bathymetry values

[fereshteh]

Hello every one
i could make coastline and boundary file for my closed domain by seagrid. And now i am preparing its bathymetry file. With regard to Grid generation page of ROMS site, that say for a 'typical' application by roms, the bathymetry can not have values of 0 or any negative values; Most users would set some minimum 'clipping' depth and not allow the values of h to be less than that depth, i must replace zero values by a value less than minimum depth of my domain (min=-.3m)?
Please help me to generate my appropriate bathymetry file.
Best regard
fereshte

[kate]

If you are using the WET_DRY option, then the sometimes-wet areas will be protected from going all the way to zero thickness. If the permanently dry areas are at -.3, that's probably OK. They just can't be exactly zero. If you have a zero, you will know right away.

[fereshteh]

hi dear Kate
thanks for your attention but sine i haven't used ROMS model, i don't know application of WET-DRY (unfortunately). therefore i did not completely catch you. You mean that I must replace zero value in the sea area with value less that minimum depth or not?
Also i become grateful to receive reply for another question: the first one is that for grid generation with seagrid, boundary file have to coincident with coast line (for generation boundary fitted mesh)? if it have, as you know controlling of orthogonality will be difficult. and if it have not, which value must be used for land area (zero or negative)?
sincerely yours
fereshte

[kate]

There is a static mask allowing your grid to cover land areas. The grid boundary does not have to follow the coastline. Within the areas that are always land, the bathymetry can be anything but zero (or NaN or Inf).

The same is true of the water areas unless you use the WET_DRY option. That applies a minimum to the water depth so it will never go to zero - adding water if necessary. I assumed you were using WET_DRY when you talk about a minimum depth of -.3 m, since the water thickness should be positive and since tides can be larger than that in many areas.

If you truly want a minimum depth of less than a meter, beware of the bathymetry steepness limits. We smooth our bathymetry so that the depth from one grid to the next changes by no more than a factor of about two. Getting from 1 m to 1000 m therefore requires at least 10 grids.

[fereshteh]

Dear Kate
first of all i am extremely appreciate for you complete reply. With regard that i want to nest me coarse domain with coastal domain (nearshore region) i have bathymetry value less that 1 m. Base on your reply, i found that when e have depth value less that 1 m,we must notice to bathymetry steepness and at first we must smooth our bathymetry (is not it?). And now, which way you suggest for smoothing our bathymetry?
Thanks so much
fereshte

[kate]

That is the age-old question. Personally, I use a modified Shapiro filter, just smoothing points with a high rx0. I then follow with a regular Shapiro filter everywhere to knock off the 2 dx stuff.

[fereshteh]

Dear
i have questions about smoothing. I have got that smoothing applied on bathymetry file which is 2d array (h(x,y)) not on grid file(if my Thought is wrong please tell me). but in your answer you said

I then follow with a regular Shapiro filter everywhere to knock off the 2 dx stuff.

1- What is that mentioned dx? Becouse just in grid file user define dx not in bathymery file
2- are regular and modified shapiro filters that you said, mentioned to shapiro02.m which located in rgrd folder of rnt toolbox? If they are different i become grateful to receive link for downloading them.
Checking the folders contains shows that there is 1 script named smooth_bath.m (in matlab/utility)
3-Is it possible to use that script instead of shapiro filter? OR smoothgrid.m in rnt toolbox/nesting_tools?
4-if there are another way to smoothing , i become so appropriate to receive its link?
Also i request that you introduced someones to me for asking such basic questions of her/him (if it is possible)
Thanks for your patient answers which help me always
cheers

[fereshteh]

Dear user
my requested hydrography data (for Caspian sea) have been provided from GEBCO (1 minute resolution).its minimum depth is 1 m and also some points have zero value (attached picture show these points). with regard that for generating grid, there should not be zero value, eliminating them that located in the sea or near coastline, will arise problem in future steps?
what is your recommendation about solving it?
thanks in advance for receiving your experiences
cheers
fereshte

[kate]

Are you using WET_DRY? If not, you need to clip the bathymetry to be greater than or equal to some hmin larger than zero.

I don't know anything about the Matlab bathymetry smoothing options.

[fereshteh]

dear kate
thanke so mush
yes, because my domain has land as well as sea, i will use WET_DRY option.
so with this option can clip bathymetry value.is not it?
thanks again
cheers
fereshte

[kate]

I haven't ever really used WET_DRY, but hope to soon in a new application. Anyway, what happens if you try it?

[fereshteh]

dear kate
be sure that i will aware you about what happen, after applying that option
thanks for your patient to answer me
fereshte