here are a set of comments from stephane on mark's reply to my
review of stephane's memo 372.  i haven't formalized the header
on it yet - oh well...

	-bryan


> "Mark Holdaway"  wrote:
>> "Bryan Butler"  wrote:
>>    'decorrelation estimates' are included in item 6.  If we have
>>    data that is significantly decorrelated on the timescale of 
>>    fast switching or WVR calibration, then we are in real trouble.
>>    That is the point of having those calibrations - to avoid the
>>    decorrelation, i.e., make it possible to track accurately the 
>>    phase variations on short timescales and correct them in the data.
>
> There will be decorrelation in fast switching.  In order to select
> the details of how to fast switch, I optimize the sensitivity of
> the observations, including time lost to the calibration process and
> sensitivity lost to decorrelation (caused by residual atmospheric errors
> and to some extent thermal errors in the calibrator's phases).
> You can see a tradeoff here:
> you can calibrate like crazy to get the residual phase errors very small,
> but will then have low sensitivity because you only spend 10% of the
> time on target source -- conversely, you could spend most time on target,
> and get a poor calibration with lots of decorrelation.
>
> I have stated this approach for several years, and have asserted the
> need for a decorrelation correction in fast switching.  Perhaps that
> is the wrong approach, but people have not taken issue with it in the
> past.

S.Guilloteau comment:
---------------------
Mark is right on that. We cannot totally suppress de-correlation, 
whatever scheme is put in place (WVR or Fast Switching). Compensating 
for decorrelation is a MUST for ALMA. In the Fast Switching scheme, by 
spending as much time on the calibrator than on the source, you expect 
to get a statistically representative sample of the decorrelation on 
the calibrator, but that is paying a high price. With the WVR, the 
inherent noise of the WVR may be higher than the residual errors in the
prediction scheme which derives pathlength from WVR measurement. In 
this case, decorrelation is dominated by the WVR noise (which is a bad 
thing, since you would rather like to be limited by the atmosphere 
itself) and it should be predictable (which is a good thing because you 
can correct for it). There is some work to be done in this area...


>> p.7: In discussion on quasars as primary calibrators: 'At the
>>    Chajnantor site, given the variety of hour angle and declination,
>>    one of them will be available at anytime for bandpass calibration.'
>>    We may find that we want the bandpass calibrator near the source
>>    of interest, not just anywhere in the sky.  This is certainly the
>>    case at the VLA.  It will depend on the stability of the bandpass
>>    with time, temperature, antenna motion, etc...
>
> The VLA did observe for its first 15 years not caring where on the sky
> the bandpass calibrator was located, and now only demanding 
> observations need to worry about where the bandpass calibrator is.
>
> At the VLA, where actual astronomers create the observation schedules,
> there is often no quantitative understanding of what is required in
> terms of calibration, and astronomers, being conservative creatures,
> will opt for over-calibration, determining the instrumental parameters
> to much higher accuracy than is required.  

S.Guilloteau comment:
---------------------
I have seen the same tendency at Plateau de Bure, where people asked to 
spend 1 hour to calibrate the bandpass to 0.1 % accuracy for a 5 sigma 
detection of high-z sources...

> Given the more demanding calibration environment of ALMA (ie, more 
> calibrations need to be made), we likely will need to optimize each 
> calibration, and not over-calibrate by too much, in order to have 
> time on source.

S.Guilloteau comment:
---------------------
Yes, we cannot afford to overcalibrate, but we must be careful not to 
undercalibrate either. Also, the need to maintain a relatively even 
quality in the database of observations should be considered. 
Calibration needs may not be decided only by the observing project.