Ocean Modeling Discussion

kate

It has been YEARS since I got scrip to work. Should I take down all reference to it? Should I just take down pyroms entirely as unsupported? Anyway, follow the xesmf example instead.

kate

What winds are you using? I have heard that ERA5 can get unphysically large.

Since I wrote that prior message I switched to a Large and Yeager bulk flux scheme, then I changed ocean models.

kate

It is far better to post here than in a PM. Please don't send me PMs asking about ROMS. I am no longer using ROMS and I'm certainly not up to date on the ROMS ice. I don't even know - is Hernan's new ice now available in COAWST and is that what you are using?

kate

If you feel that using the local velocity would be better in your situation than the phase speed, you should try coding it up and seeing how it behaves. I expect that was tried at some point and found wanting, but we only had a limited number of domains when trying this stuff.

kate

For each field using RadNud, the phase speed is computed locally, at each point on the boundary. If you look, that phase speed can be quite noisy, in time and space. You could try using ambient velocity instead, but the point is that you are trying to radiate away waves, which don't necessarily move ...

kate

It's clear enough, right? You need to look at what line number, what variables it is talking about. I often
add print statements to find out the shapes of things. You might also learn to use breakpoint() to get
into the debugger to query your variables.

kate

I don't know what to make of the log file. Yes, your domain size could cause it to run out of memory and get killed.

kate

With xesmf, you don’t run make_remap_weights_file.py. I should delete that.

kate

You wrote to me privately asking this, but I'm quite rusty on all this. I would make sure that there is a pCO2 id in varinfo.dat (is that still how it is done?) and I would hack at all the routines that write all these files. There are plenty of examples with the other fields being written.

kate

MOM6 can handle this case, with an open boundary inside the domain. ROMS is not set up to do that.

kate

My latest in Python rivers are at https://github.com/ESMG/pyroms/tree/pyt ... ill_runoff

kate

Rivers can be tricky! Each uv-source is either a u-source or a v-source. Each u-source is either going to the east or to the west. You want one of each, so one is positive, one is negative. When following a realistic coastline, I have some Python code to figure this all out for each source (used to ...

kate

The place to check is in the build directory. Look at step3d_t.f90 there to see if your conditional code is there.

For the output you were looking at, there is another file that needs to be modified to get that statement added: checkdefs.F

kate

I am not running METROMS and I'm not able to help out either. Out of the box, CICE5 has no open boundary conditions. The Norwegians have added some to their code and they wrote a paper about it, but the last I heard, Elizabeth Hunke was hoping for a pull request from them for the code. For the ...

kate

Arctic_HYCOM_GLBy

kate

This is using scrip, right? I don't use scrip any more. Use the newer ESMF examples.

kate

This is the forum for the Rutgers ROMS code. I don't know if the person running a 20 km Arctic is reading this forum. Do you have a way to contact them directly?

These big files are a real pain to host and transfer.

kate

I wrote a manual https://github.com/kshedstrom/roms_manu ... manual.pdf, but I haven't updated it in five years.

There is also a wiki: https://www.myroms.org/wiki/Documentation_Portal and I don't know how out of date that is. ROMS keeps changing!

kate

I'm getting an error too: Traceback (most recent call last): File "zlook.py", line 18, in <module> proj=proj, \ File "/import/home/kshedstrom/pyroms/pyroms_toolbox/pyroms_toolbox/zview.py", line 152, in zview resolution='h', area_thresh=5.) File "/import/c1/AKWATERS/kate/.conda/envs/snowdrifts/lib ...

kate

It is has been said by he who wrote the thing that this should be enough documentation: def zview(var, tindex, depth, grid, filename=None, \ cmin=None, cmax=None, clev=None, clb_format='%.2f', \ fill=False, contour=False, d=4, range=None, fts=None, \ title=None, clb=True, pal=None, proj='merc ...

kate

Yes, exactly.

kate

It has been years since I ran that. Did you recompile CICE in debug mode too?

kate

All there is is in the README. The current default branch is named "python3" and does work with Python 3.

kate

According to your input file: Lm == 70 ! Number of I-direction INTERIOR RHO-points Mm == 70 ! Number of J-direction INTERIOR RHO-points N == 30 ! Number of vertical levels This is not the same size grid as your boundary file: netcdf lisound_bry { dimensions: xi_rho = 57 ; xi_u = 56 ; xi_v = 57 ; eta ...

kate

No, the rain file is fine. What file does get_data want to read after that for *your* case?

kate

I'm afraid I'm not the expert on bathy_smoother, but my understanding is that there's more than one way to smooth bathymetry in there. I use an external program that filters bathymetry in a way to target only the roughest patches first, then I finish up with one pass of a Shapiro filter. There's no ...

