Hi,

 I am runing a extreme high simulation with 20 m (a one-way nesting from 7.5 km 1.5 km 500 m 100 m down to 20 m) reoslution and NBQ turning on. However, I have tried a number of tests, but it always blows up with noise growing in the deep layers. Please see the figure below.  Does someone have any idea what the problem is? The details of the configuration is also shown below. Some other information may be helpful:

(1) The model runs safely with 3s time step if I just turn NBQ off. (2) Time steps I have used: 0.1s and 0.3s. All failed with the same problem.

Here is a figure of u in the deep layers (very close the bottom, 150 layers in 5000 m water depth)

REGIONAL KuroSI MPI NBQ DIAG_CFL OBC_EAST OBC_WEST OBC_NORTH OBC_SOUTH W_HADV_TVD W_VADV_TVD GLS_MIX2017_3D CURVGRID SPHERICAL MASKING NEW_S_COORD SOLVE3D UV_COR UV_ADV SALINITY NONLIN_EOS SPLIT_EOS UV_HADV_TVD UV_VIS2 UV_VIS_SMAGO UV_MIX_S UV_VADV_TVD TS_HADV_WENO5 TS_DIF2 TS_MIX_ISO TS_VADV_WENO5 SPONGE GLS_MIXING GLS_KEPSILON CANUTO_A ANA_DIURNAL_SW QCORRECTION SFLX_CORR FRC_BRY Z_FRC_BRY M2_FRC_BRY M3_FRC_BRY T_FRC_BRY OBC_NBQ OBC_M3ORLANSKI OBC_TORLANSKI ANA_BSFLUX ANA_BTFLUX OBC_M2CHARACT M2FILTER_NONE TRACERS TEMPERATURE M3FAST NBQ_IMP HZR NBQ_PERF NBQ_GRID_SLOW NBQ_HZCORRECT OBC_NBQORLANSKI OBC_WORLANSKI VAR_RHO_2D VIS_COEF_3D TS_MIX_IMP TS_MIX_ISO_FILT NTRA_T3DMIX SPONGE_GRID SPONGE_DIF2 SPONGE_VIS2

NF_CLOBBER