1: %\documentclass[manuscript]{aastex}
2: \documentclass[12pt,preprint]{aastex}
3:
4: %\usepackage[]{natbib}
5: %% \newcommand{\vdag}{(v)^\dagger}
6: %% \newcommand{\degr}{$^\circ$}
7: \newcommand{\msun}{M$_\odot$}
8: \newcommand{\kms}{km s$^{-1}$}
9: \newcommand{\HI}{\ion{H}{1}}
10: \newcommand{\HII}{\ion{H}{2}}
11: \newcommand{\cmmb}{cm$^{-2}$}
12: \newcommand{\mo}{M$_\odot$}
13:
14:
15: \shorttitle{\HI\ in NGC~5253}
16: \shortauthors{Kobulnicky \& Skillman}
17:
18: \begin{document}
19:
20: \title{Inflows and Outflows in the Dwarf Starburst Galaxy
21: NGC~5253: High-Resolution \HI\ Observations }
22:
23: \author{Henry A. Kobulnicky}
24: \affil{Department of Physics \& Astronomy \\ 1000 E. University \\
25: University of Wyoming \\ Laramie, WY 82071
26: \\ Electronic Mail: chipk@uwyo.edu}
27:
28: \author{Evan D. Skillman}
29: \affil{University of Minnesota \\ Department of Astronomy \\
30: 115 Church St. SE \\ Minneapolis, MN 55455
31: \\ Electronic Mail: skillman@astro.umn.edu}
32:
33: %\author{Draft of 2007 October 31}
34:
35: \author{Accepted for Publication in {\it The Astronomical Journal} }
36:
37: %\vskip 1.cm
38:
39: \begin{abstract}
40:
41: {\it Very Large Array } and Parkes 64 m radiotelescope 21-cm
42: observations of the starburst dwarf galaxy NGC~5253 reveal a
43: multi-component non-axisymmetric \HI\ distribution. The
44: component associated with the stellar body
45: shows evidence for a small amount of rotational
46: support aligned with the major axis, in agreement with
47: optically measured kinematics and consistent with the small
48: galaxian mass. Approximately 20--30\% of the \HI\ emission
49: is associated with a second component, an \HI\ ``plume''
50: extending along the optical minor axis to the southeast.
51: We consider outflow, inflow, and tidal origins for this feature.
52: Outflow appears improbable, inflow is a possibility, and
53: tidal debris is most consistent with the observations.
54: Thus, kinematics of
55: the \HI\ that include this feature are not indicative of the
56: dynamical mass or the local velocity dispersion of the cold
57: gas. These \HI\ observations also reveal a filamentary
58: third component that includes an 800 pc diameter \HI\ shell
59: or bubble to the west of the nucleus, coinciding with an
60: H$\alpha$ shell. The mass of \HI\ in the shell may be as
61: large as $\sim$$4\times10^6$ \msun. This large mass,
62: coupled with the lack of expansion signatures in the neutral
63: and ionized gas ($v<30$ \kms), suggests that this feature
64: may be an example of a starburst-blown bubble stalled by
65: interaction with a massive neutral envelope. Many other
66: \HI\ kinematic features closely resemble those seen in
67: H$\alpha$ emission from the ionized gas, supporting the
68: interpretation of neutral and ionized gas outflow at
69: velocities of $\sim$30 \kms. Comparison between extinction
70: estimates from the Balmer emission-line decrement and the
71: \HI\ column densities suggest a gas-to-dust ratio 2--3 times
72: the Galactic value in this low-metallicity ($Z=1/4~Z_\odot$)
73: galaxy.
74: \end{abstract}
75:
76: \keywords{
77: galaxies: individual (NGC 5253) ---
78: galaxies: ISM ---
79: galaxies: kinematics \& dynamics ---
80: galaxies: starburst }
81:
82: \section{Introduction }
83:
84: NGC~5253 is a remarkable starbursting dwarf galaxy with a
85: prominent minor axis dust lane located in the nearby
86: Centarus A/M~83
87: galaxy complex \citep{kara} . At visual wavelengths, the
88: impression of NGC~5253 is dominated by its network of
89: ionized gas filaments which extend across its minor axis and
90: reach beyond the stellar distribution \citep{hodge,
91: graham81, caldwell, marlowe, martin}. X-ray observations
92: reveal several small starburst-heated bubbles but no
93: monolithic superbubble or galactic wind \citep{strickland,
94: summers}. It harbors an extremely luminous, compact,
95: obscured site of nuclear star formation containing several
96: star clusters with masses from $10^5$ -- $10^6$ $M_\odot$
97: having ages $<10^7$ yrs \citep{gonzalez, calzetti97, vanzi,
98: martin-h}. Much of the nucleus is obscured by high levels
99: of patchy dust extinction up to $A_V=25$ mag \citep{aitken,
100: moorwood, alonso}. Radio-wave continuum and recombination
101: line studies have revealed a dominant luminous ($10^9$
102: $L_\odot$), compact (r=1 pc), young ($<2.3$ Myr) star
103: cluster that appears to be gravitationally bound and
104: potentially destined to evolve into a globular cluster
105: \citep{beck, turner98, gorjian, mohan}. The centimeter-wave
106: spectral energy distribution is consistent with an
107: optically-thick thermal Bremsstrahlung
108: source\footnote{\citet{turner00} term this object a
109: ``supernebula'' while \citet{kj} use the name ``ultra-dense
110: \HII\ regions'' (UDHII) by analogy with the Galactic
111: ultra-compact \HII\ regions \citet{wood}.} having densities
112: exceeding $10^4$ cm$^{-3}$ and emission measures $> 10^8$
113: cm$^{-6}$ pc powered by $> 10^{52}$ Lyman continuum photons
114: s$^{-1}$. \citet{caldwell} studied the star formation
115: history of NGC~5253 and determined that the light was
116: dominated by a young component with an age between 10$^8$
117: and 10$^9$ yrs. Recent studies have concluded that all of
118: the UV-selected star clusters in NGC~5253 have
119: estimated ages $<$20 Myr, indicating that either the star
120: formation episode began very recently or else that the
121: timescale for dynamical disruption of star clusters is very
122: short \citep{tremonti, harris}. The overall gas-phase
123: metallicity of NGC~5253 is 12+$\log(O/H)\simeq8.16$ (about
124: 1/4 solar), but some of the central regions contain an
125: excess of nitrogen, consistent with local ``pollution'' from
126: Wolf-Rayet star winds \citep{wr, k97, ls07}.
127:
128: When observed at low resolution, the distribution and
129: kinematics of its neutral interstellar medium suggest that
130: the \HI\ in NGC~5253 either rotates about the minor axis or
131: flows radially along the minor axis
132: \citep{ks}. \citet{turner97} and \citet{meier} report the
133: $^{12}CO$ detection of molecular gas clouds having
134: kinematics consistent with infall along the minor axis dust
135: lane. These peculiar gas kinematics may hold clues to the
136: mechanism that triggered the present burst of star
137: formation.
138:
139: In this paper we present 21-cm aperture synthesis
140: observations from the National Radio Astronomy Observatory's
141: {\it Very Large Array}\footnote{The National Radio Astronomy
142: Observatory is operated by Associated Universities,
143: Inc. (AURA) under cooperative agreement with the National
144: Science Foundation.} (VLA) and the Parkes 64 m radio
145: telescope to investigate the conditions in the interstellar
146: medium that have produced such extreme star formation.
147: NGC~5253 is an ideal environment to seek the causes of
148: extreme star formation because the burst is so young and,
149: presumably, feedback from the energetic star clusters has
150: not yet had a dramatic impact on the surrounding ISM.
151:
152: The distance to NGC~5253 is somewhat controversial and has been measured
153: in many studies including
154: \citet{saha} (4.1 $\pm$0.5 Mpc),
155: \citet{gibson} (3.3$\pm0.3$ Mpc),
156: \citet{freedman} (3.25$\pm$0.2 Mpc),
157: \citet{karachentsev} (3.9$\pm$0.5 Mpc),
158: \citet{thim} (4.0$\pm$0.3 Mpc),
159: and \citet{sakai} (3.8 $\pm$0.2 Mpc).
160: The distances based on Cepheid variables strongly depend on the sample selection
161: \citep{gibson, thim} and so we prefer the TRGB distance of
162: 3.8 Mpc \citep{sakai} which implies an
163: angular scale of 18.4 pc arcsec$^{-1}$.
164:
165:
166: \section{Observations}
167:
168: \subsection{VLA}
169:
170: NGC~5253 was observed in the 21-cm line of neutral hydrogen
171: using the $VLA$ for approximately 6 hours in the BnA
172: configuration on 1995 September 25, 3 hours in the CnB
173: configuration on 1996 January 28, and 2 hours in the DnC
174: configuration on 1996 June 10. The 2IF correlator
175: configuration with Hanning smoothing yielded a total
176: bandwidth of 1.56 MHz (327 \kms) and 128 12.2 kHz (2.58
177: \kms) spectral channels. Intermittent observations of the
178: compact radio source 1316-336 were used to phase calibrate
179: the data, and observations of 3C286 were used for flux
180: calibration. The data were reduced with NRAO's $AIPS$
181: package following standard procedures for editing bad time
182: ranges and performing phase, flux, and bandpass
183: calibrations. The radio continuum flux was subtracted in
184: the UV plane using offline channels to produce a
185: continuum-free 21-cm line UV dataset. Data from each
186: configuration were calibrated separately, and then the UV
187: data were combined after subtracting continuum. These
188: combined data were then used to map the emission from the
189: 1420 MHz transition of neutral hydrogen with greatly
190: enhanced angular resolution and sensitivity compared to the
191: 45-minute DnC configuration observations presented by
192: \citet{ks}.
193:
194: The UV data were mapped to the image plane at two different
195: resolutions and CLEANed \citep{clark, hogbom} with resulting
196: clean synthesized beamsizes of 9.0\arcsec$\times$7.6\arcsec\
197: and 17\arcsec$\times$15\arcsec\ to emphasize both the small
198: scale structures and diffuse \HI\ features, respectively.
199: The RMS noise levels in these maps are 1.2 mJy beam$^{-1}$
200: and 1.7 mJy beam$^{-1}$, respectively, slightly larger than
201: the theoretical noise of 1.0 mJy. Because the angular
202: extent of \HI\ in NGC~5253 ($\sim2$\arcmin; \citet{ks}) is
203: much smaller than the $VLA$ primary beam size at 21 cm, no
204: primary beam correction was applied. The low-resolution
205: cube was used as a template in which regions having emission
206: exceeding 2$\sigma$ in at least two adjacent channels were
207: marked. The low resolution cube was then used to blank the
208: higher resolution cube on a channel-by-channel basis outside
209: the areas of probable emission. This approach preserves
210: low-level 1--2 $\sigma$ emission peaks, many of them real,
211: in the final high-resolution data cube used for subsequent
212: analysis.
213:
214:
215: \subsection{Parkes}
216:
217: NGC~5253 was observed with the Parkes 64 m radio telescope
218: in the 21-cm line of neutral hydrogen on 1997 February 12
219: using a total bandwidth of 8 MHz (1600 \kms) and 2048
220: channels of 0.0039 MHz (0.82 \kms) centered at $v_\odot=250$
221: \kms. The primary beam angular diameter is 16\arcmin. We
222: interleaved on-band spectra with frequency-switched off-band
223: spectra centered at $v_\odot= 880$ \kms\ to obtain
224: emission-free \HI\ profiles. A 9-pointing (3x3) grid with
225: 14\arcmin\ spacings centered on NGC~5253 was used to search
226: for extended emission surrounding NGC~5253, but only the
227: central position showed a strong 21-cm detection. The data
228: for left and right circular polarizations were reduced by
229: subtracting the frequency-switched reference spectrum from
230: the source and fitting a low-order polynomial to remove any
231: residual baseline. The system temperature averaged 33 K and
232: the mean rms noise in the resulting spectra is 0.06 K per
233: channel. We observed the Galactic \HII\ region S8 as a flux
234: calibration standard and scaled the observed temperatures to
235: match the 76~K reported by \citet{kalberla}.
236:
237:
238: \section{Analysis of the New \HI\ Observations}
239:
240: \subsection{Total \HI\ and Line Profiles}
241:
242: Figure~\ref{oned} shows the 21-cm emission line profile (top
243: panel) of NGC~5253 from the Parkes 64 m telescope (dotted
244: line), the $VLA$ D configuration short-baseline data (solid
245: line) and the $VLA$ B+C configuration high-resolution data
246: cube (dashed line) as described above. The lower panel
247: shows the 21-cm absorption line profile from the longest
248: baselines of the $VLA$ B configuration against the compact
249: thermal radio continuum source associated with the super
250: star cluster in the nucleus \citep{turner98}. The
251: integrated single-dish \HI\ flux is 43.9 Jy \kms, equivalent
252: to a neutral hydrogen mass of $1.42\times10^8$ \mo\ at a
253: distance of 3.8 Mpc, assuming,
254:
255: \begin{equation}
256: {{M_{HI}} \over {M_\odot}}= 2.36\times10^5 [D(Mpc)]^2 \int S(Jy) \delta v (km~s^{-1}).
257: \end{equation}
258:
259: The $VLA$ D configuration data recovers 77\% of this total,
260: implying the presence of some $3\times10^7$ \mo\ of neutral
261: hydrogen distributed smoothly on scales larger than
262: $\sim7$\arcmin, the largest structures visible in 2 hours to
263: the $VLA$ at 21 cm in the D configuration \citep{ulvestad}. The
264: high-resolution data cube contains 42\% of the single-dish flux, or about
265: $6.0\times10^7$ \mo. Therefore, \HI\ column densities measured
266: from this cube will be {\it lower limits} to the true values
267: which may be larger by a factor of order $\lesssim 2$.
268:
269: The 21-cm absorption spectrum in the lower panel of
270: Figure~\ref{oned} shows a broad absorption line spanning at
271: least 80 \kms\ with a maximum optical depth of $\tau=0.55$
272: near $v_\odot=409$ \kms. The velocity of maximum optical
273: depth is redshifted $\sim10-20$ \kms\ compared to the 21-cm
274: emission profile peak at $v_\odot\simeq395$ \kms. This is
275: consistent with the central radio supernebula lying behind the
276: bulk of the cool atomic gas and the latter having a net
277: radial velocity toward the nucleus, as suggested from the
278: kinematics of the cold molecular clouds by \citet{meier}.
279: Adopting \citep{dl},
280:
281: \begin{equation}
282: N_H ={{ \tau~\Delta{v}~T_s} \over {5.2\times10^{-19}}},
283: \end{equation}
284:
285: \noindent where $\tau$ is the 21-cm optical depth,
286: $T_s$ is the spin temperature and
287: $\Delta{v}$ is the velocity full width at half maximum of
288: the 21-cm absorption line, we can estimate the column
289: density, $N_H$, of neutral hydrogen in front of the compact
290: nuclear radio supernebula. Because the solid angle
291: probed by the $VLA$
292: absorption and emission spectra are vastly different, we do not
293: have an estimate of the spin temperature for this gas, so we
294: adopt a lower limit of $T_s>50$ K \citep{dl}. The overall
295: width of the feature is consistent with an abundance of cold
296: \HI\ clouds in front of the nucleus over a range of
297: velocities $v_\odot = 370 - 435 $ \kms. Therefore, we adopt
298: a lower limit of $\Delta{v}>10$ \kms. The resulting \HI\
299: column densities are $N_H > 5\times 10^{20}$ \cmmb\ at this
300: location.
301:
302: \subsection{The \HI\ Distribution at High Resolution}
303:
304: Figure~\ref{mom0} shows the integrated \HI\ column density
305: (contours) overlaid on a 6450 \AA\ optical continuum image
306: of NGC~5253 (greyscale). The inset at lower left depicts
307: the 9.0\arcsec$\times$7.6\arcsec\ synthesized beam.
308: Contours show 21-cm line fluxes of 0.04, 0.08, 0.12, 0.16,
309: 0.24, 0.36, and 0.40 Jy \kms\ corresponding to beam-averaged
310: \HI\ column densities of 6, 12, 18, 24, 36, 52, and 64
311: $\times10^{20}$ \cmmb. A cross marks the location of the
312: dominant embedded star cluster and thermal radio supernebula
313: \citep{turner98} where
314: \HI\ appears in absorption. The \HI\ column density peaks near
315: 6.4$\times10^{21}$ \cmmb\ at $\sim$15\arcsec\ (270 pc)
316: south-southeast of radio supernebula and near the dust lane
317: which coincides with the position of several molecular
318: clouds \citep{meier}.
319:
320: As noted by \citet{ks}, the \HI\ distribution in NGC~5253
321: does not closely follow the optical morphology. Rather, the
322: \HI\ is extended along the minor axis to the southeast and
323: northwest. Note that at this higher resolution, it is clear
324: that the extended gas is not symmetrically placed along the
325: minor axis. The extension to the southeast is a much more
326: prominent feature, and we designate this as the \HI\
327: ``plume''. These new data also reveal arc-like extensions
328: of \HI\ to the west of the nucleus, suggestive of a partial
329: shell or bubble of neutral hydrogen. There are also
330: extensions of \HI\ to the NE and SW that generally follow
331: the optical major axis.
332:
333: Figure~\ref{mom0HA} depicts the integrated \HI\ column
334: density (contours) overlaid on an $H\alpha$ image
335: (greyscale). Here we note that the morphology of the \HI\
336: shell to the west corresponds closely to the morphology of
337: the $H\alpha$ shell at position angle of about $-$70\degr\
338: as seen from the central radio supernebula. The morphology of
339: the \HI\ to the east appears to mimic the H$\alpha$ emission
340: in the inner parts of the galaxy where the \HI\ peak column
341: densities coincide with the H$\alpha$ peaks. Also in the
342: east, similar to in the west, there is an H$\alpha$
343: extension that appears to be outlined in \HI\ (at a position
344: angle of about 100\degr ). However, further from the center,
345: this correlation breaks down. Specifically, the bright \HI\
346: plume, at a position angle of about 140\degr\ does not have
347: a corresponding strong H$\alpha$ emission feature.
348:
349: Figure~\ref{MOM0V2} depicts the integrated \HI\ column
350: density (contours and greyscale) at a lower resolution of
351: 16\arcsec\ useful for identifying lower column density
352: features. A few additional features appear in this figure.
353: The \HI\ extension to the west now appears more clearly and
354: shows the same open shell geometry as the H$\alpha$. At
355: this lower resolution, it is now possible to see that there
356: is \HI\ at a large radius from the center but on the same
357: position angle as the center of the western open shell.
358: This looks reminiscent of simulations of bubbles that
359: achieve break-out \citep{maclow}. Also, at lower
360: resolution, the eastern extension looks more like the same
361: open shell morphology and also follows the H$\alpha$.
362:
363: In summary, the integrated \HI\ maps reveal three neutral
364: gas components that will become more clear in the velocity
365: channel maps below:
366: 1) extensions to the NE and SW that follow the optical major axis,
367: 2) an \HI\ plume to the SE that extends well beyond the optical extent of the galaxy, and
368: 3) open shells in the west and east, and
369:
370: \subsection{The \HI\ Kinematics}
371:
372: Figure~\ref{greypanels} shows 21-cm maps of NGC~5253 in
373: every 2.58 \kms\ velocity channel containing emission from
374: $v_\odot=459$ to $v_\odot=358$ \kms. The greyscale range
375: shows beam-averaged column densities of 1.0$\times10^{20}$
376: \cmmb\ (white), equivalent to 2$\times$ the rms noise, to
377: 5.0$\times10^{20}$ \cmmb\ (black). The cross marks the
378: location of the central super star cluster and radio
379: supernebula. Ellipses indicate the positions of the
380: molecular clouds A through E (from left to right, in
381: decreasing Right Ascension) identified by \citet{meier}. In
382: Figure~\ref{greypanels}, the southeast \HI\ plume is the
383: dominant feature at higher velocities (450 - 420 \kms ),
384: while the \HI\ aligned with the stellar distribution is
385: found primarily at velocities between 420 and 380 \kms .
386:
387: The kinematics of individual \HI\ features relative to the
388: ionized gas can be better seen in Figure~\ref{panels}
389: which shows \HI\ column density channel maps (contours)
390: overlaid on an $H\alpha$ image (greyscale). Every second
391: 2.58 \kms\ channel is plotted. The contours show fluxes of
392: 2.4, 3.6, 6.0, and 8.4 mJy beam$^{-1}$, equivalent to
393: beam-averaged \HI\ column densities of 1.0, 1.5, 2.5, and
394: 3.5 $\times10^{20}$ \cmmb. Again, it can be seen that the
395: highest velocity gas at $v_\odot = 420-455$ \kms\ (upper 2
396: rows of panels) is located to the southeast of the nucleus.
397: Gas at intermediate velocities $v_\odot=400-420$ \kms\ near
398: the systemic velocity of $v_\odot\simeq400$ \kms\ has a
399: circularly symmetric morphology similar to that of the
400: $H\alpha$. At these velocities there are also lower column
401: density \HI\ extensions along the major axis of the galaxy
402: to the northeast, southwest and also in the direction of the
403: H$\alpha$ shell to the west. At the lowest velocities,
404: $v_\odot < 400$ \kms, the \HI\ morphology becomes elongated
405: along the galaxy's major axis.
406:
407: Figures~\ref{mom0HA} and \ref{panels} also show
408: \HI\ extensions to the northeast and southwest along the galaxy's
409: major axis. The \HI\ protrusion to the northeast appears at
410: velocities $v_\odot=415-380$ \kms\ and does not
411: unambiguously form a coherent shell or bubble. The \HI\
412: extension to the southwest appears over a similar velocity
413: range with more gas closer to the systemic velocity. The
414: spatial and kinematic symmetry of these extensions along the
415: major axis is consistent with low amplitude rotation,
416: in the canonical manner, about the galaxy's minor axis, but
417: the pattern is, by no means, obvious.
418:
419: Figure~\ref{spider} shows a 21-cm intensity-weighted
420: velocity map ( 1st moment; contours) overlaid on a stellar
421: continuum 6450 \AA\ image (greyscale). Isovelocity contours
422: are labeled in \kms. This figure shows a steep kinematic
423: velocity gradient along the southeast \HI\ plume.\footnote{The
424: gradient is nearly parallel to the projected direction toward
425: the much larger group member M~83, and \citet{vdbergh}
426: suggested a recent interaction between the two. However,
427: the line-of-sight separation exceeds 1~Mpc given the current
428: distance determinations of 3.8$\pm$0.2 Mpc for NGC~5253 and
429: 5.16$\pm0.41$ Mpc for M~83 \citep{kara}, effectively
430: precluding any recent interaction.} The contours show the
431: largest radial velocities of $v_\odot\simeq450$ \kms\ at a
432: distance 1\arcmin\ to the south-southeast of the nucleus.
433: Radial velocities approach the systemic velocity of
434: $v_\odot\simeq395$ \kms\ near the nucleus. Along the major
435: axis to the northeast and southwest of the nucleus, the
436: velocity field is disordered and shows a small velocity
437: spread, ranging between $v_\odot\simeq385$ in the northeast
438: and $v_\odot\simeq405$ \kms\ in the southwest.
439:
440: \citet{ks} concluded, on the basis of low-resolution (45\arcsec)
441: \HI\ maps, that the neutral atomic medium
442: in NGC~5253 rotates about the {\it major} axis, as in polar
443: ring galaxies. However, these new high-resolution \HI\ data
444: reveal a strong asymmetry in the \HI\ distribution and
445: velocity field, making the rotation scenario improbable.
446: These new data are more consistent with an either an inflow or
447: outflow of neutral gas along the minor axis from the south
448: and east of the nucleus. Inflow, was considered by
449: \citet{ks} and propounded by \citet{turner97} and
450: \citet{meier} on the basis of $^{12}CO$ observations.
451:
452: Figure~\ref{spider2} shows the 21-cm intensity-weighted
453: velocity (1st moment) map as in Figure~\ref{spider}
454: (contours) overlaid on an \HI\ total intensity map (0th
455: moment; greyscale). The scale bar shows \HI\ surface
456: brightness in units of Jy beam$^{-1}$ m s$^{-1}$. A peak
457: flux of 400 Jy beam$^{-1}$ m s$^{-1}$ corresponds to \HI\
458: column densities of 6.4$\times10^{21}$ \cmmb. The peak \HI\
459: column densities occur where the velocity gradient is the
460: steepest. The
461: majority of the \HI\ by mass is located to the
462: south-southeast and in the circumnuclear region.
463: \HI\ extensions to the west, southwest, and northeast have
464: comparatively low total column densities and low
465: line-of-sight velocity dispersions, consistent with gas
466: associated with the main body of the galaxy.
467:
468: Position-velocity diagrams allow another perspective on the
469: observations. Figure~\ref{lvmajor} shows a
470: position-velocity diagram of the \HI\ emission taken along
471: the optical major axis at position angle 50\degr. Although
472: the diagram is quite complex, two main features stand out.
473: First, there is a large velocity range in the central parts.
474: Here one can appreciate the asymmetry in the field.
475: Redshifted \HI\ gas extends out to 50 \kms\ from the
476: systemic velocity, while blueshifted gas extends out only 30
477: \kms. The second notable component is the gas at roughly
478: the systemic velocity showing a slow drift in central
479: velocity across the spatial distribution with an amplitude
480: of roughly 20 \kms, consistent with that
481: expected for such a low-mass galaxy.
482:
483: Figure~\ref{lvminor} shows a position-velocity diagram of
484: the \HI\ emission taken along the optical minor axis at
485: position angle 140\degr. Here we again see a large range in
486: velocities over a relatively small spatial extent. It is
487: interesting that in this projection, the gas appears
488: continuous in velocity from the extreme of the redshifted
489: gas to the extreme of the blueshifted gas.
490:
491: \section{Interpretation of Major \HI\ Components in NGC~5253}
492:
493: Here we will try to draw a coherent picture of the three
494: major components of the \HI\ distribution:
495: the \HI\ associated with the stellar distribution,
496: the southeast \HI\ plume,
497: and the open shells to the west and east.
498:
499: \subsection{\HI\ Along the Major Axis}
500:
501: It is interesting to compare these new \HI\ observations of
502: NGC~5253 to those of NGC~625 \citep{cannon04}, another dwarf
503: starburst galaxy at a similar distance. \citet{cote}
504: showed that NGC~625 also has
505: the dominant \HI\ velocity gradient along the {\it minor} axis.
506: \citet{cannon04} showed that this large velocity gradient
507: was due to outflowing \HI\ gas from the central starburst.
508: This outflowing gas is also detected in the warm and hot
509: phases of the ISM \citep{cannon05}.
510:
511: This similarity with NGC~625 motivates a second look at
512: the \HI\ component in NGC~5253 that is aligned with the optical
513: galaxy. In NGC~625, the signature of rotation along the
514: major axis was swamped by the presence of the strong
515: \HI\ outflow. However, the position velocity diagrams
516: revealed the low amplitude rotation curve present in
517: the \HI. It is probable that we are seeing
518: the same effect in NGC~5253. Although the \HI\ in NGC~5253
519: is quite disturbed, the overall trend for higher velocities
520: in the southwest and lower velocities in the northeast
521: suggest a pattern of rotation. This is also seen in the
522: stellar velocities measured by \citet{caldwell} (their
523: Figure 10) and the H$\alpha$ major-axis velocity map of
524: \citet{martin} (her Figure 9). It is not straightforward to
525: calculate the expected amplitude of the rotation curve for
526: NGC~5253; comparing to the luminosity requires a significant
527: correction for the small M/L ratio expected for a starburst galaxy.
528: Nonetheless, after applying a small ($\le$ 20\%) correction for
529: inclination, the amplitude of $\sim$ 10 \kms\ is not out of
530: line with what could be expected for NGC~5253.
531: Note that the \HI\ study of the similar dwarf starburst galaxy
532: NGC 1569 by \citet{stil} also found only small evidence
533: of normal rotation in the
534: inner parts of NGC~1569, but that the \HI\ along the major
535: axis in the outer parts confirmed this rotation signature.
536:
537: \subsection{The \HI\ Plume}
538:
539: These new high resolution \HI\ observations preclude
540: rotation along the minor axis as a viable
541: interpretation for the southeast plume. Either inflow, outflow,
542: or some kind of tidal interaction remnant feature
543: seem more likely. Based on CO observations,
544: \citet{turner97} favored infall of metal poor gas.
545: The radial velocities of the CO clouds $\sim$200 pc east of the
546: nucleus are slightly higher than the systemic velocity, consistent with
547: inflow or possibly rotation about the major axis within the
548: innder few hundred pc. The \HI\ velocities at this location could
549: also be consistent with major-axis rotation within the nuclear
550: regions. Many galaxies are known to have counter-rotating or
551: decoupled dynamical components in the inner regions, and close
552: grvitational interactions are often invoked as the causes of such
553: kinematics.
554:
555: That
556: an accretion event may have triggered the current starburst
557: has been suggested several times \citep{vdbergh, graham81,
558: caldwell}. \citet{turner97} identify three points in favor
559: of the infall hypothesis. First, the weak CO emission
560: associated with a strong starburst favors low metallicity
561: gas. Second, the non-central position of the CO suggests a
562: transient feature. Third, the redshift of the CO relative
563: to the galaxian systemic velocity could be indicative of
564: inflow.
565: \citet{meier} further note that since the dust lane that is
566: associated with the CO and \HI\ emission in the southeast plume
567: appears to be in front of the galaxy, the geometry
568: (redshift) favors infall.
569:
570: A close look at the relative locations of the CO clouds,
571: dust lane, and \HI\ peaks along the southeast minor axis
572: reveals, in fact, two different gaseous components in
573: this direction. These two components are resolved both
574: spatially and kinematically by the
575: new \HI\ observations, and a comparison with the
576: extinction maps of \citet{calzetti97} shows that these are
577: intrinsically different components --- not two maxima of
578: the same feature.
579:
580: Figure~\ref{AV} shows a greyscale
581: representation of the dust reddening ($A_V$) computed from
582: the $H\alpha/H\beta$ ratio maps of \citet{calzetti97}
583: assuming $A_V=3.1 E(B-V)$. The inferred extinctions range
584: from $A_V(H\alpha/H\beta)$ near zero to
585: $A_V(H\alpha/H\beta)\simeq2$. The cross and ellipses show
586: the locations of the central thermal radio supernebula and the
587: CO clouds, as in previous figures. Contours in
588: Figure~\ref{AV} show the extinction derived from the neutral
589: hydrogen column densities assuming a Galactic ratio of gas
590: to dust \citep{bohlin},
591:
592: \begin{equation}
593: \left< {{(N_{HI} +2N_{H_2})} \over {E(B-V)}} \right> = 5.8\times10^{21}
594: atoms~cm^{-2}~mag^{-1}
595: \end{equation}
596:
597: \noindent For a standard Galactic ratio of total to selection extinction,
598: $R\equiv A_V/E(B-V) =3.1$, this is equivalent to,
599:
600: \begin{equation}
601: 3.1 \left< {{(N_{HI} +2N_{H_2})} \over {5.8\times10^{21}
602: atoms~cm^{-2}~mag^{-1}}} \right> = A_V
603: \label{dustgas}
604: \end{equation}
605:
606: Here, we have ignored, for the moment, any contribution from
607: molecular gas, and we remind the reader that the \HI\ column
608: densities inferred from the high-resolution interferometer
609: maps may be underestimates of the true column densities by
610: up to a factor of two, owing to the insensitivity of the
611: $VLA$ to emission on large angular scales. Hence, the
612: estimated \HI\ column densities and extinctions are lower
613: limits. The contours in Figure~\ref{AV} correspond to
614: $A_V(HI)=$4, 3.5, 3, 2.5, 2, 1.5, 1 and 0.5 mag.
615:
616: Figure~\ref{AV} shows that the extended region of high
617: extinction indicated by the Balmer decrement coincides with
618: a local maximum in the neutral hydrogen column density and
619: the locus of the positions of the CO detections. The ratio
620: $A_V(HI)/A_V(H\alpha/H\beta)$ in this region $\sim
621: 3.5/2=1.75$, consistent with the expectation that some of
622: the \HI\ lies behind the region of ionized gas, and with the
623: expectation that some of the diffuse \HI\ may be unseen in
624: the interferometer maps. More precise estimates are not
625: possible because of the large disparity in beamsize between
626: the $VLA$ \HI\ data (8\arcsec) and the \citet{calzetti97}
627: $HST$ maps (0.3\arcsec). Nevertheless, the correlation
628: between both extinction measures at this location is
629: consistent with the majority of the \HI\ (and molecular
630: gas?) lying in the foreground of the ionized medium. The
631: high \HI\ column density in this feature probably continues
632: into the region surrounding the central radio supernebula and
633: super star cluster, but measurements are unreliable at this
634: location owing to the strong \HI\ absoprtion.
635:
636: By contrast, the maximum \HI\ column density, corresponding
637: to $A_V(HI)=4$ mag, is located in the \HI\ plume $\sim$15\arcsec\ to the southwest
638: of the region with the largest $A_V(H\alpha/H\beta)$. At
639: this location there is no pronounced peak in
640: $A_V(H\alpha/H\beta)$ and there are no CO detections.
641: Figure~\ref{greypanels} shows that the gas responsible for
642: this primary peak in N(\HI) occurs at velocities
643: $v_\odot>410$ \kms\ and is kinematically and spatially
644: distinct from the neutral gas at the location of the
645: molecular clouds. {\it Thus, what we are calling the \HI\ plume
646: appears to have no physical relation to the region of high
647: extinction in the dust lane}, with the possible exception of
648: a tenuous connection at velocities $v_\odot<225$ \kms.
649:
650: The very high ratio of \HI\ column density to extinction in the
651: plume implies that either the plume gas is located on the
652: far side of NGC~5253, or that the gas is extremely
653: metal-poor and has a very large gas-to-dust ratio. The
654: metal-poor ISM in dwarf galaxies does exhibit gas-to-dust ratios
655: 2--5 times higher than in the Milky Way \citep{draine}, so
656: equation~\ref{dustgas} could require a correction factor of
657: this order. However, the mean ratio
658: $A_V(HI)/A_V(H\alpha/H\beta)=3.75/0.5=7.5$ in the vicinity
659: of the plume compared to $A_V(HI)/A_V(H\alpha/H\beta)=3.25$
660: near the dust lane suggests that the plume and the dust lane
661: are distinct entities. Even if we allow that, statistically,
662: half of the neutral medium may lie behind the star-forming
663: ionized gas, the inferred extinctions are factors of 2--3
664: larger from the \HI\ column density than from the Balmer
665: decrement. The real ratio may be even larger because, as
666: described above, the \HI\ column density estimates should be
667: regarded as lower limits and a molecular component is not
668: included.
669:
670: The true nature of this low-extinction, high-column-density,
671: redshifted \HI\ plume is difficult to pin down. One
672: possible explanation is that the \HI\ plume may be
673: outflowing gas such as commonly observed in dwarf starbursts
674: like NGC~1705 \citep{meurer98} and NGC~625 \citep{cannon04}.
675: The main difference between NGC~5253 and NGC~625 is the
676: asymmetry of the \HI\ distribution seen here. The
677: symmetrical \HI\ distribution in NGC~625 was taken as strong
678: evidence against infall of gas, but the \HI\ plume in
679: NGC~5253 does not show a similar
680: symmetry. Figure~\ref{lvminor} shows that there is
681: continuity from the strongly redshifted gas ($\sim$ +50 km
682: s$^{-1}$) through to blueshifted gas ($\sim$ $-$30 km
683: s$^{-1}$), which would be the signature of somewhat
684: asymmetric outflow. The ``spur'' of outflowing gas in
685: NGC~1705 is one-sided, and \citet{meurer98} note that this
686: is to be expected when the source of the outflow is
687: displaced relative to the midplane \citep{maclow89}.
688: However, the plume in NGC~5253 contains 20--30\% of the \HI\
689: seen in the interferometer maps, depending on the spatial
690: and velocity boundaries adopted. Such a fractionally
691: massive outflow of neutral gas over a fairly small solid
692: angle without any ionized gas counterpart makes the outflow
693: interpretation less plausible in this galaxy.
694:
695: Radial inflow, suggested by \citet{meier} on the basis of
696: the kinematics of CO clouds within the minor axis dust lane,
697: is another possibility. The geometry of the isovelocity
698: contours in the redshifted \HI\ plume can help distinguish
699: whether the inflow could be on the near or far side of NGC~5253.
700: The most redshifted contours (455 \kms) lie farthest from
701: the galaxy and are $\sim$50 \kms\ larger than the systemic
702: velocity of 400 \kms. An infalling stream is expected to
703: accelerate as it approaches the galaxy. Thus, if the plume
704: were infalling gas on the {\it near} side, the isovelocity
705: contours closer to the galaxy ought to exceed 455 \kms.
706: Instead, we observe that they approach the systemic
707: velocity. Inflow from the {\it far} side of NGC~5253 might,
708: however, produce the observed signature. Such a background
709: geometry would simultaneously explain the lack of extinction
710: seen toward the high column density peaks in the plume.
711:
712: Finally, and perhaps most plausibly, the \HI\ plume may be a
713: kinematically distinct remnant from a recent gravitational
714: interaction with another member of the Centarus A subgroup
715: of the Centarus A/M~83 complex. It is difficult to prove or
716: rule out this possibility. In this scenario, the \HI\ asymmetry is easily
717: understood as a chance projection against the main body of
718: NGC~5253, and there is no need for the tidal tail to be
719: symmetrically placed in either space or velocity.
720: The galaxy density within the Centarus A subgroup
721: is high, and interactions are to be expected.
722: Many of these systems are dynamically disturbed systems,
723: suggestive of a history fraught with close encounters.
724: A perturbing, but not destructive, close encounter
725: would simultaneously provide an explanation for the
726: \HI\ plume as a tidal feature and provide a cause for the
727: recent burst of star formation seen throughout the galaxy,
728: most exceptionally in the nucleus.
729:
730: Karachentsev et al.\ (2007) have provided a three dimensional
731: picture of the Cen A/M83 complex, and NGC~5253 is located in the
732: periphery. Using radial velocities and accurately measured distances,
733: Karachentsev et al.\ (2007) have identified NGC~5253 with the Cen A
734: sub-group based on a positive tidal index. This is because, even though
735: NGC~5253 is within 2 degrees of M~83 in projection on the sky, it is
736: 1.4 Mpc closer than M83 and nearly identical in distance to Cen A.
737: Note, however, that NGC~5253 is separately by roughly 12 degrees in
738: projection from Cen A, corresponding to roughly 800 kpc. Thus NGC~5253
739: is currently far from the strong gravitational influence of either
740: of the largest galaxies in the complex. Despite being a member of
741: a very populous complex, NGC~5253 appears to be relatively isolated.
742: In fact, the next nearest known galaxy, ESO 383-087, appears to be
743: over 400 kpc away. It is not clear which member of the
744: group might have interacted recently with NGC~5253.
745:
746: In some ways, the \HI\ plume in NGC~5253 is reminiscent of
747: some of the extended filaments of gas surrounding IC~10
748: \citep{ss89, wm98} or the giant \HI\ cloud in Virgo, \HI\ 1225+0146
749: \citep{mcmahon}. In IC~10, \HI\ ``plumes'' and
750: ``streamers'' have velocity gradients that show little relation
751: to the overall rotation of the galaxy. The true nature
752: of the extended \HI\ features in IC~10 is also not yet understood,
753: but interaction with these \HI\ features is generally thought
754: to have provoked the present starburst. \citet{taylor}
755: finds that starbursting dwarf galaxies like NGC~5253
756: are found preferentially to have close companions, so the interaction
757: hypothesis is a reasonable one.
758:
759: One very important point is that the large range of
760: velocities observed in the \HI\ in NGC~5253 is associated
761: with the southeast plume. It is common practice in
762: the literature to apply the FWHM of the gas velocity
763: distribution as indicative of internal dynamics to estimate
764: dynamical mass \citep{harris} or a stellar crossing time
765: \citep{calzetti97}. Regardless of whether the large-scale
766: velocity gradient seen in the \HI\ plume is due to infall,
767: outflow, or tidal debris, the resulting large velocity dispersion should not
768: be used as an internal dynamical indicator.
769:
770:
771: \subsection{The \HI\ Shells and Bubbles}
772:
773: The main question concerning the open shells on the west and
774: the east is whether those shells represent events centered
775: at the currently indicated positions or if they are also the
776: result of centrally energized outflows that give rise to
777: shell geometries as they emerge from the galaxy.
778: The centers of the western and eastern open shells
779: lie well outside of the main stellar distribution and far
780: from the current, centrally concentrated star formation.
781: Comparison with simulations of outflows \citep{maclow} and
782: real outflows observed in H$\alpha$ and X-rays
783: \citep{strickland04} show that open-shell morphologies can
784: be created where the centers of the shells are far from the
785: source of the energy. Is it possible that this is what we
786: are seeing in NGC~5253?
787:
788: How likely is it that star formation events capable of
789: creating large \HI\ shells could have occurred on the
790: periphery of the present stellar distribution? Our view of
791: the spatially resolved star formation history of NGC~5253
792: has evolved with time \citep{vdbergh, caldwell, calzetti97,
793: tremonti, harris}. The most recent work by \citet{harris}
794: emphasizes the very young ages ($\le$ 20 Myr) for all of the
795: clusters that they studied
796: \citep[in agreement with the work by][]{tremonti}. However,
797: \citet{harris} required detection in the UV for inclusion,
798: so their sample should be strongly biased
799: toward the youngest clusters (only the brightest of the
800: older clusters would be detected - see their Figure 15).
801: The earlier study by \citet{calzetti97} derived ages of
802: order 50 Myr for some of the clusters, the colors of the
803: core required starburst activity within the last 100 Myr,
804: and \citet{caldwell} deduced a range in cluster ages with an
805: upper limit of 10$^9$ years. From the above, it is not
806: clear whether strong star formation events in the periphery
807: of its stellar distribution in the recent past are
808: consistent with the star formation history of NGC~5253. A
809: star formation history derived from the field stars, while
810: challenging owing to the distance and the presence of
811: differential extinction, would be a valuable study.
812:
813: Figure~\ref{mom0HA} depicts the integrated \HI\ column
814: density (contours) overlaid on an $H\alpha$ image
815: (greyscale). The morphology of the \HI\ shell to the west
816: corresponds closely to the morphology of the $H\alpha$ shell
817: at position angle of about $-$70\degr\ as seen from the
818: central radio supernebula. The shell has an angular
819: diameter of $\sim$45\arcsec\ or about 830 pc. This
820: H$\alpha$ feature lies along echelle slit \#2 of
821: \citet{martin} who noted the absence of kinematic splitting
822: in $H\alpha$ that would normally demarcate an expanding
823: ionized shell. \citet{martin} characterizes this as a
824: quiescent shell. The mass of neutral hydrogen in this \HI\
825: feature is $\sim4.8\times10^6$ \msun\ as measured by summing
826: emission in the high-resolution data cube from this feature.
827: Such a significant mass of neutral gas could be responsible
828: for stalling the shell's expansion and preventing
829: ``blowout''. \citet{calzetti04} show emission-line ratio
830: maps for the ionized gas in this vicinity, and they
831: concluded on the basis of high [S~II]/H$\alpha$ ratios that
832: the numerous shells and radial filaments to the south and
833: west are shock ionized, consistent with starburst powered
834: shells. However, a softer radiation field, characteristic
835: of the diffuse ionized gas at large distances from star
836: forming regions, could also explain these line ratios
837: \citep{martin97,hunter}.
838:
839: In Figure~\ref{panels} it can be seen that the
840: \HI\ coincident with the western $H\alpha$ shell
841: spans a range of velocities $v_\odot=380 - 410$ \kms.
842: However, the fragmentary morphology of the neutral gas at
843: these velocities does not provide clear evidence of a single
844: coherent structure. The velocity range (30 \kms) of the
845: \HI\ near this H$\alpha$ shell is comparable to that found
846: in the ionized gas, consistent with slow or stalled
847: expansion. The lack of X-ray emission inside this bubble
848: \citep{summers} is also consistent with it being an older,
849: cooler, quiescent remnant from bygone era of star formation.
850: If we take the mass of the \HI\ bubble to be $4.8\times10^6$
851: \mo\ and the radial expansion speed to be 20 \kms\, the
852: kinetic energy contained in this structure is
853: $1.7\times10^{52}$ erg. The implied expansion timescale
854: for this $\sim$1 kpc diameter structure is
855: $<$40 Myr, making it unlikely that the current
856: 3 Myr old burst produced this bubble. It is probably
857: a product of the older episodes of star formation
858: from the previous 50 Myr.
859: \citet{marlowe} estimate the
860: mechanical energy injection rate for NGC~5253 to be
861: $3.6\times10^{40}$ erg s$^{-1}$, or about $5\times10^{54}$
862: erg over the last 5 Myr. Even though star formation and its
863: effects are distributed throughout NGC~5253, this energy
864: injection rate is more than sufficient to power the observed
865: \HI\ shell, presuming that such an injection rate has persisted for
866: at least 50 Myr---the ages of the clusters studied by \citet{calzetti97}.
867:
868: \section{Comparison of the \HI\ with Other Gas Phases}
869:
870: \subsection{Comparison with the Molecular Gas}
871:
872: Figure~\ref{greypanels} shows the positions of CO detections
873: identified by \citet{meier} as molecular clouds A through E
874: (from left to right, in decreasing Right Ascension).
875: Molecular cloud D is nearly coincident with the central
876: radio supernebula (marked by a cross). \HI\ appears
877: in absorption at this location, so it is not possible to
878: identify molecular cloud D with a particular 21-cm feature.
879: Molecular clouds B ($v_{LSR}=421$ \kms) and C ($v_{LSR}=419$
880: \kms) have probable \HI\ counterparts seen in
881: Figure~\ref{greypanels} between $v_{\odot}=430$ \kms\ and
882: $v_{\odot}=405$ \kms.\footnote{In the direction of NGC~5253,
883: the radial velocity in the heliocentric velocity frame is
884: nearly identical (within 1 \kms) to the Local Standard of
885: Rest frame, so we do not explicitly transform our 21-cm data
886: into the LSR frame used by \citet{meier}.} Molecular cloud
887: A ($v_{LSR}=363$ \kms) has no apparent 21-cm counterpart in
888: a similar velocity range. However, \HI\ emission is present
889: at that location over the velocity range $v_{\odot}=418-395$
890: \kms. Molecular cloud E ($v_{LSR}=424$ \kms) has no
891: corresponding \HI\ counterpart in a similar velocity range.
892: Detectable 21-cm emission is present at velocities
893: $v_{\odot}=410-380$ \kms. \citet{meier} note that clouds D
894: and E are tentative detections at the 3--5 $\sigma$ level of
895: significance, and they may not be real. The lack of an \HI\
896: counterpart at the velocity of cloud E is consistent with it
897: being noise in the CO data.
898:
899: The peak \HI\ column density of 6.4$\times10^{21}$ \cmmb\ measured here
900: can be used together with the peak $H_2$ column density measurement of
901: \citet{meier} to make a better estimate the extinction, $A_V(HI+H_2)$,
902: toward the dust lane, as described above.
903: The column density of molecular hydrogen can be estimated
904: from the integrated $^{12}CO$ emission, $I_{CO}$, assuming that
905: $I_{CO(2-1)}/I_{CO(1-0)}\simeq1$ \citep{meier01}, and adopting
906: \citep{bloemen, scoville},
907:
908: \begin{equation}
909: N_{H_2} (cm^{-2}) \simeq 3\times10^{20} I_{CO} (K~km~s^{-1}).
910: \end{equation}
911:
912: \noindent In terms of the CO flux, $F_{CO}$,
913:
914: \begin{equation}
915: N_{H_2} (cm^{-2}) \simeq 3\times10^{20} G~(K~Jy^{-1}) ~F_{CO} (Jy~km~s^{-1}),
916: \end{equation}
917:
918: \noindent where $G$ is the effective gain of the aperture
919: synthesis beam, approximately 0.46 K Jy$^{-1}$ for the
920: beamsize and frequency of the \citet{meier} {\it Owens
921: Valley Radio Observatory} ($OVRO$) data at 230 GHz. The peak
922: CO fluxes for the four molecular clouds to the east of the
923: nucleus measured by
924: \citet{meier} range from 6.9 -- 16.9 Jy \kms, corresponding to
925: molecular hydrogen column densities $N_{H_2}=0.95$ --
926: $2.33\times10^{21}$ \cmmb. This may be a lower limit to the
927: true molecular column density if the CO-to-$H_2$ conversion
928: factor in the low-metallicity gas of NGC~5253 is larger than
929: in the Milky Way, i.e., $X_{NGC~5253}/X_{MW}>1$
930: \citep[e.g.,][]{maloney, verter, israel}.
931: The maximum implied visual extinctions, based on peak
932: $\sim$8\arcsec\ beam-averaged
933: \HI\ and $H_2$ column densities are $A_V=4.4$ -- $5.9$.
934: However, as noted by \citet{calzetti97} and evidenced by
935: Figure~\ref{AV}, the extinction derived from optical
936: emission line ratio reddening maps is highly variable over
937: small scales. These values range from $A_V\simeq0.5$ mag to
938: $A_V=2$ mag toward most of the star clusters. $A_V$ may be $>9$ mag
939: toward the central embedded cluster, consistent with
940: estimates of $8$ mag $<A_V<25$ mag from the 9.7 $\mu$m silicate feature
941: optical depth \citep{aitken}. Because the \HI\ and CO
942: interferometer measurements provide only beam-averaged
943: extinction estimates over $\sim100$ pc scales, localized
944: regions may exhibit larger extinctions than the 4.4 -- 5.9
945: mag obtained from this approach.
946:
947: \subsection{Comparison with the Hot Ionized Gas}
948:
949: Figure~\ref{color} is a composite 3-color plus contour image
950: of NGC~5253 depicting the optical 6450 \AA\ stellar
951: continuum (blue), 0.3-8 keV X-rays (green), H$\alpha$ (red)
952: and total \HI\ column density (contours, as in
953: Figure~\ref{mom0}). The X-ray data, previously published in
954: \citet{summers}, was retrieved from the {\it Chandra X-ray
955: Observatory} archives (proposal number 02600546) and
956: adaptively smoothed to a minimum signal-to-noise ratio of 2
957: for presentation. Figure~\ref{color} reveals the relative
958: morphologies of the stars and the cold, warm ionized, and
959: hot phases of the ISM in NGC~5253. The X-ray, H$\alpha$,
960: and \HI\ emission extend preferentially to the south and
961: west of the nucleus, suggesting expansion of starburst
962: heated material along these directions. If the \HI\ plume is
963: inflowing or interacting from the southeast, it may inhibit expansion
964: in that direction.
965:
966: \section{Conclusions}
967:
968: Radio-wave aperture synthesis observations in the 21-cm line
969: of neutral hydrogen at 8\arcsec\ resolution have revealed a
970: complex non-axisymmetric gas distribution and velocity field
971: in NGC~5253. The total neutral atomic content of the galaxy
972: is $1.4\times10^8$ \mo\ from single-dish observations for a
973: distance of 3.8 Mpc. The neutral medium in NGC~5253 may be
974: characterized as consisting of three components:
975: a component containing the majority of the neutral
976: medium distributed along the stellar distribution and
977: showing evidence for low-amplitude rotation at the level of
978: 15--20 \kms;
979: a redshifted \HI\ plume extending to the south-southeast
980: along the minor axis containing 20--30\% of the \HI;
981: and kinematically cold \HI\ shells or
982: filaments extending along the minor axes and to the
983: southwest, containing $\sim$10\% of the neutral gas mass.
984: Taken together, these different components result in
985: very complex velocity structures. Here we will try to
986: summarize out best understanding of these structures,
987: starting with the relatively enigmatic \HI\ plume.
988:
989: \subsection{The Nature of the \HI\ Plume}
990:
991: The kinematically distinct \HI\ plume consists of a significant fraction
992: of the total \HI\ emission from NGC~5253. In this paper we have
993: discussed three possible origins for the \HI\ plume: outflowing gas,
994: inflowing gas, and a separate \HI\ cloud which is interacting with NGC~5253.
995:
996: The possibility of outflow was motivated by similarity with
997: velocity fields observed in other dwarf starburst galaxies
998: showing \HI\ outflow (e.g., NGC 1569, NGC 625, NGC 1705).
999: However, the large fraction of \HI\ ($\approx$25\%) in the
1000: anamalous velocity gas in NGC~5253 is exceptional.
1001: Furthermore, strong outflows are usually accompanied by
1002: H$\alpha$ and X-ray emission which is not observed at this location.
1003: Such a massive outflow confined to a small solid angle
1004: without corresponding signatures in the ionized gas seems
1005: unlikely.
1006:
1007: The possibility of inflow was motivated by the radial velocities of
1008: CO clouds within the minor-axis dust lane \citep{meier}.
1009: Our high resolution \HI\ observations
1010: indicate that the plume is a kinematically distinct feature from the \HI\
1011: associated with the molecular gas. The most compelling argument for
1012: inflow is the near coincidence in velocity of the \HI\
1013: plume with the galaxy's systemic velocity in the region where the two
1014: are spatially coincident. If the \HI\ plume is infalling,
1015: it must be from the far side of NGC~5253 given the plume's
1016: velocity field (i.e., the most redshifted gas lies farthest from the galaxy),
1017: consistent with the lack of extinction observed in this vicinity despite
1018: the high \HI\ column density.
1019:
1020: Finally, we have the possibility that the \HI\ plume is a
1021: distinct structure, perhaps a remnant from a recent interaction, seen in
1022: projection against the main body of NGC~5253. In this
1023: scenario, a recent gravitational encounter, perhaps with
1024: another dwarf galaxy in the same group, produced the
1025: \HI\ plume and provided the trigger for the current extreme
1026: burst of star formation. Other dwarf galaxies, both
1027: bursting and non-bursting are accompanied by pure \HI\
1028: structures. Given the relatively large gas mass associated
1029: with the \HI\ plume, this would appear to be the most likely
1030: explanation.
1031:
1032: \subsection{A Comprehensive View of the \HI\ in NGC 5253}
1033:
1034: Our high resolution view of NGC~5253 has shown us that NGC~5253 has
1035: many similarities to the best studied dwarf starburst galaxies.
1036: \HI\ concentrations near the minor axis dust lane are
1037: coincident in location and velocity with at least two of the
1038: molecular clouds seen in $^{12}$CO \citep{meier}. The \HI\
1039: extension to the south-southeast, designated here the \HI\
1040: plume, appears to be spatially and kinematically distinct
1041: from the gaseous medium surrounding the dust lane molecular
1042: clouds along the southeast minor axis.
1043: We present evidence that the redshifted \HI\ plume
1044: is likely to be either a dynamically distinct tidal remnant
1045: or an inflow on the far side of the galaxy.
1046: \HI\ extensions to the east, west, and southwest
1047: show remarkably similar morphologies as the
1048: H$\alpha$ shells and filaments. In particular, there is a
1049: close morphological similarity between the ``quiescent''
1050: H$\alpha$ shell to the west of the nucleus noted by
1051: \citet{martin} and an 800 pc diameter shell-like feature
1052: seen in \HI. This may be an example of a starburst-powered
1053: shell stalled by the mass of surrounding neutral gas.
1054:
1055: These complex \HI\ features and kinematic signatures, when
1056: considered together with distributions of the molecular,
1057: ionized, and hot coronal media, are consistent with a
1058: starbursting low-mass galaxy in the initial phases of
1059: generating a galactic wind. Unlike more massive galaxies
1060: such as M~82 or NGC~3079 which show monolithic,
1061: well-collimated outflows \citep[e.g.,][and references therein]
1062: {sh07,cbv02} several non-co-spatial star formation
1063: events over the last $\sim20$ Myr appear to be required to
1064: explain the multiple outflows observed in NGC~5253
1065: \citep{summers, strickland, caldwell}. The initiating and
1066: sustaining cause of the recent star formation activity may
1067: be an inflow or interaction of atomic gas along the
1068: ``plume'', identified here for the first time. As such,
1069: NGC~5253 joins a long list of starburst galaxies where
1070: recent gravitational interactions are implicated as the
1071: trigger for extreme star formation bursts.
1072:
1073:
1074:
1075: \acknowledgments
1076:
1077: H.~A.~K. was supported through NRA-00-01-LTSA-052.
1078: E.~D.~S. is grateful for partial support from NASA LTSARP grant NAG5-9221 and from
1079: the University of Minnesota. We thank an anonymous referee
1080: for incentive to investigate the
1081: NGC~5253--M~83 (lack of) connection in more detail.
1082: We thank Crystal Martin for supplying us with the continuum and $H\alpha$
1083: images (way back in 1995), David Meier for
1084: helpful exchanges regarding the molecular clouds,
1085: and Daniela Calzetti for providing an electronic version of
1086: the Balmer decrement reddening map.
1087: This research has made use of NASA's Astrophysics Data System Bibliographic Services.
1088:
1089: {\it Facilities:} \facility{VLA ()}, \facility{Parkes ()}
1090:
1091:
1092:
1093: \begin{thebibliography}{99}
1094: \bibitem[Aitken et al.(1982)]{aitken} Aitken, D.~K., Roche, P.~F.,
1095: Allen, M.~C., \& Phillips, M.~M. 1982, MNRAS, 199, 31P
1096: \bibitem[Alonso-Herrero et al.(2004)]{alonso} Alonso-Herrero, A., Takagi, T.,
1097: Baker, A., Rieke, G.~H., Rieke,
1098: M.~J., Imanishi, M., \& Scoville, N.~Z. 2004, ApJ, 612, 222
1099: \bibitem[Beck et al.(1996)]{beck} Beck, S.~C., Turner, J.~L., Ho, P.~T.~P.,
1100: Lacy, J.~H., \& Kelly, D.~M. 1996, ApJ, 457, 610
1101: \bibitem[Bloemen et al.(1986)]{bloemen} Bloemen, J.~B.~G.~M., Strong, A.~W., Blitz, L.,
1102: Cohen, R.~S., Dame, T.~M., Grabelsky, D.~A., Hermsen, W., Lebrun, F.,
1103: Mayer-Hasselwander, H.~A., \& Thaddeus, P. 1986, A\&A, 154, 25
1104: \bibitem[Bohlin et al.(1978)]{bohlin} Bohlin, R.~C., Savage, B.~D., \& Drake, J.~F.
1105: 1978, ApJ, 224, 132
1106: \bibitem[Caldwell \& Phillips(1989)]{caldwell} Caldwell, N., \&
1107: Phillips, M.~M.\ 1989, \apj, 338, 789
1108: \bibitem[Calzetti et al.(1997)]{calzetti97} Calzetti, D., Meurer, G.~R.,
1109: Bohlin, R.~C., Garnett, D.~R.,
1110: Kinney, A.~L., Leitherer, C., \& Storchi-Bergmann, T. 1997, AJ, 114, 1834
1111: \bibitem[Calzetti et al.(1999)]{calzetti99} Calzetti, D., Conselice, C.~J.,
1112: Gallagher, J.~S. III, \& Kinney, A. 1999, AJ, 118, 797
1113: \bibitem[Calzetti et al.(2004)]{calzetti04} Calzetti, D., Harris,
1114: J., Gallagher, J.~S., III, Smith, D.~A., Conselice, C.~J., Homeier, N., \&
1115: Kewley, L.\ 2004, \aj, 127, 1405
1116: \bibitem[Cannon et al.(2003)]{cannon03} Cannon, J.~M.,
1117: Dohm-Palmer, R.~C., Skillman, E.~D., Bomans, D.~J., C{\^o}t{\'e}, S., \&
1118: Miller, B.~W.\ 2003, \aj, 126, 2806
1119: \bibitem[Cannon et al.(2004)]{cannon04} Cannon, J.~M., McClure-Griffiths, N.~M.,
1120: Skillman, E.~D., \& C{\^o}t{\'e}, S.\ 2004, \apj, 607, 274
1121: \bibitem[Cannon et al.(2005)]{cannon05} Cannon, J.~M., Skillman,
1122: E.~D., Sembach, K.~R., \& Bomans, D.~J.\ 2005, \apj, 618, 247
1123: \bibitem[Cecil et al.(2002)]{cbv02} Cecil, G.,
1124: Bland-Hawthorn, J., \& Veilleux, S.\ 2002, \apj, 576, 745
1125: \bibitem[Clark(1980)]{clark} Clark, B.~G. 1980, A\&A, 89, 377
1126: \bibitem[C{\^o}t{\'e} et al.(2000)]{cote} C{\^o}t{\'e}, S.,
1127: Carignan, C., \& Freeman, K.~C.\ 2000, \aj, 120, 3027
1128: \bibitem[Dickey \& Lockman(1990)]{dl} Dicky, J.~M., \& Lockman, F.~J. 1990, ARAA, 28, 215
1129: \bibitem[Draine et al.(2007)]{draine} Draine, B.~T., Dale, D.~A., Bendo, G.
1130: et al. 2007, ApJ, in press
1131: \bibitem[Freedman et al.(2001)]{freedman} Freedman, W., Madore, B.~F., Gibson, B.~K.,
1132: Ferrarese, L., Kelson, D.~ D., Sakai, S., Mould, J.~R., Kennicutt, R.~C., Jr.,
1133: Ford, H.~C., Graham, J.~A., Huchra, J.~P., Hughes, S.~M.~G.,
1134: Illingworth, G.~D., Macri, L.~M., Stetson, P.~B. 2001, ApJ, 553, 47
1135: \bibitem[Gibson et al.(2000)]{gibson} Gibson, B.~K., et al. 2000, ApJ, 529, 723
1136: \bibitem[Gonzalez-Riestra et al.(1987)]{gonzalez} Gonz\'alez-Riestra, R.,
1137: Rego, M., \& Zamorano, J. 1987, A\&A, 186, 64
1138: \bibitem[Gorjian et al.(2001)]{gorjian} Gorjian, V., Turner, J.~L., \&
1139: Beck, S.~C. 2001, 554, L29
1140: \bibitem[Graham(1979)]{graham79} Graham, J.~A. 1979, ApJ, 232, 60
1141: \bibitem[Graham(1981)]{graham81} Graham, J.~A.\ 1981, \pasp, 93, 552
1142: \bibitem[Harris et al.(2004)]{harris} Harris, J., Calzetti, D., Gallagher, J.~S.~III,
1143: Smith, D.~A., \& Conselice, C.~J. 2004, ApJ, 603, 503
1144: \bibitem[Hodge(1966)]{hodge} Hodge, P.~W.\ 1966, \apj, 146, 593
1145: \bibitem[{H\"o}gbom(1974)]{hogbom} {H\"ogbom}, J. 1974, ApJS, 15, 417
1146: \bibitem[Huchtmeier \& Bohnenstengel(1981)]{hucht} Huchtmeier, W.~K., \&
1147: Bohnenstengel, H.-D. 1981, AA, 100, 72
1148: \bibitem[Hunter \& Gallagher(1990)]{hunter} Hunter, D.~A. \& Gallagher, J.~S. III, 1990, \apj, 362, 480
1149: \bibitem[Israel \& Burton(1986)]{israel} Israel, F.~P., \& Burton, W.~B. 1986, A\&A, 168, 369
1150: \bibitem[Kalberla et al.(1982)]{kalberla} Kalberla, P.~W.~M., Mebold, U., \&
1151: Reif, K. 1982, A\&A, 106, 190
1152: \bibitem[Karachentsev et al.(2007)]{kara} Karachentsev, I.~D., et al. 2007, AJ, 133, 504
1153: \bibitem[Karachentsev et al.(2002)]{karachentsev} Karachentsev, I.~D., Sharina, M.~E.,
1154: Dolphin, A.~E., Grebel, E.~K., Geisler, D., Guhathakurta, P., Hodge, P.~W.,
1155: Karachentseva, V.~E., Sarajedini, A., Seitzer, P. 2002, A\&A, 385, 21
1156: \bibitem[Kobulnicky \& Skillman(1995)]{ks} Kobulnicky, H.~A., \& Skillman,
1157: E.~D. 1995, ApJ, 454, L121
1158: \bibitem[Kobulnicky \& Johnson(1999)]{kj} Kobulnicky, H.~A., \&
1159: Johnson, K.~E. 1999, ApJ, 454, 527, 154
1160: \bibitem[Kobulnicky et al.(1997)]{k97} Kobulnicky, H.~A., Skillman, E.~D.,
1161: Roy, J--R., Walsh, J., \& Rosa, M. 1997, ApJ, 477, 679
1162: \bibitem[L{\'o}pez-S{\'a}nchez et al.(2007)]{ls07}
1163: L{\'o}pez-S{\'a}nchez, {\'A}.~R., Esteban, C., Garc{\'{\i}}a-Rojas, J.,
1164: Peimbert, M., \& Rodr{\'{\i}}guez, M.\ 2007, \apj, 656, 168
1165: \bibitem[Mac Low et al.(1989)]{maclow89} Mac Low, M.-M., McCray, R., \& Norman, M.~L.\ 1989, \apj, 337, 141
1166: \bibitem[Mac Low \& Ferrara(1999)]{maclow} Mac Low, M.-M., \&
1167: Ferrara, A.\ 1999, \apj, 513, 142
1168: \bibitem[Maloney \& Black(1988)]{maloney} Maloney, P., \& Black, J.~H. 1988, ApJ, 325, 389
1169: \bibitem[Marlowe et al.(1995)]{marlowe} Marlowe, A.~T., Heckman, T.~M.,
1170: Wyse, R.~M., \& Schommer, R. 1995, ApJ, 438, 563
1171: \bibitem[Martin(1997)]{martin97} Martin, C.~L. 1997, ApJ, 491, 561
1172: \bibitem[Martin(1998)]{martin} Martin, C.~L. 1998, ApJ, 506, 222
1173: \bibitem[Martin \& Kennicutt(1995)]{martin95} Martin, C.~L., \& Kennicutt, R.~C., Jr.\
1174: 1995, \apj, 447, 171
1175: \bibitem[Mart{\'{\i}}n-Hern{\'a}ndez et al.(2005)]{martin-h}
1176: Mart{\'{\i}}n-Hern{\'a}ndez, N.~L., Schaerer, D., \& Sauvage, M.\ 2005,
1177: \aap, 429, 449
1178: \bibitem[McMahon et al.(1990)]{mcmahon} McMahon, R.~G., Irwin, M.~J.,
1179: Giovanelli, R., Haynes, M.~P., Wolfe, A.~M., \& Hazard, C.\ 1990, \apj, 359, 302
1180: \bibitem[Meier et al.(2002)]{meier} Meier, D.~S., Turner, J.~L., \& Beck, S.~C.
1181: 2002, AJ, 124, 877
1182: \bibitem[Meier et al.(2001)]{meier01} Meier, D.~S., Turner, J.~L., Crosthwaite, L.~P.,
1183: \& Beck, S.~C. 2001, AJ, 121, 740
1184: \bibitem[Meurer et al.(1998)]{meurer98} Meurer, G.~R., Staveley-Smith, L., \& Killeen,
1185: N.~E.~B. 1998, \mnras, 300, 705
1186: \bibitem[Mohan et al.(2001)]{mohan} Mohan, N.~R., Anantharamaiah, K.~R., \&
1187: Goss, W.~M. 2001, ApJ, 557, 659
1188: \bibitem[Moorwood \& Glass(1982)]{moorwood} Moorwood, A.~F.~M., \& Glass, I.~S.
1189: 1982, A\&A, 115, 84
1190: \bibitem[Saha et al.(1995)]{saha} Saha, A., Sandage, A.,
1191: Labhardt, L., Schwengeler, H., Tammann, G.~A., Panagia, N., \& Macchetto,
1192: F.~D.\ 1995, \apj, 438, 8
1193: \bibitem[Sakai et al.(2004)]{sakai} Sakai, S., Ferrarese, L.,
1194: Kennicutt, R.~C., Jr., \& Saha, A.\ 2004, \apj, 608, 42
1195: \bibitem[Scoville et al.(1987)]{scoville} Scoville, N.~Z., Yun, M.~S.,
1196: Clemens, D.~P., Sanders, D.~B., \& Waller, W.~H. 1987, ApJS, 63, 821
1197: \bibitem[Shostak \& Skillman(1989)]{ss89} Shostak, G.~S., \&
1198: Skillman, E.~D.\ 1989, \aap, 214, 33
1199: \bibitem[Stil \& Israel(2002)]{stil} Stil, J.~M., \& Israel,
1200: F.~P.\ 2002, \aap, 392, 473
1201: \bibitem[Strickland \& Heckman(2007)]{sh07} Strickland,
1202: D.~K., \& Heckman, T.~M.\ 2007, \apj, 658, 258
1203: \bibitem[Strickland \& Stevens(1999)]{strickland} Strickland, D.~K., \&
1204: Stevens, I.~R. 1999, MNRAS, 306, 43
1205: \bibitem[Strickland et al.(2004)]{strickland04} Strickland, D.~K., Heckman, T.~M.,
1206: Colbert, E.~J.~M., Hoopes, C.~G., \& Weaver, K.~A.\ 2004, \apj, 606, 829
1207: \bibitem[Summers et al.(2004)]{summers} Summers, L.~K., Stevens, I.~R.,
1208: Strickland, D.~K., \& Heckman, T.~M. 2004, MNRAS, 351, 1
1209: \bibitem[Taylor(1997)]{taylor} Taylor, C.~L.\ 1997, \apj, 480, 524
1210: \bibitem[Thim et al.(2003)]{thim} Thim, F., Tammann, G.~A., Saha, A., Dolphin, A.,
1211: Sandage, A., Tolstoy, E., Labhardt, L. 2003, ApJ, 590, 526
1212: \bibitem[Tremonti et al.(2001)]{tremonti} Tremonti, C.~A., Calzetti, D.,
1213: Leitherer, C., \& Heckman, T.~M. 2001, ApJ, 555, 322
1214: \bibitem[Turner, Beck, \& Ho(2000)]{turner00} Turner, J.~L., Beck, S.~C., \&
1215: Ho, P.~T.~P. 2000, ApJ, 532, L109
1216: \bibitem[Turner et al.(1998)]{turner98} Turner, J.~L., Ho, P.~T.~P., \&
1217: Beck, S.~C. 1998, AJ, 116, 1212
1218: \bibitem[Turner, Beck, \& Hurt(1997)]{turner97} Turner, J.~L., Beck, S.~C., \& Hurt, R.~L.
1219: 1997, ApJ, 474, L11
1220: \bibitem[Ulvestad et al.(2006)]{ulvestad} Ulvestad, J.~S, Perley, R.~A., \& Taylor, G.~B. 2006,
1221: The VLA Observational Status Summary, %http://www.vla.nrao.edu/astro/
1222: \bibitem[van den Bergh(1980)]{vdbergh} van den Bergh, S.\ 1980, \pasp, 92, 122
1223: \bibitem[Vanzi \& Sauvage(2004)]{vanzi} Vanzi, L., \& Sauvage, M. 2004, A\&A, 415, 509
1224: \bibitem[Verter \& Hodge(1995)]{verter} Verter, F., \& Hodge, P. 1995
1225: \bibitem[Walsh \& Roy(1989)]{wr} Walsh, J.~E., \& Roy, J-R. 1989, MNRAS, 239, 297
1226: \bibitem[Wilcots \& Miller(1998)]{wm98} Wilcots, E.~M., \&
1227: Miller, B.~W.\ 1998, \aj, 116, 2363
1228: \bibitem[Wood \& Churchwell(1989)]{wood} Wood, D.~O., \& Churchwell, E.~B. 1989, ApJS, 69, 831
1229: \end{thebibliography}{}
1230:
1231:
1232: \clearpage
1233:
1234:
1235:
1236:
1237: \begin{figure}
1238: \plotone{f1.eps}
1239: \caption{21-cm emission line profiles (top) from the Parkes 64 m telescope
1240: (dotted line), the $VLA$ D configuration data (solid line) and the $VLA$
1241: B+C configuration data cube (dashed line). The lower panel shows the
1242: 21-cm absorption line profile from the longest $VLA$ baselines
1243: against the compact thermal
1244: radio continuum source associated with the super star cluster
1245: in the nucleus \citep{turner98}. \label{oned} }
1246: \end{figure}
1247:
1248:
1249: \begin{figure}
1250: \plotone{f2.eps}
1251: \caption{NGC~5253 \HI\ total intensity map (contours) with
1252: optical 6450 \AA\ continuum image (greyscale). Contours levels
1253: show \HI\ beam-averaged
1254: column densities of 6, 12, 18, 24, 36, 52, and 64
1255: $\times10^{20}$ \cmmb. An 8\arcsec\ diameter (148 pc)
1256: cross marks the location of the
1257: dominant compact central star cluster and radio ``supernebula''
1258: \citep{turner98}. Note that the contours at the position of this
1259: feature indicate absorption.
1260: Circles mark the positions of CO clouds
1261: reported by \citet{meier}. The inset at lower left shows
1262: the size of the 9.0\arcsec$\times$7.6\arcsec\ synthesized $VLA$ beam.
1263: \label{mom0} }
1264: \end{figure}
1265:
1266: \clearpage
1267:
1268: \begin{figure}
1269: \plotone{f3.eps}
1270: \caption{\HI\ column density of NGC~5253 (contours)
1271: as in Figure~\ref{mom0} overlaid on an $H\alpha$
1272: image (greyscale). Note the correspondence between the H$\alpha$ shell
1273: and \HI\ arcs to the west of the nucleus.
1274: \label{mom0HA} }
1275: \end{figure}
1276:
1277: \begin{figure}
1278: \plotone{f4.eps}
1279: \caption{\HI\ column density of NGC~5253 (contours and greyscale)
1280: at a lower resolution than in Figure~\ref{mom0} in order to highlight
1281: the lower column density features. Contours correspond to beam-averaged
1282: \HI\ column densities of 2, 4, 8, 16, and 32 $\times10^{20}$ cm$^{-2}$.
1283: Note the \HI\ extensions to the east and west
1284: that coincide with $H\alpha$ filaments.
1285: \HI\ extensions
1286: along the major axis to the southwest also have $H\alpha$ counterparts.
1287: \label{MOM0V2} }
1288: \end{figure}
1289:
1290:
1291: \begin{figure}
1292: \plotone{f5a.eps}
1293: \caption{\HI\ 21-cm maps of NGC~5253 in every 2.58 \kms\
1294: velocity channel containing emission. The greyscale
1295: range shows beam-averaged column densities of 1.0$\times10^{20}$ \cmmb\ (white),
1296: equivalent to 2$\times$ the rms noise,
1297: to 5.0$\times10^{20}$ \cmmb\ (black).
1298: The cross marks the location of the central super star cluster and the ellipses are the
1299: molecular clouds A through E (from left to right) identified by \citet{meier}.
1300: \label{greypanels} }
1301: \end{figure}
1302: \clearpage
1303:
1304: \setcounter{figure}{4}
1305:
1306: \begin{figure}
1307: \plotone{f5b.eps}
1308: \caption{continued}
1309: \end{figure}
1310:
1311: \clearpage
1312:
1313: \begin{figure}
1314: %\plotone{f6.eps}
1315: \caption{\HI\ column density channel maps of NGC~5253 (contours)
1316: overlaid on an $H\alpha$ image (greyscale).
1317: The contours show fluxes of 2.4, 3.6, 6.0, and 8.4
1318: $\times10^{-3}$ Jy beam$^{-1}$, equivalent to beam-averaged \HI\
1319: column densities of 1.0, 1.5, 2.5, and 3.5 $\times10^{20}$
1320: \cmmb. \label{panels} }
1321: \end{figure}
1322:
1323: \clearpage
1324:
1325:
1326: \begin{figure}
1327: \plotone{f7.eps}
1328: \caption{A 21-cm intensity-weighted velocity map of NGC~5253
1329: (i.e., 1st moment; contours) overlaid on an optical
1330: 6450 \AA\ image (greyscale). Heliocentric isovelocity
1331: contours are labeled in \kms. Note the large velocity
1332: gradient in the southeast of the galaxy. Note also the
1333: lack of a strong gradient along the major axis of the galaxy.
1334: \label{spider} }
1335: \end{figure}
1336:
1337: \clearpage
1338:
1339: \begin{figure}
1340: \plotone{f8.eps}
1341: \caption{A 21-cm intensity-weighted velocity map of NGC~5253
1342: (i.e., 1st moment; contours) overlaid on a 21-cm total
1343: intensity image (0th moment; greyscale). Heliocentric isovelocity
1344: contours are labeled in \kms. The scale bar shows
1345: \HI\ surface brightness in units of Jy beam$^{-1}$ m s$^{-1}$.
1346: \label{spider2} }
1347: \end{figure}
1348:
1349: \clearpage
1350:
1351: \begin{figure}
1352: \plotone{f9.eps}
1353: \caption{A position-velocity diagram taken along the major axis
1354: at position angle 50$^\circ$. Note the overall trend from left to right
1355: of more positive velocities, and the large velocity range near the center
1356: of the galaxy.
1357: \label{lvmajor} }
1358: \end{figure}
1359:
1360: \clearpage
1361:
1362: \begin{figure}
1363: \plotone{f10.eps}
1364: \caption{A position-velocity diagram taken along the minor axis
1365: at position angle 140$^\circ$. Note the large, continuous,
1366: asymmetric velocity range near the center of the galaxy.
1367: \label{lvminor} }
1368: \end{figure}
1369:
1370: \clearpage
1371:
1372: \begin{figure}
1373: %\plotone{f11.eps}
1374: \caption{Extinction, $A_V$, as determined from the \HI\ column density (contours)
1375: and the Balmer emission-line decrement (greyscale with colorbar; from \citet{calzetti97}).
1376: The labeled contours correspond to $A_V(HI)=$4, 3.5, 3, 2.5, 2, 1.5, 1 and 0.5 mag
1377: assuming a Galactic gas-to-dust ratio. The extended region of
1378: high extinction in the dust lane southeast of the nucleus
1379: has $A_V(H\alpha/H\beta)\simeq1.5$ and is coincident with the second highest peak
1380: in the \HI\ column density at $A_V(HI)\simeq3.5$. The \HI\ column density is unreliable in the vicinity of the
1381: central radio supernebula (marked by a cross) owing to \HI\ absorption.
1382: \label{AV} }
1383: \end{figure}
1384:
1385: \clearpage
1386:
1387: \begin{figure}
1388: %\plotone{f12.eps}
1389: \caption{A 3-color plus contour view of NGC~5253 showing
1390: the stars (green), H$\alpha$ (red),
1391: 0.3-8 keV X-rays (blue) and total \HI\ column density (contours, as in Figure~\ref{mom0}).
1392: \label{color} }
1393: \end{figure}
1394:
1395: \clearpage
1396:
1397:
1398:
1399: \end{document}
1400: