Scheduled Downtime
On Friday 21 April 2023 @ 5pm MT, this website will be down for maintenance and expected to return online the morning of 24 April 2023 at the latest

Regarding SST update with 3km

This post was from a previous version of the WRF&MPAS-A Support Forum. New replies have been disabled and if you have follow up questions related to this post, then please start a new thread from the forum home page.

EMMANUEL

Member
I have updated SST daily while running an atmosphere run. However, when I tried to visualize SST from the history file each time frame using ncview I noticed that SST does not change and remains the same also next frame SST looks distorted. Can you kindly suggest to resolve this issue
However other parameters looks fine
thanking you
 
The SST field is updated in a step-wise fashion, so if your SST update interval is 24 hours, then we would expect the SST field to be constant for the first 24 hours of the simulation, then update to the second state in your SST update file and remain at that state for the next 24 hours, and so on. (Note that there is a namelist option, config sstdiurn update, that applies a diurnal cycle to the SST field, but that option is disabled by default.)

Have you tried plotting the SST field at all times from your SST input file to see whether those fields look correct, or whether they show the same "distorted" pattern that you're seeing in your model history files?
 
mgduda said:
The SST field is updated in a step-wise fashion, so if your SST update interval is 24 hours, then we would expect the SST field to be constant for the first 24 hours of the simulation, then update to the second state in your SST update file and remain at that state for the next 24 hours, and so on. (Note that there is a namelist option, config sstdiurn update, that applies a diurnal cycle to the SST field, but that option is disabled by default.)

Have you tried plotting the SST field at all times from your SST input file to see whether those fields look correct, or whether they show the same "distorted" pattern that you're seeing in your model history files?

I have checked SST input file(grib file) however there is not much diff in SST. Although the image looks fine in input for each time steps but output has distorted file
 
It's good to know that the SST fields in the GRIB files look fine. Can you check the SST fields in the NetCDF input file that you prepared with config_init_case=8 using the init_atmosphere_model program (i.e., the same NetCDF file -- possibly named something like x1.65536002.sfc_input.nc -- from which the model is reading)? An important question is whether the SST fields were correctly interpolated and written by the init_atmosphere_model program; knowing the answer to this will indicate whether the problem is likely to be in the init_atmosphere_model program or in the atmosphere_model program.

Can you also attach a plot of the "distorted" SST field from one of your model history files? It's sometimes helpful to see what the erroneous field actually looks like.
 
I tried updating SST but I have an issue still. can you kindly suggest on this. First time step looks fine while next time step is distorted. While all the other variables are fine and has no issue

I have noticed that in SST surface update each day sst plot the axis are -180 to 180 and -90 to 90 whereas in the output the axes are 0 to 180 to 0 .
Is it being different axes the issue caused?


mgduda said:
It's good to know that the SST fields in the GRIB files look fine. Can you check the SST fields in the NetCDF input file that you prepared with config_init_case=8 using the init_atmosphere_model program (i.e., the same NetCDF file -- possibly named something like x1.65536002.sfc_input.nc -- from which the model is reading)? An important question is whether the SST fields were correctly interpolated and written by the init_atmosphere_model program; knowing the answer to this will indicate whether the problem is likely to be in the init_atmosphere_model program or in the atmosphere_model program.

Can you also attach a plot of the "distorted" SST field from one of your model history files? It's sometimes helpful to see what the erroneous field actually looks like.
 

Attachments

  • date1_sst.jpg
    date1_sst.jpg
    28.5 KB · Views: 787
  • date9_sst.jpg
    date9_sst.jpg
    28.8 KB · Views: 787
In the first plot you've attached, it looks like the SST field is reasonable over water points and only "distorted" over land points (where the SST isn't used). If the input SST field to the init_atmosphere_model program is masked out over land and contains only valid values over water, if I recall correctly, the init_atmosphere_model program will extrapolate to land points from the nearest water points, giving rise to patterns like the one you've shown.

If you ignore the land points, and if the SST field over water is as you expect, then I think your simulations may be fine. As a test, you could modify the SST field in your surface update file so that the SST field is, say, 0 K over land; re-running a test simulation should show no change in results, confirming that the SST field over land points is in fact ignored by the model.
 
I have done the simulation again, it shows no change with 0 K. So it is ok with the simulation.
Apart from this, I would like to know is it ok to use the same gwdo static file (which I created initially for uniform 3km mesh, generally we use the same in case of WRF, would like to know is it the same case in MPAS too) for different IC dates having the same resolution?


mgduda said:
In the first plot you've attached, it looks like the SST field is reasonable over water points and only "distorted" over land points (where the SST isn't used). If the input SST field to the init_atmosphere_model program is masked out over land and contains only valid values over water, if I recall correctly, the init_atmosphere_model program will extrapolate to land points from the nearest water points, giving rise to patterns like the one you've shown.

If you ignore the land points, and if the SST field over water is as you expect, then I think your simulations may be fine. As a test, you could modify the SST field in your surface update file so that the SST field is, say, 0 K over land; re-running a test simulation should show no change in results, confirming that the SST field over land points is in fact ignored by the model.
 
Thanks for following up, and it's good to confirm that the model ignores the SST field over land points.

All fields in the static gwdo file are time-invariant, so you can re-use the same static gwdo file for different IC dates. In that sense, MPAS-Atmosphere's static gwdo file is similar to a "geo_em' file in the WRF modeling system.
 
Thank you for sharing the details precisely.


mgduda said:
Thanks for following up, and it's good to confirm that the model ignores the SST field over land points.

All fields in the static gwdo file are time-invariant, so you can re-use the same static gwdo file for different IC dates. In that sense, MPAS-Atmosphere's static gwdo file is similar to a "geo_em' file in the WRF modeling system.
 
Top