1: \documentclass{aa}
2: \usepackage{graphicx}
3: \usepackage{txfonts}
4:
5: \begin{document}
6:
7: \title{Magnetic fields and gas in the cluster--influenced spiral\\
8: galaxy NGC\,4254
9: -- II. Structures of magnetic fields
10: }
11:
12: \author{Krzysztof T. Chy\.zy}
13: \institute{Astronomical Observatory, Jagiellonian
14: University, ul. Orla 171, 30-244 Krak\'ow, Poland
15: }
16:
17: \date{Received / Accepted }
18:
19: \titlerunning{Magnetic fields and gas in NGC~4254}
20: \authorrunning{K.T. Chy\.zy}
21:
22: \abstract
23: {}
24: {
25: The origin of asymmetric radio polarized emission in the Virgo Cluster
26: spiral NGC\,4254 is investigated and the influence of cluster environment
27: on the properties of magnetic fields is explored.
28: }
29: {Structures of magnetic fields are analyzed with the concept of ``magnetic
30: maps'', presenting distributions of different magnetic field components (total,
31: regular, and random) over the entire galaxy, free of Faraday rotation and
32: projection effects. A number of different physical
33: phenomena influencing the magnetic field are modeled analytically and
34: confronted with the galaxy's depolarization pattern and distribution
35: of magnetic field strength obtained from multifrequency polarimetric radio observations.
36: }
37: {
38: The study of orientation of intrinsic magnetic field vectors in NGC\,4254
39: indicates that their dramatic variation (from $0\degr$ to more than $40\degr$)
40: throughout the galaxy cannot arise from the dynamo process alone, but
41: must be dominated by effects such as density waves and local gas flows. We
42: determine within the galaxy the relation between the strength of
43: total magnetic field and the local star-formation rate (SFR) as a power-law with
44: an index of $+0.18\pm 0.01$. We find the opposite sense of the relation between
45: magnetic field regularity and SFR ($-0.32\pm 0.03$), and suggest that it results from efficient
46: production of random field with rising turbulence in the regions with
47: actively-forming stars. The distribution of Faraday rotation measures in NGC\,4254 indicates
48: a perturbed axisymmetrical mean-field dynamo mode or a mixture of axisymmetrical
49: and bisymmetrical ones with regular field directed outwards from the disk, which is
50: contrary to most observed galaxies. The galaxy's northern magnetic arm, located
51: on the upstream side of the local density wave, with regular field strength of
52: about $8\,\mu$G and the total one of $17\,\mu$G, much resembles those observed
53: in other galaxies. But the magnetic field within two outer arms (shifted
54: downstream of a density wave) is much stronger, up to $13\,\mu$G in the regular
55: field component and $20\,\mu$G in the total field.
56: Our modeling of cluster influence on different magnetic field components
57: indicates that within the outer magnetic arms the dynamo-induced magnetic
58: fields are modified by stretching and shearing forces rather than by cluster
59: ram pressure. Those forces, which are likely triggered by the galaxy's
60: gravitational interaction, produce an anisotropic component of the regular
61: field and enhance the polarized emission.
62: We also show that the magnetic energy within the large interarm regions and
63: the galaxy's outskirts exceeds the gas thermal and turbulent
64: energy, likely becoming dynamically important.
65: }
66: {}
67: \keywords{galaxies: general -- galaxies: ISM: magnetic fields -- galaxies:
68: magnetic fields -- galaxies: interactions -- galaxies: individual: NGC4254 --
69: radio continuum: galaxies -- ISM: magnetic fields}
70:
71: \maketitle
72:
73: \section{Introduction}
74: \label{intro}
75:
76: Disk galaxies in galaxy clusters are particularly affected by the cluster
77: environment (Roediger \& Hensler \cite{roediger05}) and can be even entirely transformed from
78: one Hubble type to another (Moore et al. \cite{moore96}). Interstellar medium (ISM) within
79: a cluster galaxy is altered in a different manner in individual galactic
80: regions, as can be seen, e.g., from H$\alpha$ (Koopmann \& Kenney \cite{koopmann04})
81: and \ion{H}{i} (Cayatte et al. \cite{cayatte90}) distributions. The magnetic field in cluster galaxies
82: as one of the ISM ingredients is also expected to be significantly modified,
83: mainly by ram-pressure (compression) of hot intracluster medium (ICM)
84: or by tidal (stretching/shearing) forces due to galactic encounters.
85: Investigation of influence of such external forces on magnetic field gives an independent
86: insight into the galaxy evolution within the cluster (Soida et al. \cite{soida06}).
87: Exploring magnetic field structures can also allow for determining how tangling
88: of magnetic field within star-forming regions alters different magnetic field components, or to
89: what extent the gaseous streamlines perturbed in cluster galaxies can tune
90: the dynamo-induced galactic magnetic fields (Shukurov \cite{shukurov98}, Elstner \cite{elstner05}).
91: To trace all those processes locally, we need sensitive radio polarimetric data of at
92: least kpc-scale resolution, which are available now only for the Virgo Cluster spirals.
93:
94: In this paper we present the first detailed analysis
95: of magnetic field, rotation measure, and depolarization patterns of a
96: cluster spiral -- the perturbed Virgo Cluster galaxy NGC\,4254.
97: In Paper I (Chy\.zy et al. \cite{chyzy07}) we presented our radio polarimetric
98: observations (of the VLA of
99: NRAO\footnote{National Radio Astronomy Observatory is a facility of National
100: Science Foundation operated under a cooperative agreement by Associated
101: Universities, Inc.} and 100-m Effelsberg telescope\footnote{The
102: 100--m telescope at Effelsberg is operated by the Max--Planck--Institut f\"ur
103: Radioastronomie (MPIfR) on behalf of the Max--Planck--Gesellschaft.})
104: and X-rays observations (from XMM-Newton\footnote{
105: XMM--Newton is an ESA science mission with instruments and contributions directly
106: funded by ESA Member States and NASA (Jansen et al. \cite{jansen01})}),
107: together with the data reduction process, as well as a global outline
108: of the galaxy's radio and polarimetric properties in
109: relation to other spectral bands. We found that in the radio domain the galaxy
110: has an asymmetrical distribution extending to the north, which resembles
111: disturbed distributions in the optical, UV, and X-rays bands, dominated
112: by a one-arm spiral structure. The polarized intensity maps at 8.46\,GHz
113: and 4.86\,GHz show an unusual strong ridge in the southern disk portion,
114: outside of a heavy optical spiral arm, with observed magnetic field
115: vectors directed along it (Fig.~\ref{f:intro}, see also Fig.~2 in Paper I).
116: Similar, but weaker, structures shifted off optical
117: features are also visible in other parts of the galaxy disk. Some polarized
118: features are strong enough to appear within optical filaments even in the
119: highly turbulent spiral arm regions. These mixed magnetic field patterns
120: are in sharp contrast to the previous studied grand-design spirals, which
121: typically show either interarm or arm-dominated polarized structures (like in
122: NGC\,6946 -- Beck \& Hoernes \cite{beckhoernes96} and M\,51 -- Fletcher et al.
123: \cite{fletcher04b}, respectively). Moreover, the star-formation rate (SFR) derived
124: from radio thermal emission is higher in NGC\,4254 than in
125: other galaxies, indicating some effects of an external (cluster) agent.
126:
127: The pattern of magnetic fields discovered in NGC\,4254 is investigated in the
128: present paper in a detailed manner. In Sect.~\ref{s:rm}, we present the
129: Faraday rotation measure and depolarization maps and derive
130: a Faraday-free intrinsic structure of magnetic field. Next, we introduce
131: ``magnetic maps'' (Sect.~\ref{s:magmaps}) -- a novel concept of presenting
132: magnetic field strength in different magnetic field components in the form of
133: distributions across the entire galaxy, free of Faraday rotation and projection
134: effects. In Sect.~\ref{s:orient}, we study the orientation (pitch angle) of
135: magnetic field vectors within structures of optical arms and gas filaments.
136: We investigate regularity of magnetic field over
137: regions of different star formation rates (Sect.~\ref{s:sfr}) and model
138: various Faraday depolarization effects (Sect.~\ref{s:depo}). The impact of
139: external compression and stretching effects on the different components of
140: galactic magnetic field is analytically modeled in Sect.~\ref{s:ani}
141: and finally discussed in Sect.~\ref{s:cluster}.
142:
143: \begin{figure}
144: \includegraphics[width=8.9cm]{f1.eps}
145: \caption{Optical DSS B image of NGC\,4254 and polarized intensity
146: at 4.86\,GHz in $15\arcsec$ resolution in contours. The contours are at $9.7\times (4, 9,
147: 16, 25, 40)\,\mu$G. Three main optical spiral arms (SW, N and NE) are
148: schematically marked. Letters a-e denote positions of contrast derivations in
149: Sect.~\ref{s:cluster}.
150: }
151: \label{f:intro}
152: \end{figure}
153:
154: \section{Results}
155: \subsection{Distribution of Faraday rotation and depolarization}
156: \label{s:rm}
157:
158: For the purpose of analyzing magnetic field in NGC\,4254 we use the radio
159: polarimetric observations described in detail in Paper I. The
160: observed polarized component of synchrotron emission results from
161: the regular (ordered) component of magnetic field. It can contain a mixture
162: of coherent (unidirectional) and anisotropic (incoherent) magnetic
163: fields. Only the coherent field contributes to the observed Faraday
164: rotation because the contributions from an anisotropic field with
165: vectors of opposite directions cancel when averaged over the telescope beam.
166: To construct the Faraday rotation measure ($RM$) map we take the
167: distributions of polarization angles at 4.86 and 8.46\,GHz in
168: $15\arcsec$ resolution. The observed Faraday rotation also includes
169: the contribution from the Galaxy's foreground rotation measure.
170: The spiral NGC\,4254 lies at $270.4\degr$ Galactic longitude and $75.2\degr$
171: Galactic latitude, where Johnston-Hollitt et al. (\cite{johnston04})
172: show only low values of foreground $RM$, changing sign from place
173: to place with absolute values $<25$\,rad\,m$^{-2}$. The average value of $RM$
174: estimated from our data over the entire galaxy is also small: about
175: +15\,rad\,m$^{-2}$. Some reliable measure of foreground $RM$ can be obtained
176: from a polarized background source visible to the SW from NGC\,4254
177: (${\rm RA}_{2000}=12^{\rm h}18^{\rm m}42\fs1$, ${\rm Dec}_{2000}=+14\degr
178: 23\arcmin 07\arcsec$, Fig~\ref{f:rm}a).
179: Within the uncertainties in measurement of about $3\degr$, the source
180: shows no difference in polarization position angle between 8.46 and 4.86\,GHz,
181: thus suggesting a low RM ($< 20$\,rad\,m$^{-2}$). Between 4.86 and 1.43\,GHz we
182: measured the position angle difference of $5.5\degr$, which provides very
183: small $RM$ of 2.5\,rad\,m$^{-2}$. Therefore, we do not apply any correction
184: to the data for the foreground Faraday rotation.
185:
186: \begin{figure*}
187: \includegraphics[width=8.9cm]{f2.eps}
188: \includegraphics[width=8.9cm]{f3.eps}
189: \caption{
190: a) Faraday rotation measure distribution in NGC\,4254 in $15\arcsec$
191: resolution in colors, computed between 8.46\,GHz and 4.86\,GHz with contours
192: of total radio intensity at 4.86\,GHz. The galaxy major axis is marked. b)
193: Faraday depolarization distribution in colors and contours of soft X-ray
194: emission from our XMM-Newton observations.
195: }
196: \label{f:rm}
197: \end{figure*}
198:
199: The $RM$ map for NGC\,4254 is presented in Fig.~\ref{f:rm}a together
200: with contours of total radio emission. Over many galactical regions the
201: absolute values of $RM$ are typically small, of the order of
202: 50-70\,rad\,m$^{-2}$ and exceed 100\,rad\,m$^{-2}$ only within small regions.
203: The variation of $RM$ over the disk can reflect various physical processes
204: (see Sect.~\ref{s:rm}) and local changes of magnetic field
205: strength as well as its orientation with respect to the observer. In general,
206: the western part of the galaxy has negative $RM$ values, while the
207: eastern part has positive ones that are slightly
208: larger in the absolute sense. This demonstrates a substantial coherence of the
209: sign of $RM$ and hence of the magnetic field orientation. Such coherence of
210: magnetic field strongly suggests a global galactic dynamo in action (Beck et al.
211: \cite{beck96}, Widrow \cite{widrow02}).
212:
213: Close to the center of NGC\,4254 there are two sudden jumps of $RM$ oriented
214: at 90$\degr$ to each other (${\rm RA}=12^{\rm h}18^{\rm m}51\fs0$, ${\rm Dec}
215: =+14\degr 24\arcmin 31\arcsec$; and ${\rm RA}=12^{\rm h}18^{\rm m}52\fs0$,
216: ${\rm Dec}=+14\degr 25\arcmin 20\arcsec$). In these places, our map keeps track
217: of $RM$ values up to about $\pm 400\,$rad\,m$^{-2}$. Such jumps usually
218: mean that the internal Faraday rotation angle exceeds $90\degr$ and indicate a
219: Faraday-thick regime (Sokoloff et al. \cite{sokoloff98}). The jumps do not
220: correspond to any feature in the radio or \ion{H}{i} emission. However, the
221: southern jump is close to a CO extension going radially outwards from the
222: nuclear ring (see Fig.~3 in Sofue et al. \cite{sofue}).
223:
224: The region of the SW optical arm where the strongest polarized emission
225: in NGC\,4254 is observed shows a rather small amount of Faraday rotation:
226: around its eastern polarized peak (marked as "a" in Fig.~\ref{f:intro}) the
227: $RM$ is about $-30\pm 7$\,rad\,m$^{-2}$. Around the western peak (marked as "b" in
228: Fig.~\ref{f:intro}) the $RM$ has the opposite sign and a value of about
229: $+20\pm 6$\,rad\,m$^{-2}$. This may indicate a change of magnetic field
230: orientation with respect to the observer. In fact, the two peaks are on both
231: sides of the galaxy's minor axis (see Fig.~\ref{f:rm}a with marked the galaxy's
232: major axis at the position angle of $68\degr$ envisaged from the \ion{H}{i}
233: gas kinematics, Phookun at al. \cite{phookun93}).
234:
235: Faraday rotation effects not only change the orientation of actual
236: magnetic field vectors but also modify (typically reduce) the observed degree of
237: polarization. We measure the wavelength-dependent depolarization $DP$ as the ratio
238: of the degree of polarization ($p$) of nonthermal emission at 4.86\,GHz
239: (6.2\,cm) and 8.46\,GHz (3.5\,cm):
240: \begin{equation}
241: DP=p_{4.86}/p_{8.46}.
242: \end{equation}
243: We separated the nonthermal emission at both frequencies from the total radio
244: intensity with our observations at 1.43\,GHz, 4.86\,GHz, and 8.46\,GHz, assuming
245: a constant nonthermal spectral index $\alpha_{\rm nth}=1.0$ (cf. Paper I).
246: The map of depolarization of NGC\,4254 presented in Figure~\ref{f:rm}b
247: exhibits globally an asymmetric pattern, as in the case of the $RM$ map. The
248: southern and western parts of the galaxy show very little depolarization
249: ($DP>0.9$). This is in agreement with substantially-polarized emission observed here
250: at a low frequency of 1.43\,GHz (Paper I), which is an independent indication of
251: Faraday-thin regime.
252:
253: In the eastern galaxy part depolarization is stronger, with $DP$ values of
254: about 0.6-0.8. The strongest depolarization effects occur at the nucleus,
255: where locally $DP\approx 0.2$, and around $RM$ jumps, where $DP<0.2$. These
256: low $DP$ values must have been produced by a strong Faraday rotation within
257: the emitting region (Sokoloff et al. \cite{sokoloff98} and Sect.~\ref{s:depo}).
258: All those central regions are actually associated with strong H$\alpha$ and
259: hot gas (soft X-ray) emissions. As expected, they also show rather weak
260: polarized intensity at 1.43\,GHz (less than 0.03\,mJy) and a weak degree
261: of polarization ($<1\%$).
262:
263: We attempt to identify main processes to affect polarization in NGC\,4254 by
264: performing a mutual comparison between $DP$, $RM$, and radio thermal emission
265: in 136 beam-independent regions within the galaxy. There is no statistically
266: important global correlation of the thermal gas emission with the $DP$ or
267: $RM$ values: the Pearson correlation coefficient $r$ is around $-0.10$.
268: The relation of $DP$ with $RM$ is also weak ($r=-0.02$). Thus,
269: no single process associated with the galactic thermal gas or the regular
270: magnetic field is capable of accounting for the $DP$ distribution. Therefore, a more
271: detailed analysis in separate regions in the galaxy is needed (see
272: Sect.~\ref{s:depo}).
273:
274: \begin{figure}
275: \includegraphics[width=8.9cm]{f4.eps}
276: \caption{
277: Distribution of rotation measures between 8.46 and 4.86\,GHz along the
278: azimuthal angle in 5 different rings of NGC\,4254. The rings of 1.2\,kpc width
279: start from $30\arcsec$ (2.4\,kpc) distance from the galaxy center (ring 1, the
280: most upper one). The last ring (bottom) ends at distance of 8.5\,kpc. The
281: best-fitted dynamo modes are also shown: for the ring 3 the bisymmetric m=1 mode
282: (dashed) and superposition of axisymmetric and bisymmetric m=0/1 modes
283: (dotted); and for ring 4 the axisymmetric m=0 mode (dashed), see Table~\ref{t:modes}.
284: }
285: \label{f:rmsect}
286: \end{figure}
287:
288: \begin{figure}[t]
289: \includegraphics[width=9.0cm]{f5.eps}
290: \caption{
291: Actual (Faraday-free) structure of magnetic field vectors in NGC\,4254 obtained
292: from the combined VLA and single-dish (Effelsberg) polarized data at 8.46\,GHz
293: and 4.86\,GHz. The lengths of vectors are proportional to the degree of
294: polarization at 4.86\,GHz. The regions with inner and outer magnetic arms are
295: designated. Overlaid (in colors) is the H$\alpha$ image (from Knapen et al. \cite{knapen04}).
296: }
297: \label{f:true}
298: \end{figure}
299:
300: \subsection{Genuine structure of magnetic field}
301: \label{s:genuine}
302:
303: The same sign of $RM$ values over large regions in NGC\,4254 (Sect.~\ref{s:rm},
304: Fig.~\ref{f:rm}a) most probably originates from a dynamo-generated magnetic field.
305: In order to recognize a dominant dynamo mode in action we determined a
306: distribution of $RM$ within the galactic plane along rings around the disk
307: center. The rings of radial width of 1.2\,kpc (corresponding to the beam
308: size of 15\arcsec) were split along azimuthal angle into sectors of $15\degr$ width.
309: For each sector we calculated the average rotation measure and we present the
310: distribution of $RM$ in five rings in Fig.~\ref{f:rmsect}.
311: Then, in the most characteristic rings 3 and 4, encompassing
312: the southern polarized ridge, we fit different periodic variations
313: of $RM$ resulting from dynamo modes and their superpositions. The best fits
314: are presented in Table~\ref{t:modes} and Fig.~\ref{f:rmsect}. The $RM$ distributions
315: partly resemble single-periodic variations, thus the axisymmetric m=0 mode.
316: The fitted phase shifts also correspond well to the observed magnetic pitch
317: angles (about $20\degr$) in the first and the second quadrant of the azimuthal angle.
318: However, a strong disturbance breaks the dynamo symmetry in the third quadrant
319: (the azimuth in the range of $150\degr$ -- $270\degr$). The $RM$ instead of
320: achieving, as predicted by the model, a negative minimum of about
321: $-200\,\mathrm{rad}\,\mathrm{m}^{-2}$ manifests a more complicated pattern
322: involving a local field reversal. This deviation corresponds
323: to a strongly-polarized large interarm region between SW/N optical arms, with
324: a quite smooth magnetic
325: field pattern. Similarly, the bisymmetric (m=1) dynamo mode alone cannot account for
326: the $RM$ distribution as it yields too strong a maximum at azimuth of $210\degr$ and
327: too weak one at $40\degr$. However, adding it as a secondary component to the
328: axisymmetric mode (with roughly half of its amplitude) significantly improves the
329: fits. Higher dynamo modes and other mixed modes do not provide any better fits.
330: Surprisingly, it is also true for a quadrupolar (m=2) mode and a
331: mixture m=0/2 modes, which were suggested to explain the phenomena of
332: magnetic arms in NGC\,6946 (Beck \cite{beck07}, Rohde et al.
333: \cite{rohde99}).
334:
335: \begin{table}[t]
336: \caption[]{Parameters of the single (m=0) double periodic (m=1) cosine
337: functions and their superposition (m=0/1) fitted to the rotation
338: measure variations in the two rings of NGC\,4254 (cf. Fig.~\ref{f:rmsect}).
339: Amplitude, phase shift, background offset, and the Theil~U coefficient of
340: the quality of fit are given.
341: }
342: \begin{center}
343: \begin{tabular}{lrrrr}
344: \hline
345: \hline
346: Dynamo mode & ampl. & ph. shift & offset & Theil U$^a$ \\
347: \hline
348: RING 3\\
349: m=0 & 79 & 25 & 23 & 0.58 \\
350: m=1 & 53 & 43 & 41 & 0.83 \\
351: m=0/1 & 79/51 & 34 & 33 & 0.34 \\
352: \hline
353: RING 4\\
354: m=0 & 82 & 21 & 2 & 0.57 \\
355: m=1 & 48 & 48 & 17 & 0.94 \\
356: m=0/1 & 71/36 & 33 & 9 & 0.36 \\
357: \hline \\
358: \label{t:dynamo}
359: \end{tabular}
360: \end{center}
361: $^a$~The Theil~U coefficient is limited to a range from
362: zero to one (Theil \cite{theil72}). Zero means a perfect fit.
363: \label{t:modes}
364: \end{table}
365:
366: Using the $RM$ and \ion{H}{i} data we can estimate the preferred radial
367: direction of magnetic field, e.g., inward or outward of the galactic center (cf.
368: Krause \& Beck \cite{krause98}).
369: The kinematical properties of the spiral structure in
370: NGC\,4254 can provide information on the sense of the galaxy's rotation. As
371: argued by Phookun at al. (\cite{phookun93}), the spiral arms
372: in NGC\,4254 are generally of trailing type, although a small fraction of stars rotate
373: in reverse. From the \ion{H}{i} velocity field we know that the western
374: part of the galaxy is in approaching motion, thus the southern part of the disk is the
375: nearest side. As both the radial velocity and $RM$ have the same sign on both ends
376: of the galaxy's major axis (in particular, magnetic fields are directed outwards from the
377: observer in the western part), we conclude that the regular (coherent) field in
378: NGC\,4254 is oriented outwards from the disk center, which is contrary to most observed
379: galaxies (Krause \& Beck \cite{krause98}). The outwards directed field was observed to
380: date only in the disk of M\,51. Thus, in both galaxies, outwards fields could possibly
381: arise from the external interaction with a companion.
382:
383: With the distribution of $RM$ available at hand, we can reconstruct the intrinsic
384: position angle of the magnetic field over the whole galaxy, free of Faraday rotation.
385: We assume a simple quadratic relation of $RM$ to wavelength, which should be
386: an appropriate approximation for the high frequency data used. The actual
387: magnetic field structure thus corrected is presented in Fig.~\ref{f:true}
388: with vector lengths proportional to the degree of polarization at
389: 4.86\,GHz. The magnetic structure is smooth and of spiral pattern, even
390: within the galaxy inner $1\arcmin$ region, where the optical (H$\alpha$) spiral
391: structure is unclear.
392:
393: The visible magnetic arms in Fig.~\ref{f:true} are mainly interlaced with
394: optical (H$\alpha$) ones. However, we distinguish `inner' magnetic arms,
395: which are displaced off the optical ones to their inner sides and the
396: `outer arms', which are shifted off outwards, to the downstream side
397: of a density wave. In the next section, we
398: show that both kinds of magnetic arms differ in magnetic properties, which
399: must have been due to some differences regarding their origin. In general,
400: the pattern of magnetic arms in NGC\,4254 apparently follow the structure of
401: local optical features, which indicates
402: a substantial influence of density waves on magnetic pitch angle
403: (see Sect.~\ref{s:orient} and \ref{s:cluster}).
404:
405: \subsection{Magnetic maps}
406: \label{s:magmaps}
407:
408: \begin{figure*} [t]
409: \centering
410: \includegraphics[width=18.5cm]{f6.eps}
411: \caption{Magnetic maps for NGC\,4254 with a resolution of $15\arcsec$ (in colors): total (a), regular (b),
412: random (c) magnetic field strength, and field regularity
413: (d) with contours of \ion{H}{i}, H$\alpha$, infrared (24\,$\mu$m), and UV emission,
414: respectively. The method of the maps construction is given in
415: (Sect.~\ref{s:magmaps}).
416: }
417: \label{f:magmaps}
418: \end{figure*}
419:
420: Currently, magnetic properties of galaxies are most often presented using the field
421: strength estimated in a small number of individual regions within a galaxy or as the average
422: for the whole disk. In the case of complicated magnetic field structures as, e.g.,
423: in the Antennae system of interacting galaxies proper interpretation of polarimetric
424: data requires a comparison of field strength in the regular and random components in
425: many different regions over the entire system (Chy\.zy \& Beck \cite{chyzy04}). The
426: magnetic field in the perturbed spiral NGC\,4254 is quite complicated too,
427: showing unusual polarization features and $RM$ pattern, different kind of magnetic arms
428: in various parts of the galaxy, and thus requires a similar mode of analysis.
429: We, therefore, propose here to present for the first time galactic magnetic field in a form
430: of complete ``magnetic maps'' that show the distributions of strengths of
431: intrinsic magnetic field components in all available regions in the galaxy disk,
432: with the galaxy projection and Faraday rotation effects removed.
433:
434: \begin{table*}[t]
435: \caption[]{
436: The intrinsic magnetic field strength of total $B_{\rm{tot}}$ and regular $B_{\rm{reg}}$
437: components as well as the field regularity in different regions of NGC\,4254,
438: denoted in Fig.~\ref{f:magmaps} (top-right). The given uncertainties express random noise
439: in the maps used. The errors, including possible systematic uncertainties due
440: to applied assumptions, can reach 25\% of the values obtained. The presented
441: magnetic fields strengths are taken directly from magnetic maps described
442: in (Sect.~\ref{s:magmaps}).
443: }
444: \begin{center}
445: \begin{tabular}{lrrr}
446: \hline
447: \hline
448: & $B_{\mathrm{tot}}$ & $B_{\mathrm{reg}}$ & ${B_{\mathrm{reg}}}/{B_{\mathrm{ran}}}$ \\
449: & [$\mu$G] & [$\mu$G] & \\
450: \hline
451: Total & $16 \pm 1$ & $7 \pm 0.2$ & n.a. \\
452: 1) Southern outer magnetic arm & $19.7 \pm 0.5$ & $12.6 \pm 0.3$ & $0.83 \pm 0.03$ \\
453: 2) Western outer magnetic arm & $18.9 \pm 0.8$ & $12.3 \pm 0.4$ & $0.86 \pm 0.09$ \\
454: 3) Northern inner magnetic arm & $16.9 \pm 0.2$ & $ 7.8 \pm 0.3$ & $0.52 \pm 0.02$ \\
455: 4) Central part & $24.3 \pm 0.5$ & $ 4.8 \pm 0.7$ & $0.20 \pm 0.03$ \\
456: 5) SW optical spiral arm & $22.2 \pm 0.3$ & $ 7.4 \pm 0.2$ & $0.35 \pm 0.02$ \\
457: 6) NE optical spiral arm & $20.3 \pm 0.7$ & $ 6.5 \pm 0.8$ & $0.34 \pm 0.05$ \\
458: \hline\\
459: \label{t:mf}
460: \end{tabular}
461: \end{center}
462: \end{table*}
463:
464: The strength of {\em intrinsic} regular
465: component of magnetic field $B_{\rm{reg}}$, corrected for galaxy
466: inclination and varying angle between the line of sight and the field line, is
467: derived from the observed radio polarized intensity. We assume that the regular
468: magnetic field lies only in the galaxy plane, which permits derivation
469: of the intrinsic (in the disk plane) magnetic field position angle from the
470: derived sky-projected (but Faraday rotation free) polarization position angle
471: (Sect.~\ref{s:rm}) and the galaxy orientation. Following Phookun et al.
472: (\cite{phookun93}) we adopt values of $42\degr$ and $68\degr$, respectively,
473: for the galaxy inclination and the major axis position angle. The intrinsic
474: (de-projected) total magnetic field strength $B_{\rm{tot}}$ is calculated
475: through a similar procedure using, in this case, the total nonthermal
476: (synchrotron) intensity. The random magnetic field component $B_{\rm{ran}}$
477: is derived by subtracting the reconstructed intrinsic regular field from the
478: total field.
479:
480: In our calculations, we apply the energy ratio $k=100$ of relativistic
481: protons and electrons, and a 300\,MeV cutoff in the cosmic rays (CR)
482: proton spectrum (e.g., Ehle \& Beck \cite{ehle93}; Chy\.zy \& Beck
483: \cite{chyzy04}). We assume the equipartition energy condition between CRs
484: and magnetic fields and apply a constant nonthermal spectral index
485: $\alpha_{\rm nth}=1.0$ (c.f. Paper I). The magnetic field is supposed to be
486: located within the plane of the galaxy disk (without a vertical
487: component) of about 1\,kpc unprojected thickness. The whole
488: described procedure of magnetic maps' derivation was implemented as a task in the
489: AIPS\footnote{Astronomical Imaging Processing System of the National
490: Radio Astronomy Observatory}. The largest uncertainties in the obtained
491: magnetic field strengths are most likely to be the systematic ones due to
492: the assumed above values for the unknown parameters, rather than random
493: errors due to uncertainties in observed total and polarized radio
494: intensities (see below).
495:
496: The first magnetic map of NGC\,4254 presented in Fig.~\ref{f:magmaps}a
497: shows the distribution of {\em total} magnetic field strength with contours of
498: \ion{H}{i} intensity for reference sake. In the central disk region the strength reaches
499: 25\,$\mu$G, while decreasing to about 10\,$\mu$G at the outer galaxy parts (see
500: also Table~\ref{t:mf}). The strong localized signal in the disk northern
501: periphery (RA=$12^{\rm h}18^{\rm m}46^{\rm s}$,
502: Dec=$14\degr26\arcmin 45\arcsec$) coming from a confusing
503: background source is masked in this and following magnetic maps.
504: The mean strength of total magnetic field over the
505: whole galaxy (without the background source) is $16\pm1$~$\mu$G, which is in agreement with the
506: low-resolution estimation by Soida et al. (\cite{soida96}). This value is
507: higher than typical ones (of about $10\,\mu$G) found in a large sample of
508: galaxies (Beck et al. \cite{beck96}) but closer to the mean field strength
509: in the disks of the Antennae galaxies (about $20\,\mu$G, Chy\.zy \& Beck
510: \cite{chyzy04}). As can be seen from Fig.~\ref{f:magmaps}a (and
511: Table~\ref{t:mf}), the sites of star formation
512: and effective production of CRs in optical spiral arms possess a strong magnetic
513: field of about $16-22\,\mu{\rm G}$. In the interarm regions, the total field
514: weakens typically to $14-17\,\mu{\rm G}$. We mention that in some analyses of
515: galaxies of Hubble type similar to NGC\,4254 a disk thickness of 0.5\,kpc is
516: assumed. If, by way of comparison, this value is applied to NGC\,4254, the
517: corresponding total magnetic field in the optical arms rises to
518: $19-25\,\mu{\rm G}$ and to $30\,\mu{\rm G}$ in the disk core.
519:
520: The second magnetic map of intrinsic {\em regular} field is presented
521: in Fig.~\ref{f:magmaps}b with contours of H$\alpha$ emission. The
522: typical strength of regular field in the N and NE optical arms is about
523: $5\,\mu$G. In the inner magnetic arm (region 3) it rises up to
524: $7.5-8\,\mu$G. In the SW optical spiral arm, the regular field reaches
525: $6-7\,\mu$G, while outside it, in the southern outer magnetic arm (in
526: the strongly-polarized ridge, region 1) it is enhanced up to $13\,\mu$G. The large
527: interarm region between SW and N optical arms maintains over a large area, a strong
528: regular field of about $10\,\mu$G, stronger than in the northern inner magnetic
529: arm. In the western outer magnetic arm (region 2), it reaches a value even as
530: strong as the regular field in the southern polarized ridge: $12-13\,\mu$G.
531:
532: The distribution of intrinsic regular field shows a global asymmetry in the
533: SW-NE direction that was partly hidden due to projection effects in the
534: distribution of polarized intensity, which shows just a strong N-S asymmetry
535: (Fig.~\ref{f:intro}, Paper I). The strong regular field, manifested not only in
536: the southern galactic outskirts, but also in the western outer magnetic arm
537: (region 2) {\em inside} the galactic disk, must have profound consequences.
538: This introduces a question of the origin of the field in both the regions, and
539: their possible connection.
540:
541: A map of the turbulent component of magnetic field
542: (Fig.~\ref{f:magmaps}c) shows an excellent
543: correlation with the dust emission (in contours)
544: observed in mid-infrared (MIR) by Spitzer (see Paper I). This
545: can indicate a strong production of turbulent field or a strong tangling
546: (de Avillez \& Breitschwerdt \cite{deavillez05}) in
547: sites of star-forming regions revealed without extinction by the MIR
548: emission. Surprisingly, in the place of strongly-increased regular field
549: in the southern outer magnetic arm no such
550: enhancement is visible in the random field. We model this behavior of
551: different magnetic field components in Sect.~\ref{s:ani}.
552:
553: The ratio of derived (projection free) regular to random magnetic field
554: components -- the field regularity ${B_{\mathrm{reg}}}/{B_{\mathrm{ran}}}$
555: -- is a useful measure of the net
556: production of regular field independent of magnetic field strength,
557: as well as a sensitive tracer of local processes in magnetized plasma. We
558: present the distribution of field regularity as yet another ``magnetic
559: map'' in Fig.~\ref{f:magmaps}d, together with the contours
560: of UV emission associated with population of young stars and the regions of
561: most turbulent ISM. In fact, along the three main optical spiral arms and in
562: the galaxy center the field regularity is low ($0.2-0.5$). Both the magnetic outer
563: arms (region 1 and 2) have an exceptionally high-field regularity of about
564: 0.8, in contrast to a typical value of 0.5 in other interarm regions
565: (including region 3). As it could be expected, the regularity is also larger when
566: going outside turbulent regions into the galaxy outskirts. Hence, such a
567: highly-regular magnetic field in the galactic halo is likely to enter the
568: Virgo Cluster's medium.
569:
570: With the magnetic maps at hand, we also looked at where the magnetic energy
571: dominates the total energy in the thermal gas component. We derived the
572: distribution of the latter quantity from the emission measure estimated from
573: radio thermal emission (Paper I). The magnetic energy exceeds
574: the thermal one across almost the entire galaxy (except for just a few regions
575: of very intensive star formation). Moreover, it exceeds the turbulent gas energy
576: (estimated from the dispersion in the \ion{H}{i} velocity field, Paper I) in
577: the southern polarized ridge, in the large western interarm region,
578: and in the whole polarized galactic outskirts visible at 4.86\,GHz and
579: 1.4\,GHz. Thus, in those regions the magnetic field can be even dynamically
580: important (as in NGC\,6946; Beck \cite{beck07}).
581:
582: In the above estimations of the magnetic field strength, we assumed a constant
583: synchrotron spectral index $\alpha_{\rm nth}=1.0$ across the galaxy, while
584: it may attain smaller values in star-forming regions and larger ones in the
585: galaxy's outskirts. Such variations, if known, could easily be incorporated
586: in the procedures calculating the magnetic maps. Fortunately, the strength of
587: the magnetic field only weakly depends on $\alpha_{\rm nth}$. For example,
588: we performed an additional separation of thermal and nonthermal radio emission
589: assuming $\alpha_{\rm nth}=0.8$, which is more suitable for spiral arms. Then,
590: we derived the magnetic maps with the new $\alpha_{\rm nth}$ and compared
591: them with the previous ones. In spiral arms and in the galactic centre the
592: differences are up to about $2\,\mu$G. In particular, in the optically-bright
593: region in the NE spiral arm at RA=$12^{\rm h}18^{\rm m}53^{\rm s}$,
594: Dec=$14\degr25\arcmin 51\arcsec$ the total field strength is smaller by
595: $1.3\,\mu$G, the regular one by $0.58\,\mu$G, and the random one by $1.4\,\mu$G,
596: giving the field regularity of 0.22, instead of 0.30 before.
597: Due to energy losses of the CR electrons, the field strength can be
598: underestimated in the outer part of the galaxy due to a higher
599: ratio $k$ of proton to electron energy. We estimated that effect by setting
600: $k$ to 200. Differences to the original maps are in this case up to about $3\,\mu$G.
601: In the tip of western outer magnetic arm at
602: RA=$12^{\rm h}18^{\rm m}44^{\rm s}$, Dec=$14\degr26\arcmin 21\arcsec$
603: the total field is stronger now by $2.4\,\mu$G, the regular one by $2.1\,\mu$G,
604: and the random one by $0.6\,\mu$G. The field regularity rose from 1.25
605: to 1.41. Such variations of parameters across NGC\,4254, as discussed above,
606: would not change conclusions of this paper.
607:
608: \section{Discussion}
609: \label{s:discussion}
610:
611: \begin{figure*}
612: \includegraphics[width=18.5cm]{f7.eps}
613: \caption{A phase diagram of the regular component of magnetic fields in
614: dependence of natural logarithm of distance and galactocentric azimuthal
615: angle, as measured from the northern tip of the major axis counterclockwise.
616: The vectors are proportional to the polarized intensity at 4.86\,GHz, which is
617: also presented in contours. In the frame of this representation, a logarithmic
618: spiral arm would be a straight line inclined by the pitch angle. The galactic
619: spiral structure is shown by the distribution of UV emission (in color) with
620: additionally marked (the straight lines) approximate orientations of the main
621: three optical spiral arms. All the presented values were averaged in sectors of
622: $7\degr$ width in the azimuth.
623: }
624: \label{f:orient}
625: \end{figure*}
626:
627: The intrinsic magnetic field vectors in NGC\,4254 seem to follow the perturbed
628: optical spiral pattern (Fig~\ref{f:true}). This suggests that apart from the
629: dynamo action, an additional influence of gas flows is involved.
630: The $RM$ distribution also reveals a perturbed axisymmetric dynamo mode or
631: a mixture of two modes and an unusual outward directed orientation
632: of the magnetic field (Sect~\ref{s:genuine}).
633: According to the magnetic maps, the total magnetic field is stronger
634: in NGC\,4254 than in a typical field galaxy (Fig.~\ref{f:magmaps}, Table~\ref{t:mf}).
635: Such higher magnetic activity can be associated with the high SFR
636: manifested by this galaxy (Paper I). However, several magnetic arms with strong
637: regular fields located on the downstream and upstream sides of local density waves
638: cannot be readily accounted for and different physical processes
639: could be involved. The strong regular field (of about $12\,\mu$G) revealed
640: not only in
641: the southern galactic outskirts, but also in the western outer magnetic arm,
642: well inside the galactic disk, suggests a possibility that both the magnetic
643: structures could have arisen from the same phenomena.
644: To address all these problems, below we analyze the
645: correspondence of the orientation of the magnetic field with the optical spiral arms
646: and the connection of the regular field with star-forming regions.
647: Subsequently, we model depolarization and magnetic field components, according to
648: different physical phenomena.
649:
650: \subsection{Orientations of magnetic field vectors}
651: \label{s:orient}
652:
653: We can explore the relation between the magnetic and perturbed optical spiral
654: arm pattern in NGC\,4254 by investigating the orientation of magnetic field vectors
655: with respect to the main optical spiral arms. We constructed a
656: phase diagram of orientations of such vectors, $RM$--corrected and de-projected,
657: as a function of radius $R$ (in natural logarithm) and galactocentric
658: azimuthal angle.
659:
660: Throughout large galaxy domains the magnetic field vectors show a similar pitch
661: angle (Fig.~\ref{f:orient}). In the inner disk part ($R<50\arcsec$) two magnetic patterns
662: are observed. The dominant one has a coherent large pitch angle ($>25\degr$), regardless
663: of the chaotic pattern of H$\alpha$ and UV emitting gas.
664: This pattern is broken in two regions (around the azimuthal angle of
665: $150\degr$ and $340\degr$) with almost azimuthal magnetic field. Small pitch angles
666: could be associated with the central bar-like structure seen in high-resolution
667: ($7\farcs 5$) radio maps (Paper I). Indeed, the position angle of the bar visible
668: in high-resolution radio observations (see Fig.~3 in Paper I)
669: is $55\degr$ giving the azimuthal angles of the bar ends in the regions of low
670: values of the magnetic pitch angle. A similar configuration of magnetic field (with azimuthal
671: field close to the bar ends) is visible in the strongly-barred galaxy NGC\,1365
672: (Moss et al. \cite{moss07}). However, the performed MHD simulation was unable to
673: reproduce this magnetic pattern in detail (the mean difference between
674: observed and simulated pitch angles is $30\degr$). In order to fully explain the
675: magnetic field within the inner disk of NGC\,4254, an advanced MHD simulation of a weakly
676: barred galaxy is needed.
677:
678: In the southern polarized ridge (radius
679: $60\arcsec-100\arcsec$ and azimuth $45\degr-150\degr$), the regular magnetic
680: field changes its pitch angle and the polarized emission is shifted outside the
681: local UV-maxima. The magnetic vectors seem to be aligned with the optical (and UV)
682: SW spiral arm, which has an oscillating pattern around the pitch angle of about $20\degr$
683: (marked in Fig.~\ref{f:orient}).
684:
685: At larger azimuth ($>180\degr$), the N and NE optical spiral arms have
686: different average pitch angle than the SW arm ($30\degr$ and $25\degr$, respectively),
687: providing further evidence for the perturbed spiral pattern of NGC\,4254.
688: The N optical arm is heavily curved at large distances ($r>30\arcsec$)
689: and its pitch angle is reduced to zero. The $\vec{B}$-vectors
690: trace the orientation of both the optical arms and deviate from them only slightly
691: (up to $20\degr$), even at the tip of the N arm. The polarized maxima are shifted
692: from the optical spiral arms, especially on the downstream side of the N arm. In this
693: region (between SW and N arms), the magnetic pitch angle reaches a locally large value
694: of $50\degr$. We showed in Sect.~\ref{s:magmaps} that in this part of the galaxy
695: the energy density of the ISM can be dominated by the magnetic component, thus probably
696: allowing the magnetic field to form its own configuration.
697:
698: The following of the perturbed optical spiral arms by the vectors of
699: the regular magnetic field suggests a physical association of both patterns.
700: Similar correlations has been observed for some grand-design spirals,
701: e.g., in M\,51 (Fletcher et al. \cite{fletcher04b}; Patrikeev et al.
702: \cite{patrikeev06}). The alignment of $\vec{B}$-vectors with CO pitch angles in M\,51 was
703: suggested to results from the shock compression along the upstream side of
704: spiral density waves. The northern part of NGC\,4254 is much more
705: flocculent than the spiral structure in M\,51, which implies that the
706: suggested interaction must still hold for weak density waves and for a
707: perturbed disk. However, the downstream magnetic arms in NGC\,4254 cannot be
708: explained by this process.
709:
710: Locally, large magnetic pitch angle correlated with orientation of spiral
711: arms in NGC\,4254 is not attainable in the classical MHD dynamo model (Elstner
712: et al. \cite{elstner00}). Relatively-large, magnetic pitch angles (up to
713: $40\degr$) could still be obtained in MHD dynamo by allowing an increase in the
714: correlation time of interstellar turbulence (Rohde \& Elstner \cite{rohde97};
715: Rohde et al. \cite{rohde99}). However, the magnetic pitch angle was found
716: in this case not to depend on the pitch angle of the gaseous arms.
717: Furthermore, this approach was criticized by Shukurov (\cite{shukurov05}),
718: who argued for shorter correlation time in spiral arms, contrary to the authors above.
719: MHD simulations by Elstner et al. (\cite{elstner00}), which include an enhancement
720: of turbulent diffusion in spiral arms and density wave-alike gas velocity
721: patterns, produced the influence of gas flows on dynamo-generated magnetic fields.
722: However, the generated magnetic pitch angle in the
723: interarm regions are smaller than within spiral arms. This is not actually
724: observed in NGC\,4254: in the large interarm region between SW and N
725: optical arms, the pitch angle is the largest in the whole galaxy (up to $50\degr$).
726: Further investigation by MHD modeling to show whether gas flows or
727: high turbulent diffusion are able to maintain such a large pitch angle
728: in NGC\,4254 is highly needed.
729:
730: Addressing the problem of observed phase shifts between the magnetic and
731: optical arms in NGC\,4254, we mention the work of Shukurov (\cite{shukurov98}),
732: who suggested that a similar shift observed in NGC\,6946 can result from a
733: certain time lag between an enhancement in turbulence within density waves and
734: the response in dynamo coefficients. This cannot account for the outer
735: magnetic arms in NGC\,4254; the lag in these cases would be negative.
736: There is another possibility given by Shukurov explaining
737: magnetic arms in NGC\,6946 as a suppression of dynamo efficiency in the
738: gaseous arms by an enhanced turbulent
739: magnetic diffusivity. However, such a shift becomes smaller along the SW arm
740: and disappears in the western part of the arm (azimuth $>140\degr$), contrary
741: to the expectations, since a larger distance from the galaxy center should result in
742: a smaller dynamo number and a higher shift (as observed, e.g., in M\,51, Fletcher
743: et al. \cite{fletcher04b}). Hence, the magnetic arms of NGC\,4254 and the large
744: magnetic pitch angle indicate a strong influence of gaseous flows on
745: the dynamo-induced magnetic fields, and possibly
746: some external cause, such as compression or shear, at work (see below).
747:
748: \subsection{Magnetic field regulation by SFR}
749: \label{s:sfr}
750:
751: \begin{figure*}
752: \centering
753: \includegraphics[width=9.1cm]{f8.eps}
754: \includegraphics[width=9.1cm]{f9.eps}
755: \includegraphics[width=9.1cm]{f10.eps}
756: \includegraphics[width=9.1cm]{f11.eps}
757: \caption{
758: Total magnetic field strength (a), field regularity (b), random
759: (c), and regular (d) magnetic field strengths versus
760: IR-based SFR within NGC\,4254. From among all regions (marked by
761: diamond sign) magnetic arms are denoted separately: southern outer
762: magnetic arm by crosses; western outer magnetic arm by
763: rectangles; and northern inner magnetic arm by triangles.
764: The lines fitted by bisector method to all regions are drawn (except
765: d-panel, see Sect.~\ref{s:sfr}).
766: }
767: \label{f:btsfr}
768: \end{figure*}
769:
770: Magnetic maps enable analyzing the correlation between the strength of
771: various components of the magnetic field and the star formation activity over
772: different regions of
773: NGC\,4254. In Fig.~\ref{f:btsfr}a, we present such a relation as based on
774: 24\,$\mu$m SFR (cf. Paper I) and the magnetic map of total field strength
775: (Fig.~\ref{f:magmaps}a). The observed relation for 277 beam independent
776: ($15\arcsec$ spaced) regions over the galaxy is a single power-law
777: throughout almost three orders of magnitude in SFR. The relation is tight,
778: with the Pearson correlation coefficient of 0.93, and has the
779: form fitted by the bisector method:
780: \begin{equation}
781: \log B_{\rm{tot}}=0.18\,(\pm 0.01) \log {\rm SFR}+1.56\,(\pm 0.01).
782: \end{equation}
783: This relation is also another example of the well-known radio/IR
784: correlation discussed in Paper I and reflects a process of scaling the
785: production of magnetic field and CRs with star formation activity.
786:
787: We also separated regions of S, W, and N magnetic arms and marked them all
788: in Fig.~\ref{f:btsfr}. We repeated the fitting procedures for them. The
789: resulting slopes $a$, offsets $b$, correlation coefficients $r$,
790: and number of used regions $N$ are given in Table~\ref{t:slopes}. The
791: regions of outer magnetic maps have relatively-high total field strength
792: as compared to the inner N magnetic arm.
793:
794: The intensive star formation is generally considered to produce an increased
795: level of turbulence and to influence the regular magnetic field. However,
796: this process has not been quantitatively investigated to date. Now, for the
797: first time, this can be done with the constructed magnetic maps for
798: NGC\,4254. The relation between SFR and the regular field's strength scaled
799: by the random field -- i.e., the field regularity -- is presented in
800: Fig.~\ref{f:btsfr}b for 234 independent galaxy regions. The relation is well
801: described by a power-law that we estimated by the bisector method, which
802: yields:
803: \begin{equation}
804: \log \frac{\mathrm{B_{reg}}}{\mathrm{B_{ran}}}=-0.32\,(\pm 0.01) \log {\rm SFR}-0.90\,(\pm0.03).
805: \end{equation}
806: It shows that contrary to the total magnetic field strength, the field
807: regularity strongly decreases with the star-formation level: the observed
808: anticorrelation is $-0.71$. Several processes can explain the observed
809: trend: 1) field tangling by some processes related to the star formation;
810: 2) production of turbulent fields proportional to SFR (by a turbulent dynamo
811: or any other turbulent field amplification mechanisms); and 3) production of
812: regular field anti-proportional to SFR. The latter can be achieved with the
813: mean-field dynamo effects anti-proportional to SFR, e.g., by the correlation
814: length of turbulence anti-proportional to SFR, as proposed by Rohde et al.
815: (\cite{rohde99}), or by the turbulent diffusivity enhanced by shear, as
816: proposed by Moss et al. (\cite{moss01}) and Moss et al. (\cite{moss07}),
817: or by the suppression of mean-field dynamo action in spiral arms
818: (Shukurov \cite{shukurov05}). As all the above concepts give
819: different predictions for scaling of regular and random magnetic field
820: with SFR, we present separately the strength of both the magnetic components
821: versus SFR in Figs.~\ref{f:btsfr}c-d and the fitted power-laws in
822: Table~\ref{t:slopes}. Strikingly enough, there is a strong correlation of
823: random field $B_{\rm{ran}}$ with SFR ($r=0.91$ for all regions) and no such
824: correlation for regular field ($r=0.08$). This suggests that the effective
825: production of random magnetic field in the star-forming regions (i.e., the
826: second possibly underlying process from those listed above) is the principal
827: cause of the strong anticorrelation of field regularity and SFR
828: (Fig.~\ref{f:btsfr}b).
829:
830: However, the $B_{\rm{reg}}-$SFR relation shows a two-way behavior
831: (Fig.~\ref{f:btsfr}d) for stronger regular fields with an upper envelope
832: slightly raising for weak SFR, and declining for the large SFR. The dividing
833: value is $\mathrm{SFR}\approx 0.016\,\mathrm{M}_
834: {\sun}\,\mathrm{yr}^{-1}\,\mathrm{kpc}^{-2}$, which corresponds to
835: $B_{\rm ran}\approx 14\,\mu$G and to the maximum of $B_{\rm reg}$ of about
836: $12.5\,\mu$G. This two-way trend is apparent in the data in
837: Table~\ref{t:slopes} as positive and negative correlation coefficients for
838: western and southern magnetic arms, respectively, and an insignificant
839: correlation when all galactic regions are taken together.
840: Hence, the production of regular field in NGC\,4254
841: could have been suppressed {\em in the most vivid star-forming regions} (process 3)
842: or the regular field could have been disrupted by an efficient field tangling
843: (process 1 above). Further analyses of magnetic structures in other galaxies
844: and advanced simulations of MHD dynamo involving various SFRs are needed to
845: fully account for these relations.
846:
847: \begin{table}
848: \caption[]{The fitted relations between total $B_{\rm{tot}}$,
849: random $B_{\rm{ran}}$, and regular $B_{\rm{reg}}$ magnetic
850: field strength, and the field regularity
851: ${B_{\mathrm{reg}}}/{B_{\mathrm{ran}}}$, versus star-formation
852: rate SFR for all regions of NGC\,4254 and for different magnetic arms.
853: }
854: \label{t:slopes}
855: \centering
856: \begin{tabular}{lllll}
857: \hline\hline
858: regions & $a$ & $b$ & $r$ & $N$ \\
859: \hline
860: &\multicolumn{2}{l}{$\log B_{\rm{tot}}=a
861: \log \mathrm{SFR} + b$}\\
862: all & $0.18\pm0.01$ & $1.56\pm0.01$ & $+0.93$ & 277\\
863: S arm & $0.13\pm0.01$ & $1.50\pm0.02$ & $+0.89$ & 25\\
864: W arm & $0.17\pm0.01$ & $1.59\pm0.02$ & $+0.98$ & 16\\
865: N arm & $0.19\pm0.03$ & $1.56\pm0.05$ & $+0.72$ & 31\\
866: \hline
867: &\multicolumn{2}{l}{$\log B_{\rm{ran}}=a
868: \log \mathrm{SFR} + b$}\\
869: all & $0.26\pm0.01$ & $1.64\pm0.02$ & $0.91$ & 234\\
870: S arm & $0.22\pm0.02$ & $1.57\pm0.03$ & $0.91$ & 25\\
871: W arm & $0.30\pm0.03$ & $1.73\pm0.06$ & $0.95$ & 16\\
872: N arm & $0.20\pm0.03$ & $1.55\pm0.05$ & $0.72$ & 31\\
873: \hline
874: &\multicolumn{2}{l}{$\log B_{\rm{reg}}=a
875: \log \mathrm{SFR} + b$}\\
876: all &\multicolumn{2}{c}{two-way trend} & $0.08$ & 234\\
877: S arm & $-0.28\pm0.04$ & $0.52\pm0.07$ & $-0.49$ & 25\\
878: W arm & $0.17\pm0.07$ & $1.43\pm0.15$ & $0.46$ & 16\\
879: N arm & $0.47\pm0.16$ & $1.69\pm0.05$ & $0.28$ & 31\\
880: \hline
881: &\multicolumn{2}{l}{$\log \frac{B_{\rm{reg}}}{B_{\rm{ran}}}=a
882: \log \mathrm{SFR} + b$}\\
883: all & $-0.32\pm0.01$ & $-0.90\pm0.03$ & $-0.71$ & 234\\
884: S arm & $-0.42\pm0.05$ & $-0.90\pm0.08$ & $-0.78$ & 25\\
885: W arm & $-0.30\pm0.04$ & $-0.71\pm0.11$ & $-0.75$ & 16\\
886: N arm & $-0.56\pm0.21$ & $-1.41\pm0.39$ & $-0.21$ & 31\\
887: \hline
888: \end{tabular}
889: \end{table}
890: The regions of S and W outer magnetic arms reveal a distinct behavior from
891: the other galactical regions. Their regular field strength reaches the highest
892: attainable values, which places them on the top of $B_{\rm{reg}}-$SFR relation
893: (Fig.~\ref{f:btsfr}d). In contrast, their random field is average,
894: situating them close to the fitted power-law for all the galactic regions
895: (Fig.~\ref{f:btsfr}c). Thus, the stronger total field for
896: the outer magnetic maps mentioned above is due to the stronger regular field and not the
897: random one.
898:
899: This behavior is well rendered by the field regularity data. On the field
900: regularity-SFR diagram (Fig.~\ref{f:btsfr}b) the outer magnetic arms approach
901: large values at the upper envelope of the relation. In contrast, the regions
902: of the inner N magnetic arm are distributed mostly below the average trend
903: determined for all the galaxy regions (the fitted line in
904: Fig.~\ref{f:btsfr}b). This can indicate that the regular field in the outer
905: magnetic arms is probably not fully controlled by the star formation, as for
906: instance in the southern outer magnetic arm, which shows a high-field regularity
907: and does not avoid the high-SFR regions. Apparently, some additional process
908: to regularize magnetic fields or a more efficient production of regular field
909: component must be at work within the outer magnetic arms (which we explore
910: in Sect.~\ref{s:ani}).
911:
912: We also investigate how the applied assumption of constant nonthermal
913: spectral index $\alpha_{\rm nth}$ in derivation of magnetic maps
914: in Sect.~\ref{s:magmaps} could influence the discussed above relations.
915: We use from Sect.~\ref{s:magmaps} a rough estimation of possible
916: changes of the total magnetic field in the NE spiral arm and in the
917: tip of the western outer magnetic arm.
918: We predict a possible slight decrease of the slope in $B_{\rm{tot}}-$SFR
919: relation which would, however, give a power-law fit within the current
920: spread of points in Figure~\ref{f:btsfr}a. Similar considerations predict
921: a possible small decrease of the slope of $B_{\rm{ran}}-$SFR relation
922: (Fig.~\ref{f:btsfr}c) and an increase of the anticorrelation in the field
923: regularity-SFR relation (Fig.~\ref{f:btsfr}b), again within the spread of
924: the data points. The variations of $\alpha_{\rm nth}$
925: across the disk would induce an increase of regular field in galaxy
926: outskirts and a decrease in star-forming regions, which we suspect could
927: lead to a slight predominance of anticorrelation in the $B_{\rm{reg}}-$SFR
928: relation (Fig.~\ref{f:btsfr}d).
929:
930: \begin{table*}
931: \caption[]{
932: Observed and modeled depolarization of synchrotron emission
933: in several regions of NGC\,4254 due to differential Faraday rotation, Faraday
934: dispersion, and gradients in $RM$. The regions are marked in Fig.~\ref{f:magmaps}b.
935: }
936: \begin{center}
937: \begin{tabular}{lrrrr}
938: \hline
939: \hline
940: & obs. $DP$(3.6/6.3) & Diff. FR & Far. Disp. & Grad. $RM$ \\
941: \hline
942: 1) Southern outer magnetic arm & $1.08\pm0.02$ & $ 0.99\pm0.01$ & $0.99\pm0.01$ & $0.99\pm0.01$ \\
943: 2) Western outer magnetic arm & $1.14\pm0.14$ & $ 0.97\pm0.02$ & $1.00\pm0.01$ & $0.98\pm0.01$ \\
944: 3) Northern inner magnetic arm & $1.04\pm0.17$ & $ 0.99\pm0.01$ & $1.00\pm0.01$ & $0.98\pm0.01$ \\
945: 4) Central part & $0.57\pm0.18$ & $ 0.66\pm0.28$ & $0.70\pm0.04$ & $0.91\pm0.11$ \\
946: 5) SW optical spiral arm & $0.89\pm0.04$ & $ 0.99\pm0.01$ & $0.86\pm0.06$ & $0.99\pm0.01$ \\
947: 6) NE optical spiral arm & $0.72\pm0.15$ & $ 0.91\pm0.07$ & $0.89\pm0.15$ & $0.94\pm0.04$ \\
948: \hline \\
949: \label{t:depo}
950: \end{tabular}
951: \end{center}
952: \end{table*}
953:
954: \subsection{Modeling depolarization}
955: \label{s:depo}
956:
957: To establish the origin of magnetic field in NGC\,4254, and the processes that
958: underlie the observed regularity of magnetic field, it is necessary to understand
959: the depolarization processes that influence the observed properties of polarized
960: emission. In this section, we model various frequency-dependent
961: depolarization effects in NGC\,4254 by constructing distributions (maps) of
962: expected depolarization and comparing them with observed distribution (Fig.~\ref{f:rm}b).
963:
964: The line-of-sight component of coherent magnetic field embedded
965: within the synchrotron emitting region and mixed with ionized gas can
966: cause the differential Faraday rotation (Burn \cite{burn66}). To obtain the
967: modeled values of the depolarization in NGC\,4254, we use the formula
968: (Sokoloff et al. \cite{sokoloff98}):
969: \begin{equation}
970: DP_{dFR}=\left| \frac{\lambda^2_2}{\lambda^2_1}
971: \frac{\sin(2 RM \lambda^2_{1})}{\sin(2 RM \lambda^2_{2})}\right|
972: \label{e:dfr}
973: \end{equation}
974: where $\lambda_1=0.062$\,m, $\lambda_2=0.035$\,m, and the rotation measure $RM$
975: values we take from the constructed $RM$ map (Fig.~\ref{f:rm}a).
976:
977: The modeled depolarization, due to this effect, and observed depolarization are
978: presented in Table~\ref{t:depo} for some characteristic places within NGC\,4254.
979: Uncertainties of depolarization values estimated from their standard deviations in
980: the measured regions are also given in Table~\ref{t:depo}. Throughout almost
981: the entire galaxy the calculated depolarization is weak, $DP_{dFR}$ values are
982: close to about 0.95. A larger modeled depolarization (with large statistical
983: variations) appears in the central disk region, where $DP_{dFR} \approx 0.7\pm0.3$,
984: and in the regions of $RM$ jumps, where
985: $DP_{dFR}<0.2$, and where observed $DP$ are also similarly low. However, variations
986: of depolarization are large in those regions.
987: The differential Faraday rotation does not account for the observed depolarization
988: in the most NE disk portion (around 0.7, Fig.~\ref{f:rm}b), as it typically gives
989: values around 0.9.
990:
991: As the second process of depolarization, we consider the internal Faraday
992: dispersion produced by a turbulent magnetic field and thermal gas within
993: the emitting volume (Burn \cite{burn66}; Sokoloff et al. \cite{sokoloff98}).
994: The fractional depolarization resulting from this process is given by:
995: \begin{equation}
996: DP_{disp}=\frac{\lambda^4_{2}}
997: {\lambda^4_{1}}
998: \frac{1-\exp(-2\sigma^2_{RM} \lambda^4_{1})}
999: {1-\exp(-2 \sigma^2_{RM} \lambda^4_{2})}.
1000: \end{equation}
1001: The dispersion of rotation measure $\sigma_{RM}$ can be described using
1002: the ``random walk'' approach:
1003:
1004: \begin{equation}
1005: \sigma_{RM}=0.812 B_r n_c d \sqrt{N}
1006: \end{equation}
1007: where $B_r$ is the component of turbulent magnetic field in the direction of the
1008: observer, $d$ is the turbulent cell/cloud size (correlation length) of the
1009: random magnetic field, $n_c$ is the electron density within the cloud, and
1010: $N$ is the cloud number along the line of sight. If the ionized diffuse gas
1011: has an extent $L$ toward the observer, and has a filling factor $f$, then its
1012: mean electron density $<n>=f n_c$ and $N=L f d^{-1}$.
1013: Similar results (with $\sigma_{RM}$ larger by up to two in Eq. 6) are
1014: obtained using different approach (Beck et al. \cite{beck03}).
1015:
1016: The magnetic maps of random magnetic field strength allow us to derive a map
1017: of its line-of-sight component $B_r=B_{\rm ran}/\sqrt{3}$.
1018: Taking typical values of parameters: $d=50$\,pc, $L=1$\,kpc, and $f=0.5$
1019: (cf. Beck \cite{beck07}),
1020: we can adjust the only free parameter -- the cloud density $n_c$ -- to get
1021: the best correspondence of the modeled $DP_{disp}$ with the actual map of
1022: observed depolarization within the spiral arms, where the Faraday dispersion
1023: is expected to be most pronounced. We find the best value
1024: $n_c=0.08$\,cm$^{-3}$ which yields $<n>=0.04$\,cm$^{-3}$, close to typical values used
1025: (e.g. $<n>=0.03$\,cm$^{-3}$ for NGC\,6946; Beck \cite{beck07}). Within these numbers Faraday
1026: dispersion can almost fully explain the actual low values of depolarization
1027: in the center of NGC\,4254 giving at the same time good correspondence
1028: within the magnetic arms (Table~\ref{t:depo}). In the most NE part of the galaxy
1029: with rather weak thermal emission the obtained Faraday dispersion
1030: ($DP_{disp}>0.95$) cannot account for the observed values (about 0.7).
1031:
1032: We can also obtain the mean electron density from the emission measure $EM$ of
1033: diffuse thermal gas: $<n>=\sqrt{f<n^2>}=\sqrt{f EM L^{-1}}$ which yields
1034: \begin{equation}
1035: \sigma_{RM}=0.812 B_r \sqrt{EM d},
1036: \end{equation}
1037: independent of the filling factor.
1038: The emission measure $EM$ can be estimated from our map of radio thermal
1039: emission of NGC\,4254 at 8.46\,GHz.
1040: Following Walterbos (\cite{walterbos00}), we assume that in NGC\,4254,
1041: as in the Galaxy and in M\,31, about 20\% of the total ionized
1042: gas is diffuse, and that only this ISM phase contributes to $RM$ and
1043: to dispersion of $RM$. To make the modeled depolarization map
1044: consistent with observations, we must assume a
1045: correlation size $d$ in Eq.~(7) of about 0.5\,pc. However, this value is much
1046: smaller than applied above in Eq.~(6), and which is usually assumed
1047: for other galaxies (see e.g., Fletcher et al. \cite{fletcher04a};
1048: Beck \cite{beck07}). For regions of high $EM$ ($>1000$\,pc\,cm$^{-3}$)
1049: it also evokes unreasonable small filling factor of the diffuse gas
1050: $f$, below 0.04. Either the used assumption in Eq.~(7) of 20\% diffuse
1051: gas is wrong, which would mean that NGC\,4254 is a quite different
1052: galaxy from the Milky Way and M\,31 due to, e.g., a much higher SFR,
1053: or our depolarization model is insufficient. An independent way to
1054: determine locally the amount of the diffused ionized emission in
1055: NGC\,4254 is needed to solve this puzzle. Hence, fully understanding
1056: Faraday depolarization effects (as modeled by Eqs. 6 and 7) seems
1057: essential in building a consistent model of the ISM in galaxies.
1058:
1059: The third process, we will explore is depolarization $DP_{gFR}$ by
1060: gradients in the $RM$ across the observed synthesized beam of $\theta$ width,
1061: which occurs in a foreground screen (Sokoloff et al. \cite{sokoloff98}).
1062: We consider linear $RM$ gradients that are resolved with
1063: the observed beam. If the $RM$ has unresolved fine structure (within the
1064: beam size) then depolarization could be larger. Hence, in our approach
1065: the calculated $DP_{gFR}$ values represent an upper limit of depolarization.
1066: We constructed a relevant depolarization model map for NGC\,4254 using
1067: AIPS task BDEPO, which use the following approximation:
1068: \begin{equation}
1069: DP_{gFR}=exp\left[\frac{-(g_{RM} \theta)^2 (\lambda^4_1 - \lambda^4_2)}{2\ln 2}\right]
1070: \end{equation}
1071: where $g_{RM}$ is the gradient in $RM$. We estimate this gradient by applying
1072: the Sobol operator to the observed $RM$ map with the help of the AIPS
1073: task NINER.
1074:
1075: The modeled $DP_{gFR}$ values outside the core of NGC\,4254 are typically
1076: larger than 0.95 (Table~\ref{t:depo}). Relatively low values of 0.75-0.85 are predicted
1077: for the NE outskirts of the galaxy, where observed depolarization is at a similar
1078: level and which could not be accounted for by other depolarization effects. As expected,
1079: gradients in $RM$ cause a large amount of depolarization in the $RM$ jump regions, where
1080: $DP_{gRM}$ goes down to 0.2-0.4.
1081:
1082: In summary, various depolarization processes dominate in different regions of
1083: NGC\,4254, which explain the conclusion of Sect.~\ref{s:rm}, where we found no global
1084: correlation of depolarization with $RM$ or thermal emission.
1085: The depolarization observed in the central part of the galaxy is well accounted
1086: for by Faraday dispersion together with differential Faraday rotation
1087: (Table~\ref{t:depo}), while the large depolarization observed around $RM$
1088: jumps can be successfully explained by differential Faraday rotation and
1089: depolarization due to $RM$ gradients. The gradients in $RM$ can also underlie
1090: the depolarization observed in the NE part of the galaxy.
1091:
1092: None of the modeled processes predict an enhanced level of depolarization
1093: in the regions of both outer magnetic arms in agreement with observations.
1094: In fact, a completely different set of data at 1.43\,GHz shows that all
1095: those regions are still polarized at this frequency (Fig.~4 in Paper I) and
1096: must be Faraday-thin, thus avoiding strong Faraday effects.
1097: As both outer magnetic arms manifest depolarization properties similar to the
1098: N inner magnetic arm, they cannot arise from reduced depolarization.
1099: Therefore, outer magnetic arms require an additional process
1100: that enhances regular component of the magnetic field.
1101:
1102: \subsection{Modeling of magnetic field components}
1103: \label{s:ani}
1104:
1105: In Sect.~\ref{s:sfr}, we demonstrated that the outer magnetic arms in NGC\,4254
1106: involve a statistically more regular magnetic field than the other
1107: galaxy regions with a comparable level of star-forming activity. We keep in
1108: mind that the
1109: regular field strength derived from the polarized intensity (Sect.~\ref{s:rm})
1110: contains coherent and anisotropic components.
1111: Gas compression and stretching (with shearing) are well-recognized processes
1112: that can modify the coherent field and produce anisotropic field from
1113: isotropic (random) one (Sokoloff et al. \cite{sokoloff98}, Beck et al.
1114: \cite{beck05}). In Paper I, we suggested that the southern polarized ridge may
1115: indeed be affected by stretching/shearing forces of tidal origin. As less
1116: likely, but still possible, we considered compressional forces acting in
1117: the southern disk portion due to the ram pressure of hot cluster gas.
1118: We propose here an analytical model to discriminate between stretching and
1119: ram-pressure alternatives as well as to explain the origin of outer magnetic arms.
1120:
1121: In this model, we describe the galactic magnetic field in the cylindrical polar
1122: coordinates ($R$, $\phi$, $z$) with ($R$, $\phi$) aligned with the galaxy
1123: plane and with azimuthal angle $\phi$ counted counterclockwise from the
1124: northern edge of the galaxy major axis. The observer (sky) plane is inclined
1125: to the galactic plane by the galaxy inclination angle $i=42\degr$
1126: (Paper I). The discussed compression hypothesis
1127: involves external processes to act only in the outskirts of the galaxy, while
1128: the western outer magnetic arm, which is inside the galaxy disk, would remain
1129: unexplained. Therefore, we concentrate on modeling the southern magnetic arm.
1130:
1131: As the initial properties of the modeled southern arm (without stretching or
1132: compression) we take the magnetic properties of the N arm (Table~\ref{t:mf}),
1133: which we expect to be undistorted by tidal (stretching) or ram-pressure (compression)
1134: forces and to contain no anisotropic field. We notice that in the N arm (exactly
1135: region 3, Table~\ref{t:mf}) magnetic field vectors are aligned very close with the
1136: direction of the galaxy major axis (the difference is about $10\degr$, Fig.~\ref{f:true}).
1137: In the southern magnetic arm (region 1, Table~\ref{t:mf}) they are oriented almost perpendicularly
1138: to the galaxy minor axis (difference of about $12\degr$).
1139: Therefore, without any correction for projection effects we can approximate
1140: the coherent field in the southern arm by its azimuthal component only
1141: of strength equal to the regular field observed in the N arm.
1142: Hence, following the approach of Sokoloff et al. (\cite{sokoloff98}) the
1143: components of magnetic field $\vec B_1$ in the modeled southern
1144: arm are initially:
1145: \begin{eqnarray}
1146: \nonumber B_{\phi,1}&=&\bar{B}+b_{\phi,1}\\
1147: B_{R,1}&=&b_{R,1} \\
1148: \nonumber B_{z,1}&=&b_{z,1}
1149: \end{eqnarray}
1150: where $\bar{B}$ denotes volume (ensemble) averaged magnetic field, thus its
1151: coherent part $\bar{B}=\left<B_{\phi,1}\right>$ and $b_{\phi,1}$, $b_{R,1}$,
1152: $b_{z,1}$ are random magnetic field components, with $\left<b_{\phi,1}\right> =
1153: \left< b_{R,1}\right> = \left<b_{z,1}\right> = 0$. Assuming at the beginning
1154: a fully isotropic random field, we demand $\left<b^2_{\phi,1}\right>=
1155: \left< b^2_{R,1}\right>=\left<b^2_{z,1}\right>=\sigma^2$, where $\sigma$ is
1156: the standard deviation of a random field in one dimension.
1157:
1158: We assume that the stretching/shearing is along the southern magnetic arm. This should
1159: keep constant the strength of the initial coherent field and produce an excess
1160: in the form of azimuthal random component in the azimuthal direction, thus providing for
1161: the field anisotropy. We approximately scale the strength of
1162: this process by a single {\em stretching factor} $\eta$, giving the field
1163: components in response to stretching in the form:
1164: \begin{eqnarray}
1165: \nonumber B_{\phi,2}&=&\bar{B}+\eta b_{\phi,1} \\
1166: B_{R,2}&=&b_{R,1}\\
1167: \nonumber B_{z,2}&=&b_{z,1}
1168: \end{eqnarray}
1169: In our approach, $\eta=1$ means no stretching/shearing, which corresponds to
1170: the shear strength $K=0$ in the work of Beck at al. (\cite{beck05}).
1171:
1172: With these magnetic field strengths, we can derive two non-dimensional quantities
1173: in the observer frame: the ratio $S_2/S_1$ of the observed total synchrotron intensities
1174: from the southern arm after ($S_2$) and before stretching ($S_1$); and the degree of
1175: polarization $p_2$ of synchrotron emission after stretching. In
1176: these derivations, we generally follow the approach of Sokoloff et al.
1177: (\cite{sokoloff98}) for the case of magnetic field and cosmic ray energy
1178: equipartition. In this case, the number density of CRs $N_{\rm CR}$ scales
1179: with the magnetic field as $B^2$. However, due to the CRs diffusion, we do
1180: not expect a complete spatial correlation between both quantities
1181: to hold locally, i.e., on scales below 1\,kpc (cf. discussion in Beck at al.
1182: \cite{beck03}). We then treat both the quantities as locally
1183: statistically independent variables whose ensemble
1184: averages satisfy $\left< N_{CR}\right>\propto \left< B^2\right>$.
1185: Under this condition we get:
1186: \begin{equation}
1187: \frac{S_2}{S_1}=\frac{\left< B^2_2 \right> \left< B^2_{\perp,2} \right> }
1188: {\left< B^2_1 \right> \left< B^2_{\perp,1} \right> }
1189: =\frac{\left[ \bar{B}^2+\sigma^2 \left( \eta^2+2\right) \right] \left[
1190: \bar{B}^2+ \sigma^2 \left( \eta^2+1\right) \right]}{\left( \bar{B}^2+3\sigma^2 \right) \left( \bar{B}^2+2\sigma^2
1191: \right)}
1192: \end{equation}
1193: where $B_{\perp,1}$ and $B_{\perp,2}$ are perpendicular to the line of sight
1194: components of the magnetic field in the southern arm before and after stretching,
1195: respectively, and $B_1$ and $B_2$ are the corresponding total field strengths.
1196:
1197: \begin{table}
1198: \caption[]{
1199: Modeling of the southern magnetic arm of NGC\,4254 by stretching/shearing and
1200: compression processes. The radio properties of the N magnetic arm are
1201: regarded as the initial conditions for the S magnetic arm (region 3 and 1 in
1202: Table~\ref{t:mf}, respectively). The compression and stretching factors were
1203: adjusted so as to yield the observational values (given in italics) of
1204: synchrotron emissivity ratio $S_{2}/S_2$ or the degree of polarization in
1205: the southern arm $p_{2}$ (see Sect.~\ref{s:ani}).
1206: }
1207: \begin{center}
1208: \begin{tabular}{lccc}
1209: \hline
1210: \hline
1211: & stretching $\eta$ or & $S_{2}/S_1$ & $p_{2}$ \\
1212: & compression $\zeta$ factor & \\
1213: \hline
1214: Observed & -- & $2.0\pm 0.2$ & $0.37\pm 0.03$ \\
1215: \hline
1216: Stretching & 1.53 & \em{2.0} & 0.39 \\
1217: Stretching & 1.47 & 1.8 & \em{0.37} \\
1218: \hline
1219: Compression & 1.24 & {\em 2.0} & 0.29 \\
1220: Compression & 1.84 & 8.1 & {\em 0.37} \\
1221: \hline \\
1222: \label{t:compr}
1223: \end{tabular}
1224: \end{center}
1225: \end{table}
1226: For the degree of polarization after stretching we obtain:
1227: \begin{equation}
1228: p_2=p_0\frac{\bar{B}^2+\left( \eta^2-1\right) \sigma^2 }{ \bar{B}^2+\left(
1229: \eta^2+1 \right) \sigma^2}
1230: \end{equation}
1231: where $p_0\approx 0.75$ is the maximum degree of polarization. We notice that
1232: $S_2/S_1$ as well as $p_2$ do not depend on the galaxy inclination, as expected
1233: for the orientation of magnetic vectors in the southern arm, perpendicular to the minor
1234: axis.
1235:
1236: In our modeling of the southern arm, we adjust the stretching factor $\eta$ to get
1237: exactly the observed value of the ratio $S_2/S_1$ (given in Table~\ref{t:compr})
1238: and then compare the modeled $p_2$ to the observed value (Table~\ref{t:compr}).
1239: In the second approach, we adjust $\eta$ to get for $p_2$ its observational value
1240: and then we compare modeled $S_2/S_1$ to the actual value. The results
1241: presented in Table~\ref{t:compr} show that the modeled values fit very well
1242: the observed ones within their uncertainties. So a tidal stretching can
1243: very effectively produce the anisotropic magnetic field in NGC\,4254. About
1244: 60\% of the observed polarized intensity of the southern magnetic arm
1245: comes in this case from the anisotropic random field. The determined
1246: stretching factor $\eta\approx1.5$ is rather small when compared, e.g., to the
1247: typical shear strengths in strongly-barred galaxies (about 10 in NGC\,1097,
1248: Beck et al. \cite{beck05}).
1249:
1250: The second process -- compression -- would produce enhanced magnetic field
1251: components tangent to the compression plane, leaving the component
1252: normal to it unchanged (see Beck et al. \cite{beck05}). We suspect that in the case of
1253: NGC\,4254, compression might act in the plane perpendicular to the galaxy disk,
1254: from the southeastern side of it. We assume that due to compression the
1255: tangent regular and random
1256: field components are $\zeta$ times stronger than the original ones. This {\em
1257: compression factor} can be interpreted as the ratio of gas densities
1258: ($\zeta=\rho_2/\rho_2$) before ($\rho_1$) and after compression ($\rho_2$).
1259: Hence, the magnetic field in the southern arm after compression is:
1260: \begin{eqnarray}
1261: \nonumber B_{\phi,2}&=&\zeta \,\bar{B}+\zeta\, b_{\phi,1} \\
1262: B_{R,2}&=&b_{R,1}\\
1263: \nonumber B_{z,2}&=&\zeta \, b_{z,1}
1264: \end{eqnarray}
1265: Similar to the analysis of stretching, we derive in case of compression, the
1266: synchrotron emission ratio:
1267: \begin{eqnarray}
1268: \frac{S_2}{S_1}&=&\frac{\left< B^2_2 \right> \left< B^2_{\perp,2} \right> }
1269: {\left< B^2_1 \right> \left< B^2_{\perp,1} \right> }
1270: =\frac{\left[ \zeta^2 \bar{B}^2+\sigma^2 \left( 2\zeta^2+1\right) \right]}
1271: {\left( \bar{B}^2+3\sigma^2 \right)} \times \\
1272: \nonumber &\times & \frac{\left\{ \zeta^2 \bar{B}^2+\sigma^2 \left[\zeta^2 \left(1+\sin^2 i \right)
1273: +\cos^2 i \right] \right\}}{ \left( \bar{B}^2+2\sigma^2 \right)}
1274: \end{eqnarray}
1275: and the degree of polarization in the southern arm:
1276: \begin{equation}
1277: p_2=p_0\frac{\zeta^2 \bar{B}^2+\sigma^2 \left( \zeta^2 -1\right) \cos^2 i }
1278: {\zeta^2 \bar{B}^2+\sigma^2 \left[ \zeta^2 \left(1+\sin^2 i\right) +\cos^2 i\right] }.
1279: \end{equation}
1280:
1281: According to the above expressions, we model that $\zeta\approx 1.2$ gives a
1282: synchrotron emissivity ratio $S_2/S_1$ equal to the observed value
1283: (Table~\ref{t:compr}), but the corresponding degree of polarization
1284: of 0.29 is lower than the observed one ($0.37\pm0.03$). A higher
1285: compression value, which yields a stronger
1286: anisotropic field and a sufficient degree of polarization, results in a
1287: synchrotron emission ratio (8.1) much higher than the one actually observed
1288: ($2.0\pm0.2$). In comparison to
1289: the stretching process, for the same production of anisotropic field,
1290: compression produces a larger amount of isotropic random field, resulting
1291: in a lower degree of the modeled polarization, which is not consistent with observations.
1292:
1293: We checked that our conclusions are not too sensitive to the galaxy orientation:
1294: lowering the galaxy inclination by 2 degrees (to $i=40\degr$) enhances the
1295: degree of polarization by only 1\%.
1296: We also performed an analysis of
1297: compression in two other cases: the equipartition assumption with a full local
1298: spatial correlation between the density of CRs and the magnetic field energy, and
1299: the case of no equipartition, i.e., constant density of CRs.
1300: Although we regard these assumptions as less probable than those actually adopted
1301: (no local correlation), we mention that in the case of constant density
1302: of CRs for the value of $\zeta=1.5$, the synchrotron emission ratio $S_2/S_1$
1303: attains the observed value and the degree of polarization reaches $p_2=0.33$,
1304: which is quite consistent with the observed $p_2=0.37\pm0.03$. However,
1305: the higher compression amount ($\zeta=1.8$), while yielding a sufficient degree
1306: of polarization, results in the ratio $S_2/S_1=2.9$, inconsistent with
1307: observations. The case of equipartition with local spatial correlation
1308: gives similar results. The reason why different assumptions on equipartition
1309: (with and without local correlation) give close results in model fitting
1310: is the low factor $\zeta$ that
1311: is needed to explain the enhancement of the magnetic field in the southern arm
1312: of NGC\,4254. Due to nonlinear influence of $\zeta$ on synchrotron emission
1313: ratio (Eq. 14), much larger differences in results are expected for
1314: shocks with larger $\zeta$. Such cases (with full local spatial correlation)
1315: were studied by Beck et al. (\cite{beck05}), see their Fig.~21 for a rough comparison.
1316:
1317: We conclude that the stretching hypothesis does likely provide a proper
1318: explanation for the high-field regularity in the southern magnetic arm.
1319:
1320: \subsection{The influence of the cluster environments}
1321: \label{s:cluster}
1322:
1323: One of the most characteristic features of magnetic structure in NGC\,4254 is
1324: the well-aligned magnetic field in the two outer (southern and western) magnetic
1325: arms. Although, on the whole, the field regularity seems to be dominated by
1326: the production of random fields in star-forming regions (Sect.~\ref{s:sfr}), in
1327: outer magnetic arms the regularity attains larger values than in other disk regions of
1328: similar SFR, and thus of similar random fields.
1329:
1330: According to our analysis of depolarization (Sect.~\ref{s:depo}),
1331: the outer magnetic arms are marginally depolarized, likewise the N
1332: arm, and their high magnetic regularity cannot result
1333: from a reduced depolarization. This conclusion corresponds with
1334: similar results on magnetic arms in NGC\,6946 (Beck \cite{beck07}).
1335: However, in the case of NGC\,4254, two strong (outer) magnetic arms are
1336: located downstream of nearby density waves, which was not observed in
1337: other galaxies. Even the northern part of NGC\,4254 shows (inner) magnetic arms
1338: on the upstream side of spiral arms. It appears that the unusual properties
1339: of outer magnetic arms in NGC\,4254 are instead associated with some external
1340: forces and the cluster environment.
1341:
1342: Using the modeling of magnetic field components in the southern magnetic
1343: arm (Sect.~\ref{s:ani}), we have shown that the compression due to the
1344: external ram-pressure of hot cluster gas does not provide a sufficiently-high
1345: magnetic field regularity for the observed level of polarized emission.
1346: In Paper I, we also enumerated other difficulties of this idea, e.g., as
1347: low depletion of \ion{H}{i} gas and globally-enhanced, instead of
1348: spatially-truncated, star formation in NGC\,4254.
1349:
1350: There are known examples of galaxies with a distinct action of strong
1351: compressional forces and enhanced magnetic fields. Such forces were
1352: argued to be present in the barred galaxies (Beck et al. \cite{beck05}), and
1353: assessed by the density ratio of pre- and post-shock gas. We searched for such
1354: a compressional evidence in the contrast ratio of polarized, nonthermal,
1355: and \ion{H}{i} emission in NGC\,4254, measured between magnetic or optical
1356: arms and their outside locations separated by 1.5 beams. We inspected five
1357: such positions indicated in Fig.~\ref{f:intro}, and present the results in
1358: Table~\ref{t:contrast}. The polarized ridge (locations a and b) and the W
1359: outer magnetic arm (c) do not reveal any distinctly large change in the
1360: analyzed distributions as compared with the other regions.
1361:
1362: \begin{table}[t]
1363: \caption[]{
1364: Contrast ratio in and out of the spiral arms in NGC\,4254 in polarized and
1365: total synchrotron as well as \ion{H}{i} emission. The positions of the inspected
1366: five regions are denoted in Fig.~\ref{f:intro}.
1367: }
1368: \begin{center}
1369: \begin{tabular}{lrrrrr}
1370: \hline
1371: \hline
1372: & a & b & c & d & e \\
1373: \hline
1374: Polarized intensity & 5.3 & 6.1 & 12.9 & 1.9 & 16.6 \\
1375: Nonthermal intensity & 3.7 & 4.2 & 1.8 & 2.7 & 3.3 \\
1376: \ion{H}{i} emission & 2.1 & 1.8 & 2.3 & 2.5 & 1.4 \\
1377: \hline \\
1378: \label{contrast}
1379: \end{tabular}
1380: \end{center}
1381: \label{t:contrast}
1382: \end{table}
1383:
1384: In view of the data currently available and the performed modeling of magnetic
1385: field components (Sect.~\ref{s:ani}), we suggest that the most likely hypothesis
1386: for the enhanced regular field in the outer magnetic arms are stretching forces of gravitational
1387: (tidal) origin. This prediction can be further checked by $RM$ data, not sensitive to
1388: anisotropic field. The $RM$ analysis cannot be performed for the southern outer
1389: magnetic arm (region 1), since the regular field component lies almost in the
1390: plane of the sky. However, as the strengths of different components of
1391: magnetic field for both the outer arms are very similar (Table~\ref{t:mf}),
1392: we suspect that both the arms are enriched in an anisotropic field of similar
1393: magnitude and origin. Hence, we perform $RM$ analysis for the western magnetic
1394: arm. The line-of-sight component of coherent magnetic field
1395: $<B_{coh||}>$ can be estimated from the actual value of rotation measure:
1396: \begin{equation}
1397: RM=0.812 <B_{\rm coh||}> N_e L\,.
1398: \label{e:rm}
1399: \end{equation}
1400: The mean thermal electron density $N_e$ over the galaxy thermal disk pathlength
1401: $L$ can be estimated from the thermal radio emission (the emission measure
1402: $EM$), the fraction of the diffuse gas in the total ionized gas (see
1403: Sect.\ref{s:depo}), and the filling factor $f$. The inner N magnetic arm
1404: (region 3, Fig.~\ref{f:magmaps}b) has the observed ${RM}=-43\pm17\,
1405: {\rm rad}\,{\rm m}^{-2}$. Supposing for N and W magnetic arms the same
1406: values for $L$, $f$ and for the fraction of the diffuse gas, we get from (16)
1407: a prediction of rotation measure in the W arm (region 2) of
1408: about 150\,rad\,m$^{-2}$, assuming that the entire regular field in that region
1409: (12.3\,$\mu$G) is coherent. This is much more than the observed
1410: $+40\pm17\,\rm{rad}\,\rm{m}^{-2}$. Even taking into account
1411: large statistical uncertainties in the measured $RM$ values, this
1412: disagreement can suggest that the polarized emission in the outer western
1413: arm comes not only from the regular coherent field traced by $RM$ data, but to
1414: a large extent from an anisotropic field, in agreement with the modeling of stretching effect.
1415:
1416: The stretching and shearing origin of the high magnetic regularity in outer magnetic arms
1417: is also in accord with the study of magnetic field
1418: orientations (Sect.~\ref{s:orient}) indicating a strong influence of gaseous
1419: streamlines on the entire magnetic structure in
1420: that galaxy. In the Antennae merging system, where such stretching forces are
1421: expected in the NE portion of NGC\,4038, at the base of the tidal tail, there
1422: is indeed a polarized ridge (Chy\.zy \& Beck \cite{chyzy04}) of high field
1423: regularity and with magnetic field vectors aligned with local gas flows.
1424:
1425: Recently, Kantharia et al. (\cite{kantharia08}) detected the radio
1426: envelope of NGC\,4254 at frequencies down to 240\,MHz. They argue that
1427: the envelope is a signature of the ram-pressure gas, stripped from the
1428: galactic disk and now expanding. However,
1429: radio envelopes were also observed around non-cluster spirals (cf. Paper I)
1430: and their existence cannot be used in favor of ram-pressure effects. The
1431: argument that the envelope has a steep synchrotron spectrum
1432: ($\alpha_{\rm nth}<-1.0$) is not convincing either, as
1433: field galaxies also show spectral steepening at the disk outskirts, which
1434: results from CRs diffusion and energy losses
1435: (Sect.~\ref{s:sfr}). Besides, comparing the ram pressure with the magnetic
1436: field pressure, they assumed several times lower magnetic field strength
1437: ($3\,\mu$G) than is actually observed (Fig.~\ref{f:magmaps}a). Next,
1438: their argument that the high star-formation in NGC\,4254
1439: is due to ICM wind is surprising as the Virgo Cluster spirals show the
1440: opposite effect: a lower star formation for galaxies that experienced
1441: ICM-ISM stripping events (Paper I, Koopmann \& Kenney \cite{koopmann04}).
1442:
1443: Although NGC\,4254 seems to be a 'young' Virgo Cluster member (Paper I), it
1444: reveals distortions that instead of a ram pressure, they could be
1445: attributed to a gravitational
1446: harassment (Moore et al. \cite{moore96}). It is visible in the galaxy's
1447: perturbed spiral pattern, disturbed \ion{H}{i} emission, and enhanced level
1448: of star formation. We also propose that the enhanced total and anisotropic
1449: magnetic field, the strong southern polarized ridge, and the complex magnetic
1450: field morphology, not found in the field galaxies, could also be the
1451: manifestation of the harassment process. We expect that the dramatic
1452: morphological transformations of gas, as well as magnetic field, are still
1453: ahead of this object, when it is going to pass the Virgo Cluster core,
1454: and will experience severe ram pressure stripping and global tides.
1455:
1456: Recent observations of other Virgo spirals (Vollmer at al. \cite{vollmer07},
1457: We\.zgowiec et al. \cite{wezgowiec07}) show that asymmetrical polarized emission
1458: are common among Virgo Cluster galaxies. The modeling of magnetic field
1459: components as applied to NGC\,4254 in Sect.~\ref{s:ani} can also be performed
1460: for those galaxies. This could at last allow for a statistical analysis of
1461: diverse influence of ICM on magnetic field structures in cluster spirals.
1462:
1463: \section{Summary and conclusions}
1464: \label{s:summary}
1465:
1466: We present the first detailed investigation of magnetic field,
1467: Faraday rotation measures, and depolarization of synchrotron emission in a spiral
1468: galaxy -- NGC\,4254 -- embedded in the Virgo Cluster medium. In Paper I, we
1469: described our radio (VLA and Effelsberg) and X-ray (XMM-Newton) observations and data reduction,
1470: along with a comprehensive discussion of the galaxy's properties in different wavelengths.
1471: The complex polarization properties discovered in different regions of the galaxy
1472: are analyzed here, with the help of the newly developed concept of presenting magnetic
1473: field in the form of various ``magnetic maps'' (Sect.~\ref{s:magmaps}) and with
1474: analytical modeling of Faraday effects (Sect.~\ref{s:depo}), as well as modeling of
1475: compression and stretching forces on different magnetic field components (Sect.~\ref{s:ani}).
1476:
1477: The performed study enabled us to obtain the following results:
1478: \begin{itemize}
1479:
1480: \item[-] The distribution of $RM$ in NGC\,4254 reveals large areas of coherent
1481: (unidirectional) magnetic field, which resemble a
1482: perturbed axisymmetric dynamo mode or a mixture of axisymmetric
1483: and bisymmetric modes. The magnetic field vectors are oriented outwards from the galaxy
1484: disk, contrary to other galaxies.
1485:
1486: \item[-] The constructed ``magnetic maps'' show that the total magnetic field
1487: in NGC\,4254 is stronger than in typical field spirals, reaching
1488: 17\,$\mu$G in the interarm regions, 22\,$\mu$G in spiral arms, and even
1489: 25\,$\mu$G in the disk core (assuming 1\,kpc disk thickness).
1490: The magnetic energy exceeds the thermal one in almost
1491: all regions of the galaxy disk. It surpass even the turbulent gas energy in the
1492: southern polarized ridge, the western interarm region, and at the disk
1493: outskirts, and may become dynamically important.
1494:
1495: \item[-]
1496: The regular field strength reaches the largest value of 13\,$\mu$G
1497: and the highest field regularity of about 0.8 in the two (southern and western)
1498: outer magnetic arms. The inner N magnetic arm involves a weaker regular field
1499: (8\,$\mu$G) and less field regularity (0.5). While the strongest regular
1500: field occurs in magnetic arms, the total and random field strengths are at the highest in
1501: the optical arms.
1502:
1503: \item[-] The Faraday-free pitch angles of magnetic field vectors, which change
1504: throughout the galaxy from zero (in the polarized ridge) to more than $40\degr$ (in
1505: the north), always keep their orientation close to the optical filaments
1506: (within $20\degr$). The dynamo-generated magnetic fields must be
1507: significantly modified in NGC\,4254 by density waves and gas flows.
1508:
1509: \item[-] The distribution of depolarization cannot be easily explained by any
1510: individual process (Sect.~\ref{s:rm}), and thus various Faraday effects must be at work in
1511: different galaxy regions. Our modeling demonstrated (Sect.~\ref{s:depo}) the
1512: dominant role of Faraday dispersion and differential Faraday rotation within the
1513: spiral arms and in the galaxy center, where the observed depolarization is most
1514: significant ($DP\approx 0.8-0.6$). None of Faraday depolarization effects influence
1515: the outer or inner magnetic arms in agreement with the observed
1516: depolarization there.
1517:
1518: \item[-] The strength of total magnetic field does correlate very well with the
1519: star-formation rate of individual regions in the galaxy disk (the
1520: correlation coefficient $r=0.95$), giving a well-fitted power-law relation
1521: with a slope $+0.18\pm0.01$. However, in the same regions,
1522: the regularity of magnetic field shows a high level of {\em anticorrelation}
1523: ($r=-0.67$) with the star formation activity, which mainly results from
1524: efficient production of random field in vivid star-forming regions. For
1525: the first time, we show that this relation is well approximated by a power-law
1526: with the slope of $-0.32\pm0.01$. The enhanced field regularity in the outer
1527: magnetic arms are, however, not fully controlled by the star formation.
1528:
1529: \item[-] Our modeling of magnetic field components in magnetic arms
1530: indicates that the enhanced magnetic field regularity
1531: in the outer magnetic arms can be produced by stretching and shearing forces,
1532: likely of tidal origin, which produce an anisotropic magnetic component
1533: and enhance polarized emission. This is supported by
1534: the analysis of depolarization and Faraday rotation measures, while being
1535: in accord with the field regularity-SFR relation. Hence, modeling of
1536: magnetic field components seems to be a great tool to discriminate between
1537: different physical phenomena acting in cluster environment.
1538:
1539: \end{itemize}
1540:
1541: The validity of the proposed stretching origin of enhanced magnetic field in
1542: the outer arms of NGC\,4254 could be further explored by some advanced MHD
1543: modeling, able to discriminate between magnetic field components and
1544: reproduce the perturbed structure of the galactic spiral arms. This kind of
1545: modeling of magnetic field components performed for NGC\,4254 can also be
1546: applied to other Virgo Cluster spirals and would allow for a statistical
1547: analysis of the magnetic field evolution within galaxies influenced by
1548: the cluster medium.
1549:
1550: \begin{acknowledgements}
1551: I would like to thank Dr Rainer Beck and Prof. Marek Urbanik for helpful
1552: discussions and stimulating comments, especially on depolarization effects.
1553: This work was supported by the Polish
1554: Ministry of Science and Higher Education, grant 2693/H03/2006/31.
1555: \end{acknowledgements}
1556:
1557: \begin{thebibliography}{}
1558: \bibitem[1996]{beckhoernes96} Beck, R., \& Hoernes P. 1996, Nature, 379, 47
1559: \bibitem[1996]{beck96} Beck, R., Brandenburg A., Moss, D., Shukurov, A., \& Sokoloff, D. 1996, \araa, 34, 155
1560: \bibitem[2003]{beck03} Beck, R., Shukurov, A., Sokoloff, D., Wielebinski, R. 2003, A\&A, 411, 99
1561: \bibitem[2005]{beck05} Beck, R. Fletcher, A., Shukurov, A., Snodin, A., Sokoloff, D. D.,
1562: Ehle, M., Moss, D., \& Shoutenkov, V. 2005, A\&A, 444, 739
1563: \bibitem[2007]{beck07} Beck, R. 2007, A\&A, 470, 539
1564: \bibitem[1966]{burn66} Burn, R. 1966, MNRAS, 133,67
1565: \bibitem[1990]{cayatte90} Cayatte, V., van Gorkom, J. H., Balkowski, \& C., Kotanyi, C. 1990, AJ, 100, 604
1566: \bibitem[2004]{chyzy04} Chy\.zy, K. T., \& Beck, R. 2004, A\&A, 417, 541
1567: \bibitem[2007]{chyzy07} Chy\.zy, K. T., Ehle, M., \& Beck, R. 2007, A\&A 474, 415
1568: \bibitem[2005]{deavillez05} de Avillez, M.~A., \& Breitschwerdt, D.\ 2005, \apjl, 634, L65
1569: \bibitem[2000]{elstner00} Elstner, D., Otmianowska-Mazur, K., von Linden, S., \& Urbanik,
1570: M. 2000, A\&A, 357, 129
1571: \bibitem[2005]{elstner05} Elstner, D. 2005, in The Magnetized Plasma in Galaxy Evolution,
1572: ed. Chy\.zy, K. T. et al. ( OAUJ, Krakow) 117
1573: \bibitem[1993]{ehle93} Ehle, M., \& Beck, R. 1993, A\&A, 273, 45
1574: \bibitem[2004a]{fletcher04a} Fletcher, A., Berkhuijsen, E.~M., Beck, R.,
1575: \& Shukurov, A.\ 2004, A\&A, 414, 53
1576: \bibitem[2004b]{fletcher04b} Fletcher, A., Beck, R., Berkhuijsen, E. M., Horellou, C., Shukurov, A.,
1577: 2004, in How does the galaxy work?, ed. W.J. Alfaro et al. (Dordrecht: Kluwer) 299
1578: \bibitem[2001]{jansen01} Jansen, F., Lumb, D., Altieri, B. et al. 2001, A\&A, 365, L1
1579: \bibitem[2004]{johnston04} Johnston-Hollitt, M., Hollitt, C.~P., \& Ekers,
1580: R.~D.\ 2004, The Magnetized Interstellar Medium, ed. B. Uyaniker et al.
1581: (Katlenburg:Copernicus) 13
1582: \bibitem[2008]{kantharia08} Kantharia, N.~G.,
1583: Rao, A.~P., \& Sirothia, S.~K.\ 2008, \mnras, 381, 173
1584: \bibitem[2004]{knapen04} Knapen, J. H., Stedman, S., Bramich, D. M., Folkes, S. F., \& Bradley, T. R.
1585: 2004, A\&A, 426, 1135
1586: \bibitem[2004]{koopmann04} Koopmann, R. A., \& Kenney, J. D. P. 2004, ApJ, 613, 851
1587: \bibitem[1998]{krause98} Krause, F., \& Beck, R. 1998, A\&A 335, 789
1588: \bibitem[1996]{moore96} Moore, B., Kartz, N., Lake, G., Dressler, A., \&
1589: Omler, A., Jr. 1996, Nature, 379, 613
1590: \bibitem[2001]{moss01} Moss, D., Shukurov, A.,
1591: Sokoloff, D., Beck, R., \& Fletcher, A.\ 2001, \aap, 380, 55
1592: \bibitem[2007]{moss07} Moss, D., Snodin, A.~P., Englmaier, P., Shukurov, A., Beck, R.,
1593: \& Sokoloff, D.~D.\ 2007, \aap, 465, 157
1594: \bibitem[2006]{patrikeev06} Patrikeev, I., Fletcher, A., Stepanov, R.,
1595: Beck, R., Berkhuijsen, E.M., Frick, P., \& Horellou, C. 2006, A\&A 458, 441
1596: \bibitem[1993]{phookun93} Phookun, B., Vogel S. N., \& Mundy L. G. 1993, ApJ, 418, 113
1597: \bibitem[1989]{rand89} Rand, R. J., \& Kulkarni, S. T. 1989, ApJ, 343,760
1598: \bibitem[2005]{roediger05} Roediger, E., \& Hensler, G. 2005, \aap, 433, 875
1599: \bibitem[1997]{rohde97} Rohde, R., Elstner, D.,
1600: {\ R\"uml}diger, G.\ 1997, Acta Cosmologica, 23, 205
1601: \bibitem[1999]{rohde99} Rohde, R., Beck, R., \& Elstner, D. 1999, A\&A, 350, 423
1602: \bibitem[1998]{shukurov98} Shukurov, A. 1998, MNRAS, 299, L21
1603: \bibitem[2005]{shukurov05} Shukurov, A. 2005, In: Cosmic Magnetic Fields, eds.
1604: R.\,Wielebinski, \& R.\,Beck (Heidelberg: Springer) 111
1605: \bibitem[2003]{sofue} Sofue, Y., Koda, J., Nakanishi, H., \& Hidaka, M. 2003, PASJ, 55, 75
1606: \bibitem[1998]{sokoloff98} Sokoloff, D. D., Bykov, A. A., Shukurov, A. et al. 1998,
1607: MNRAS, 299,189
1608: \bibitem[1996]{soida96} Soida, M., Urbanik, M., \& Beck, R. 1996, A\&A, 312, 409
1609: \bibitem[2006]{soida06} Soida, M., Otmianowska-Mazur, K., Chy{\.z}y, K., \& Vollmer, B.\ 2006, \aap, 458, 727
1610: \bibitem[1972]{theil72} Theil, H. 1972, Statistical Decomposition Analysis (Amsterdam/ London: North-Holland)
1611: \bibitem[2007]{vollmer07} Vollmer, B., Soida, M., Beck, R., Urbanik, M., Chy{\.z}y, K.~T.,
1612: Otmianowska-Mazur, K., Kenney, J.~D.~P., \& van Gorkom, J.~H.\ 2007, \aap, 464, L37
1613: \bibitem[2000]{walterbos00} Walterbos, R.A.M. 2000, in The Interstellar
1614: Medium in M31 and M33, ed. E.M. Berkhuijsen et al. (Shaker, Aachen) 99
1615: \bibitem[2007]{wezgowiec07} We\.zgowiec, M., Urbanik, M., Vollmer, B., Beck, R., Chy\.zy, K.~T., Soida, M., \& Balkowski,
1616: C.\ 2007, A\&A, 471, 93
1617: \bibitem[2002]{widrow02} Widrow, L.~M.\ 2002, Reviews of
1618: Modern Physics, 74, 775
1619:
1620: \end{thebibliography}
1621:
1622: \end{document}
1623: