4DVAR misfit_initial have many zero values

[xwu]

Hi all,

Following the WC13 instructions, I set up a RBL4DVAR test on a model grid receiving realistic atmospheric forcing. The nonlinear model has run for one model year and the model outputs were used to create the standard deviation files for 4DVAR. For my RBL4DVAR test case, the time window is three days. The observation data have T and S from UK Met Office Hadley Center. Both the tangent and adjoint models can run successfully. However, when I checked the 'misfit_initial' and 'misfit_final', I found out that 'misfit_initial' have nonzero values during the first six hours, then 'misfit_initial' becomes all ZERO (see panel c below). This is different from the WC13 RBL4DVAR results, wherein 'misfit_initial' shows nonzero values within the entire time window (see panel a below).

I also checked the increment of the model initial conditions in my test case. The results show that the initial values of (zeta, u, v, T, S) all have zero increment, whereas initial values of (taux, tauy and Q) have Nonzero increments (see below). This is also different from the WC13 RBL4DVAR results as the initial (zeta, u, v, T, S) were all adjusted in WC13.

I double checked my observation input file and confirmed that all the observation (lon,lat) are within the model grid. The 4DVAR log file ('log_rect_rbl.txt'), the roms input file ('roms_rutger.txt') and header file ('crect.h') were attached here.

I am wondering, why the 'misfit_initial' is nonzero at first and then becomes all zeros? I want to ask for any suggestions to help track and resolve this issue. I can provide more information if needed.

thanks for your time,

Xiaodong

[arango]

Please stop sending me personal emails. They will go directly to the trash bin. I don't have the time to reply to specific emails from users when something doesn't work. ROMS has thousands of users, and our policy is not to answer such particular inquiries. All the communications about ROMS should be in this Forum so others can use such valuable information.

If you were a little more curious, you would find out in log_rect_rbl.txt all the observations after 2021-11-02 23:40:00.00 are rejected:

Code: Select all

Number of State Observations Processed:  ObsTime =    8341.9861,  2021-11-02 23:40:00.00

          Variable          IstrObs    IendObs      Count   Rejected

          temp                 9162       9187         26         26
          salt                 9188       9203         16         16

          Total                                        42         42
          Obs Tally                                    42         42
 

Thus, the problem is in your file /home/eb/Projects/Esea/Rect/RBL4DVAR/Input/rect_ts_obs_3days_bound.nc. You need to figure out why the observations are rejected. Check your data and processing scripts. There is nothing we can do, nor can any user in this community explain it to you!

ROMS provides an extraordinary amount of information in the standard output file, and it is your responsibility to examine and understand such information before you post it in this Forum or by personal email when something is not working.