hour or day

[jafar1979]

Dear Prof.Arango
i guess that time unit of force file which created by d_ecmwf2roms.m should be hours not days . My reason is :created force file have time size equal to 3hourly ecmwf data (both of them have 2920 time steps for one year), Therefore their units should be equal too?is not it?

Code: Select all

S.Variables(4).Attributes(iatt).Value = ['days since ' datestr(mybasedate,31)];

Please if i am on wrong tell me?
Happy New Year

jafar

[arango]

Nope. The time variable (sms_time, srf_time, shf_time, or swf_time) in the NetCDF file created by the script d_ecmwf2roms.m is in days The number of time records in a NetCDF file is independent of the units of the time coordinate. In ROMS, the default is to provide the units for the forcing fields as days since a reference time.

You are confusing the units given by the ERA file:

Code: Select all

dimensions:

        time = 2920 ;

variables:

        int time(time) ;
                time:units = "hours since 1900-01-01 00:00:0.0" ;
                time:long_name = "time" ;

In a leap year, you will have 2928 time records. Otherwise, you will get 2920 time records.

You cannot use directly the file generated by ERA into ROMS We need to process the ERA fields to get the correct force for ROMS. Just check the time variable for the ROMS forcing NetCDF files. For example, we can have:

Code: Select all

	double sms_time(sms_time) ;
		sms_time:long_name = "surface momentum stress time" ;
		sms_time:units = "days since 2000-01-01 00:00:00" ;
		sms_time:calendar = "gregorian" ;

...

data:

 sms_time = 0.0625, 0.1875, 0.3125, 0.4375, 0.5625, 0.6875, 0.8125, 0.9375, 
    1.0625, 1.1875, 1.3125, 1.4375, 1.5625, 1.6875, 1.8125, 1.9375, 2.0625, 
    2.1875, 2.3125, 2.4375, 2.5625, 2.6875, 2.8125, 2.9375, 3.0625, 3.1875, 
    3.3125, 3.4375, 3.5625, 3.6875, 3.8125, 3.9375, 4.0625, 4.1875, 4.3125, 
    4.4375, 4.5625, 4.6875, 4.8125, 4.9375, 5.0625, 5.1875, 5.3125, 5.4375, 
...

[jafar1979]

Dear Prof. Arango
many thanks for your complete information.
i know that you are so busy but if it is possible for you would you please tell me what kind of process on the ERA fields to get the correct force for ROMS is needed.
Because i use ERA into ROMS directly.
Cheers
jafar

[kate]

This is where some computer hacking comes into play. Hernan might have a script for this in matlab, but you need to gain some skills in dealing with NetCDF files for this sort of pre- and post-processing of ROMS files. Pick a scripting language and get comfortable with it. There are many options: Matlab, Python, Ferret and NCL all have their champions in the ROMS community.

Another option is to hack ROMS so that it accepts a time units attribute in hours and does the right thing.

[jafar1979]

Dear Kate
thanks as usual.
It is possible to sent mentioned Matlab script for me or sent its link . I usually use Matlab as a scripting language.
Thanks
have a nice holiday
jafar

[arango]

What are you talking about?

The Matlab script to process ECMWF ERA data for ROMS is d_ecmwf2roms.m and it is in the Matlab repository Some of ERA surface fluxes are time-averaged cumulative sums. So you have to process them first, before they can be used in ROMS. You cannot use the ERA data directly into ROMS because it will be completely WRONG

Read in the information in d_ecmwf2roms.m. How are you going to use a atmospheric product in ROMS without reading all the information first about such forcing? There is plenty of documentation on the ECMWF website...