HPC Climate modelling
are three reasons why we need a lot of about Blue Crystal is that it is so big so we can work thousands of times faster. This could
computational power. Firstly, we might need run up to 100 simulations at a time.’ potentially bring huge benefits, it may be
to represent features accurately, for example if ‘Before this HPC facility came along we possible to move away from modelling over
you are trying to model the outlet glaciers of just had to use best estimates of the boundary 20,000 years to being able to model over one
the Greenland Ice Sheet, that obviously takes conditions. This new HPC facility actually million years.’
a lot of computational power. makes the previous work we did look
‘Secondly, the coupling between the ice meaningless in comparison.’ Oceans 2025
sheets, water and atmosphere takes up a lot The centre is currently trying to work No, it’s not yet another sequel to a rehashed
of computational resource and, finally, we out what happened between a time around Hollywood movie, Oceans 2025 is actually
are running experiments by running these 20,000 years ago when the Earth’s northern concerned with modelling the Earth’s oceans.
different models many times.’ hemisphere was blanketed with ice, to how Researchers at the National Oceanography
This holy trinity of accuracy, computational the planet became the relatively ice-free world Centre at Southampton (NOCS) are part
bulk and repeatability is echoed by climate we live in today. Blue Crystal is allowing the of the much larger Oceans 2025 project,
modellers around the world, with the final researchers to look into this time frame in which seeks to increase knowledge of
point being particularly important. By much more detail than ever before, as Lunt the marine environment. The project,
changing the initial conditions within each explains: ‘We used to do one simulation collaboration between seven UK marine
model and running simulations multiple times research centres, will run from 2007 to 2012
concurrently, scientists can produce a mean with a £120m budget. Specifically, the NOCS
result. This means, in the case of the Mike group will aim to deliver the ocean models
Barnsley Centre, finding the most likely effect needed for the next decade of UK marine
of the melting glaciers on the planet. Murray science.
adds: ‘This leads to more robust answers, The French/European NEMO ocean model
rather than just running one simulation with will be used for the majority of the ocean
one set of initial conditions.’ modelling carried out by NOCS. NEMO
At the University of Bristol, climate (Nucleus for European Modelling of the
modellers are simulating how the planet’s Ocean) allows ocean-related components
climate changed tens of thousands of years (for example sea-ice, biochemistry, ocean
ago. Modelling the entire planet over such dynamics and so on) to work either together
immense timescales requires an immense or separately. Unlike previous ocean models
supercomputer so the university opened used by UK researchers, the NEMO model
its £7m facility, Blue Crystal, last year. The has not been specifically optimised for use on
supercomputer can carry out more than 37 Visual interpretation of climate change data, UK supercomputers.
trillion calculations per second. aided by Blue Ice. Dr Andrew Coward, researcher at NOCS,
While most current weather simulations believes the ability to add complexity to
of the UK or Europe may look at a resolution where the ice was at its maximum thickness the model is important and says: ‘Many
of a few kilometres across and timescales of a (around 20,000 years ago), and another at the uncertainties exist about the role of ocean
few days, for such planet-size, thousand-year minimum thickness (the present day).’ biogeochemistry in the climate system.
timescale simulations the Earth is broken ‘Now we have moved away from such Models of ocean biogeochemistry are
down into areas hundreds of kilometres across. snapshots and have transient models themselves complex but life in the ocean is
For example, the UK is represented by six or simulating the ice changing over the whole also critically dependent on details of the
seven boxes. But the simulations can cope 20,000 years. This simply was not possible five physical environment that it finds itself in.
with such relatively small timesteps, as Dr or so years ago.’ ‘Thus complex biogeochemistry models
Dan Lunt, research fellow at the School of But the code needed to be re-jigged to cope need to be coupled with high resolution
Geographical Science at the University of with the heavyweight hardware it must run ocean circulation models in order to provide
Bristol, says: ‘The new facility has helped our on, as Lunt explains: ‘A lot of the climate investigative tools. Such tools are needed
research no end. We used to model timescales model code was written 30 years ago, and it’s to elucidate the biological feedbacks in the
of a century or two; now we simulate over beginning to show its age because it simply climate system. NEMO provides the ideal
thousands of years and can break those was not written with hundreds or thousands framework for studies of this type but the
simulations down to timesteps of every 30 of processors in mind. At the moment, we can sheer size and complexity of the coupled
minutes.’ only break the Earth down into up to about 50 ocean-biogeochemistry system will continue
As Professor Murray mentioned, another chunks – beyond that you have a saturation to provide a supercomputing challenge.’
big boost to such climate change research point where the code does not know how to
is that the scientists can map out many parallelise the work over multiple processors.’ Eye of the storm
simulations at once to accommodate all the Lunt adds: ‘We are now developing code Over the pond, a US centre called the
different variables that have to be taken into that is more parallelisable, which can run over CHAMPS lab (Coastal Hydroscience
account. Lunt adds: ‘One of the great things thousands of processors and correspondingly Analysis, Modelling and Predictive
www.scientific-computing.com SCIENTIFIC COMPUTING WORLD february/march 2009
35
SCWfeb09 pp34-36 weather.indd 35 4/2/09 10:44:40
Page 1  |  Page 2  |  Page 3  |  Page 4  |  Page 5  |  Page 6  |  Page 7  |  Page 8  |  Page 9  |  Page 10  |  Page 11  |  Page 12  |  Page 13  |  Page 14  |  Page 15  |  Page 16  |  Page 17  |  Page 18  |  Page 19  |  Page 20  |  Page 21  |  Page 22  |  Page 23  |  Page 24  |  Page 25  |  Page 26  |  Page 27  |  Page 28  |  Page 29  |  Page 30  |  Page 31  |  Page 32  |  Page 33  |  Page 34  |  Page 35  |  Page 36  |  Page 37  |  Page 38  |  Page 39  |  Page 40  |  Page 41  |  Page 42  |  Page 43  |  Page 44