3D Visualization SoftwareIn climate models, the physical domain of the atmosphere and the ocean is represented by three-dimensional computational grids. The finer the computational grid is, the more realistically can many important phenomena be simulated. Typical data sets with results of climate models are therfore three-dimensional, multivariate and time-dependent. The increasing use of ensemble simulation techniques, where simulations are repeated several times with slightly modified starting values, enables capturing the climate variability and hence adds another dimension to the data. The resulting huge and complex data sets can hardly be evaluated without up-to-date data analysis and visualization tools


To facilitate the interactive work with the model data, DKRZ provides its users with

  • access to powerful GPU nodes that are an integral part of the supercomputer
  • interactive visualization at home or in your office via remote 3D-rendering
  • domain specific data visualization software
  • appropriate technical advice and assistance in solving visualization problems
  • hands-on tutorial courses

GPU Nodes

GPU nodes of the supercomputer MistralGPUs ("graphics processing unit") are dedicated graphics processors that initially were developed for the rendering of computer generated imagery. Due to their high degree of parallelism, today's GPUs are extremely powerful special purpose processors, which are used for interactive 3D rendering of complex scenes, but, however, also to speed up machine learning and simulation applications (General Purpose Computation on Graphics Processing Units, GPGPU).

The GPU partition of our supercomputer system is currently mainly used for interactive data visualization of large time-dependent 3d model data as well as for application of machine learning techniques in the climate sciences. The GPU nodes are equipped with high-end NVidia GPUs and, in comparison to the CPU nodes, a large main memory, which is essential for the interactive visual analysis of large data. As the parallel file system of the supercomputer is directly mounted, the data doesn't need to be explicitely transferred for visualization.

The combination of powerful GPUs and a remore 3D rendering technology enables convenient remote use of these resources by the users of DKRZ

Visualization Software

Paraview is a public domain 3D visualization solution wich enables the direct 3D visualization of model variables as well as to interactively explore the relations between different quantities of multi varaite data sets by means of linking and brushing.

Met.3D is an open-source visualisation tool for interactive, three-dimensional visualisation of numerical ensemble weather predictions and similar numerical atmospheric model datasets.

With respect to data analysis, additional processing and 2D visualization solutions such as NCL and CDO are also available on the system.

For more details on the available visualization software look here.



In our gallery you will find many example pictures and animations illustrating the use of visualization in climate research.