I have no idea, beyond the usual advice about how to pick target densities.
If we were configuring a Caspian Sea model today, we would almost certainly use fixed sigma-Z layers (levels) down to the shelf break. Based on nsigma=14, this is currently set at 54.76 m:
dp0k( 1) = 1.00 m thkns = 1.00 m depth = 1.00 m
dp0k( 2) = 1.19 m thkns = 1.19 m depth = 2.19 m
dp0k( 3) = 1.42 m thkns = 1.42 m depth = 3.61 m
dp0k( 4) = 1.68 m thkns = 1.68 m depth = 5.29 m
dp0k( 5) = 2.00 m thkns = 2.00 m depth = 7.29 m
dp0k( 6) = 2.38 m thkns = 2.38 m depth = 9.68 m
dp0k( 7) = 2.84 m thkns = 2.84 m depth = 12.52 m
dp0k( 6) = 2.38 m thkns = 2.38 m depth = 9.68 m
dp0k( 7) = 2.84 m thkns = 2.84 m depth = 12.52 m
dp0k( 8) = 3.38 m thkns = 3.38 m depth = 15.89 m
dp0k( 9) = 4.02 m thkns = 4.02 m depth = 19.91 m
dp0k(10) = 4.78 m thkns = 4.78 m depth = 24.69 m
dp0k(11) = 5.69 m thkns = 5.69 m depth = 30.37 m
dp0k(12) = 6.76 m thkns = 6.76 m depth = 37.14 m
dp0k(13) = 8.05 m thkns = 8.05 m depth = 45.18 m
dp0k(14) = 9.58 m thkns = 9.58 m depth = 54.76 m
dp0k(15) = 10.00 m thkns = 10.00 m depth = 64.76 m
dp0k(16) = 10.00 m thkns = 10.00 m depth = 74.76 m
dp0k(17) = 10.00 m thkns = 10.00 m depth = 84.76 m
dp0k(18) = 10.00 m thkns = 10.00 m depth = 94.76 m
dp0k(19) = 10.00 m thkns = 10.00 m depth = 104.76 m
dp0k(20) = 10.00 m thkns = 10.00 m depth = 114.76 m
dp0k(21) = 10.00 m thkns = 10.00 m depth = 124.76 m
dp0k(22) = 10.00 m thkns = 10.00 m depth = 134.76 m
dp0k(23) = 10.00 m thkns = 10.00 m depth = 144.76 m
dp0k(24) = 10.00 m thkns = 10.00 m depth = 154.76 m
dp0k(25) = 10.00 m thkns = 10.00 m depth = 164.76 m
You should check to see that ~55m is a good value for the shelf break (this is below the Caspian Sea surface, not below global mean sea level).
The Gulf of Mexico Reanalysis is an example of this aproach, although its target densities are sigma2. See
ftp://ftp.hycom.org/datasets/GOMu0.04/expt_50.1/conf/blkdat.input_501This has nsigma=16 and the sill depth is about 83 m (note that isotop is set to 83.0). also note that the terrain following dp00s's are all 1 m and the top deep layer, dp00(1), is 1 m.
Alan.
On Thursday, May 3, 2018 at 7:08:42 AM UTC-4, samad r wrote: