Before doing any realistic runs, I am trying initializing with a horizontally uniform (but vertically varying) stratification appropriate to my modelling region to find optimal theta_s and theta_b for minimizing spurious currents. I've found that theta_s = 3 and theta_b = 0.7 works reasonably well for horizontal currents, however I'm getting w up to ~4cm/s in the bottom layer over steepest topography (rvalue <0.2 everywhere); this decreases to ~1mm/s in the second to bottom layer and continues decreasing in shallower layers.
I have tried drastically decreasing my timestep, changing my bottom drag type and increasing theta_b, all with minimal effect.
Is this a part of a physical process that I don't yet understand, or is there a problem with my setup of the model?
All thoughts and comments welcome!
Thanks in advance,
Wendy
large w in bottom layer
I am still having this problem; I appreciate that it is nobody's job to help me sort out my issues, but if someone could tell me if this is a normal occurence and to be expected from the model, or if this is abnormal and I should continue to look for the problem, it would be so helpful!
To summarize, I have horizontally uniform initial conditions and no forcing. After ~1hr model time, I have w in the bottom layer around 4cm/s near steepest topography. The n=2 layer (and subsequent layers) has w around 1mm/s. I am curious about the 4cm/s w.
I am using version 2.2.
Thanks,
Wendy
To summarize, I have horizontally uniform initial conditions and no forcing. After ~1hr model time, I have w in the bottom layer around 4cm/s near steepest topography. The n=2 layer (and subsequent layers) has w around 1mm/s. I am curious about the 4cm/s w.
I am using version 2.2.
Thanks,
Wendy
some things to try:
- how are you initializing the currents and zeta? Start the model with u and v and ubar and vbar = 0.
- what turbulence closure are you using?
- "w" is at the grid cell interfaces, so there are N+1 values. Is the value at k=0 (ie at the sediment water interface) small, and then w is large at k=1?
- Are both "w" and "omega" large, or is there a conversion error?
- as the 'w" increases, what happens to the stratification and AKv and AKt?
- how are you initializing the currents and zeta? Start the model with u and v and ubar and vbar = 0.
- what turbulence closure are you using?
- "w" is at the grid cell interfaces, so there are N+1 values. Is the value at k=0 (ie at the sediment water interface) small, and then w is large at k=1?
- Are both "w" and "omega" large, or is there a conversion error?
- as the 'w" increases, what happens to the stratification and AKv and AKt?
thanks for getting back to me; to answer your questions:
-I am initializing with u, v, ubar, vbar and zeta all =0
-I am using the GLS. k-epsilon turbulence closure
-Sorry about that, I mean w is large at k=0 and then decreases at k=1....
-Omega is small! Omega is zero at k=0 and at the surface, and less than 1mm/s everywhere in between. If, as found in wvelocity.F, w=Hz*omega/(m*n), then this should not be possible, unless I'm missing something?
-AKv, AKt (and AKs) all increase with depth (from 10^-5 at the surface to 10^-4 at k=1), until the k=0 level at which point they all sharply decrease back to 10^-5.
-Calculation of buoyancy frequency for the initial conditions shows that my initial profile is stable. but becomes more stable with time. My initial stratification consists of a linear change in salinity and temperature near the surface followed by an exponential drop-off at depth. Looking at profiles of change in temperature and salinity over time, the greatest changes seem to be occurring at the transition depth of salinity between the linear and exponential regimes for any given profile, but this effect is most pronounced where this transition occurs near the total cell depth and where the local topography is steep.
I have tested with a purely exponential salinity profile, and found that the problem decreased, but did not disappear.
-I am initializing with u, v, ubar, vbar and zeta all =0
-I am using the GLS. k-epsilon turbulence closure
-Sorry about that, I mean w is large at k=0 and then decreases at k=1....
-Omega is small! Omega is zero at k=0 and at the surface, and less than 1mm/s everywhere in between. If, as found in wvelocity.F, w=Hz*omega/(m*n), then this should not be possible, unless I'm missing something?
-AKv, AKt (and AKs) all increase with depth (from 10^-5 at the surface to 10^-4 at k=1), until the k=0 level at which point they all sharply decrease back to 10^-5.
-Calculation of buoyancy frequency for the initial conditions shows that my initial profile is stable. but becomes more stable with time. My initial stratification consists of a linear change in salinity and temperature near the surface followed by an exponential drop-off at depth. Looking at profiles of change in temperature and salinity over time, the greatest changes seem to be occurring at the transition depth of salinity between the linear and exponential regimes for any given profile, but this effect is most pronounced where this transition occurs near the total cell depth and where the local topography is steep.
I have tested with a purely exponential salinity profile, and found that the problem decreased, but did not disappear.