NIRCam Standard Subpixel Dithers
JWST NIRCam subpixel dithers are small pointing offsets between exposures that mitigate an undersampled PSF and bad detector pixels.
On this page
NIRCam subpixel dithers are small pointing offsets performed between exposures in imaging or wide field slitless spectroscopy (WFSS) observations. They include subpixel offsets to improve the spatial resolution of the final combined (drizzled) image mosaic of all exposures.
This is especially important below 2 µm in the short wavelength channel (0.6–2.3 µm) and below 4 µm in the long wavelength channel (2.4–5.0 µm). Below these Nyquist wavelengths, NIRCam's detectors undersample the JWST PSF (the FWHM is less than 2 pixels). This undersampling can be most severe in the short wavelength channel, up to a factor of ~2 with F070W. (Wavefront errors may broaden the PSF below 2 µm, but PSF features will still scale as λ/D.) Therefore, subpixel dither patterns were designed to provide optimal benefits for the ~0.031" pixels in the short wavelength channel (Anderson 2009). Detailed background information on principles of dithered observations with JWST are also described in Koekemoer & Lindsay (2005), Anderson (2011), Anderson (2014), and Coe (2017).
Words in bold are GUI menus/
panels or data software packages;
bold italics are buttons in GUI
tools or package parameters.
Each STANDARD subpixel dither step includes shifts along both detector axes (x and y) to mitigate bad pixels. Each shift is an integer plus a fractional pixel (subpixel) step (e.g., 2.5 pixels). The integer component mitigates bad detector pixels and flat field uncertainties, while the fractional subpixel step improves PSF sampling and achievable spatial resolution. These patterns are detailed below. Small patterns (up to 0.41" across) are designed individually for up to 9 dither positions. Subsets of a single larger pattern (1.74" across) are used when requesting 12–64 dither positions.
More compact subpixel SMALL-GRID-DITHER patterns are also available for imaging. They are performed more quickly (lower overheads) using the fine steering mirror instead of slewing the telescope. They also includes fewer integer pixel steps, reducing their ability to mitigate bad pixels or flat field uncertainties. The fractional pixel steps are exactly preserved for up to 9 dither positions.
NIRCam subpixel dithers are smaller than the primary dithers used to cover detector gaps and mitigate flat field uncertainties on larger scales. Primary dithers do offer some sub-optimal subpixel sampling.
Note when performing coordinated parallel observations, additional custom subpixel dither patterns are available that achieve optimal pixel phase sampling for both the prime and parallel instruments. NIRCam's small grid dithers are not available for use with parallels as they are not large enough for other instruments.
A complete set of primary and secondary subpixel dither patterns (specified in arcseconds) is available as ASCII tables in a .zip file:
This .zip file contains a collection of separate files for all the NIRCam-related dither patterns. Specifically relevant to the subpixel dither patterns described in this article, it contains these subpixel imaging dither patterns (mapping their names to the APT pattern names):
Filename | APT dither pattern type and pattern name |
|---|---|
| NircamImagingSmallSubpixel.txt | Subpixel Dither Type: SMALL-GRID-DITHER |
| NircamImagingSubpixel.txt | Subpixel Dither Type: STANDARD |
Subpixel dither offsets may be modified slightly as measurements of the NIRCam detector pixel scales are refined.
All pointing offsets are relative to the selected aperture's reference position in that aperture's ideal coordinate frame (X, Y). All NIRCam apertures are nearly aligned (to ~1° of rotation) with both the JWST coordinate system (V2, V3) and all detector rows and columns.
Patterns for 2–9 positions
Eight small (up to 0.41" across) STANDARD subpixel dither patterns were designed with 2–9 positions (Figure 1). The 4-point pattern samples the pixel-phase space with twice the resolution of the NIRCam pixels. The 9-point pattern samples 3 times the pixel resolution, which would be most useful for F070W observations.
All 8 patterns are kept small enough (within a 13 × 13 pixel box, 0.41" on a side) to ensure coherent subpixel shifts for all pixels on the detector, given geometric distortion. Allowing for 2% deviations in pixel scale at the detector edges, a 13-pixel shift at the detector center might differ by 0.26 pixel at the edges. This subpixel deviation maintains coherent subpixel shifts across the detector.
APT provides options for N = 1, 2, 3, 4, 5, 7, 9 Subpixel Positions (excluding 6 and 8 for simplicity). N = 1 means no subpixel dithering.
Click on the figures for a larger view.
Each row shows a dither pattern defined for use with a different number of Subpixel Positions: N = 2–9. The two central columns show the full dither patterns in units of short wavelength pixels (0.031") and long wavelength pixels (0.063"), respectively. The outer two columns show the phase spacing, or fractional pixel offset, of each dither position. The open circles show the phase spacing pattern repeated outside the central pixel. Reproduced from Anderson 2009, Figures 17 and 18.
Pattern for 12–64 positions
The 9-point pattern provides the finest spatial sampling required. Finer details would be blurred by the PSF, even in F070W imaging. Additional subpixel dithers cannot improve image resolution, but they can be used in a larger pattern to further mitigate bad pixels and flat-field uncertainties.
A larger 64-point pattern (Figure 2), designed within a 55 × 55 pixel box (1.74" on a side), consists of 16 sets of 4-point dither patterns, each fitting within a 10 × 10 pixel box (0.32" on a side) and doubling the native pixel sampling. The full 16-set pattern spirals outward from the center to cover a larger area with each set of 4 dither points. This strategy is similar to that used when imaging the Hubble Ultra Deep Field (HUDF).
APT provides options for N = 12, 16, 20, 24, 30, 36, 48, and 64 Subpixel Positions. Selecting 12, for example, executes the first 12 points of the 64-point pattern below.
Click on the figure for a larger view.
The larger 64-point dither pattern designed for use with 12 or more Subpixel Positions. (The first N points of the pattern are executed.) Top: pixel phase shifts. Bottom: full shifts. Left: short-wavelength channel in 0.031" pixel units. Right: long-wavelength channel in 0.063" pixel units. Reproduced from Anderson 2009, Figure 20.
References
Anderson, J. 2009, JWST-STScI-001738
Dither Patterns for NIRCam Imaging
Anderson, J. 2011, JWST-STScI-002199
NIRCam Dithering Strategies I: A Least Squares Approach to Image Combination
Anderson, J., 2014, JWST-STScI-002473
NIRCam Dithering Strategies II: Primaries, Secondaries, and Sampling
Coe, D. 2017, JWST-STScI-005798
More Efficient NIRCam Dither Patterns
Koekemoer, A. M. & Lindsay, K. 2005, JWST-STScI-000647
An Investigation of Optimal Dither Strategies for JWST