[Difx-users] atmospheric corrections

[Geoff Crew]

Chris is correct; (I was avoiding entering into a complete exposition, 
but have changed the Subject line since details seem to be of interest):

Yes, one may turn off atmospheric corrections on Calc and the tools 
(calcif2 and difxcalc) have many knobs.

The situation at ALMA is that they have atmospheric corrections turned 
on and have multiple weather stations on the plateau to feed it with 
current conditions (temperature, pressure, &c--I do not know the details 
and cannot change the way it is operated in any case.  They have 
recently established that Calc11 works as well as Calc9.)

Thus the signal recorded at ALMA for VLBI has an atmospheric correction 
already in place, and i did not immediately find a knob on calcif2 (or 
difxcalc) to turn the correction off for ONE antenna (i.e. ALMA).  If 
one does nothing, one ends up applying the atmospheric correction twice 
and there is a noticeable affect, but it is also not large.

The quickest remedy worked well enough:  startdifx accepts environment 
variables DIFX_CALC_PROGRAM and DIFX_CALC_OPTIONS so you can use any 
program in place of the default and feed it any options you like.  So I 
introduced calcifMixed which accepts many of the same options and also a 
list of antennas to not atmosphere correct (and being a lazy sort, the 
default to the -a ... argument is to apply the treatment for the Aa 
station.)  All calcifMixed does is run the normal program (specified via 
a -c argument, so either calcif2 or difxcalc may be used), read the 
output, and undo the corrections for stations in the -a .. argument 
list.  It leaves model files lying around so that one can check the math.

It's only been tested for calcif2 (ALMA has not yet recorded data for 
VLBI with Calc11) and the results were encouraging enough to suggest the 
double-atmosphere correction problem was resolved.  I can imagine the 
process could be improved.  (In the infinite resource approximation much 
of the data at ALMA is likely captured somewhere and one could then in 
principal recalculate and correct everything--but that is a massive 
amount of work for no clear gain, so I opted to move on to other issues.)

ALMA will use Calc11 in Cycle 8, so calcifMixed will get vetted for use 
with difxcalc sometime next year.

On 8/13/21 2:24 AM, Reynolds, Cormac (S&A, Kensington WA) wrote:
> On a technical note, how exactly is that (un)correction of the
> atmosphere done for ALMA? calc11 has knobs to switch the dry and wet
> atmosphere components off, but I don't see how to do it for just one
> antenna (i.e. you have them on for all antennas or off for all). How do
> you get ALMA to be treated differently to the rest of the array?
>
> I ask because we had a case recently where the atmospheric corrections
> were accidentally left in for the ATCA tied array. I came up with a
> kludged method to switch them off in DiFX for just ATCA, but if there
> is a more standard solution that would be much preferable.
>
> thanks,
> Cormac.
>
> On Thu, 2021-08-12 at 11:27 -0400, Geoff Crew via Difx-users wrote:
>> Yes, one effectively has an average signal with some average of
>> corrections.
>>
>> (Fortunately no one has requested us to go to second order to sort
>> that
>> out, as it's intractable.)
>>
>> On 8/12/21 8:43 AM, Rodrigo Amestica via Difx-users wrote:
>>> if in ALMA the DRY and WET corrections are applied per antenna, how
>>> could
>>> them be uncorrected from the phased VLBI station signal? Perhaps an
>>> average
>>> value can be calculated for the ALMA station as a whole?
>>>
>>> ps: I hope that my question helps for completness, without steering
>>> away
>>> from the original posting.
>>>
>>> On Thu, 12 Aug 2021 08:08:58 -0400 (34 minutes, 3 seconds ago),
>>> Geoff Crew via Difx-users <difx-users at listmgr.nrao.edu> wrote:
>>>> For completeness:  an array using Calc (e.g. ALMA) for its delay
>>>> model
>>>> ends up having made these corrections; these must then be
>>>> uncorrected
>>>> in the VLBI correlation.
>>>> On 8/11/21 11:02 PM, Adam Deller via Difx-users wrote:
>>>>>    * DRY and WET refer to specific sub-components of the delay
>>>>> model
>>>>> (the dry and wet troposphere, respectively.) You don't need to
>>>>> worry
>>>>> about those - they are just there in case for some reason a
>>>>> user
>>>>> wants to undo the default calculations of these contributions
>>>>> to the
>>>>> delay model in post-processing and re-apply their own. (Mostly
>>>>> used
>>>>> by geodesists.)
>>>> -- 
>>>> Geoff Crew
>>>> gbc at haystack.mit.edu
>>>> _______________________________________________
>>>> Difx-users mailing list
>>>> Difx-users at listmgr.nrao.edu
>>>> https://listmgr.nrao.edu/mailman/listinfo/difx-users
>>> _______________________________________________
>>> Difx-users mailing list
>>> Difx-users at listmgr.nrao.edu
>>> https://listmgr.nrao.edu/mailman/listinfo/difx-users
>>>
>>>
>>>
>
> !DSPAM:61161041216184077114980!

-- 
Geoff Crew
gbc at haystack.mit.edu