O3.2.4 Moist Held-Suarez Perfect-Model Calibration

[nefrathenrici]

The Climate Modeling Alliance

Software Design Issue 📜

Purpose

We will implement a pipeline to calibrate on the moist held-suarez experiment in ClimaAtmos. We will calibrate on the horizontal temperature gradient.
Once this is complete, we will generalize to a pipeline that can handle other cases.

Cost/Benefits/Risks

Primary risk is developer time, but this will save us a lot of time for manual calibration in the long run.

People and Personnel

@szy21 @trontrytel

Components

• EKP interface: Configure EKP, construct the prior distribution, update ensemble
• HPC interface: this should be GPU and CPU compatible, ideally running on the cluster and the a100 server.
• Atmos interface: Configure and run the atmosphere model.
• Data processing utilities: Needed for constructing the observation map
• Visualization utilities: Used for comparing model output and determining useful information for the observation map

Inputs

• EKP configuration: ensemble and iteration sizes
• Prior distribution TOML
• Observation map
• Steady-state restart file

Results and Deliverables

• Fully working and reproducible perfect-model calibration for moist held-suarez
• Simple experiment configuration interface with tests
• Inverse problem for the HS case

Task Breakdown And Schedule

• Set up repository structure, github actions, Julia project #5
• Create interface
• Moist HS observation map
• Demonstrate moist HS perfect model calibration

SDI Revision Log

CC

@tapios @simonbyrne @cmbengue

[tapios]

Looks good to time. Let's target running this on GPUs right away, to increase throughput. This may then also necessitate some GPU compatibility work for EKP (if it isn't already compatible).