creating dqdsst

General scientific issues regarding ROMS

Moderators: arango, robertson

Post Reply
Message
Author
pooran
Posts: 70
Joined: Fri Jan 31, 2014 5:54 pm
Location: Khoramshahr Marine Science and Technology Universi

creating dqdsst

#1 Post by pooran » Mon Dec 01, 2014 12:22 pm

dear
i want to create dqdsst for first time by my own data. Base on different posts get_dqdsst.m is a good idea for this aim which needs sea surface temperature (Celsius), sea surface atmospheric temperature (Celsius), atmospheric density (kilogram meter-3), wind speed (meter s-1)and sea level specific humidity. I become so so grateful if you answer 2 questions of mine (if you have time):
1- is it correct to use Tair_era.nc (one output of d_ecmwf2roms.m script) as a sea surface atmospheric temperature?
2-Is follow formula good for creating sea level specific humidity:

Code: Select all

ew = 6.1121*(1.0007+3.46e-6*Pa).*exp((17.502*Ta)./(240.97+Ta));  % in mb
q  = 0.62197*(ew./(Pa-0.378*ew));                   % mb -> kg/kg
which Ta is : 2 metre dewpoint temperature (c) and Pa is surface pressure (mb).
Many thanks for your attention and time, and sorry for such questions :oops:
cheers
pooran

pooran
Posts: 70
Joined: Fri Jan 31, 2014 5:54 pm
Location: Khoramshahr Marine Science and Technology Universi

Re: creating dqdsst

#2 Post by pooran » Tue Dec 02, 2014 9:59 am

would you please guide me that It is correct to use 2 meter dewpoint temperature (c) and surface pressure (mb) or 2 meter temperature (Tair_era.nc (c)) and mean seal level pressure (Pair_era.nc (mb)) as Ta and Pa for computing saturate specific humidity (in above formula)? Because Their calculated specific humidity were different. bulk_flux.F used Tair_era.nc and Pair_era.nc but ECMWF's supporter suggest 2 meter dewpoint temperature (c) and surface pressure (mb).
http://old.ecmwf.int/research/ifsdocs/CY36r1/index.html ( at equations 7.4 and 7.5 in part IV, physical processes)
Many thanks for your help

pooran
Posts: 70
Joined: Fri Jan 31, 2014 5:54 pm
Location: Khoramshahr Marine Science and Technology Universi

Re: creating dqdsst

#3 Post by pooran » Tue Dec 09, 2014 11:04 am

1 – does dqdsst's relation with sst and wins speed is directly and Reverse, respectively.
2- Pair (mb),Tair (c), qsea (kg/kg), air density (kg/m3), sst (c)( which 2 first one created by d_ecmwf2roms.m and third one by first post's equation and fix value (1.29) as a air density) have benn used for computing dQdSST. it is some time less that -80 and Its mean is about -20 and. With regard that there is not any previous data in my studding region (Caspian sea) for comparing, i become appreciate if you have any recommend for me , help me.

d.kobashi
Posts: 60
Joined: Tue Sep 28, 2010 11:59 pm
Location: Texas A&M University

Re: creating dqdsst

#4 Post by d.kobashi » Wed Dec 10, 2014 6:04 am

Why don't you use COADS or ICOADS dataset? I personally don't think you need atmospheric model outputs (e.g. ERA interim) to create dqdsst. Climatology should be fine. Some use a constant dqdsst over their entire model domain.

COADS
http://iridl.ldeo.columbia.edu/SOURCES/.COADS/

ICOADS
http://icoads.noaa.gov/

Read the following paper to understand the concept of dqdsst.
1. Barnier, B., Siefridt, L. & Marchesiello, P. Thermal forcing for a global ocean circulation model using a three-year climatology of ECMWF analyses. J. Mar. Syst. 6, 363–380 (1995).

Hope this helps.

DJ

pooran
Posts: 70
Joined: Fri Jan 31, 2014 5:54 pm
Location: Khoramshahr Marine Science and Technology Universi

Re: creating dqdsst

#5 Post by pooran » Sun Dec 14, 2014 11:35 am

Dear kobashi
many thanks for your reply. I have checked those databases before but they don not have enough data in Caspian sea region :oops: because of this i was computing data from different formula and functions.
So, IS there any recommended constant value for dqdsst (especially in in Caspian sea).
I become so appreciate to get it
cheers
pooran

pooran
Posts: 70
Joined: Fri Jan 31, 2014 5:54 pm
Location: Khoramshahr Marine Science and Technology Universi

Re: creating dqdsst

#6 Post by pooran » Tue Dec 16, 2014 10:03 am

Dear Kobashi
i have studied mentioned article and will chose -32 as dqdsst (my studding area is located between 36N and 47N which has seasonal variation of dQdSSt). i become appreciate if you help me to understand more about it by answering follow question:
1-i found that using SST and dQdSST will help to get better model's SST. Is this process some things like data assimilation to correct model's SST base on climatological SST?
2- base on assimilating concept, it seems that it is better to use longer time interval for SST (such as daily or one for each 5 days). So, If ROMS do data assimilation, what time interval do you recommend for SST (Base on you experiment on ROMS )?
Many many thanks for any reply and time

User avatar
kate
Posts: 3772
Joined: Wed Jul 02, 2003 5:29 pm
Location: IMS/UAF, USA

Re: creating dqdsst

#7 Post by kate » Tue Dec 16, 2014 6:14 pm

Having dQdSST and SST does act as a nudging to that SST, but it was designed to give better heat fluxes when your SST differs from that used when computing Q.

As with other forcings, I'd go with more frequent if you have them.

pooran
Posts: 70
Joined: Fri Jan 31, 2014 5:54 pm
Location: Khoramshahr Marine Science and Technology Universi

Re: creating dqdsst

#8 Post by pooran » Wed Dec 17, 2014 10:17 am

So if it acts as a nudging to that SST:
1-does it required to define TCLIMATOLOGY and TCLM_NUDGING or QCORECTION is enough?
2- will this nudging do in each step or base Nudging/relaxation time scale (TNUDG)? Or TNUDG will applied whenever TCLIMATOLOGY and TCLM_NUDGING are defined?

User avatar
kate
Posts: 3772
Joined: Wed Jul 02, 2003 5:29 pm
Location: IMS/UAF, USA

Re: creating dqdsst

#9 Post by kate » Wed Dec 17, 2014 5:27 pm

It is a nudging with the nudging timescale set by dQdSST. It is completely separate from TCLIMATOLOGY and doesn't use that machinery at all. It only acts as a nudging because that term in the Taylor series expansion of Q about SST acts as a feedback term, making your SST more closely match that used in computing Q. TNUDG is not invoked.

The atmospheric forcing is applied at each timestep, using linear time interpolation between your forcing records.

If you think about the longwave out term in the surface heat balance, it depends on temperature in Kelvin to the 4th power. The hotter the surface, the more heat you radiate. If Q was computed using a warmer temperature than you have, it will suck heat out of your ocean too fast. The dQdSST term is there to compensate for that extra heat loss - it will be a nudging to that warmer SST.

Post Reply