neg. dp (m) near the strong current at the open boundary

[Jeongin Kim]

Dear Alan, 

I always appreciate your work and fast response on this HYCOM forum.

I'm running 0.01 x 0.01 deg simulation for the Florida Current region, nested in the North Atlantic Ocean simulation that I've run (Not the HYCOM reanalysis data).

But for some reason, my model crashes after a few days or a few weeks near that red boxed region, regardless of any batrop and baclin combination.

My CFL minimum is about 3.68, so I'm trying batrop within 2 to 2.5, and baclin within 25 to 40, but not all combinations are working smoothly.

I got a similar issue when I tried to run in the North Atlantic Ocean domain, where the strong current is located near the open boundary, and I resolved it by reducing the domain size and removing that problematic region.

So I guess this might be because of the strong current at the open boundary, but I'm not sure. 

My e-folding dates are from 0.1 to 10, and I've also double-checked it by plotting rmu_invday.

Can you give me any clue about this neg. dp problem?

Best regards,

Jeongin

[Alan Wallcraft]

One problem is:

 -0.04166667      'bnstfq' = number of days between baro nesting archive input
 -0.04166667      'nestfq' = number of days between 3-d  nesting archive input

The negative sign indicates the input is mean archives (archm files), but you actually have snapshot archives (archv files):

ln -sf /storage/home/hcoda1/6/jkim3990/scratch/HYCOM-examples/FLCa0.01/expt_01.2/../datasets/subregion/archv.2011_001_01.a archm.2011_001_01.a

The nesting program can read either kind of archive with either naming convention, but the reason both are allowed is because mean archives are nominally at the midpoint of their span.  So daily means are nominally at 12Z and daily snapshots are at 00Z.  So setting nestfq negative indicates that the archive is archm and its date is for a mean.

There is no such thing as hourly means because they should nominally be in the center of each hour and HYCOM's archives don't include the minutes (always 0 minutes into the hour).  So I'm not sure what setting nestfq to minus one hour actually does.

Since you have archv files, change to:

 0.04166667      'bnstfq' = number of days between baro nesting archive input
 0.04166667      'nestfq' = number of days between 3-d  nesting archive input

And keep the original archv name.

I don't know if this is the only issue, but see if this helps.

Since you have hourly nesting files, the e-folding time could be less than 0.1 days:

W, E boundaries: 80 grid pts with 0.1-9 day e-folding time

Typically signals longer than 5x the e-folding time will be reproduced.

Or you could reduce the reading load by reading every 3 hours (say)

 0.04166667      'bnstfq' = number of days between baro nesting archive input
 0.125                  'nestfq' = number of days between 3-d  nesting archive input

Alan.