...
Spherical harmonic transform available in NCL and pyNGL for lat/lon data.
Interpolate EAM native grid output to a (N+1)x2N lat/lon cap grid with TempestRemap’s “highorder” (native SE shape function) algorithm.
For cubed sphere grid with resolution NE, take N>=NE*6 (degrees of freedom, pole to pole)
Need instantaneous vorticity, divergence and PS on the GLL grid (not the PG2) grid. PS is needed if interpolating to pressure levels
in EAM, use: ‘VOR:I','DIV:I','DYN_PS:I'
Need ‘DYN_PS’ instead of ‘PS’ to get PS on the GLL grid instead of the PG2 grid.
For smoother smooth results, need spectra from ~60 snapshots. 2x per day for 1 month will produce nice results average over at least 2 months with 12h snapshots. Running longer rather than with higher frequency I/O results in smoother spectra in Aqua planet. What about hourly over 3 days? Are there seasonal effects in Real planet?
At NE256, interpolating and computing the spectra for each snapshot takes about 5min and 30GB of memory. At NE1024 this will probably require ~500GB of memory and could take 320min per snapshot.
...
The spectra above from single snaphots have a lot of noise which can be removed by time averaging. We next analyze how much averaging is necessary from an aqua planet NE256 simulation with 1 month of daily snapshots. The legend indicates how many snapshots were used when averaging E(k). Left plot is with daily snapshots, right plot is with 3h snapshots. With 1 month of averaging, the data is still a touch noisy, with either daily or 3h output. There is also not much difference between daily or 3h data, so a longer simulation is more important than higher frequency outputThe 3h data is only slightly less noisy than daily (comparing 233 snapshots of 3h data to 28 snapshots of daily data). 2 months of 12h data would probably be sufficient. In the plots below, each curves are all shifted downward so they would not be plotted on top of each other.
...
Software needed
...