Norwegian Service Centre for Climate Modelling -> NoSerC project description 2003-2004 -> Project Description 2000-2002 -> Background
1. Background
The projects initiated in the area of climate research are among the most compute intensive projects in Norway. The access to state-of-the-art high performance computers is of vital importance to projects such as RegClim [1] and other similar projects. The climate models not only require huge amounts of computing resources, but also effective storage solutions for and access of Terabytes of data, both input, temporary and output data.
Today, the scientists in the climate projects spend much of their time on porting of models and handling of data, such as
- Porting models to the computer system they have access to;
- Obtaining input data from various sources (other compute centres/climate projects, national and international organizations, internal databases etc.). Some of these data are difficult and time consuming to access, due to slow networks and security restrictions;
- Converting input data to formats appropriate for their models;
- Copying between local resources and national high performance computing centres;
- Converting output data to formats appropriate for the data analyses tools involved;
- Archiving data. This also includes negotiating with system administrators to obtain space for data on systems where the data can easily be retrieved (disk, on-line tapes etc.);
- Searching for and maintaining data analyses tools, such as software for visualization and statistical analyses;
With the proposed increase in the climate modelling activities in Norway, as reflected in the tender from the Research Council dated 24.03.2000 [2], the need and the value of a national climate service centre will be even more important.
The scientists involved usually have good skills in computers science, and thus they are able to take care of these tasks. Nevertheless, more of their valuable time could be used for research if some of these tasks were carried out in co-operation with computer engineers with background in data handling. Furthermore, a Norwegian service centre for handling of climate data will increase the standardization and reduce the duplicate work within the climate projects in Norway. Finally, the scientists rarely have time to optimise their models for efficiency on high performance computers. Although they may receive some assistance from the high performance computing centres, this task could be further improved by establishing a service centre.
DNMI has a strong record in both climate modelling and data handling and archiving. The Research and Development department and the Climate division have a relevant and leading role in RegClim. The Research and Development department has also the responsibility for management of the UNECE/EMEP emission database (see http://www.emep.int). The UNECE/EMEP database is maintained and developed at DNMI to serve the needs of the international parties of the convention on Long Range Transboundary Air Pollution (LRTAP). The IT-division operates several systems for handling meteorological data, such as a historical climate database and an archiving system for observations and model output fields. The Climate division has a strong expertise on climate observations and quality of data. Combined with the experience of NOTUR in the area of system operation and support of users across Norway, we have a good background for operating a national service centre for climate modelling.
Send comments to webmaster