You are here: Home » Services » DYAMOND Initiative

The DYAMOND Initiative

Simulation examples of the DYAMOND Initiative

Figure: Simulation examples of the DYAMOND initiative (simulated day Aug 4th 2016). Can you tell which one is observation? By clicking on the image you can get a larger version (attention 20 MB)

DYAMOND stands for DYnamics of the Atmospheric general circulation Modeled On Non-hydrostatic Domains. This initiative project describes a framework for the intercomparison of an emerging class of atmospheric circulation models that, through their resolution of the major modes of atmospheric heat transport, endeavor to represent the most important scales of the full three-dimensional fluid dynamics of the atmospheric circulation.
The DYAMOND protocol (pdf) specifies a simulation of forty days and forty nights, beginning at 1 August 2016, using global models with a storm resolving grid spacing of 5 km or less. As initial and boundary data, daily sea-surface temperatures as well as global meteorological analysis taken from the European Center for Medium Range Weather Forecasts (ECMWF) are provided. The DYAMOND protocol was kept as simple as possible to encourage participation and ensure a fast turn-around.
At the moment, nine different groups from six national entities across three continents submitted and demonstrate that such simulations are nowerdays practical.

DYAMOND addresses two main questions:

  • How sensitive are the simulations to a particular implementation?
  • What are performance and analysis bottle necks associated with global storm-resolving models?

Due to the joint interest in high-resolution simulations, ESiWACE strongly supports and interacts with DYAMOND.


Participating Groups and Models

The idea of DYAMOND arose in October 2017 and started as a joint initiative between the Max Planck Institute for Meteorology (MPI-M, Bjorn Stevens) and University of Tokyo (Masaki Satoh). Further impulse was given through the involvement of US participants (Chris Bretherton, University of Washington), the support of Deutscher Wetterdienst (DWD) through the Hans Ertel Centre for Weather Research (HErZ) project (Daniel Klocke). Within the scope of the ESiWACE project the German Climate Computing Center (DKRZ) agreed to host the output and serve as an analysis hub for the community (Joachim Biercamp, Philipp Neumann).

Participating models are listed below along with the models' main scientific/technical contact.

Model Contact Person

Ryosuke Shibuya



Input Data

Initial Data  for 1 August 2016
A IFS grib file with initial data for 1 August 2016 is provided here. To process the data, you may use cdo or eccodes.
A NetCDF file, basically equivalent with the grib input, is provided here.
Alternatively, the initial data can be directly retrieved from the MARS database at ECMWF using this request.
Data for Sea Surface Temperature and Sea Ice
IFS data (7 day means) are provided here.



 Further news are published within our Newsletter of ESiWACE and ESiWACE2.


General contact

To scope out details on the work, interested groups please contact the .

Upcoming Events
PASC 2019 Minisymposia Jul 12, 2019 04:55 PM (Europe/Vienna) — Zuerich (CH)
Workshop on Machine learning for Weather and Climate Models Sep 02, 2019 09:00 AM (Europe/Vienna) — Oxford (UK)
Container Hackathon for Modellers Dec 03, 2019 09:00 AM (Europe/Vienna) — Lugano (CH)
Previous events…
Upcoming events…
© Copyright ESiWACE 2015