IRAC Flat Selection

IRAC Flat Field Selection and Scheduling (8/13/02)

The IRAC flat-field will be derived by looking at a region of the sky with high zodiacal background. Because the zody is relatively faint at the shortest IRAC wavelengths, we really do need to be near the peak. Furthermore, SIRTF will have to observe the flat-fields at fixed times set by the scheduling of the instrument campaigns. Unfortunately, because of SIRTF's pointing constraints, we cannot choose a single fixed point near the ecliptic, since any given point in the ecliptic is only visible for some 40 days, and in case the brightness of the zody is local and tied to the sun, and not the celestial sphere.

There are two approaches that could be used:

At any given time there is a "best" allowed place to point. Once could figure out where the sun is when a flat is to be taken, and point at +80 degrees (SIRTF cannot point closer to the sun than 80 degrees). However, this means that you are always pointed somewhere different when each flat is taken.
One could establish a fixed network of flat-field sites along the ecliptic. Then the best one available (i.e. meeting pointing constraints) is chosen.

I am opting for the latter. The primary advantage is that it ensures greater uniformity of calibration. Flat-fielding IRAC is already problematic due to variation in the sky. By selecting a finite number of flat-field locations we can pre-select these to ensure that they minimize the number of bright stars and offensive extended structure (such as significant cirrus).

To see what the other instruments have done, you can go to:

Jeonghee's page of MIPS fields. (internal SSC access only)
Vassilis's page of IRS fields.

Selection

I have selected 24 points spaced at 15 degree intervals along the ecliptic. This ensures that there is one spot that is between 0 and 15 degrees away from the best location allowed. As many six of spots are visible at any one given time. The spacing between these points sets the maximum distance one can be from the ideal point. However, we don't want to make them too close or we will have defeated the whole point of having a finite number in the first place. Conversely, they must be no more than 40 degrees apart to ensure visibility. A 15 degree shift from the "best" location is roughly a drop of 20% in intensity, and seems like a good compromise.

Furthermore, after talking to Jeonghee I have shifted the points 5 degrees off the ecliptic itself to try and avoid zody structure. This is a small loss in intensity of a few percent.

I then examined the ISSA plates using the visualization tool in SPOT and chose either the plus or minus 5 degree position based on the number/flux of 12um sources and the degree of cirrus structure seen at 100um. In some cases I tweaked the positions further to better optimize them for the above. Note that I have not looked at the 2MASS data for these regions yet to further optimize them. Volunteers?

Name	Eclip. Long.	Eclip. Lat.
zody_field_1	0.00	-5.00
zody_field_1	0.00	-5.00
zody_field_2	15.00	-5.00
zody_field_3	30.00	-5.00
zody_field_4	45.00	+5.00
zody_field_5	60.00	-5.00
zody_field_6	75.00	+5.00
zody_field_7	90.00	-5.00
zody_field_8	105.00	+5.00
zody_field_9	120.00	-5.00
zody_field_10	135.00	-5.00
zody_field_11	150.00	-4.00
zody_field_12	165.00	-5.00
zody_field_13	180.00	+5.00
zody_field_14	195.00	+5.50
zody_field_15	210.00	+7.00
zody_field_16	225.00	-5.30
zody_field_17	240.00	+5.00
zody_field_18	255.00	-5.20
zody_field_19	270.00	+5.00
zody_field_20	285.00	+5.00
zody_field_21	300.00	-7.00
zody_field_22	315.00	-5.80
zody_field_23	330.00	+4.30
zody_field_24	345.00	+5.00

Table with FITS images for all of the above fields.

Humu (slow, but accessible)
SSC-ET (fast, but local access)

Scheduling

The in-flight procedure would then be as follows:

The OPS groups expects for each campaign a single "calibration" program to be delivered. This will consist of all the calibration AORs (darks flats, standard stars, etc.) along with instructions on their scheduling. I propose that we simply include all of the flat-field AORs and give them the instructions on how to pick the field, rather than try to pick the fields ourselves. The alternative is for us to pick the field we think is the right one to use, assuming that they can tell us fixed times for when the campaign will be scheduled, and deliver it as part of the calibration program.
To pick the field, someone will look up the SIRTF-centric position of the sun via the ephemeris for the time that the flats will be scheduled.
The correct field is the one whose longitude is closest to and greater than (solar longitude + 80), modulo 360. The +80 position is chosen over -80 for efficiency reasons.

The AORs for the above are attached here. Note that these are the 3x3 (9 dither) pattern planned for the mask-and-coadd reduction and which was prototyped previously. Also included at the end is the aor for the NEP dark which goes along with this observation. I expect to take the flat and it's own dark back-to-back. The regular sky dark does not take as many images as this special dark, and hence would be a noticeable contributor to the noise in the derived flat. I have yet to figure out with the OPS group how we will specify the constraints.