EGU21-2513, updated on 31 Mar 2021
EGU General Assembly 2021
© Author(s) 2021. This work is distributed under
the Creative Commons Attribution 4.0 License.

Skip high-volume data transfer and access free computing resources for your CMIP6 multi-model analyses

Maria Moreno de Castro1, Marco Kulüke1, Fabian Wachsmann1, Regina Kwee-Hinzmann1, Stephan Kindermann1, Paola Nassisi2, Guillaume Levavasseur3, Sandro Fiore4, Charlotte Pascoe5, Martin Juckes5, Sophie Morellon3, and Sylvie Joussaume3
Maria Moreno de Castro et al.
  • 1DKRZ German Climate Computer Center, Germany
  • 2Fondazione Centro Euro-Mediterraneo sui Cambiamenti Climatici (CMCC), Italy
  • 3Institut Pierre-Simon Laplace (IPSL), France
  • 4Department of Information Engineering and Computer Science, University of Trento, Italy
  • 5Science and Technology Facilities Council (STFC), UK

Tired of downloading tons of model results? Is your internet connection flakey? Are you about to overload your computer’s memory with the constant increase of data volume and you need more computing resources? You can request free of charge computing time at one of the supercomputers of the Infrastructure of the European Network of Earth System modelling (IS-ENES)1, the European part of Earth System Grid Federation (ESGF)2, which also hosts and maintains more than 6 Petabytes of CMIP6 and CORDEX data.

Thanks to this new EU Comission funded service, you can run your own scripts in your favorite programming language and straightforward pre- and post-process model data. There is no need for heavy data transfer, just load with one line of code the data slice you need because your script will directly access the data pool. Therefore, days-lasting calculations will be done in seconds. You can test the service, we very easily provide pre-access activities.

In this session we will run Jupyter notebooks directly on the German Climate Computing Center (DKRZ)3, one of the ENES high performance computers and a ESGF data center, showing how to load, filter, concatenate, take means, and plot several CMIP6 models to compare their results, use some CMIP6 models to calculate some climate indexes for any location and period, and evaluate model skills with observational data. We will use Climate Data Operators (cdo)4 and Python packages for Big Data manipulation, as Intake5, to easily extract the data from the huge catalog, and Xarray6, to easily read NetDCF files and scale to parallel computing. We are continuously creating more use cases for multi-model evaluation, mechanisms of variability, and impact analysis, visit the demos, find more information, and apply here:


How to cite: Moreno de Castro, M., Kulüke, M., Wachsmann, F., Kwee-Hinzmann, R., Kindermann, S., Nassisi, P., Levavasseur, G., Fiore, S., Pascoe, C., Juckes, M., Morellon, S., and Joussaume, S.: Skip high-volume data transfer and access free computing resources for your CMIP6 multi-model analyses, EGU General Assembly 2021, online, 19–30 Apr 2021, EGU21-2513,, 2021.


Display file

Comments on the display

to access the discussion