reversed time variable in forcing file

[flcastej]

Dear all,

I have realised of a issue about the time variable in my forcing file and I will like to share it with you to get your feedback about it.

I have prepared my forcing file and ncdump -v sms_time mysimulation_frc.nc gives me:

double shf_time(shf_time) ;
shf_time:long_name = "surface heat flux time" ;
shf_time:units = "seconds since 0001-01-01 00:00:00" ;
shf_time:cycle_length = 1. ;


shf_time = 63445507200, 63446889600 ;

And when is read by ROMS, it gives:

GET_2DFLD - surface net heat flux, t = ***** 00:00:00
(Rec=0000001, Index=2, File: marmenor_frz.nc)
(Tmin= 734323.0000 Tmax= 734339.0000)
(Min = 0.00000000E+00 Max = 0.00000000E+00)

My mysimulation.in file have :

DSTART = 734323 ! days
TIDE_START = 0.0d0 ! days
TIME_REF = 00010101.00 ! yyyymmdd.dd

But, when I am trying to run the model, I get:

SET_2DFLD - current model time exceeds ending value for variable: shflux
TDAYS = 734323.0000
Data Tmin = 734323.0000 Data Tmax = 734339.0000
Data Tstr = 734339.0000 Data Tend = 734323.0000
TINTRP1 = 734338.0000 TINTRP2 = 734323.0000
FAC1 = 0.0000 FAC2 = -15.0000

It seems like Tstr and Tend are reversed. If I prepare the forcing file with shf_time reversed :

shf_time = 63446889600, 63445507200 ;

The model runs without problems. Although I can run my simulation, I will like to know what I am doing wrong.

Thanks a lot !!!

Thanks a lot.

[flcastej]

Finally, after reading several times the forum and review the code, I have been able to fixed it and set the forcing file properly. I have done two things:

1.) Change the reference time from 0001-01-01 00:00:00 to 1968-05-23 00:00:00 in order to work with smaller number and prevent any precision or floating point problem.

2.) My forcing file has a cycle_length=1. I have removed the cycle_length variable in my forcing file. I think that it was the reason why I needed to put the date reversed.

I hope it could be useful to someone in the future.