1: \documentclass[12pt,preprint]{aastex}
2: %\usepackage{mkfig}
3: \begin{document}
4:
5: \title{\bf Dependence of Heliospheric Ly$\alpha$ Absorption on the
6: Interstellar Magnetic Field}
7:
8: \author{Brian E. Wood\altaffilmark{1}, Vladislav V.
9: Izmodenov\altaffilmark{2,3}, Jeffrey L. Linsky\altaffilmark{1},
10: Dmitry Alexashov\altaffilmark{3} }
11:
12: \altaffiltext{1}{JILA, University of Colorado, 440 UCB, Boulder, CO
13: 80309-0440; woodb@origins.colorado.edu, jlinsky@jila.colorado.edu.}
14: \altaffiltext{2}{Lomonosov Moscow State University, Dept. of
15: Aeromechanics and Gas Dynamics, Moscow 119899, Russia; izmod@ipmnet.ru.}
16: \altaffiltext{3}{Institute for Problems in Mechanics RAS, Prospekt
17: Vernadskogo 101-1, Moscow 117526, Russia; and Space Research
18: Institute (IKI) RAS}
19:
20: \begin{abstract}
21:
22: We use newly developed 3D kinetic MHD models of the heliosphere
23: to predict heliospheric H~I Ly$\alpha$ absorption for various lines
24: of sight. These predictions are compared
25: with actual Ly$\alpha$ spectra from the {\em Hubble Space Telescope},
26: many of which have yielded previous detections of heliospheric
27: absorption. We find that the absorption predicted by the models
28: is weakly affected by both the magnitude and orientation of the
29: assumed ISM magnetic field. Models with $B=1.25-2.5$~$\mu$G and
30: an angle of $\alpha=15-45^{\circ}$ with respect to the upwind direction
31: of the ISM flow generally provide the best fits to the data,
32: but the sensitivity of the Ly$\alpha$ absorption to many model
33: input parameters makes it difficult to fully characterize the
34: region of parameter space allowed by the data. We also use the models
35: to assess the degree to which heliospheric asymmetries induced by the
36: ISM field should be apparent in Ly$\alpha$ absorption. An ISM field
37: that is skewed with respect to the ISM flow vector results in substantial
38: azimuthal asymmetries in both the hydrogen wall and heliosheath,
39: but only the heliosheath asymmetries yield potentially detectable
40: asymmetries in Ly$\alpha$ absorption; and then only in downwind
41: directions, where comparison with the data is complicated by few
42: actual absorption detections and an insufficient model grid extent.
43:
44: \end{abstract}
45:
46: \keywords{MHD --- solar wind --- interplanetary medium --- ultraviolet:
47: stars}
48:
49: \section{INTRODUCTION}
50:
51: The interaction region between the solar wind and ambient ISM has
52: been the subject of hydrodynamic modeling efforts \citep{enp61,enp63}
53: since around the time of the first {\it in situ} observations of the
54: solar wind by {\em Mariner~2} \citep{mn62} and by
55: {\em Luna~2} \citep{kig60}. As shown in Figure~1, this interaction
56: results in a large scale structure for the heliosphere that consists of
57: three boundaries: the termination shock (TS), where the solar wind is
58: shocked to subsonic speeds; the bow shock (BS), where the ISM flow
59: is shocked to subsonic speeds; and in between the two the
60: heliopause (HP), which separates the plasma flows of the fully
61: ionized solar wind and partially ionized ISM. Reviews of the
62: history of heliospheric modeling include \citet{teh89}, \citet{vbb90},
63: \citet{gpz99}, and \citet{vbb06}.
64:
65: {\em Voyager~1} recently encountered the TS at a distance
66: of 94~AU from the Sun in roughly the upwind direction relative to the
67: ISM flow \citep{ecs05}. However, the locations of the
68: more distant HP and BS remain observationally uncertain.
69: In general, there are very few observational constraints for the
70: properties of the heliosphere beyond the TS. One of the
71: few exceptions is heliospheric Ly$\alpha$ absorption,
72: which is observable in {\em Hubble Space Telescope} (HST) spectra of
73: nearby stars.
74:
75: Unlike the ionized component of the ISM, the neutrals in the ISM
76: can penetrate into all regions of the heliosphere. Charge exchange
77: processes involving these neutrals create populations of hot H~I
78: that permeate the heliosphere, and it is these neutrals that produce
79: absorption signatures in stellar Ly$\alpha$ lines observed by HST.
80: For most lines of sight, the absorption is dominated by H~I in the
81: so-called ``hydrogen wall'' region in between the HP and BS
82: \citep{vbb91,bew05b}, but in downwind
83: directions absorption from heliosheath neutrals, created by charge
84: exchange between the TS and HP, can be paramount
85: \citep{vvi99,bew07}. The heliospheric absorption
86: is only detectable when the ISM absorption for the observed line of sight
87: is not too broad to obscure the absorption. In upwind directions the
88: interstellar H~I column density (in cm$^{-2}$) must be
89: $\log N({\rm H~I})<18.2$ to detect heliospheric absorption, but in
90: downwind directions one must have $\log N({\rm H~I})<17.8$ \citep{bew05a}.
91: Astrospheric absorption from the wind-ISM interaction region surrounding
92: the observed star can also sometimes be detected.
93:
94: Starting with \citet{kgg97}, there have been many attempts
95: to use the Ly$\alpha$ absorption observations to test heliospheric
96: models. The hydrodynamic models are generally quite successful in
97: reproducing the observed amount of absorption, especially in upwind
98: directions where the hydrogen wall accounts for most of it
99: \citep{kgg97,vvi99,vvi02,bew00}. The Ly$\alpha$ absorption therefore
100: represents a convincing detection of the hydrogen wall, and a validation
101: of the models that predicted it even before it was detected by HST.
102:
103: However, the exact amount of absorption predicted by the models is
104: dependent on the parameters that are assumed for the Local
105: Interstellar Cloud (LIC) in which the Sun resides \citep{rl92}. Some
106: aspects of the ambient ISM are known very well, such as the LIC flow
107: speed and direction \citep[e.g.,][]{mw04,em04}, but
108: others are not known as precisely. Thus, there has been hope that
109: the Ly$\alpha$ absorption can help constrain certain properties of
110: the ISM. \citet{vvi02}, for example, experimented with
111: numerous different models assuming different combinations of ISM
112: proton and H~I densities. The absorption predicted by the models
113: does vary with the input parameters, but the absorption diagnostic
114: seems to have only a modest sensitivity to most input parameters
115: of interest, making it difficult to simply define a range of
116: parameters that are consistent with the data. The dependence of
117: the predicted absorption on the nature of the hydrodynamic code
118: used in the modeling is also a problem \citep{bew00,vvi02}.
119: The source of this difficulty lies in
120: the complexity of H~I velocity distributions within the
121: heliosphere, which are non-Maxwellian \citep[e.g.,][]{vi01}
122: and therefore can only be modeled with fully kinetic or
123: complex multi-fluid codes.
124:
125: All of the models that have been compared with the data in the
126: past have been 2-dimensional, axisymmetric models. Recently,
127: 3-dimensional MHD models have become available that are capable of
128: considering the effects of the ISM magnetic field on heliospheric
129: structure, while still maintaining a sufficiently sophisticated
130: treatment of the neutrals to properly consider them and the plasma
131: in a self-consistent manner \citep{vi05,vvi06,nvp06}.
132: Figure~1 presents results of
133: calculations made with a 3D kinetic MHD model of the heliosphere by
134: \citet{vi05}. It shows that the inclusion of even a modest
135: ISM field can indeed affect the shape of the global heliosphere.
136: We will determine here whether this also has significant effects
137: on the Ly$\alpha$ absorption. The nature of the magnetic field in
138: the ISM immediately outside the heliosphere is poorly known, so we
139: will also assess the sensitivity of the Ly$\alpha$ absorption to
140: changes in the assumed ISM field strength and orientation. In
141: doing so, we consider many more HST-observed lines of sight than
142: have been used in prior data-model comparisons.
143:
144: \section{THE CHOSEN SAMPLE OF HST Ly$\alpha$ OBSERVATIONS}
145:
146: The amount of heliospheric Ly$\alpha$ absorption depends greatly
147: on the direction of the observed line of sight. The greatest
148: spatial dependence is on the poloidal angle $\theta$ between the
149: line of sight and the upwind direction of the ISM flow. Clearly
150: it is advantageous to consider many different lines of sight with
151: a wide variety of $\theta$'s in comparing the heliospheric
152: absorption predicted by models with the data. Considering a
153: variety of directions is even more important when testing 3D MHD
154: models, which can yield heliospheric structures that are not
155: axisymmetric and therefore will have absorption predictions that
156: are dependent on the azimuthal angle as well as being dependent on
157: $\theta$ (see Fig.~1). Past data-model comparisons considered no
158: more than six HST-observed lines of sight, which individually have
159: either provided real detections of heliospheric absorption or
160: merely upper limits \citep{bew00,vvi02}.
161: This is a rather small number of lines of sight even for testing
162: axisymmetric models, let alone the 3D ones. However, the number of
163: heliospheric absorption detections has recently increased
164: significantly \citep{bew05b,bew07}, so it is
165: well worthwhile to reassess the sample of available HST data to
166: select a larger sample of spectra to test the 3D kinetic MHD
167: models.
168:
169: \citet{bew05b} provide a complete list of HST-observed
170: Ly$\alpha$ spectra that are appropriate for our purposes, all of which
171: have been analyzed to measure ISM H~I column densities, to search for
172: evidence of heliospheric/astrospheric absorption, and to measure stellar
173: Ly$\alpha$ fluxes corrected for the contaminating ISM absorption.
174: Figure~2 is a sky map in ecliptic coordinates of lines of sight
175: with a stellar Ly$\alpha$ line observed by HST. All these spectra have
176: sufficient spectral resolution to permit a reasonably precise search for
177: heliospheric absorption. The boxes indicate the 11 lines of sight that
178: actually yield detections of heliospheric absorption. All the other
179: lines of sight yield nondetections.
180:
181: Many of the detections are clustered around the upwind direction of
182: the ISM flow. The advantageous nature of upwind lines of sight for
183: detecting heliospheric absorption is consistent with model predictions,
184: which suggest that the deceleration of H~I in the hydrogen wall relative
185: to the ISM flow should be largest in these directions. This results in
186: a greater separation of the heliospheric absorption from that of the ISM,
187: thereby making it easier to detect heliospheric absorption in upwind
188: lines of sight \citep{bew05b}.
189:
190: There is also a cluster of three detections very close to the
191: downwind direction. Initial analysis of these Ly$\alpha$ spectra did
192: {\em not} yield detections \citep{bew05b}. However, we have found
193: that the stellar Ly$\alpha$ profiles reconstructed for
194: $\theta>160^{\circ}$ lines of sight are systematically blueshifted from
195: the stellar rest frames, indicating the presence of very broad, shallow
196: absorption on the red side of the Ly$\alpha$ profiles \citep{bew07}.
197: This is the exactly the sort of absorption signature one expects from
198: heliosheath neutrals (as opposed to hydrogen wall neutrals). Since
199: very downwind lines of sight looking down the tail of the heliosphere
200: will have very long path lengths through the heliosheath, it is
201: in the most downwind lines of sight where one might expect to see this
202: broad absorption.
203: %, which is harder to discern than the sharper hydrogen
204: %wall absorption signature that accounts for the upwind detections.
