Space Telescope Science Institute
Table of Contents Previous Next Index Print

NICMOS Data Handbook > Chapter 4: Anomalies and Error Sources > 4.5 Count Rate Non-Linearity

4.5 Count Rate Non-Linearity
The NICMOS STScI team has determined that NICMOS has a significant count rate dependent nonlinearity that also depends on wavelength. This is a different nonlinearity from the well-known total count dependent nonlinearity described in Section 4.4. The nonlinearity in amounts to 0.05-0.10 mag offset per dex change in incident flux for the shortest wavelength (F090M and F110W), about 0.03 mag/dex at F160W and less than that at longer wavelengths (see figure below). Details about how the nonlinearity was quantified can be found in the NICMOS ISR 2006-001 and NICMOS ISR 2006-002 Instrument Science Reports (ISRs).
Figure 4.8: Figure 4.8: Graph of non-linearity by camera, top line NIC1, middle line NIC2, and bottom line NIC3
A python routine to correct NICMOS data for the nonlinearity has been
developed. The routine rnlincor is available in STSDAS as part of the NICMOS package and can be used as an independent program outside STSDAS as well. The routine corrects NICMOS count-rate images assuming the nonlinearity follows a power-law behavior. The wavelength dependence of the nonlinearity is interpolated between the measured points of de Jong et al. (2006, NICMOS ISR 2006-001) and Bohlin et al. (2006, NICMOS ISR 2006-002) if necessary.
While the exact cause of the count rate nonlinearity is not know, it is believed that it depends on the incoming flux only. Therefore the routine should be run on images that have processed through the pipeline as much as possible with all instrumental effects removed. Most notably, the quadrant bias should be removed with pedsky or pedsub before performing the count rate nonlinearity correction. The pedsky routine subtracts the sky value, but the nonlinearity script detects this and will add that value back in (as it is part of the flux that determines the nonlinearity) before doing the correction. The routine can be run on MULTIDRIZZLED images, but this is only correct if the sky value did not vary a lot relative the sources of interest and by making sure a typical sky value is set in the final MULTIDRIZZLED image. The routine can only be used on imaging data (pixels in grism observations receive flux from a broad range of wavelengths) and should not be applied to calibration reference files (e.g., dark frames, flat fields, etc.).
The rnlincor routine requires calibration tables specifying the wavelength dependence of the non-linearity and an offset to make sure the photometric zero point matches the NICMOS calibration values in the header. The filenames of these calibration tables need to be specified in the header of the images to be corrected (keywords are zprattab and rnlcortb). Images retrieved from the archive before this new routine was introduced will not have these header items. These header items can be added with the IRAF routine hedit or, alternatively, one can retrieve the images again from the archive.
The usage of the routine is very straightforward as it has no free parameters to influence its result:
If no output file is specified it will be created with a _ncl extension based on the input file name.
Little is known about the temporal variation of this effect, but the NICMOS team has a monitoring program in place. Please check for the latest information on the rate nonlinearity at:

Table of Contents Previous Next Index Print