Recently, I have been getting requests about how to access the LASSO-CACTI data set on ARM’s Cumulus cluster. This is a great way to work with the data when you need to access a large number of files, since you do not need to download the data to a local computer.
Below is a summary of how to find the data once you have an account established to access Cumulus. If you have not already had an account set up for you, refer to the related chapter in the LASSO-CACTI technical documentation for more information on that process.
At present, the LASSO-CACTI simulation data is pre-staged on Cumulus at /gpfs/wolf/cli120/world-shared/d3m088/cacti/staged_runs. The runs are organized in a folder structure by date/ensemble/physics config/scale. Once you find the simulation you want to use, there are directories for the raw wrfout files (out_dX where X is the domain number), the subset files (subset_dX), and the run directories.
It would be really helpful if you can keep track of the runs and types of files you use and then report that back to me at lasso@arm.gov for tracking purposes. By accessing the data directly on Cumulus, we bypass the automated tracking done when downloading the data from the Bundle Browser. ARM is required to track this level of detail by DOE, so self-reporting your usage is important.
At some point in the near future, we will likely have to remove the run (with the input data) and out_dX directories from the online staging directory. We have to free up space to work on the new LASSO-ENA scenario. So, I recommend working with the subset files as much as possible. However, that won’t work for some applications, so don’t be afraid of working with the wrfout files. You just might need to let me know not to remove certain ones, copy them to your work location (we would rather not have multiple copies online though 
), or download them from the Bundle Browser.
Let me know if you have questions or difficulty with anything. For example, we have found an occasional corrupt file that I have had to fix. I can also help with understanding the modeling details and how best to use the simulations for particular applications.
-Bill (LASSO PI)