of heat and freshwater caused by
global warming may have important consequences
for its future behaviour
21
The Thermohaline Circulation
22
Consequences?
Some of the less likely consequences of a
shut-down of the THC The Day After Tomorrow
While western Europe and north America may cool,
globally the earth will warm, and a new ice-age
is unlikely
23
Shutdown in the model
THC On (top left) THC Off (bottom
left) Temperature consequences (below)
24
Study design

• We identified two parameters affecting the
  freshwater budget of the Atlantic
• Atlantic-to-Pacific zonal transport
• Atmospheric diffusivity (meridional transport)
• Using a portal to a Condor pool, simulations
  using different values of these parameters were
  varied
• The results of these simulations were then used
  to feed new simulations to examine classic
  bistability

25
961 member ensemble
26
The initial ensemble
27
9 x 961 member ensemble
28
Identifying bistable region
29
Scientific and e-Scientific conclusions
The work has allowed us to determine the region
of parameter space over which the model THC is
bistable With this information, we have been
able to study climate feedbacks in more detail,
and work out the minimum duration of Greenland
icesheet melting that can shutdown the THC (88
years!) The use of e-Science allowed us
unprecedented total simulation duration (42
million years) with time-efficiency of 1 order
of magnitude
30
The GENIE Team

• Coordinator
• Tim Lenton CEH Edinburgh
• Principal investigator
• Paul Valdes Bristol
• Research Team and Collaborators
• James Annan FRSGC, Japan
• Chris Brockwell UEA Norwich
• David Cameron CEH Edinburgh
• Peter Cox Hadley Centre (UKMO)
• Neil Edwards Bern, Switzerland
• Murtaza Gulamali London e-Science Centre
• Julia Hargreaves FRSGC, Japan
• Phil Harris CEH Wallingford
• Dan Lunt Bristol
• Bob Marsh SOC
• Andrew Price Southampton e-Science Centre
• Andy Ridgwell UBC, Canada

• Management Team
• Melvin Cannell CEH Edinburgh
• Trevor Cooper-Chadwick Southampton e-Sci.
  Centre
• Simon Cox Southampton e-Sci. Centre
• John Darlington London e-Science Centre
• Richard Harding CEH Wallingford
• Steven Newhouse London e-Science Centre
• Tony Payne Bristol
• John Shepherd SOC
• Andrew Watson UEA Norwich