search

trondkr / ERA5-ROMS

Download, and create ERA5 data forcing for ROMS
17 stars 13 forks source link

Problem with using output to run ROMS #3

Closed trondkr closed 2 years ago

trondkr commented 3 years ago

When testing the current output of this package for an Arctic ROMS model it seems that ROMS has either 1) a problem with one of the variables that prevent seasonality to emerge in the ROMS solution 2) the internal interpolation in ROMS is broken.. This needs to be further investigated

lushilima commented 3 years ago

Hi Trond, I found the same issue, I supose. When we use ERA5 to run ROMS.

` Initial domain volumes: TotVolume = 3.2449563409E+16 m3 MinCellVol = 1.1849078930E+07 m3 MaxCellVol = 1.4712066248E+11 m3 Max/Min = 1.2416210859E+04

NL ROMS/TOMS: started time-stepping: (Grid: 01 TimeSteps: 000000012097 - 000000038880)

GET_2DFLD   - surface u-wind component,                        1998-01-15 00:00:00.00
               (Grid=01, Rec=0000113, Index=2, File: era5_u10_year_1998_roms.nc)
               (Tmin=      18263.0000 Tmax=      18627.8750)      t =      18277.0000
               (Min =  1.00000000E+35 Max = -1.00000000E+35)      regrid = T
GET_2DFLD   - surface v-wind component,                        1998-01-15 00:00:00.00
               (Grid=01, Rec=0000113, Index=2, File: era5_v10_year_1998_roms.nc)
               (Tmin=      18263.0000 Tmax=      18627.8750)      t =      18277.0000
               (Min =  1.00000000E+35 Max = -1.00000000E+35)      regrid = T
GET_2DFLD   - surface air pressure,                            1998-01-15 00:00:00.00
               (Grid=01, Rec=0000113, Index=2, File: era5_msl_year_1998_roms.nc)
               (Tmin=      18263.0000 Tmax=      18627.8750)      t =      18277.0000
               (Min =  1.00000000E+35 Max = -1.00000000E+35)      regrid = T`

Did you found the same issue, after the first time-step, all the atm data gets crazy? In the ROMS forum, I realized that someone else also had the same error and also used the ERA5-roms.

lushilima commented 2 years ago

Hi Trond, I solved the problem that I described above. This happened due the latitude was inverted. And consequently, the model couldn't read the data correctly.

trondkr commented 2 years ago

@lushilima Great! Would you mind posting the code for inverting the latitude and where you did this (or do a pull request)? Then I can fix it too. Thanks

lushilima commented 2 years ago

Well, I tested this changing the data with NCO... So now I am testing and trying to change on the script. Also, I have changed the time config because my simulations use seconds instead of hours as a pattern.