ed2-mandifore

The goal of ed2-mandifore is to run ED2 with Setaria in already grown, complex ecosystems using weather data from MANDIFORE sites.

Reproducibility

This project uses renv to manage package dependencies. This is essential for reproducing this work as the version of PEcAn.ED2 used is installed from a pull request that will likely never be merged. Using a different version of PEcAn.ED2 will result in errored runs. Run renv::restore() to install dependencies.

Setup scripts

Scripts 00 and 01 have already been run and have generated the data in data/. They do not need to be run again.

1. Start with sourcing 02_setup-runs.R to generate files in the transect/ directory.
2. In the terminal, navigate to a particular run (e.g. ./transect/MANDIFORE-SEUS-352/pine) and start the job as a background process with ./run.sh.
3. Follow the checklist below to check that the job is running correctly

Job Start Checklist

Do all of this before starting the next job!

• [ ] Is the R output being saved to workflow.Rout?
• [ ] Are all the expected run/ and out/ folders created locally?
• [ ] Find settings_checked.xml and look through it.
• [ ] Are all the expected run/ and out/ folders created on the HPC?
• [ ] Once the job starts on the HPC, record the SLURM job ID. It is not printed anywhere in the logs, so you will need to manually copy and paste it somewhere (e.g. into the pid.nohup file that has the local PID)
• [ ] Spot-check the log files for multiple runs to see that simulation has started

...now you can start another job.

Job analytics

If you want to know how long a job took on the HPC you can use:

sacct -j <jobid> -o Start,End,Elapsed

To check remaining compute hours:

va