By: Edward Olszewski, David Sand, Jill Bechtold, Dave Trilling
Last update: Jan. 12, 2007

Summary: This page describes the background material and philosophy of the calibration. Go to the Calibration OB page to access OB's for any particular night.


We will need faint standard stars and astrometric fields for LBC: V<20 or so.

While indeed the commissioning will give us these astrometry numbers, they can be obtained nightly, for free, from standard fields. Also, there is work being carried out in the Dewar
between commissioning and the start of SDT on Jan. 12. We need to check the astrometry whenever the dewar and instrument are worked on.

I. Bias frames in the afternoon.

The bias level varies with time so will be subtracted using the overscan region. Every day, before sunset, we'll take a set of bias frames, to check that the camera is functioning, and to check that the read noise is what we expect it to be.

II. Flats

We'll take twilight flats, and construct dark sky flats for all filters. The best scheme for flat-fielding is TBD.

II. Astrometry:

The USNO-B catalog is sufficient for many purposes. The main problem will be that USNO B is only good brighter than 19th mag or so. USNO B catalog is at One clicks on "Data and Software", then on "USNOFS Image and Catalogue Archive."

Use the NOMAD catalog, which is a compendium of catalogs, after consultation with Dave Monet about problems deriving decent astrometry from USNO-B in heavily crowded globular cluster fields like 47 Tuc. Experience with CCD array astrometry from KPNO 4m, CTIO 4m, Steward 90prime, Burrell Schmidt, Magellan IMACS, and other arrays is that one can derive coordinates to 0.2 arcsec or better with little trouble. This accuracy is good enough that one can place fibers on stars for spectroscopy, and one can coadd dithered images without residual distortions.

USNO-B covers the entire sky, so any field is an astrometry field.

Short exposures will be needed for astrometric fields to avoid saturation and trailing. If all goes normally, the distortions derived from these fields will be divided into two forms: first, higher order terms which are unchanging; second, zeropoint and linear terms that take into account where you're pointed and the differential refraction. In IRAFspeak, there are routines for deriving good astrometry from a WCS in the header and those corrections (where I mean good enough for regular astronomers, not good enough for real astrometrists, say, at the 0.2 arcsec level, which is fine for almost all applications). It is certainly the case that these terms are filter (color) dependent.

While there is real astrometry for M67 and other clusters, it buys you little since there are many 10th mag stars in M67 unless you're doing specific projects that need that sort of astrometry. If the main purpose is to map camera distortions, the USNO-B and fainter standard fields, are preferable .

Many science fields, and some of the calibration fields are also SDSS fields, and their astrometry can be used as well.

II. Photometry:

The largest compilation of faint BVRI (where RI are Cousins R I ) photometry is by Stetson, who extended Landolt fields by reducing many CCD observations of those same fields.

Stetson's list is found at

One then clicks on "Stetson" in the upper left. The important columns, besides RA/DEC, are the size of the field in arcmin, and the number of BVRI standards. Many of these fields are of clusters, so they're not the ideal set of evenly spaced, uncrowded, standards, but they're quite decent.

Again, all of these fields are also USNO-B astrometry fields.

Saha (PASP 117 827) has good fields to V=21 in NGC 2419 and Pal 4 and Pal 14.

U band photometry and SDSS GR photometry is more problematic. For SDSS, all I can suggest is that many northern fields are also in the SDSS data releases (though not all). The problem is that a random SDSS field is only good to 2% or so, and the heavily observed fields have not been published.

U is always problematic. No one's U filter is well matched to others' (so there are color terms and luminosity-class terms).

It looks to me like the best U photometry is still Landolt. The problem here is few stars per field (SA 92 has 15). A 1-sec will not saturate U but will saturate some stars on some in some other bands. In SA 92, for example, 10 of the 15 stars would be saturated in a 1 sec V exposure, so obtaining U band color terms is pretty much impossible. This will be worse with the new U filter!

For the January SDT, we've chosen Stetson/Landolt fields to observe.

The goal of the photometric standards is to determine:

zero point
color term

On photometric nights, a set of standards with different airmass will be taken at the beginning of the night, and the end of the night, and possibly the middle of the night.

At some point, we will set up the OB to move a standard field around to get the same stars on each chip.

On non-photometric nights, we will still start with a standard field, since it's a good check that everything is right.

III. Filters:

At the moment, LBC Blue has UBV filters and SDSS GR filters. It's a mixed set.

IV. Exposure times:

Using the LBC ETC, assuming:

0.8 arcsec seeing (one wants to know the time for stds for the best seeing,

  1. so one doesn't saturate)
  2. 7 day moon
  3. airmass 1.3

10 sec in B gives saturation of 15.1 in B
1 sec in B gives saturation of 12.6 (2.5 mag brighter, of course)

10 sec in V saturates at 15.5
1 sec in V saturates at 13.00

10 sec in SDSS saturates at G 15.9
1 sec in SDSS G saturates at 13.4

10 sec in SDSS R saturates 15.4
1 sec in SDSS R saturates 12.9

10 sec in U 13.7
1 sec in U saturates at 11.2

In all filters, one gets better than 2% photometry per star at 19th mag. So the Stetson fields are fine, and there will be plenty of USNO-B stars in a 10 sec exposure in any field.

I ran USNO-B for a 20 arcmin rectangle at the North Galactic Pole. There are 270 stars with 15.0<R<19.0, and another 400 stars to 20th. This is a worst case scenario (except for star forming regions and dark clouds).

V. More on Photometry:

The quantities needed from the photometric standards are:

1. zeropoint (for each CCD)

2. color term (for each CCD)

Since the CCDs each have a different QE as a function of wavelength, each will have their own color term. The color term corrects the zero point for the fact that the individual standard stars and individual program stars have different spectral energy distributions within the filter band-pass. Observations in multiple filters can be used to derive the color terms.

3. extinction

On photometric nights, we want to derive the extinction -- i.e. the zero point as a function of airmass. One wants standards with a range of a minimum of 0.4 airmasses. If you do standards only at twilight, do several fields.

-- Main.jill - 09 Jan 2007
Topic revision: r3 - 12 Mar 2007, RichardPogge
This site is powered by FoswikiCopyright © by the contributing authors. All material on this collaboration platform is the property of the contributing authors.
Ideas, requests, problems regarding Foswiki? Send feedback