\first_batch
, \second_batch
, and \third_batch
contain bash scripts for pulling variables out of the simulation outputs and concatenating the time series.
Running these jobs worked for most variables, but not all of them.
\fix_missing_atm
, \fix_missing_ocn
, \fix_missing_5ocn
processes the variables missed in the original batch jobs.
\archiving
and \archiving_pt2
contain bash scripts for archiving the post-processed netCDF files into .tar files to store on Ranch. Variables are archived into .tar files and then archived in groups to target the ideal file size for long term storage on Ranch.
\validation
contains a script that counts the number of netCDF files outputted by the batch jobs.
\transfer
contains a script that transfers the archives from lonestar6
check_integrity.sh
checks the integrity of the .tar archives by opening them and attempting to search the contents
find_archives.sh
is for use on Ranch and allows the user to search the archive files for specific variables.
The simulation output is organized into muliple netcdf files by model component and then divided into a time series of varying size depending on the time step. The time series associated with each variable was extracted and concatenated into larger segments on the order of 10 to 20 year chunks. These concatenated time series were then archived together. These archives were further archived again into indexed groups to reduce the number of files and target the optimal file size for tape storage on Ranch.