You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
@iangrooms noticed that the current implementation of Leith and Smagorinsky viscosity defines Delta on anisotropic grids as:
Delta**2 := 2 dx**2 dy**2 / (dx**2 + dy**2)
This approach leads to small viscosities in the equatorial band due to refined meridional resolution. Ian's Leith+E implementation (USE_LEITHY) defines Delta as Max{dx, dy}, ensuring the cascade arrests at the coarser grid scale. This alternative also smooths the Ah coefficient, leading to more reasonable viscosity values, particularly in the equatorial band.
We are currently testing using USE_LEITHY=True + LEITHY_CK=0.0 (without backscatter) + LEITH_BI_CONST = 75 as a replacement for LEITH_AH = True + LEITH_BI_CONST = 128.0.
We also noticed that enabling a Laplacian backscatter (LEITHY_CK=1.0) improves the Equatorial undercurrent, so we are now testing different values of LEITH_BI_CONST with this option.
For beta05, we will use LEITHY, either with or without backscatter.
The text was updated successfully, but these errors were encountered:
@iangrooms noticed that the current implementation of Leith and Smagorinsky viscosity defines Delta on anisotropic grids as:
Delta**2 := 2 dx**2 dy**2 / (dx**2 + dy**2)
This approach leads to small viscosities in the equatorial band due to refined meridional resolution. Ian's Leith+E implementation (
USE_LEITHY
) definesDelta
asMax{dx, dy},
ensuring the cascade arrests at the coarser grid scale. This alternative also smooths the Ah coefficient, leading to more reasonable viscosity values, particularly in the equatorial band.We are currently testing using
USE_LEITHY=True + LEITHY_CK=0.0 (without backscatter) + LEITH_BI_CONST = 75
as a replacement forLEITH_AH = True + LEITH_BI_CONST = 128.0
.We also noticed that enabling a Laplacian backscatter (
LEITHY_CK=1.0
) improves the Equatorial undercurrent, so we are now testing different values ofLEITH_BI_CONST
with this option.For beta05, we will use LEITHY, either with or without backscatter.
The text was updated successfully, but these errors were encountered: