Forecasting Synchrotron Spectral Parameters with QUIJOTE-MFI2 in combination with Planck and WMAP

Abstract

We present a parametric component separation forecast for the QUIJOTE-MFI2 instrument (10-20 GHz), assessing its impact on constraining polarised synchrotron emission at $1^\circ$ FWHM and $N_{\rm side}=64$. Using simulated sky maps based on power-law and curved synchrotron spectra, we show that adding QUIJOTE-MFI2 to existing WMAP+$Planck$+MFI data yields statistically unbiased parameter estimates with substantial uncertainty reductions: improvement factors reach $\sim$10 for the synchrotron spectral index ($β_s$), $\sim$5 for the curvature parameter ($C_s$), and $\sim$43 for polarisation amplitudes in bright regions. Deep QUIJOTE cosmological fields enable $β_s$ constraints even in intrinsically low SNR regions where WMAP+$Planck$ alone remain prior-dominated. Current combined sensitivities are insufficient to detect a synchrotron curvature of $C_s=-0.052$ on a pixel-by-pixel basis, but a $2σ$ detection is achievable for $|C_s|\gtrsim 0.14$ in the brightest regions of the Galactic plane. In those deep cosmological fields, combining QUIJOTE-MFI2 with WMAP and $Planck$ reduces the median synchrotron residual at 100 GHz by a factor 2.2 (to 0.022 $μ$K$_{\rm CMB}$) and the total residual by a factor 1.8 (to 0.030 $μ$K$_{\rm CMB}$). These results demonstrate that QUIJOTE-MFI2 will provide critical low-frequency information for modelling Galactic synchrotron emission, offering valuable complementary constraints for future CMB surveys such as LiteBIRD and the Simons Observatory.

Figures (16)

• Figure 1: Independent Jeffreys prior distributions for each model parameter, computed for the three dataset combinations: WMAP+Planck, WMAP+Planck+MFI, and WMAP+Planck+MFI+MFI2. The curves show the two synchrotron models, with purple representing the power-law model (Eq. \ref{['eq:model_sync_pl']}) and green representing the curved model (Eq. \ref{['eq:model_sync_curv']}).
• Figure 2: Location of the four case study pixels on the simulated 28.4 GHz polarisation intensity map.
• Figure 3: Spectral Energy Distribution (SED) of polarisation intensity $P = \sqrt{Q^2+U^2}$ of the four case study pixels. Dashed lines show the theoretical SED assuming a power law synchrotron model, while points represent simulated noisy data with corresponding error bars. No bias correction has been applied to $P$, which is plotted here for illustration only (all fits are performed on $Q$ and $U$ Stokes parameters). At low SNR the resulting Ricean bias causes a slight deviation from the model, most evident for the low SNR pixel.
• Figure 4: Polarisation intensity SED with MCMC samples for the NPS pixel. Colored lines show 400 random samples for the total model (red), synchrotron (purple), dust (turquoise), and CMB (blue) components. The black solid lines represent the best fit total SED and its individual components (synchrotron, dust, CMB). Data points with error bars correspond to the simulated observations.
• Figure 5: Marginalised one-dimensional posterior probability density functions (PDFs) for the synchrotron parameters in the four case study pixels in the wide survey configuration, assuming synchrotron power-law spectrum. These were obtained from MCMC sampling with the three datasets: WMAP and Planck in green, with the inclusion of MFI in light blue and with the inclusion of MFI2 in purple. The black dashed lines are the true parameter values known from PySM.