Table 1. Data sets used in this analysis. The first block of maps are used as templates for fitting the synchrotron, free–free, and dust components. The second block lists the maps to which we applied the template fitting procedure.
| Telescope/survey | Frequency (GHz) | FWHM (arcmin) | σ cal (%) | Reference | Notes |
|---|---|---|---|---|---|
| Haslam | 0.408 | 51 | 10 | Remazeilles et al. (2015) | Synchrotron template |
| C-BASS | 4.76 | 44 | 3 | This work | Synchrotron template |
| WHAM Hα | N/A | 1 | 10 | Dickinson et al. (2003), Finkbeiner (2003) | Free–free template |
| Planck HFI 353 GHz | 353 | 4.7 | 5 | Planck Collaboration I (2020) | Dust template |
| Planck 353 GHz optical depth | 353 | 5.0 | 5 | Planck Collaboration XI (2014) | Dust template |
| Planck dust radiance R | – | 5.0 | – | Planck Collaboration XI (2014) | Dust template |
| IRAS100 μm | 2997 | 4.3 | 13.5 | Miville-Deschenes & Lagache (2005) | Dust template |
| FDS8 model 8 | 94 | 6.1 | – | Schlegel, Finkbeiner & Davis (1998) | Dust template |
| WMAP K band | 22.8 | 51.3 | 3 | Bennett et al. (2013) | 9-year |
| Planck LFI 30 GHz | 28.4 | 33.16 | 3 | Planck Collaboration I (2020) | PR3 |
| WMAP Ka-band | 33.0 | 39.1 | 3 | Bennett et al. (2013) | 9-year |
| WMAP Q-band | 40.7 | 30.8 | 3 | Bennett et al. (2013) | 9-year |
| Planck LFI 44 GHz | 44.1 | 28.09 | 3 | Planck Collaboration I (2020) | PR3 |
| WMAP V-band | 60.7 | 30.8 | 3 | Bennett et al. (2013) | 9-year |
| Planck LFI 70 GHz | 70.4 | 13.08 | 3 | Planck Collaboration I (2020) | PR3 |
| WMAP W-band | 93.5 | 30.8 | 3 | Bennett et al. (2013) | 9-year |
| Planck HFI 143 GHz | 143 | 7.18 | 5 | Planck Collaboration I (2020) | PR3 |
| Planck HFI 217 GHz | 217 | 4.87 | 5 | Planck Collaboration I (2020) | PR3 |
| Planck HFI 353 GHz | 353 | 4.7 | 5 | Planck Collaboration I (2020) | PR3 |
| Planck HFI 545 GHz | 545 | 4.73 | 5 | Planck Collaboration I (2020) | PR3 |
| Planck HFI 857 GHz | 857 | 4.51 | 5 | Planck Collaboration I (2020) | PR3 |
| IRAS100 μm | 2997 | 4.3 | 13.5 | Miville-Deschenes & Lagache (2005) | IRIS |
3.1 Synchrotron templates
As well as using the C-BASS map (Section 2) to trace the diffuse synchrotron emission we also use the 408-MHz radio continuum survey by Haslam et al. (1982), which has remained for decades the best tracer of diffuse Galactic synchrotron emission at WMAP and Planck frequencies. The survey was undertaken during the 1960 and 1970s and combines observations from the Effelsberg 100 m, and the Jodrell Bank 76 m Mk1 and Mk1a telescopes in the Northern hemisphere, and data from the Parkes 64 m telescope for the Southern hemisphere.
The average FWHM of the original map produced by Haslam et al. (1982) is 56 arcmin. Calibration and zero level offsets were set by an earlier 404-MHz survey (Pauliny-Toth & Shakeshaft 1962). The expected calibration uncertainty is 10 per cent. However, no deconvolution of the beam is possible, therefore we do not know how the calibration changes with angular scale. The original 408-MHz map has in recent years been improved in Remazeilles et al. (2015), which reduces striping due to instrumental systematics in the original data via Fourier filtering, and also provides an improved subtraction of point sources in the map. For this analysis, we use the destriped and desourced map provided by Remazeilles et al. (2015) from the Microwave Background Data Analysis (LAMBDA) website.[1]
3.2 Free–free template
The best currently available method for tracing free–free emission at mid-to-high Galactic latitudes is to use the Hα (λ656.28 nm) emission line. This is because for a given HII region the brightness at radio frequencies and the intensity of the Hα transition are both dependent on just the emission measure and the electron temperature of the plasma (Draine 2011).
Hα emission is particularly sensitive to dust extinction, which acts to reduce the Hα intensity along any given line of sight. Hα maps are also contaminated by stellar continuum emission that must be carefully subtracted. We use the all-sky composite Hα maps given by Dickinson et al. (2003) and Finkbeiner (2003) as templates for the free–free emission. Both combine publicly available Hα data sets but approach the correction for dust absorption and continuum removal slightly differently. The Finkbeiner (2003) data set assumes no correction for dust absorption, while for the Dickinson et al. (2003) Hα maps we use two dust mixing fractions: fd = 0 and 0.33.
3.3 Dust templates
We use the reprocessed IRAS 100-μm data (IRIS), which has a global calibration uncertainty of 13.5 per cent (Miville-Deschenes & Lagache 2005). The IRAS 100 μm (I100μ m ) is primarily a tracer of cold interstellar dust, however, it is somewhat sensitive to the local interstellar radiation field (ISRF) since it is near the peak of the thermal dust emission spectrum (Tibbs et al. 2013).
To trace the cold thermal dust component on the Rayleigh–Jeans tail of the thermal dust spectrum, we use the Planck 353-GHz map (I353) (Planck Collaboration I 2020), as well as the derived dust optical depth at 353 GHz (τ353) (Planck Collaboration X 2016). These maps are directly proportional to the dust column (Planck Collaboration XI 2014) but should be insensitive to changes in the ISRF. At frequencies around the peak of the thermal dust spectrum, the τ353 data are a better tracer of Galactic dust emission than I353 since they do not contain cosmic infrared background anisotropies.
We use the Planck-derived dust radiance map (Planck Collaboration XI 2014), which is the integrated bolometric intensity of the dust grains. The dust radiance is proportional to the amount of light absorbed by the dust and as such is directly proportional to the ISRF as well as the dust column.
Finally, to make comparisons with previous works, we also include the 94 GHz dust brightness map derived from model 8 in Schlegel
MNRAS 513, 5900–5919 (2022)
Downloaded from https://academic.oup.com/mnras/article/513/4/5900/6581346 by guest on 30 April 2023
