ETC Appendix: Inputs and Calculations

Pandeia uses datasets from multiple testing campaigns and different types of input files to support Exposure Time Calculations for Roman's Wide Field Instrument (WFI). This article outlines the files used, and explains detector and mode-specific parameters and considerations. 

This appendix summarizes the input information and data that Pandeia uses in its calculations, including:

  1. Default selections for each observing mode that are set up via the build_default_calc() function in Pandeia
  2. Instrument and detector parameters that are pre-configured by Pandeia
  3. Filter throughputs
  4. Multiaccum (MA) tables
  5. Point Spread Functions
  6. Background models

All these values are stored in a file called config.json  located under the $pandeia_refdata/roman/wfi/.

  Pandeia is the back-end engine for the web ETC and the information summarized here applies to both tools. This documentation is written for  Pandeia version 2025.9 (released on September 15, 2025).



Default Setup for Calculation

The default parameters for each observing mode, as configured by the ETC's built-in function build_default_calc(<telescope>, <instrument>, <mode>), are listed below. Users may modify these parameters after initialization in the web ETC and  Pandeia . 

Pandeia takes lowercase letters as inputs. 


Imaging Mode

The parameters initialized when a calculation is setup via build_default_calc('roman', 'wfi', 'imaging') are given in the Table of Imaging Mode Parameters.

Table of Imaging Mode Parameters 

Detector

wfi01

Pixel size

0.11''

MA table

im_135_8

Filter

f158

Strategy

Imagingapphot(Imaging Aperture Photometry)

Aperture size: 0.2''

Sky annulus: 0.4'', 0.6'' (inner, outer, respectively)

Target location

Center of the detector (X: 0'' & Y: 0'')

Scene size

5''


Imaging Aperture Photometry

The ETC performs aperture photometry, and the resulting signal-to-noise ratio (SNR) may differ from the SNR obtained through PSF photometry.


Spectroscopy Mode

The parameters initialized when a calculation is setup via build_default_calc('roman', 'wfi', 'spectroscopy') are given in the Table of Spectroscopy Mode Parameters. 

Table of Imaging Mode Parameters 

Detector

wfi01

Pixel size

0.11''

MA table

sp_300_16

Disperser

grism

Strategy

Specapphot(Aperture Spectral Extraction)

Aperture size: 0.32''

Sky annulus: 0.24'', 0.64 (inner, outer, respectively)

Reference wavelength: 1.5749 microns 


Target location

Center of the detector (X: 0'' & Y: 0'')

Scene size

5''


Reference wavelength

This parameter is applicable only in spectroscopy mode. The reference wavelength denoted the wavelength at which the source extraction is performed. The extracted SNR and flux at this wavelength are shown in the Reports tab as 'Extracted Signal-to-Noise Ratio' and 'Extracted Flux'. Wavelength-dependent extracted SNR and flux are available in the SNR(λ) tab in the Plots pane of the web ETC. 




Detector Parameters 

We provide a complete list of detector parameters used as inputs to the ETC. Certain parameters are common to all 18 detectors, whereas others are detector-specific.


Detector-Independent Parameters

These parameters are common to all 18 WFI detectors and are listed in the Table of Detector Independent Parameters. Parameters that are fixed and not subject to change are marked as 'Fixed' in the Note column.

Some of the parameters were determined from WFI Ground Testing Campaigns testing and are marked as 'TVAC2'  in the Note column. These values are based on preliminary reference files derived from an initial  analysis performed by the Science Operations Center (SOC) at STScI, using data from the second Thermal Vacuum Testing Campaign (TVAC). The SOC maintains the reference files for the Exposure Level Pipeline in units of DN; these are converted to electrons using the IPC-corrected gain. For a typical gain is 1.8 e-/DN, and applying the GSFC-determined IPC correction factor of 1.08, the resulting IPC-corrected gain is 1.67 e⁻/DN. 


The TVAC2 parameters and their derivation are as follows:

  • fullwell: For Roman, saturation is defined as the value at which the nonlinearity correction becomes unreliable, which occurs at approximately 55,000 DN. To determine the equivalent full-well depth, this value is first converted to the pre-nonlinearity-correction level (an increase of roughly 10%) and then converted to electrons using a representative IPC-corrected gain. 
  • rn_correlation: This boolean specifies whether to apply a read noise correlation matrix derived from ground-test data. The matrix gives the correlation between adjacent pixels in a read noise map, measured from the residuals of a ramp fit to the difference of two dark ramps after correction for 1/f noise (Betti et al. 2024).  


Table of Detector Independent Parameters 

Parameter

Description

Value

Note

crs

Should cosmic rays be considered in the noise calculation by default?

true


dark

Should dark current be considered in the noise calculation by default?

true


det_type

Detector type

h4rg-10

Fixed

excess

Should excess noise be considered in the noise calculation by default?

false


ffnoise

Should flat-field noise be considered in the noise calculation by default?

true


fullwell

Fullwell depth in electrons

100000

TVAC2

ipc

Should IPC convolution be considered in the noise calculation by default? 

false


mingroups

Minimum number of unsaturated groups/resultants below which a pixel is said to be fully saturated

2

Fixed

pix_size_x

Pixel size in microns. Used in cosmic ray calculation

10

Fixed

pix_size_y

Pixel size in microns. Used in cosmic ray calculation

10

Fixed

pixels_x

Used for defining chip area for multi-detector-wide fields of view

4088

Fixed

pixels_y

Used for defining chip area for multi-detector-wide fields of view

4088

Fixed

readnoise

Should read noise be considered in the noise calculation by default?

true


rn_correlation

Should the read noise be considered as correlated in the noise calculation by default? If true, uses a correlation matrix

true

TVAC2

saturation

Should calculations be able to saturate?

true


scatter

Should detector scattering be considered in the noise calculation by default? 

true


coll_area

Area of detector in cm^2

36076.8 


minimal_spectral_spacing

Maximum resolution we allow within the ETC; any input (template SED) higher than this gets resampled to R=this value

200000.0


cn_npixels

Number of pixels affected by each cosmic ray event

 5.0


cr_rate

Rate of cosmic rays in events/cm^2/s

 0.3898



Detector-Specific Parameters

Some parameters vary by detector; these are summarized in the Table of Detector Specific Parameters. Each parameter is based on preliminary reference files derived from an initial analysis performed by the Science Operations Center (SOC) at STScI, using data from the second Thermal Vacuum Testing Campaign (TVAC). The SOC maintains the reference files for the Exposure Level Pipeline in units of DN; these are converted to electrons using the IPC-corrected gain. For a typical gain is 1.8 e-/DN, and applying the GSFC-determined IPC correction factor of 1.08, the resulting IPC-corrected gain is 1.67 e⁻/DN. 

The detector-dependent parameters are:

  • dark_current: The dark current in electrons per second (e-/second). These values are adopted from the detailed analysis of Betti et al. (2025, see their table 2), based on long dark exposures with noise sources removed. The reported values are converted to electrons per second using a representative IPC-corrected gain.
  • ff_electrons: The number of e- in a hypothetical flat-field exposure, used to determine the flat-field error noise component. These values were derived from the flat-field reference file by converting the average per-pixel uncertainty (assumed to be dominated by photon noise) into the equivalent signal. 
  • rn: The read noise in e-. This single-read noise is computed as the variance of the ramp-fit residuals from the difference of two dark ramps after removal of 1/f noise (Betti et al. 2024), and converted to electrons per second using a representative IPC-corrected gain.

Table of Detector Specific Parameters  

WFI Detectordark_current (Dark current in e-/sec)ff_electrons (Number of e- in hypothetical flat; used to determine the flat-field error noise component)rn (Readnoise in e-)
WFI010.0167224000011.85
WFI020.018224000010.02
WFI03

0.0165

224000010.29
WFI040.017224000010.61
WFI050.029522400009.93
WFI060.01722400009.49
WFI070.023822400009.51
WFI080.01622400008.79
WFI090.016822400009.15
WFI100.0112224000011.06
WFI110.027722400009.86
WFI120.020322400009.72
WFI130.0118224000011.4
WFI140.017222400009.46
WFI150.01522400009.39
WFI160.016722400009.68
WFI170.015722400009.06
WFI180.017322400008.84



Filter Throughputs

The Roman effective area curves are provided as CSV files in the Roman Technical Repository, and the filter throughputs used by Pandeia are obtained by dividing the effective area curves by the telescope effective area (3.6076 m2). The throughput files are located in $pandeia_refdata/roman/wfi/filters for the eight imaging filters and $pandeia_refdata/roman/wfi/blaze for the dispersers.



MultiAccum Tables

The supported MultiAccum (MA) tables are from Revision G. As of the R2025.9 version (released on September 15, 2025), the web ETC no longer supports truncation of the MA tables. Users can explore the MA table truncation using  Pandeia but caution should be exercised when considering it. In addition, the engineering MA tables are supported in Pandeia only. For more information on the MA table, please refer to WFI MultiAccum (MA) Tables article.  



Point Spread Function (PSF)

The PSFs are pre-computed using STPSF 2.1.0 and are located in $pandeia_refdata/roman/wfi/psfs. For more information on STPSF , please refer to STPSF for Roman. 



Background Models

The background models available in Pandeia are pre-computed from the Roman Background Tool (RBT) at specific locations that are representative of three Core Community Surveys. More specifically, backgrounds are pre-computed for six locations in the High Latitude Wide Area Survey (HLWAS), one location for the High Latitude Time Domain Survey (HLTDS), and one location for the Galactic Bulge Time Domain Survey (GBTDS). Descriptions of these backgrounds can be found in the Pandeia Engine & Roman Interactive Sensitivity Tool (RIST) section in Assessing Background Levels for WFI Observations. The background models are located in $pandeia_refdata/roman/background. 



For additional questions not answered in this article, please contact the Roman Help Desk.





Latest Update



Publication

 

Initial publication of the article.