Absolute Wavelength Calibration

The goal of the JWST absolute wavelength calibration is to transform spectra from detector pixel space to physical units of wavelength. To achieve this goal and meet wavelength accuracy requirements, on-orbit measurements with internal lamps and celestial calibrators will be used.

On this page

See also: Absolute Flux Calibration, Data Calibration Considerations, JWST Data Reduction Pipeline Overview

The JWST science instruments offer a variety of spectroscopic capabilities. In the near-IR domain, NIRCam provides grism slitess spectroscopy and NIRISS offers wide field slitless spectroscopy and single object slitless spectroscopy

NIRSpec has 4 different spectroscopic modes: 

In the mid-IR MIRI delivers low-resolution slitted and slitless spectroscopy and medium resolution integral field unit (IFU) spectroscopy. For calibration purposes, this wide range of observations requires coordination: the NIR instruments can  use the same calibrators for many modes, and MIRI may need NIR data for planning purposes. Much effort is being invested now in defining the JWST wavelength calibrators list; this page will be updated when more detailed information is available.

Wavelength calibration requirements

Table 1. Accuracy requirements for each of the JWST science instruments

Science instrument



For all spectroscopic modes (MRS and LRS) the wavelengths shall be known to within 1/10 of a resolution element for an unresolved line with S/N = 50 after calibration.


No formal requirement


The wavelength solution for both WFSS grisms and orders 1 and 2 of the SOSS cross-dispersed grism shall be known to within 1/10th of a resolution element after calibration.


For all spectroscopic modes the wavelength scale shall be determined with an accuracy of better than 1/8 of a resolution element after calibration

The requirement for the NIRISS grisms is taken to refer to a bright point source. For these slitless spectroscopy modes the wavelength resolution is degraded for extended sources, and it depends on the detailed size and shape of the object in that case.

NIRSpec calibration strategy

The NIRSpec wavelength calibration is applied to data using a parametric optical model of the instrument.  The model provides a means of calculating the wavelength value at each pixel, given the disperser and aperture, by following the optical path of light from a celestial source through the instrument to the detector plane.  The model was adopted in part to provide the pipeline with a flexible way of dealing with complex aspects of wavelength calibration over the NIRSpec field of view.  One issue of particular importance is the fact that the grating wheel positioning is not strictly repeatable; this can result in dispersion shifts of up to one pixel in the detector plane between identical spectra taken before and after a grating wheel move. Any such shifts can be calculated with high accuracy using wheel tilt sensor telemetry, and a correction has been incorporated into the model calculations. 

NIRSpec is the only science instrument that has on-board calibration sources suitable for wavelength characterization. Initial calibration of the instrument model was carried out during ground testing using a combination of internal calibration lamps and an external Ar lamp. 

The internal line lamps will be used to determine a final calibration of the instrument model on-orbit, as well as provide long-term monitoring. The lamps include two different line sources:

  • "REF" lamp with erbium filter, which provides narrow absorption lines over a limited range of wavelengths.
  • "LINE" lamps with interference filters that provide broad emission features over the full wavelength range for all modes.

External observations of a celestial emission line source are still necessary in order to provide an independent check of the model. 

Considerations for celestial calibrators 

Ideally, a wavelength calibrator must comply with the following characteristics: 

  •  Line density.

  •  Different source shapes: some spatially extended for MOS (not required but currently adopted) and IFU, and others point sources (grisms).
  •  The presence of companions should be taken considered to avoid confusion.
  •  Variable targets are not desired.

 Feasibility of ground observations is also being considered: sources of  K=15 need ~1hr from Keck/Gemini/VLT. Wavelengths longer than 3 μm are mostly out of reach.


Candidate targets

Planetary nebulae are well suited wavelength calibrators; they exhibit relatively narrow emission lines with moderate density, at least in the NIR. CLOUDY models can be used to extrapolate emission strengths at wavelengths non-accessible from the ground, for planning exposure times.

  • Ground NIR observations are required to constrain the input physical parameters of the line models.
  • It is also possible to use existing Spitzer observations for MIR.

Table 2. Information on some of the celestial calibrators already selected

InstrumentCalibratorRA (J2000)DEC (J2000)TypeK magnitude
NIRSpecNGC 654317:58:33+66:38:00Planetary Nebulae8.34
NIRISSSMP LMC 5805:24:20.81-70:05:01.9Planetary Nebulae14.5
NIRCamSMP LMC 5805:24:20.81-70:05:01.9Planetary Nebulae14.5

NIRSpec target: NGC6543

The “Cat’s Eye” planetary nebulae is located in the northern CVZ, its halo is well-matched to the NIRSpec MSA FOV, and optical observations suggest “inert” kinematics (ΔVr < 5 km/s, FWHM ~ 10 km/s). IR spectroscopy is available for the of core only. A proposal has been submitted to Keck (NASA time); the main results are:

  • MOSFIRE H/K band observations have shown that halo knots are pure H_2 emission. Since CLOUDY is less applicable in this case, we have  used separate shock models instead to estimate line strengths across the NIRSpec wavelength range. 
  • Enough density of lines for  RV analysis for extrapolation to 3–5 μm.  RVs are consistent with optical measurements, and the uncertainties & dispersion are well within wavelength calibration accuracy requirement for R=2700 (~14 km/s).

NIRCam/NIRISS target: SMP LMC 58 

This nearly unresolved (~0.08”) planetary nebulae emission line source is located in the CVZ. It's K magnitude is 14.5, rendering it too faint for SOSS calibration. It has at least 3–5 lines per filter. In this case there is need ground-based NIR spectrum to verify line fluxes. We have obtained ground-based NIR spectroscopy to measure line fluxes, the data analysis is still in progress. 

MIRI (and NIRISS SOSS) targets:

Be stars are active objects that a have suitable line/continuum ratios. They also exhibit good line density across large wavelength range, lines are not too bright and they are unresolved (SOSS, LRS) . However, they may be variable and are unresolved; there is need mapping for MRS observations. We are currently vetting a small list of Be targets in the LMC using ground-based NIR spectroscopy.



Stanghellini, L., et al. 2002, ApJ 575 178S 
Optical Slitless Spectroscopy of Large Magellanic Cloud Planetary Nebulae: A Study of the Emission Lines and Morphology

Latest updates
Originally published