*Project:
Global Coastal Ocean modelling
2005-2009*
Building a computer modelling system to explore the coastal marine environment on a global scale.
What role do shallow seas play in the global climate? How will climate change affect the coastal environment? These are questions that global climate models cannot easily answer, because these seas are too small and too shallow. This project brings together physical oceanographers, ecosystem modellers and computer scientists from four leading UK research institutions to address this challenge.
What makes shelf seas so important for the Earth system?
•Most people’s direct experience of the marine environment is from coastal seas: 37 percent of the world population live within 100km of the sea. As climate change alters the marine environment this is where it will most dramatically affect people’s lives.
• Shelf seas are regions of exceptionally high biological production: around 14-30 percent of the marine plant growth occurs in shelf seas.
• Everything that travels between the land and the deep ocean goes via the continental shelves – be it nutrients, pollutants, carbon or freshwater. On the way they get mixed physically and transformed chemically.
What makes shelf seas so difficult to model?
While shelf seas only occupy 7 percent of the area they require 70 times the computer power of the deep ocean because of the small scale processes, such as currents, tides and mixing, in shelf seas. This is why we need oceanographers and computer scientists working together to find novel solutions to include shelf seas in our understanding of the Earth System.
Building models of all the continental shelves
In the past, for each shelf sea area, such as the North Sea, researchers have needed to setup model domains from scratch; often spending months gathering and configuring the data for a new model domain. We have developed an automated system to set up models of continental shelves anywhere on the planet, drawing on global data sets of such things as water depth, weather conditions, and river run off. This shortens the processes to a matter of minutes, allowing us to quickly build models of the 60 or so shelf seas around the globe.
Connecting Seas together
Only a few seas are isolated from the deep ocean so we use datasets from global ocean models around the boundaries. Additionally, most seas are also connected to other shelf-seas, and what makes this system unique is we have built in the capability for each of the domains to exchange information. This can happen either between compute nodes on the same cluster computer or between different clusters on a compute grid.
Connected super-computers
Each of these shelf sea domains requires the massive computer power available from large cluster computers. We are developing grid-computing tools to spread the considerable work load, without affecting the other users of those machines. The tools are easy to use and look likes the model is running on your desk top ever though it is actually running on a remote machine.
Outcomes New projects to:
• Explore the effects of climate change on the world’s major fisheries
• Investigate the effects Saharan dust on plankton growth in the west African up-welling regions
--
JasonHolt - 19 Sep 2007