Tide Phase Problem

Report or discuss software problems and other woes

Moderators: arango, robertson

Post Reply
Message
Author
pmaccc
Posts: 39
Joined: Wed Oct 22, 2003 6:59 pm
Location: U. Wash., USA

Tide Phase Problem

#1 Post by pmaccc » Mon May 13, 2019 5:18 pm

I have a realistic run that uses tidal forcing from TPXO7.2 at the boundaries, both SSH and UV. The tidal forcing runs perfectly in general. I make some hand adjustments the the tidal amplitude when I make the forcing, to better match observations. This also works fine. However when I try to make adjustments to the phase, adding or subtracting a constant number of degrees to tide_Ephase and tide_Cphase, the model blows up, regardless of time step. Detailed symptoms: it will run if I just add 1 degree, but not if I add (or subtract) 10 degrees or more. It also blows up if I only adjust tide_Ephase and not tide_Cphase. The confusing thing to me is that all I am doing is changing the timing of high tide, not the amplitude.

Thanks for your help!

Parker

jcwarner
Posts: 812
Joined: Wed Dec 31, 2003 6:16 pm
Location: USGS, USA

Re: Tide Phase Problem

#2 Post by jcwarner » Mon May 13, 2019 6:21 pm

basic Q: but are you adding the phase in radians? Ephase would have the most impact. are you adding to all the freqs or just some?

pmaccc
Posts: 39
Joined: Wed Oct 22, 2003 6:59 pm
Location: U. Wash., USA

Re: Tide Phase Problem

#3 Post by pmaccc » Mon May 13, 2019 6:31 pm

I am adding to all 8 tidal frequencies, and doing so in degrees in my "tides.nc" forcing file. My varinfo.dat implies to me that the input should be in degrees and that ROMS (rev. 823) converts it internally to radians:

'tide_Ephase' ! Input
'tidal elevation phase angle'
'degrees' ! [radians]
'SSH_Tphase, scalar, series'
'tide_period'
'idTzph'
'r2dvar'
0.017453292519943295 ! pi/180

jcwarner
Posts: 812
Joined: Wed Dec 31, 2003 6:16 pm
Location: USGS, USA

Re: Tide Phase Problem

#4 Post by jcwarner » Mon May 13, 2019 6:42 pm

ok. i forgot how the input went. mine does the same: tide_Ephase in degress in netcdf, varinfo converts to rads.
Do you use ramp_tides? What if you adjust tide_start in the ocean.in.
-j

pmaccc
Posts: 39
Joined: Wed Oct 22, 2003 6:59 pm
Location: U. Wash., USA

Re: Tide Phase Problem

#5 Post by pmaccc » Mon May 13, 2019 7:22 pm

Thanks for the suggestions. I stopped using ramp_tides because (1) it seemed to handle the "rude awakening" without blowing up, and (2) it meant that I always had to make 2 versions of the executable. If I were always doing long "cold start" runs I could just have the ramp_tides version, but my forecast model workflow means I need to start most of the time without ramp_tides.

Note: my vote for future development would be to move the ramp_tides decision to the .in and out of the .h.

I'll try fiddling with tide_start. Ideally I wanted to adjust the phases individually to best match observations.

PM

Post Reply