Calibration of Linear Ramp Filter Data

-John Biretta and Keith Noll (Last update: 18 May 2007)

Flat Field Calibration:

Currently, images observed with the linear ramp filters are not flat field calibrated in the calibration pipeline. Observers should obtain a flat field reference image from the HST archive, and multiply this image into their data. We recommend using a narrow band filter flat at a wavelength near to the LRF observation wavelength (see list of narrow band flats below). There are no narrow band filters near 8000 Å, here we recommend using flats in filter F791W.

An obvious question is "why not use a flat observed in the LRF filter itself, instead of a flat observed in some other narrow band filter?" If flats were used which were observed in the LRF filters, it would be necessary to have an accurate spectrum for the light source, and this is difficult. In making flat fields, both EARTHFLATS (flat fields taken while HST is pointed toward the Earth) and VISFLATS (flat fields taken with the internal VISFLAT lamp and the shutter closed) are used. For EARTHFLATS, the color of the Earth varies widely depending on whether one observes land, sea, or clouds. The color of the internal VISFLAT lamp is known to vary as a function of position in the field of view, lamp "on" time, and total number of times the lamp is used. Since the filters are far from the focal plane (the OTA beam has an approximately 33 arcsecond diameter at the filter), any dust spots and other small imperfections in the filter have essentially no effect on the data. Any large-scale variations in the filter are contained in the filter transmission curves, and are corrected during photometric calibration. Moreover, some of the ramps have pinholes which would affect flat fields specifically made for the LRFs, but they do not have any affect on science data.

At some point we will incorporate LRF flat fielding in the calibration pipeline. Software modifications are needed to allow selection of the flat as a function of wavelength.

Recommended reference files for LRF flat field calibration can be found on the WFPC2 Reference Files page.

Locating the Target in the Image:

Unlike other filters, the position of the target in the WFPC2 field of view depends on the observation wavelength. Specifically, it depends on the WAVELENGTH specified on the Phase II proposal.

If you need help remembering the WAVELENGTH on your Phase II proposal, it is recorded in the LRFWAVE and PHOTMODE keywords of your calibrated (c0h) data, which can be accessed using the 'hedit' task within IRAF/STSDAS (e.g. hedit *c0h LRFWAVE . or hedit *c0h PHOTMODE . ). You can also find the wavelength you specified in your Phase II proposal by displaying a copy of it through the PRESTO web page. Simply fill in the proposal number or the PI's name on the Program Information page, click on "Go", and then click on FORMATTED LISTING. You will be able to scroll through the proposal and find the wavelength used.

(Caution: sometimes observers work the other way, and measure the target position on the CCD, input the CCD and (X,Y) into the LRF calculator tool, and then request the wavelength at that position. While the tool gives correct results, interpreting the results can be confusing. This is because there are often several combinations of CCD and (X,Y) that will produce a certain wavelength. The HST scheduling software is designed to use only the optimal combination for a given wavelength, but by inputing the CCD and (X,Y) it is possible to discover these non-optimal settings. Hence, if you give the LRF tool the CCD and (X,Y) of some object and it tells you the wavelength, inputing this wavelength back into the LRF tool may give a completely different CCD and (X,Y) combination (i.e. the optimal setting).)

Photometric Calibration:

There are two methods available for obtaining photometric calibration -- SYNPHOT and the WFPC2 Exposure Time Calculator. The easiest method is to use the WFPC2 Exposure Time Calculator; this will give accurate results for lines which are narrow when compared to the LRF filter bandpass. More complex situations will require use of SYNPHOT, and this is described further below.

Photometric Calibration via the Exposure Time Calculator:

The easiest method to obtain photometric calibration of the LRFs is via the WFPC2 Exposure Time Calculator. This program will allow you to specify a target flux and exposure time, and then it will calculate the number of counts expected for the target. You can then use these results, and a little arithmetic, to get the target flux.

Depending on whether you want to derive a total integtrated flux for the target, or a flux per square arcsecond, you should select the Point Source Calculator, or the Extended Source Calculator, respectively.

For example, let's say you have observed the 5007 [O III] line in a redshift Z=0.555 galaxy. You used the LRF filter at WAVELENGTH = 7786 Å, and a 1200 second exposure. Let's initially guess the source total flux to be 1.0e-14 erg cm^-2 s^-1 as observed at Earth. You would choose the POINT SOURCE calculator, and fill it out as below.

NOTE: this is a sample form, do not fill in numbers here.

Object: Emission Line:

Line Flux: (erg cm^-2 s^-1)

Line: 0 III 5007 Z : (units)

Instrument Configuration:

If using LRF filter give desired Central Wavelength: Å

Exposure: Enter either S/N or Exposure Time.

Signal to Noise: Exposure Time: Sec.

There are more inputs than we have shown in the above example, but the other settings are unimportant for this purpose. They can just be left set at the initial default values.

After completing the form, click on CALCULATE. After about 15 seconds it will return a long set of results. As you scroll down through these results you will find lines like these:

The entry "object_counts = 12648 electrons" gives the number of electrons which would be observed for a target flux of 1.0e-14 erg cm^-2 s^-1 incident at Earth. To convert this to observed data numbers, you will need to divide by the nominal gain setting. Most observers will need to divide by 7.0 (default setting). In cases where ATD-GAIN=15 was specified on the proposal, divide by 14.0 instead. Hence the above 12648 electrons corresponds to 1807 DN measured on the image.

This number of counts can then be compared with the actual counts on the observed image, after subtraction of any background or continuum emission. For example, if two 1200 sec. exposures were taken and averaged to remove cosmicrays, and the averaged, flattened image showed 800 DN total in the target, then the source flux (at Earth) is:

1.0e-14 erg cm-2 s-1  (800 / 1807) = 4.4e-15 erg cm-2 s-1

The accuracy of this calibration is expected to be 5 percent or better. Most of the uncertainty results from use of pre-launch transmission curves for the Linear Ramp Filters. Our experience with other filters is that the pre-launch and on-orbit calibrations agree to 5 percent or better.

Some points to keep in mind when using the Exposure Time Calculator (ETC) for photometric calibration:

• Spectral lines are assumed to be monochromatic. If the intrinsic line width becomes comparable to the LRF bandpass width (about 1.3 percent of the central wavelength), you should probably use SYNPHOT, instead.
• Stellar sources are assumed to have a smooth spectrum. If your target has deep absorption features near the LRF wavelength, you should probably use SYNPHOT, instead.

Go to Point Source ETC.

Go to Extended Source ETC.

(Caution: be sure to use the wavelength setting from the Phase II proposal when performing photometric calibration with either the ETC or SYNPHOT. Sometimes observers will offset the wavelength in the Phase II proposal to force the observation onto a particular CCD (e.g. PC1). For example, if the spectral line is a 6636A, they may request 6630A to force the scheduling software to use the PC1 CCD. In these cases 6630A should be used during photometric calibration.)

Photometric Calibration via SYNPHOT:

Work in Progress:

• Pipeline flat field calibration. A modification to SOGS software is needed to select flats by central wavelength.