Output variables

[Sougata18]

I did a 4000 years of spin up run of the 4 deg setup. I have few questions regarding VEROS output veriables.

1. Some output variables (such as 'u' , 'v' , 'temp' , 'salt' etc.) are available in both 'averages' and 'snapshot' files. What is the difference ?
2. The 'overturning' file contains variables all having attribute 'meridional transport' , what are the differences ?
3. The time coordinate is shown in negative integers. What does that mean ?

[dionhaefner]

1. Snapshots are instantaneous values and averages are time averages, based on the chosen sampling_frequency of the averages diagnostic.

• vsf_iso: Vertical streamfunction (zonally averaged) in isopycnal coordinates
• vsf_depth: Vertical streamfunction (zonally averages) in depth coordinates
• bolus_iso / bolus_depth: Same, but only for eddy-driven velocities

3. That should not be the case (time is in days). What are you using to analyze the output?

[Sougata18]

3. I am using xarray for the analysis purpose. The time array goes like this -

array([-8960024073709551616, -8928920073709551616, -8897816073709551616, ..., -3062696515966861312, -3031592515966861312, -3000488515966861312], dtype='timedelta64[ns]')

5. I want to get the meridional transport zonally averaged for the Atlantic basin. Based on what you told about the sampling frequency of the averages diagonstics, I am little confused about how to calculate the transport. Cause the usual way of integrating the u or v component is not giving me the correct values for the transport.
6. Also the time origin is given as 01-JAN-1900 00:00:00 , is it based on the initial conditions ?

[dionhaefner]

3. I am using xarray for the analysis purpose. The time array goes like this -

array([-8960024073709551616, -8928920073709551616, -8897816073709551616, ..., -3062696515966861312, -3031592515966861312, -3000488515966861312], dtype='timedelta64[ns]')

That looks like a bug, perhaps something overflowed. I'll see if I can reproduce it.

4. I want to get the meridional transport zonally averaged for the Atlantic basin. Based on what you told about the sampling frequency of the averages diagonstics, I am little confused about how to calculate the transport. Cause the usual way of integrating the u or v component is not giving me the correct values for the transport.

Can you show what you tried? How are the values wrong?

5. Also the time origin is given as 01-JAN-1900 00:00:00 , is it based on the initial conditions ?

No, this is just an arbitrary "year 0" for the model. All times are encoded as "seconds since time origin".

[dionhaefner]

Indeed, the timedelta64[ns] data type xarray uses can only represent dates up to the year 2262. You can either install cftime or use xr.open_dataset(..., decode_times=False) to solve the problem.

[Sougata18]

4. I want to get the meridional transport zonally averaged for the Atlantic basin. Based on what you told about the sampling frequency of the averages diagonstics, I am little confused about how to calculate the transport. Cause the usual way of integrating the u or v component is not giving me the correct values for the transport.

Can you show what you tried? How are the values wrong?

So to calculate the streamfunction for meridional transport I need to integrate 'v' zonally and wrt depth , which should give me values in the range of $10^{06}$ but after the integration (for a single meridional) it gives values like -
array([[-94.94058593], [-94.95589481], [-94.97078953], [-94.9851467 ], [-94.99834973], [-95.01052151], [-95.02236012], [-95.03303989], [-95.04495352], [-95.05687063], [-95.06745418], [-95.07858525], [-95.08779227], [-95.09413547], [-95.09822458], [-95.10236002], [-95.10936345], [-95.11939067], [-95.13169373], [-95.14572238], ... [-96.36318112], [-96.36753048], [-96.37126962], [-96.37563124], [-96.37943745], [-96.38374498], [-96.38759197], [-96.39191286], [-96.39574721], [-96.39999575], [-96.40396971], [-96.40814764], [-96.41229912], [-96.41638589], [-96.42036418], [-96.4245432 ], [-96.4286091 ], [-96.43267466], [-96.43669264], [-96.44058458]])

Here for integration I followed xr.integrate ,
v=da.v[:,:,:,69:89] dv = v.fillna(0) int=dv.integrate(['xt','zt'])

I also noticed that the psi values are in correct order, but they are not a function of depth(psi (Time, yu, xu)). Can psi be calculated as a function of depth also during model integration ?

[dionhaefner]

xt is in degrees, not meters. Correcting for this should give you the correct order of magnitude.

[Sougata18]

But the unit of xt shouldn't have any effect on the values of u and v component. I tried to convert the degrees to meters but that doesn't change the order of the values.

and what about this...

I also noticed that the psi values are in correct order, but they are not a function of depth(psi (Time, yu, xu)). Can psi be calculated as a function of depth also during model integration ?

[dionhaefner]

But the unit of xt shouldn't have any effect on the values of u and v component. I tried to convert the degrees to meters but that doesn't change the order of the values.

It does change the value of the integral: u dx depends on the units of x.

I also noticed that the psi values are in correct order, but they are not a function of depth(psi (Time, yu, xu)). Can psi be calculated as a function of depth also during model integration ?

No, that is not implemented.

[Sougata18]

It does change the value of the integral: u dx depends on the units of x.

So I changed the degrees to meters and then performed integration,
xt_m=(da.xt)*(111.32 * 1000)
yu_m=(da.yu)*(111.32 * 1000)
da2=da.swap_dims({'xt': 'xt_m','yu':'yu_m'})
v=da2.v[:,:,24:25,69:89].fillna(0)
int=v.integrate(['xt_m','zt'])

But the values retunred are -
array([[-23.73514648], [-23.7389737 ], [-23.74269738], [-23.74628668], [-23.74958743], [-23.75263038], [-23.75559003], [-23.75825997], [-23.76123838], [-23.76421766], [-23.76686355], [-23.76964631], [-23.77194807], [-23.77353387], [-23.77455615], [-23.77559 ], [-23.77734086], [-23.77984767], [-23.78292343], [-23.7864306 ], ... [-24.09079528], [-24.09188262], [-24.0928174 ], [-24.09390781], [-24.09485936], [-24.09593624], [-24.09689799], [-24.09797821], [-24.0989368 ], [-24.09999894], [-24.10099243], [-24.10203691], [-24.10307478], [-24.10409647], [-24.10509105], [-24.1061358 ], [-24.10715227], [-24.10816867], [-24.10917316], [-24.11014614]])

[dionhaefner]

I'm not an expert in xarray, so I suggest you look for help in a dedicated community if you want to do it this way. All I know is that your original estimate is too small by a factor of about 1.1e5, which gives you the correct order of magnitude.

[Sougata18]

Okay! anyway thanks for the help!