zero value in tangent linear initial file

Discussion about tangent linear and adjoint models, variational data assimilation, and other related issues.

Moderators: arango, robertson

Post Reply
Message
Author
simion1232006
Posts: 60
Joined: Tue Sep 29, 2009 3:50 pm
Location: School of Environment System Engineering,UWA

zero value in tangent linear initial file

#1 Unread post by simion1232006 »

I cut part of the log file, see below:
The nonlinear model runs all right, while when the model switch to the tangent model, it creates a initial file with all zero values (NWS_200911_itl.nc). Does anyone have any experience on this? I am quite new in regional model data assimilation. Many thanks.


10740 40145 20:00:00 7.110248E-03 2.045630E+04 2.045631E+04 2.978209E+15
(238,162,24) 1.345348E-02 2.720841E-02 7.285469E-02 1.565786E+00
10770 40145 22:00:00 7.111070E-03 2.045630E+04 2.045631E+04 2.978211E+15
(238,162,23) 1.352397E-02 2.705895E-02 7.701877E-02 1.543674E+00
10800 40146 00:00:00 7.122028E-03 2.045630E+04 2.045631E+04 2.978212E+15
(237,159,22) 9.442482E-03 3.606497E-02 7.121669E-02 1.542963E+00
WRT_HIS - wrote history fields (Index=1,1) into time record = 0000061
WRT_RST - wrote re-start fields (Index=1,1) into time record = 0000001

Number of State Observations Processed: ObsTime = 40146.0000

Variable IstrObs IendObs Count Rejected

zeta 302525 303915 1391 0

Total 1391 0
Obs Tally 303915 0

Wrote NLM state at observation locations, datum = 0302525 - 0303915

WRT_FRC - wrote initial fields (Iter=0001, Index=1, Rec=0001)
WRT_FRC - wrote initial fields (Iter=0001, Index=1, Rec=0002)
DEF_HESSIAN - creating Hessian file: NWS_200911_hss_001.nc
1 1 1 1

<<<< 4D Variational Data Assimilation, Outer = 001, Inner = 000 >>>>


TL_INITIAL: Configuring and initializing tangent linear model ...

TL_DEF_INI - creating initial file: NWS_200911_itl.nc
1 1 1 1
TL_WRT_INI - wrote initial fields (Iter=0001, Index=1,1, Rec=0001)
TL_WRT_INI - wrote v-increments (Iter=0001, Index=1,1, Rec=0002)

TLM: GET_STATE - Read state initial conditions, t = 40116 00:00:00
(File: NWS_200911_itl.nc, Rec=0001, Index=1)

- free-surface
(Min = 0.00000000E+00 Max = 0.00000000E+00)
- free-surface, open boundaries
(Min = 0.00000000E+00 Max = 0.00000000E+00)
- vertically integrated u-momentum component
(Min = 0.00000000E+00 Max = 0.00000000E+00)
- vertically integrated u-momentum component, open boundaries
(Min = 0.00000000E+00 Max = 0.00000000E+00)
- vertically integrated v-momentum component
(Min = 0.00000000E+00 Max = 0.00000000E+00)
- vertically integrated v-momentum component, open boundaries
(Min = 0.00000000E+00 Max = 0.00000000E+00)
- surface u-momentum stress, adjusted tl_ustr
(Min = 0.00000000E+00 Max = 0.00000000E+00)
- surface v-momentum stress, adjusted tl_vstr
(Min = 0.00000000E+00 Max = 0.00000000E+00)
- u-momentum component
(Min = 0.00000000E+00 Max = 0.00000000E+00)
- u-momentum component, open boundaries
(Min = 0.00000000E+00 Max = 0.00000000E+00)
- v-momentum component
(Min = 0.00000000E+00 Max = 0.00000000E+00)
- v-momentum component, open boundaries
(Min = 0.00000000E+00 Max = 0.00000000E+00)
- potential temperature
(Min = 0.00000000E+00 Max = 0.00000000E+00)
- salinity
(Min = 0.00000000E+00 Max = 0.00000000E+00)
- potential temperature, open boundaries
(Min = 0.00000000E+00 Max = 0.00000000E+00)
- salinity, open boundaries
(Min = 0.00000000E+00 Max = 0.00000000E+00)
- surface net heat flux, adjusted tl_tflux
(Min = 0.00000000E+00 Max = 0.00000000E+00)
- surface net salt flux, (E-P)*SALT, adjusted tl_tflux
(Min = 0.00000000E+00 Max = 0.00000000E+00)
GET_2DFLD - surface u-momentum stress, t = 40116 00:00:00
(Rec=0001, Index=1, File: NWS_200911_fwd_000.nc)
(Tmin= 40116.0000 Tmax= 40146.0000)
(Min = -1.15952858E-04 Max = 6.20054110E-05)
GET_2DFLD - surface v-momentum stress, t = 40116 00:00:00

simion1232006
Posts: 60
Joined: Tue Sep 29, 2009 3:50 pm
Location: School of Environment System Engineering,UWA

Re: zero value in tangent linear initial file

#2 Unread post by simion1232006 »

I ran the WC13 case (the strong constrained case) and got the similar log information. In the next loop, the model updated the itl file, and the value are not zero any more. In the weak constrained case, everything is normal.

There should be nothing wrong with it. Never mind.

clemente.tanajura
Posts: 1
Joined: Sat Jun 02, 2012 2:48 pm
Location: UFBA & UCSC

Re: zero value in tangent linear initial file

#3 Unread post by clemente.tanajura »

Yes, it is correct. The tangent linear model deals with increments of the control vector and it is initialized with zero.

Good luck
Clemente A. S. Tanajura
Oceanographic Modeling and Observation Network (REMO)
UFBA, Brazil

Post Reply