kate

The questions is - what does get_data.f90 want to read after the rain field? Is there something wrong with that file? Have you looked at that file with say ncview?

kate

Maybe you need an even shorter timestep during a typhoon? The model does want a shorter timestep than you might think, especially with lots of vertical levels.

kate

My latest code is not even trying to load scrip because I commented out these two lines:
```
try:
from pyroms import scrip
```
from pyroms/pyroms/remapping/__init__.py

I got tired of all those scrip warnings! If you really, really want to try loading scrip, uncomment them out again.

kate

See conversations on github https://github.com/ESMG/pyroms/issues/47
and https://github.com/ESMG/pyroms/issues/48

kate

You should contact Jim McWilliams and ask him.

kate

I typically have LMD_SKPP and LMD_NON_LOCAL, not the LMD_BKPP.

kate

What happens without LMD_BKPP?

kate

I would guess that it's more of a vertical mixing problem than an open boundary problem. Have you looked at vertical cross sections?

kate

and OBCFAC == 0.0d0 Tnudge=0. I think this means that I didn't use the nudging. What exactly is OBCFAC == 0.0 doing in the code? Perhaps you should check. You are expecting the boundary to be clamped, so why not clamp the salinity boundary? The reason I don't clamp it is that clamping tends to lead ...

kate

You have two choices: add a second field much like precipitation, or add a bunch of individual sources.

kate

I’m sorry, I don’t remember what I had done there. I’m not at home this week to look into it either.

kate

My https://github.com/kshedstrom/roms repo has a metroms branch which I believe solves the forcing file problem.

Note that this code is completely unsupported!

kate

For river tracers, you have to provide values for all of them. Otherwise your run will be unstable, maybe not today or tomorrow, but someday. It's tricky to find good values to use, but you have to use *something*.

kate

If you want to check out another dead branch, I've got a metroms branch on my ROMS github repo. I did fuss with the forcing.

kate

There were some serious problems with the ice code of roughly seven years ago, since fixed in my branch. Maybe you should start over with METROMS, coupling a modern ROMS with CICE. Or better yet, join me on the MOM6-SIS2 adventure. The frazil ice code is the canary in the coal mine - the most ...

kate

That's a seven year old fork! I suggest you try my moldy two year old repo instead.

kate

Inside ROMS, each variable can have its own time, no problem. The preferred method (last I was aware) is that each variable in the netcdf file should point to the appropriate time variable through the time attribute. This will override the varinfo.dat file. In terms of units, I wouldn't count on ...

kate

This one is not the fault of the Makefile, but rather a bug. Looking at mod_forces, I have: # ifdef SHORTWAVE real(r8), pointer :: srflx(:,:) # ifndef ANA_SRFLUX real(r8), pointer :: srflxG(:,:,:) # endif # ifdef ALBEDO real(r8), pointer :: albedo(:,:) # ifdef ICE_MODEL real(r8), pointer :: albedo ...

kate

So the dependency list is wrong and doesn't include the needs of the ice model. I never used Hernan's build system, so there must be something missing in his Makefile. I looked in his makefile and he's pointing to sea ice modules under ROMS/Nonlinear/SeaIce. Maybe if you move the ice routines to ...

kate

You need to rebuild the dependency list. Some things need to be compiled before other things. The easiest way to do this is "make clean" or removing the build directory or whatever works for how you build the model.

kate

Is this what you are asking about? read_icepar.f90: remark #5133: The input stream is empty Are you sure you turned on the ice model? I checked and my boundary file only has ocean_time. That's how I set up my varinfo.dat: 'zeta_west' ! Input 'free-surface western boundary condition' 'meter' ! [m ...

kate

get_data_ 218 get_data.f90 What is it reading here? Are you sure it's the boundary file? It just read salinity climatology - should it read temperature climatology next? I would run it in the debugger or add a print statement to that line you highlighted. I don't know what's going on. Note: all the ...

kate

Mohamed Iskandarani asked me to post this: I started working with Dale in the early 90’s, he was just setting up shop at Rutgers with Kate Hedstrom when I joined his research group to develop an ocean model based on spectral elements. I had known of Dale from his articles on spectral methods when I w ...

kate

How many times do I have to say "unsupported"?

https://github.com/kshedstrom/roms

kate

Looking back, I want to say a few words about the state of ocean modeling when I met Dale in 1988. There was a well-known global model from GFDL which became MOM. There were a few other models, but they weren't open source. Dale wanted to create a new model with unique numerics, suitable for ...

kate

The way ROMS is compiled, you view the .F file, then an intermediate .f90 file is created off in some build directory. It is those line numbers you need to look at: read_phypar_ 1839 read_phypar.f90 You need to make sure you have an ocean.in that's compatible with the github ROMS, also varinfo.dat.

kate

Those options you mention are from my (unsupported) rogue version of ROMS, not the trunk code. Which code are you using?

kate

Dale Haidvogel played a big part in getting my career started. My PhD project involved a rotating tank with actual salt water. My first job afterwards was to do ocean modeling, but I didn't know anything about it - there were no classes in it then at Scripps. My new boss at the (now long gone ...

kate

1. I'm not really supporting that code any more. 2. Yes, turning on the ice cpp flag is all that's needed. 3. You'll get a more useful model output on seg fault if you recompile in debug mode. 4. Your problem is probably in the ocean.in, with the ice code needing different options. 5. I suspect you ...

kate

My ROMS with ice is no longer being maintained. Hernan was working on a ROMS with ice, haven't heard about it recently.

Your best bet is probably METROMS from the Norwegians.

kate

None of those things need lpsolve. It is used for a very specific type of bathymetry smoothing. Bathymetry smoothing is part of grid generation, where you interpolate bathymetry to your grid, then smooth it enough so that ROMS can run. Calling compiled code from Python has been problematic and it's ...

kate

Do you have the latest pyroms? The current README instructions say to use pip install instead.

Depending on what you plan to do with pyroms, you may not need lpsolve at all. I don't use it. What do you need pyroms for?

kate

Note that my repo hasn't changed in well over a year. I'm not keeping up!

kate

I used point sources which are not through cppdefs, but turned on with LuvSrc and LtracerSrc. I just pushed an update to pyroms for dealing with Dave Hill's runoff.

kate

I assume the wind-wave parameterization you refer to is the one activated by the LI_FOX_KEMPER flag?

Yes

kate

My recent runs have used the lateral flux river sources, not the runoff. I stole a wind-wave mixing parameterization (Li and Fox-Kemper) to add more surface mixing. We also changed the vertical profile of the input so that flow comes in below the surface too, instead of just at the surface. There ...

kate

This ROMS is fundamentally a hydrostatic model. There are other models which include the non-hydrostatic effects of "not small" vertical accelerations. MITgcm is the one I hear most about.

kate

There are two options at least: (a) change pyroms to accept the lat, lon variables as they are in the file or (b) change the file to include the fields that pyroms is looking for.

kate

You will have best results using nesting, the large domain being large enough to enclose all the southern ice. Yes, I am no longer maintaining that ROMS ice model. Yes, the Norwegians are using MetROMS and they have put ice open boundaries into their CICE, similar to the ROMS open boundaries. You ...

kate

It won't do what you want as easily as all that. You have to have 40 different lines with 200 floats each (or vice versa).

kate

I've been using gridpak for my Arctic grids. It's some crazy mess of ancient software that no one else wants to use. It has gotten worse with time - I can't compile xcoast on our current supercomputer. That's one reason the goal is to develop Python tools for all this.

kate

Rob suggests creating a GitHub issue. It’s something he hasn’t tried yet.

kate

I don't know what to say. Perhaps we should disable the building of scrip? It's not like
I use it anymore, as I've said here in these forums before. For me it builds, but doesn't load.

kate

netcdf.inc comes with the NetCDF-Fortran prerequisite. The first question is do you have it? Where is it? The second question is how do you get scrip to find it? There's a makefile in pyroms/pyroms/external/scrip/source using the NETCDF_INCDIR environment variable. Mine is currently pointing to ...

kate

It says it can't find pyroms/src/iso.f. Do you have pyroms/pyroms/pyroms/src/iso.f? I have seven Fortran files in that directory, including iso.f.

kate

You haven't told us which tools you are using to generate initial and boundary files. Impossible to say from here what's wrong.

kate

I should explain the pyroms status: * It's a package written by people who are long gone from Rutgers, no longer using ROMS. * It is only my github repo because I got into git early, I am not a major contributor. * It is no longer representing best modern Python practices. I suggest you dig into ...

kate

If you are missing one day of SSH, can you go to their server and download it directly?

kate

Have you considered asking on github? Or asking Trond? For ERA5, I am using the tools in https://github.com/jsimkins2/nwa25. The README in setup/forcing has a link to how to download files. Once they are downloaded, we do several operations on them before using them to force the model. All the codes ...

kate

You are getting further with scrip than I have in years! My advice is to switch to xesmf. I know, it has its own issues, best to make it its own conda environment.

kate

No, sorry, packing for a trip, not able to help right now.

kate

remap_bdry.py should not be run stand-alone, but rather is called by make_bdry_file.py.

kate

It's not working correctly and my husband has no idea why either. mask_t should have three dimensions, just like var. He says the "~" should be returning valid entries. We want to know which points are masked out. For me, ncview looking at the temp variable shows some white patches where it's land ...

kate

So the shape of mask_t is wrong for some reason. What does that "~" do?

kate

I meant inside Python, print out the shapes of the objects in Python.

kate

I don't know. What are the shapes of all the things?

kate

It is easy to code up and it makes sense over a flat bottom.

kate

No, no, no MIX_S_TS! It's a terrible thing to do with a sigma coordinate. Why wouldn't MIX_GEO work? My biggest problems with it involve wetting and drying and realistic bathymetry. I don't know what you are trying. I'm just suggesting trying a shorter timestep so that you can rule out CFL troubles ...

kate

Yes, there's a CFL condition. There's also the timescale of damping on the scales you want to damp (2 dx) and the timescale of damping on scales you don't want to damp. A CFL of 1.0 will be unstable. What are you shooting for?

kate

It has been a while since I could get scrip to work. The solution is to give up on it, switch to remapping with ESMPy. There's an example using xesmf in examples/Arctic_HYCOM_GLBy.

kate

They have different timestep constraints, depending on dx. You could try a shorter timestep to see if it runs.

kate

Does it run with TNU2? Is this a new domain? Are you sure you have a small enough timestep?

kate

What exactly are you running? It has been a few years since I got scrip to work for me, so I'm switching over to ESMPy. There's one HYCOM example of it.

kate

For what it's worth, this is what I'm using: LBC(isFsur) == Che Che Che Che ! free-surface LBC(isUbar) == Shc Shc Shc Shc ! 2D U-momentum LBC(isVbar) == Shc Shc Shc Shc ! 2D V-momentum LBC(isUvel) == RadNud RadNud RadNud RadNud ! 3D U-momentum LBC(isVvel) == RadNud RadNud RadNud RadNud ! 3D V ...

kate

You are using an unsupported code! I never ran with the ICESHELF, though there are those who did. Yes, the outputs will go from sea bottom to under the ice. There are diagnostics which should be zero where there's no ice that aren't zero. I wouldn't swear by diagnostics ever being right. I abandoned ...

kate

Yes, someone in Enrique's lab wrote a Python code to rename files based on the date in there. Very useful!

MOM6 files all have names starting with the date, such as 20100101.ice_daily.nc

kate

He is at least the second person to run into this issue. I ran into it with daily output over 40 years, each day in its own file. Maybe it wouldn't happen with today's code, I don't know.

kate

It's quite easy. Just change output.F so that you have:

Code: Select all

  10          FORMAT (a,'_',i5.5,'.nc')

instead of

Code: Select all

  10          FORMAT (a,'_',i4.4,'.nc')

and similar for all the other format statements.

kate

Sorry, I stopped supporting my ROMS branch, so John Warner opted to stop including the ice into COAWST.

You can use the metroms coupling to use CICE with ROMS, but CICE isn't on an Arakawa C-grid.

kate

make: *** No rule to make target `/public/home/hufang/project/ecs922/Build_roms/round.f90', needed by `/public/home/hufang/project/ecs922/Build_roms/round.o'. Stop. I am not aware of any round.F or .f90 file. I haven't updated my code in quite some time, so I'm missing things from the main trunk ...

kate

Those NE boundary conditions look ghastly. Something is going on there, I don't know what.

kate

It is bad before taking the first timestep. I would look carefully at your initial conditions.

kate

How long before the final blow-up is this snapshot? You are running COAWST? I have no experience with modes of blowing up for anything but the ocean component. For the ocean, it can go from fine to bad in the blink of an eye. A history snapshot some hours before the trouble can look great. I tend to ...

kate

The first thing to do is to track down the restart issue. You can run the model for two steps, restart it for two more, then compare that solution to a run that's four steps long. You have to restart after an even number of steps because the model doesn't save nstp, which alternates between 1 and 2 ...

kate

You have already compiled your library externally? There is a mod_eco.mod file in that directory? If so, you need to tell the compiler how to find that file. Or you could copy it into the ROMS build directory, which is what is done with the netcdf.mod and friends files.

kate

ROMS does read the varinfo.dat and scale things accordingly.

kate

I'm glad you got it to run. As for the restarts matching exactly, you do not have enough information stored in the history files to accomplish that. The nature of the ROMS timestepping requires information from more than one timelevel of some fields, plus some right-hand-side information to do ...

kate

I have run from an averages file, no problem. Do you have all the fields you need? Did you tell ROMS it was a new run with NRREC=0? If there's biology, did you really save everything you need in the history file? I've had runs die with COBALT when the averages file didn't have say the htotal field ...

kate

You can run for a bit with one output frequency, then restart with a different output frequency. Anything more fancy is going to take hacking the code.

kate

m2 is the barotropic velocity while m3 is the full barotropic plus baroclinic velocity.

kate

Your error might not be in the input files at all. It says:

DEF_VAR - unable to define variable: Tobc_in
in NetCDF file: /scratch/gh9/mx3818/waom/output_1996_2015/

Do you have permission to write to this /scratch/gh9/mx3818/waom/ directory? Does it exist?

kate

You can do the OBC via the ana_ files too. Check out the ana_m2obc.h for example.

kate

Yes, it doesn't care if you have one file or five. The variable names and dimensions all have to be appropriate to their fields.

kate

You can merge them before having ROMS read them and regrid/rotate them if they are both on lat/lon grids of the same resolution. You can copy the ROMS code for reading then regrid/rotating winds to also read/rotate/regrid say Unwind2 and Vwind2. Then you can come up with some means of merging them ...

kate

I would say that any time the answer changes for different tilings, you have a bug. To debug these things, run both tilings for just a timestep and plot the differences between solutions. Which field changes first? If you run 1x4 and 4x1, do you get a grid pattern in the diffs of three vertical ...

kate

I gave up on offline bio because it fought with wetting and drying and all my domains have wetting and drying. Sorry, I probably left it in some half-baked stage.

kate

There are models which know about the different months, but ROMS isn't one of them. You could add that feature with some effort. You could run for 31 days to average for January, then restart and run for x days for February.

kate

/home/markus/roms_project2/NORSJO2/Build_romsG/omega.f90:215

Can you show us what is on this line? Each person's omega.f90 can differ from the others due to differing cpp options.

kate

If you want thirty-day months, that is simple enough to do. Just figure out how many timesteps per month and ask for that.

I typically save daily, one record per file, rename the files to have the date in the filename, then average to get monthlies.

kate

You'd have to try a polar projection instead of Lambert conic. Is that supported by your software?

kate

I bet av2 is simply an average of the two arguments.

Your domain is similar but different from my Arctic domains. Mine cover all the Bering, but less of the Atlantic.

kate

Yes, I think that should work. There are a few ana_trc_psource.h files in https://github.com/kshedstrom/Apps, mostly (all?) for steady tracer sources. They are meant to go with my ROMS fork so I have no idea how incompatible the trunk code is these days.

kate

Is it in the roms.in or in the grid file? My grid file has:

Code: Select all

// global attributes:
		:type = "ROMS grid file" ;

kate

Is this a time when there should be full pack ice preventing the gyre? I had to have really strong nudging on the ice fields to make things sort of behave.

kate

INP_PAR:FIND_FILE - empty file name string for standard input script KeyWord: SSFNAME

Do you have a line for SSFNAME? That is a newish requirement.

kate

Yes, gridpak on github. What do you want in an Arctic grid? I've got two you could try as well.

kate

I'd think it would be easier to use a restart file. Would this be a problem for some reason?

kate

Rob Cermak has been working on https://github.com/ESMG/gridtools for making MOM6 grids, including handling polar grids. There's a ROMS-to-MOM6 grid tool and one could easily make the reverse. I make my Arctic grids using old, old grid tools which still work and still produce the psi fields along ...

kate

I assume you know what I used for a regional Beaufort domain. I would go back to the RadNud option and also use nudging to climatology. If it's blowing up due to high speeds, where is that happening? Is that the usual vertical instability needing a short, short timestep or something else? As for ...

kate

How did you make your grid? Some grid generators add more points to the outside of the model domain than others. You can just extrapolate what you need from the interior. Or if all the lat_rho, lon_rho values are zero, then there's a problem in the generation phase.

kate

I'm sure I've said this a number of times before, but it's time to develop some debugging skills. Plot the results after one timestep, try a shorter timestep just in case. Look to see what is going bad and where. If you want ice, I'd keep it on since turning it off isn't helping. The ice-free ...

kate

Another option for fussing is the bottom drag. You can make a spatially variable bottom drag and modify it to get a better match.

kate

The best place to search is the code itself. The file to look at is ROMS/Functionals/ana_srflux.h, which has options both including ALBEDO_CLOUD and not including it.

kate

If you can get xxgdb to work, wonderful! I would expect it to be fine as long as your job fits onto one core/process.

kate

I would look at this in a debugger, but lacking that, try print statements. You know it's something to do with dTdt or dTdx, so print them and the things that go into them. It's happening early enough in the run that this shouldn't be too insane. Print the i,j,k values so you know which point(s) are ...

kate

GET_2DFLD - snow fall rate, 1998-01-15 03:00:00.00 (Grid=01, Rec=0000114, Index=1, File: final_snow_1998.nc) (Tmin= 18263.0000 Tmax= 18627.8750) t = 18277.1250 (Min = 1.00000000E+35 Max = -1.00000000E+35) regrid = T You've got some crazy snow fall rates here. The ice frazil thing can be triggered ...

kate

Yes, old netcdf files had a 2 GB limit. You can make a bunch of under 2 GB files and ask ROMS to read from one, then the next, then the next. Or you can convert them to a newer netcdf format (possibly even with compression) and concatenate them into one file.

kate

How large are your wind stresses?

kate

Why don't you look through the trunk code examples to see what they use?

#undef MIX_GEO_TS
#define MIX_S_TS

I would swap these.

# define LI_FOX_KEMPER

This is not in the trunk.

kate

The trunk code does not have CCSM_FLUX, nor does it have all those albedo options. You should use the .h file consistent with your code.

kate

TIME_REF = 2000101.0d0

I think it thinks you have year 200 instead of year 2000. Try adding another 0.

kate

Woah! Where did you get that DSTART? It should be the number of days between your time origin (2000) and your starting time (2013?). That would make 13*365 plus the number of leap days in there, so 4749 days.

kate

So, do you know what to do next? You need to change your TIME_REF to match that of your forcing files. I have: ! DSTART = 29226.5d0 ! days (1980) DSTART = 42003.5d0 ! days (2015) TIDE_START = -1.0d0 ! days TIME_REF = 19000101.0d0 ! yyyymmdd.dd

kate

What about trying "ncdump -h" on one of these things? I get: netcdf JRA55DO_1.5_rain_2020 { dimensions: lon = 640 ; lat = 320 ; time = UNLIMITED ; // (2928 currently) variables: double lon(lon) ; lon:units = "degrees_east" ; lon:valid_min = 0. ; lon:valid_max = 359.4375 ; lon:long_name = "longitude ...

kate

What is the time variable like in one of these things? You need one common time origin for *everything*.

kate

You have only one month of forcing per file? You can string them together in the roms.in file. I have my forcing files with one year each and string them together from one year to the next: NFFILES == 14 ! number of forcing files FRCNAME == /center1/AKWATERS/kate/JRA55-do/JRA55DO_1.4_Pair_2015.nc ...

kate

#define MIX_S_TS

This is OK if your bottom is flat, otherwise it leads to mixing along s-surfaces which can be unrealistic. I don't see anything else wrong, but I could be missing something, who knows.

kate

So how are you forcing it? You have:

#define ANA_STFLUX
#define ANA_SSFLUX

Are these set to zero or something reasonable?

Note that I have been using BULK_FLUXES (with CCSM_FLUXES) and atmospheric reanalysis fields for over a decade now.

kate

get_idata.f90(73): error #6404: This name does not have a type, and must have an explicit type. [ING] & TIDE(ing)%pioFile, & -------------------------------^ compilation aborted for get_idata.f90 (code 1) make: *** [/nesi/project/mocean03294/NZB_3/build/get_idata.o] Error 1 Looks to me like you ...

kate

Your model is salty at the surface when it should be fresh. What are you using for fresh water forcing?

kate

It could be that ROMS only checks once when deciding whether to remap or not. The check happens at line 226 of nf_fread2d.F.

kate

Adding a strip of land-masked cells is fair amount of work.

This is why it is good to know your tools. We have landmask editing in both Matlab and Python. Also, it is worthwhile to learn to edit netcdf files in SOME programming language.

kate

Actually, it looks like some bugs have been introduced. This:

Code: Select all

INI(ng)Vid(idVwet)=Vindex

is not the proper syntax, as the compiler notes. Try

Code: Select all

INI(ng)%Vid(idVwet)=Vindex

instead.

kate

I would look for idRwet and friends in your varinfo.dat. They need to be there.

kate

An Arakawa C-grid is supposed to be staggered, with different sizes of the different variables. I haven't used panoply, but you need to teach your plotting tools about the staggered C-grid.

kate

So the same configuration runs with 16 vertical levels and blows up with 30? This is not surprising. When you add more vertical levels, the vertical advection requires a shorter timestep. Your run ran long enough that it seems fine except for the timestep limitation. Try a smaller DT. Your vertical ...

kate

You may well be right, but as I said elsewhere, I'm no longer maintaining that code.

kate

It is true that Paul Budgell put the ICE_SHOREFAST in there for the Caspian, but that was a long, long time ago. The ICE_LANDFAST is what I would use now, it's the Lemieux landfast ice parameterization which is also in CICE6. If you need it in CICE5, I can help with that. Why would you read the ...

kate

I'm sorry, I'm no longer supporting that ice code. I'm five years into a plan to switch to using MOM6 instead, finally getting somewhere.

For ice with ROMS, the Norwegians are supporting their metroms package (but don't try to learn git from their example).

kate

The vertical diffusivity is in Akt, which is dimensioned by three spatial dimensions and the NT (number of tracers) dimension. You can set some to zero. There will still be implicit changes which I don't think you can do anything about other than try different advection schemes.

kate

In modeling, there is always implicit diffusion from your advection operator. You can reduce this by going to a fourth-order advection scheme. I use the third-order one with its implicit diffusion, then turn off explicit diffusion by turning off the cppdefs for it. Vertical diffusion is trickier ...

kate

Did you provide river tracer values?

kate

As I posted elsewhere, I've given up on offline biology. The point of it was that you run the physics once, saving daily outputs. Then you read those fields in as "climalotogy" for running the biology. The dream is that you can maybe take a longer timestep when you only have the biology going on and ...

kate

If this is my code, I've given up on OFFLINE_BIOLOGY. I sort of had it working in the "main" branch, not the default "kate_svn" branch. But for me, I have wetting and drying and it wasn't doing the right thing in the dry parts, nor did it even know if it was wet or dry that day.

kate

This is when you need to become a detective. Which field is blowing up. You say it is in the land? Is that the only place going bad?

kate

I haven't used BIO_UMAINE since Hernan changed the advection scheme choices. Good catch, John. As for my github code, I have stopped updating it because updates can be disruptive and I don't want to spend the time sorting them out. Mark Hadfield is keeping his github code up to date. If you find any ...

kate

The check you are running into is in diag.F: IF (maxbio(ibt,ng).gt.max_bio(ibt)) THEN It looks like you have max_bio set to 10000. Does that make sense for the units you are using? You can change it if 10000 is a valid number for your BGC model. I probably added the check for COSINE, but then kept ...

kate

That's great that the model is running without any problems. You're all set, then, right? You need to look at your solution to see if you are happy with it. Are you happy with it? If not, there are steps you can take to possibly improve your boundary solution, but nothing is guaranteed when it comes ...

kate

How are you getting your ROMS information? There is both a manual and a wiki. The wiki page might help you. The manual has much the same information and also examples containing open boundaries.

kate

I see the wiki page on FLOATS is still empty. There is probably something about floats in the manual, though.

kate

The FLOATS option is a whole different can of worms. Is that what you want? Online or offline?

kate

What sort of open boundary do you want? In the UPWELLING case, the southern edge is in shallow water and typically has a wall. Are you changing the bathymetry? Are you providing fields on the open boundary? I never use a sponge in the ROMS sense of the word. I do use nudging to a climatology, which ...

kate

You need these as well:

Code: Select all

# define ANA_BPFLUX        /* analytical bottom passive tracers fluxes */
# define ANA_SPFLUX

unless you want the top and bottom tracer fluxes to be non-zero.

kate

First, you turn on T_PASSIVE and set NPT in roms.in. For the initial conditions, ANA_PASSIVE can be used, otherwise you'll have to put the initial conditions into a netcdf file. I've usually initialized with zero via ANA_PASSIVE, then used sources to introduce the passive dye locally. What is your ...

kate

The University of Colorado Cooperative Institute for Environmental Sciences (CIRES) jointly with the NOAA/Physical Sciences Lab is hiring an associate scientist in the area of ocean data analysis/visualization. Here is a brief position description: The chosen candidate will assist in the development ...

kate

This position is closing soon. If you intend to apply, get your application submitted now.

kate

Yes, the diagnostics are time-averaged.

kate

Make sure NBCFILES is defined. Some roms.in files have it, some do not. If using COAWST, you need to have it.

kate

For the incoming 2-D forcing, it depends. ROMS will check the dimensions of each field you provide. If they match that of your ROMS grid, it assumes Uwind and Vwind are aligned in the directions of the grid. If not, it assumes it needs to do the interpolation and the rotation to the ROMS grid.

kate

I use SODA as my initial condition, no spin-up period. You can try a spin-up and see how much drift you get.

kate

I typically add a nudging to a monthly climatology in a band around the edge of the domain, ramping to zero in the middle. Note that this will overwrite TNUDG and friends in the default ROMS code.

kate

GET_2DFLD - surface u-wind component, 2013-01-01 00:00:00.00 (Grid=01, Rec=1, Index=1, File: era5_u10_year_2013_roms.nc) (Tmin= 0.0000 Tmax= 30.9583) t = 0.0000 (Min = 1.00000000E+35 Max = -1.00000000E+35) regrid = T GET_2DFLD - surface v-wind component, 2013-01-01 00:00:00.00 (Grid=01, Rec=1 ...

kate

NETCDF_GET_TIME_1D - error while inquiring ID for variable: srf_time in input file: era5_msdwswrf_year_2013_roms.nc What does "ncdump -h era5_msdwswrf_year_2013_roms.nc" look like? Is there a time variable? What is it called? ROMS is looking for srf_time and it could have gotten this from (a) the ...

kate

Same answer as above - use a shorter timestep and look more closely at your model output files.

kate

You know your DT and you should be able to figure out how many timesteps per day from that. Say DT=120 leads to 720 steps/day. Three days of steps is therefore 2160 steps.

kate

When going coarser, you can take longer timesteps. When going coarser, are you reducing the number of grid points?

Why not use MPI?

kate

The debugger I use is called totalview. There is also ddt. Both are quite expensive, but the supercomputer center I use pays for the license. For free, there is gdb, which works if you run on just one processor. There may be a gui (ddd?) that can make it easier to use, but using these things is ...

kate

We know it's dying for some unknown reason. So maybe it didn't actually finish ana_initial. I would look in there. Did you change anything there?

Personally, I would run this in the debugger to watch it go bad.

kate

You know it got to ANA_INITIAL because it printed stuff from there. I think the next calls it would make from initial.F are these: CALL set_depth0 (ng, tile, iNLM) CALL set_depth (ng, tile, iNLM) CALL set_massflux (ng, tile, iNLM) I would add print statements before these, then after these, then ...

kate

What I would do is recompile netcdf-fortran. Did you compile it yourself before or use the one from "apt install"?

kate

I'll say what I always say: try a shorter timestep and look more closely at your solution to see where it is blowing up.

kate

Do you have USE_NETCDF4? Do you have netcdf installed properly?

kate

I think you want to have USE_MPIF90 turned on.

kate

Hz is the height of the vertical grid box, which you could call Hz=∂z/∂σ.
For consistency between the baroclinic and barotropic solutions, you also need D=ΣHz.

kate

I wanted to let everyone know that Princeton is now accepting applications for a Princeton/GFDL postdoc to further develop the regional capabilities of the MOM6 ocean model: https://puwebp.princeton.edu/AcadHire/apply/application.xhtml?listingId=19521 The position would include algorithmic ...

kate

I would not edit UPWELLING, but rather create a new name for your case. Then search the whole code for UPWELLING so you know what files to look in. For the flat bathymetry, the "else" case on ana_grid.h should take care of that for you. For the wind stress, you should be able to find that in the ...

kate

There is some discussion of the timestepping in the roms manual . Most of them point to the old/new time which alternates between 1 and 2, while 3 is always a working time level for tracers. Note: on restart, the model doesn't save these things and always assumes whatever is true for the first step ...

kate

initial_ 219 initial.f90

What is on line 219 of your initial.f90? This would be in your scratch directory.

kate

Yes, you got it.

I don't use Matlab if I can help it. You should get the script from the ROMS Matlab repository.

kate

Online nesting requires a special 1:3 or 1:5 relationship between the grids, as in Lake Jersey. Both models are run at the same time and the computer spends a lot of time communicating between the domains for you. That's what the contact stuff is for. I did this only once and it was pretty painfully ...

kate

You are talking about doing online nesting in that case. I have a lot less experience with that than with offline nesting.

kate

When doing offline nesting, each domain is a fully independent domain. You run the larger, coarser one first, then use fields from that domain to create the initial/boundary files for the smaller domain. This creation of boundary files for the child grid happens offline, for instance with pyroms ...

kate

I always have USE_MPIF90 on so that it knows where to find mpif.h.
Are you using COAWST? It is recommended to use that code when coupling to SWAN. Not that I've done it with any of the versions.

kate

All those Dout flags control the output, not the tides. I have this section in my .h file: /* tides */ #define LTIDES #ifdef LTIDES # if defined AVERAGES && !defined USE_DEBUG # define FILTERED # endif # define SSH_TIDES # define UV_TIDES # define ADD_FSOBC # define ADD_M2OBC # undef RAMP_TIDES ...

Search found 4082 matches