Due to a failure of the STIS primary (Side-1) electronics, the redundant (Side-2) electronics have been in use since July 2001. The redundant electronics allow larger fluctuations in CCD temperature, resulting in significant variations in CCD dark rate (see ISR STIS 2001-003). Eventually, dark subtraction in the archive pipeline will automatically handle dark rate variations, but for now darks can be scaled manually as a function of temperature (see ISR STIS 2001-003). As a proxy for CCD detector temperature, the CCD housing temperature (CCDHTAV) will be added to STIS science headers. Until this keyword is available, we will maintain a CCD housing temperature log. Columns in the log file contain universal time (yr.day:hr:min:sec), Julian date (days), and CCD housing temperature (degrees C). Although STIS samples the temperature at 30 second intervals, a new log entry is created only when the measured temperature changes or a new telemetry record begins.

Science data obtained after 1 July 2001 at a particular CCD housing temperature can be reprocessed manually by subtracting a superdark that has been multiplied by the scale factor: 1 + 0.07*(T - 18) where T is the temperature of the science data, and the superdark reflects the dark rate at a reference temperature of 18 degrees C. The superdark should typically be constructed from a one week period contemporaneous with the science data, but never spanning a CCD anneal. (CCD housing temperature actually drops to about 0 degrees C during an anneal.) Dark frames are obtained twice a day in calibration programs 8864 and 8901. Prior to combining these individual darks into a superdark, each dark is assigned a temperature from the CCD housing temperature log and than scaled to the reference temperature using the relationship above (i.e., divide each dark by the above scale factor, where T is the temperature of each dark). Cosmic rays are rejected while combining the scaled darks into a superdark.

Summary of steps needed to create a temperature-dependent superdark:

1. Run calstis (or basic2d) on your individual darks, with DQICORR, BLEVCORR, and BIASCORR all set to perform.

2. Divide each dark by the scale factor = 1 + 0.07*(T-18), where T is the temperature of each dark.

3. Combine the darks with cosmic-ray rejection using ocrreject.

4. Multiply the resulting superdark by the scale factor = 1 + 0.07*(T-18), where T is the temperature of your science data.