[mmaimcal] request for source types

[Bryan Butler]

hello MAC members,

there seems to be some confusion about the request i relayed in last
wednesday's phone meeting.  this confusion is entirely my fault, as
i was supposed to send out a more detailed email explaining what i
was talking about and i didn't do that.  i apologize, but i hope you
understand (my wife is now 4 days overdue for our first child!).

so, let me explain my request in more detail here.

there has been alot of discussion lately (and at some level for many,
many years) about deciding on both a philosophy for laying out the
physical locations of the antennas (the "configuration"), and the
selection of actual, real pad locations on the high site in chile.
we are now in a situation where we will probably soon need to start
surveying for possible sites, and before we do that it would be nice
to have an idea of what we should be looking for.

examples of configuration styles are leonia kogan's donuts or nested
circles, and adrian webster & john conway's logarithmic spirals.  there
are others, and the list of MMA memos related to configuration studies
is long (and i won't list it here).

now, there is a philosophical decision to be made which is tied in to
the selection of style of configuration.  this is the number of distinct
configurations, or, in fact, whether we have no distinct configurations 
at all (and are continuously moving antennas/reconfiguring).  min yun 
and leonia kogan have written a nice MMA memo on this subject which you 
can read if you like.  this decision has not been made, and is partly an
operational, partly a scientific issue, and partly a cost issue (given
pad reusability).  if you have thoughts on this particular matter, 
please make them known.

aside from this, there is the issue of how different styles of 
configurations perform with respect to imaging particular types of 
objects.  this is the issue that we are currently trying to get a 
handle on.  the europeans are also working on this topic (notably john 
conway and francois villafond).  this is a thorny topic, as it involves 
a mix of many poorly understood (or at least difficult to quantify or 
effect) separate pieces.  

i think of this myself as akin to an SNR problem.  we need to specify 
the "signal", i.e., define the sky brightness distribution (the "model")
and the antenna pad layout (the "configuration"), and we also need to 
define the "noise" (including thermal, atmospheric, pointing, and 
anything else we can think of to make it as real as possible while 
still remaining implementable).  there is then the additional issue of 
the imaging problem, i.e., once you have your simulated visibilities, 
how do you turn them back into the estimated sky brightness distribution
which you may then compare to your model, in order to make sensible 
comparisons between configurations.  this is a very complicated subject 
and we are also trying to address it with some level of effort.  
especially important with respect to these types of simulations to
compare configurations is the inclusion of the total power data in the
imaging process.  in fact, you might argue that in order to just 
compare configurations, one might not want to include any sources of 
noise.  that may be true, but in any case, we are still left with the 
task of defining the "signal", or determining the model(s) to be used 
in this comparison.

in fact, what would be ideal is for us to know _a priori_ exactly the 
types of sky brightness distributions that will be observed, at which 
resolutions, and then to obtain or create models of these sources.  
this is impossible, of course, but we hope to get some type of handle on
it, if even in the most general terms.  a trivial example is the 
statement that large objects will mostly be observed in the compact 
configurations.  extending that, one might say that the breakdown of 
time in the most compact configuration might look something like: 35% 
extragalactic, 35% galactic, 20% real stars, 5% solar system, 5% 
"other" [this is close to what the VLA breakdown is].  i think this 
last bit is going too far, but how far really _is_ too far, or how well 
can we define the types of sources that really will be observed at 
different resolutions?  a simpler question than really trying to pin 
down the exact observing strategies/sources for different resolutions
is the one of simply obtaining examples of the types of sources that we 
_know_ will be observed with ALMA.  this is the basis of my request to 
the MAC (though, of course, you are encouraged to comment on any aspect 
of this discussion).

so, what we need as a starting point in detailed configuration 
comparisons is a set of images of sky brightness distributions that 
might reasonably be expected to be observed for some significant 
fraction of time with the ALMA (one off experiments shouldn't drive the 
design).  we have been informally referring to this in our NRAO group 
as the "standard image library" (i will refer to it as the SIL 
hereafter).  this is really needed when we start comparing our 
simulations to those of the europeans, as if both sides aren't using 
the same images, we will be comparing apples to oranges.  it is ideas 
about what should go into the SIL that we would like to get from the 
MAC.  in fact, if you can point us to real images, this is even better, 
but just declaring what types of images we need is the first step.

while images from current millimeter arrays are good indicators of the 
types of things that might be looked at with ALMA, they are not, in 
fact, great candidates for the SIL.  the reason is that they have 
already been filtered both in dynamic range, and in spatial scale, by 
the observing array.  again, these are good in terms of defining the 
image types, but if they are to be used as actual populators of the 
SIL, then some caressing of the image will have to be done.  my idea 
(and this isn't mine alone, i think i first heard mark holdaway 
espousing this) is that it would be much better to obtain HST or VLA 
images which correspond to these types of sources, since they have much 
more appropriate dynamic range and range of spatial scales.  we can
probably do this but again need to know what types of images to garnish 
from these instruments.

current ideas about what should go into the SIL are: Orion ridge 
(courtesy of neal evans), a carbon star (courtesy of john beiging), M51 
(courtesy of many), and CasA (or another good SNR - possibly courtesy 
of larry rudnick).  other possibilities are the fantastic bipolar 
outflow HST images (eta carina?), one of the nice HST nebulae (eagle or 
horsehead?), VirgoA from VLA, galactic center from VLA, a planet 
(either VLA or HST), and some estimation of the HDF (although this is 
almost an entire subject by itself).

so, what i'd like from MAC members is to get ideas on both the types of
sources that you think will be observed with ALMA (and at what 
resolutions), and pointers to specific images if they exist.  please
email me (bbutler at nrao.edu) with your ideas on this.

once we have some idea of what should go into the SIL, we can start 
creating it, and iterate from there.  there will eventually be a web 
repository where the SIL and a description will be available, but that 
is for the future.

sorry for the long-windedness, but i thought that some detail was needed
to clear up any confusion that might exist.


	-bryan butler
	 assistant scientist
	 national radio astronomy observatory