Skip to content

Running the DP SCREAM Configuration

Peter Bogenschutz edited this page Nov 8, 2024 · 17 revisions

Scripts

To obtain scripts to run the DP-SCREAM please:

  1. Clone the E3SM-Project/scmlib repo.
  2. Find the DP-SCREAMv0 (fortran code) scripts in the top level directory named DP_SCREAM_scripts. Note that the scripts to run the DP-EAMxx configuration (C++ code) can be found in the DPxx_SCREAM_scripts folder. Soon the DP-EAMxx configuration will be the only supported configuration of DP mode.

Simply select your desired case and modify the required input in the header in the "USER DEFINED SETTINGS" (relating to the machine you are using and output paths etc.). All cases should then run out-of-the-box on any machine E3SM can run on.

Code

You should be running with the most up-to-date version of SCREAM master. Support for older versions of the SCREAM code base is not provided at this time.

Domain Size and Resolution Settings

All DP-SCREAM scripts are configured to run with grid spacing of 3.3 km in the horizontal (to match the global SCREAM model), with a domain size that was deemed appropriate for that case. Cases that simulate boundary layer clouds (for instance) are generally run with a domain size of 50 km x 50 km, while cases that involve deep convection are typically set up to run with a domain size of 200 km x 200 km.

Note, however, that DP-SCREAM gives you the freedom to run with whatever domain size and resolution you want. Clear instructions on how to change domain size and resolution (and to determine the appropriate time steps) are given in the script header.

Input Data

All required input data is pre-installed on supported (or commonly run) E3SM machines in addition to the E3SM input data server. Therefore, you should be able to run the DP-SCREAM out-of-the-box on any machine E3SM/SCREAM is commonly run on (i.e. NERSC, LCRC, Livermore Computing, Compy, etc.).