Astronomy & Astrophysics

All Figures

$\begin{figure} \par\includegraphics[width=8.25cm,clip]{4020f1a.ps}\hspace*{2mm} \includegraphics[width=9.15cm,clip]{4020f1b.ps} \end{figure}$ Figure 1: Rotation measure maps of the regions in Auriga ( left) and in Horologium ( right), superimposed on polarized intensity in grey scale where the maximum is 90 mJy/beam for Auriga and 110 mJy/beam for Horologium. Rotation measures are denoted by white circles, where filled circles are positive RMs. The diameter of the symbol represents the magnitude of RM, and the scaling is given in rad m^-2. Only RMs for which $\langle P \rangle \ge 5\sigma$ and reduced $\chi ^2$ of the linear $\phi (\lambda ^2)$ -relation <2 are shown, and only every second independent beam.
Open with DEXTER

In the text

$\begin{figure} \par\includegraphics[width=14.2cm,clip]{4020f2.ps} \end{figure}$ Figure 2: Illustration of how offsets can cause small-scale structure in P. Left: an model polarization angle distribution in one dimension, where P is assumed constant. Center: small-scale structure in Q (solid line) and U (dashed line) corresponding to the change in polarization angle, while P (dotted line) remains constant. Right: the interferometer response to this distribution, where P does show apparent structure.
Open with DEXTER

In the text

$\begin{figure} \par\includegraphics[width=10.7cm,clip]{4020f3.ps} \end{figure}$ Figure 3: The effect that offsets have on apparent RMs. Top: hypothetical polarization vectors at 3 adjacent positions in the sky for three values of true RM, at 5 wavelengths denoted 1 to 5. Center: results after subtracting offsets determined from the situation in the top panel, showing how the offsets can destroy the linear $\phi (\lambda ^2)$ -relation. Bottom: 4 estimates of $\Delta \phi$ for each wavelength step $\Delta \lambda ^2$ , which gives the apparent $RM = \Delta \phi /\Delta \lambda ^2$ as the slope of a linear fit of $\phi$ to $\lambda ^2$ .
Open with DEXTER

In the text

$\begin{figure} \par\includegraphics[width=16cm,clip]{4020f4.ps} \end{figure}$ Figure 4: RM distributions in separate pointings. The plots show the central row of pointings in the Auriga region. Dotted lines are Gaussian fits to the data, and the fitted $\sigma _{RM}$ are given above the plot.
Open with DEXTER

In the text

$\begin{figure} \par\includegraphics[width=14.1cm,clip]{4020f5.ps} \end{figure}$ Figure 5: Missing large-scale structure in Q (solid line) and U (dashed line) for a Faraday screen with a Gaussian RM distribution with width $\sigma _{RM}$ and P₀= 1, assuming that the offsets are constant over the field. Diamonds denote $\sigma _{RM}$ observed in the pointings of the Auriga field, squares are the pointings of the Horologium field. Offsets depend on the intrinsic polarization angle $\phi _0$ , shown are the best- and worst-case scenario for $\phi_0 = 10\mbox{$^{\circ}$ }$ ( left) and $\phi_0 = 80\mbox{$^{\circ}$ }$ ( right).
Open with DEXTER

In the text

$\begin{figure} \par\includegraphics[width=8.6cm,clip]{4020f6.ps} \end{figure}$ Figure 6: The influence of offsets on reduced $\chi ^2$ in a subfield of Auriga. Left: reduced $\chi ^2$ distribution of linear $\phi (\lambda ^2)$ -relation with best-fit offsets against values of reduced $\chi ^2$ without offsets. Right: distribution of reduced $\chi ^2$ with computed offsets (dotted line) and without offsets (solid line).
Open with DEXTER

In the text

$\begin{figure} \par\includegraphics[width=13.2cm,clip]{4020f7.ps} \end{figure}$ Figure 7: Frequency dependence of the depth of the canals, in the Auriga region ( top) and Horologium region ( bottom). P is the average over all canal-pixels, and in each plot the canals were defined in an other frequency band, at 341, 349, 355, 360, and 375 MHz from left to right. The prediction for $P(\lambda ^2)$ if all canals were caused by beam depolarization due to a change in RM is shown in dashed lines, from bottom to top for $\Delta RM \approx$ 2.1, 6.3 rad m^-2 (i.e. $\Delta\phi = 90\mbox{$^{\circ}$ }, 270\mbox{$^{\circ}$ }$ ). The prediction for $P(\lambda ^2)$ if the canals were caused by differential Faraday rotation are shown dotted for $2 RM_{\rm c} \lambda ^2 = n\pi$ for n = 1, 3 and 5.
Open with DEXTER

In the text

$\begin{figure} \par\includegraphics[width=13cm,clip]{4020f8a.ps}\vspace*{2mm} \includegraphics[width=13.15cm,clip]{4020f8b.ps} \end{figure}$ Figure 8: Distribution of $\Delta RM_{\rm c}$ across canals and the absolute $RM_{\rm c}$ at the canal position (as estimated from its neighbors). Canals are defined at frequencies 341, 349, 355, 360, and 375 MHz from left to right. Top panels show the Auriga region and bottom panels the Horologium region. Only $RM_{\rm c}$ values where $\chi ^2_{{\rm red}} < 2$ and P > 20 mJy/beam are used. In the $\Delta RM_{\rm c}$ plots, dotted vertical lines are $\Delta RM_{\rm c}$ values where $\Delta \phi$ across a canal would be $\pm$ 90 $^{\circ }$ , $\pm$ 270 $^{\circ }$ or $\pm$ 450 $^{\circ }$ . In the $RM_{\rm c}$ plots, the dotted vertical lines are values where $2 RM_{\rm c} \lambda ^2 = n\pi$ .
Open with DEXTER

In the text

$\begin{figure} \par\includegraphics[width=8.8cm,clip]{4020f9.ps} \end{figure}$ Figure 9: Predicted canal shape for a toy model with angle changes of 90 $^{\circ }$ across a canal with different gradients ( left). Convolution of Q and U gives the P distribution as shown right for the 3 gradients. The width of the beam is 12 pixels.
Open with DEXTER

In the text

$\begin{figure} \par\includegraphics[width=8.8cm,clip]{4020f10.ps} \end{figure}$ Figure 10: Measures shapes of the deepest canals. Upper plots: examples of observed one-dimensional P distributions in the Horologium region for five frequencies, where the deepest canal is observed at 341, 355 and 349 MHz respectively. Lower plots: the same P distribution for the deepest canal as above (solid) and the best fit according to the model of Fig. 9. The P profile is so steep that the change in angle that causes the canal must be on scales of an arcminute or smaller.
Open with DEXTER

In the text

$\begin{figure} \par\includegraphics[width=16cm,clip]{4020f4.ps} \end{figure}$	Figure 4: RM distributions in separate pointings. The plots show the central row of pointings in the Auriga region. Dotted lines are Gaussian fits to the data, and the fitted $\sigma _{RM}$ are given above the plot.
Open with DEXTER

$\begin{figure} \par\includegraphics[width=8.6cm,clip]{4020f6.ps} \end{figure}$	Figure 6: The influence of offsets on reduced $\chi ^2$ in a subfield of Auriga. Left: reduced $\chi ^2$ distribution of linear $\phi (\lambda ^2)$ -relation with best-fit offsets against values of reduced $\chi ^2$ without offsets. Right: distribution of reduced $\chi ^2$ with computed offsets (dotted line) and without offsets (solid line).
Open with DEXTER

$\begin{figure} \par\includegraphics[width=8.8cm,clip]{4020f9.ps} \end{figure}$	Figure 9: Predicted canal shape for a toy model with angle changes of 90 $^{\circ }$ across a canal with different gradients ( left). Convolution of Q and U gives the P distribution as shown right for the 3 gradients. The width of the beam is 12 pixels.
Open with DEXTER