NWM TWL Retrospective

[KatherinePowell-NOAA]

Install, run, and test NWM coastal coupling with SCHISM/SFINCS(?) to generate a TWL (Total Water Level) retrospective dataset.

[KatherinePowell-NOAA]

Jason (@jduckerOWP) installed ESMF libraries to a specific Python environment and validated their stability and use on the Cloud-Sandbox workflow - a prerequisite for NWM Total Water Level (TWL) retrospective. Documented the installation instructions to set up that specific Python environment on the Cloud-Sandbox head node, including supplementary Python module installations required to run the entire NWM TWL retrospective workflow.

[KatherinePowell-NOAA]

Jason completed the NWM TWL retrospective spinup simulation for SCHISM, which consisted of constant NWM streamflow values in time, tidal boundary conditions, and calm/no precip conditions over a 10-day period. These conditions were optimal / correct scientific approach to spin-up SCHISM for retrospective start date (1979-02-01 00:00:00).
Also developed preliminary shell script to execute the NWM TWL preprocessing and SCHISM execution workflow in sequential cloudflow executions. This script also streamlines dynamic changing of files based on dates in order for this workflow to run without any user interaction after the shell script executes.
Jason also completed preliminary computational efficiency evaluations of the SCHISM model on the hpc6a.48large node instance using various schemes of the MPI and OPENMP protocols. NOTE: this is a big feat because we have translated slurm/psub workload manager code for OPENMP/MPI protocols for a given model setup, which has not been done yet on the Cloud-Sandbox project. Our intention is to document and integrate options for users to request these OPENMPI/MPI features for a given model execution within the job configuration file as a new template option for the Cloud-Sandbox workflow.