205: Thus, we now consider these three lines of sight to have detections
206: of heliospheric absorption, though the nature of these detections is
207: rather different from the others.
208:
209: Our goal is to select a sample of HST-observed lines of sight
210: from Figure~2 to use for comparing observed and predicted heliospheric
211: Ly$\alpha$ absorption. Obviously we start by choosing the 11
212: detections, which actually provide quantitative measurements of the
213: absorption. We add to these detections nine nondetections (diamonds
214: in Fig.~2) that at least provide upper limits for the amount of
215: absorption that might be present in those directions. These
216: nondetections are chosen to sample parts of the sky not covered by
217: the detections. Another major selection criterion is ISM H~I column
218: density. Lines of sight with low ISM column densities are preferable
219: since they provide more restrictive upper limits on heliospheric
220: absorption. Data quality (i.e., resolution, signal-to-noise) also
221: plays a role in choosing which nondetections to consider.
222:
223: Numbered symbols in Figure~2 indicate the final sample of 20
224: lines of sight to be used in our data-model comparisons. The stellar
225: identifications of these lines of sight are indicated in Figures~3 and 4,
226: along with the Ly$\alpha$ spectra, which are displayed in order of
227: increasing $\theta$. We focus only on the red side
228: of the Ly$\alpha$ absorption profile, where the heliospheric
229: absorption resides. The fluxes are normalized to the intrinsic stellar
230: Ly$\alpha$ profile reconstructed in the original analysis of the data.
231: We refer the reader to \citet{bew05b} and references therein to
232: see the full Ly$\alpha$ spectra and descriptions of their analysis.
233: The dotted green lines in the figure show only the ISM absorption
234: based on these analyses. For the heliospheric absorption detections,
235: there is excess absorption observed beyond that from the ISM.
236: Successful heliospheric models should predict the right
237: amount of excess absorption to fit the data for these lines of sight.
238: For the nondetections, the ISM absorption fits the data reasonably well.
239: In these cases, successful heliospheric models should predict essentially
240: no significant absorption beyond that from the ISM.
241:
242: The three $\theta>160^{\circ}$ detections (\#18--\#20 in Figs.~2--4)
243: are special cases, as mentioned above. The original reconstructed stellar
244: Ly$\alpha$ profiles suggest no heliospheric absorption, but the
245: blueshifts of these profiles away from their stellar rest
246: frames implies that these profiles are inaccurate. We can infer the amount
247: of heliospheric absorption in these directions by constructing a stellar
248: profile forced to be centered on the stellar radial velocity and
249: then seeing how much of the red wing of that profile must be absorbed to
250: yield the original profile. The shaded regions in Figure~3 for these
251: three downwind lines of sight indicate this excess
252: absorption, where the uncertainties are estimated by allowing the
253: stellar radial velocity to be $\pm 3$ km~s$^{-1}$ from its measured
254: value. This excess absorption cannot be extended to lower velocities
255: closer to the center of the Ly$\alpha$ line because near line center
256: stellar Ly$\alpha$ profiles cannot be assumed to be symmetric and
257: centered on the stellar rest frame. Stellar Ly$\alpha$ profiles often
258: have self-reversals near line center, which are often asymmetric. For
259: details about all of this, see \citet{bew07}. The important thing
260: to note here is that for
261: these three very downwind lines of sight, the absorption predicted by
262: the models should {\em not} fit the data but should instead fall within
263: the shaded regions.
264:
265: Finally, the requirement that the intrinsic stellar Ly$\alpha$
266: profile be within $\pm 3$ km~s$^{-1}$ of the stellar rest frame allows
267: us to compute upper limits for the amount of broad heliosheath
268: absorption that can be present for lines of sight without detected
269: heliospheric absorption. Thick dashed lines in Figure~3 show these
270: upper absorption limits, but only for downwind lines of sight
271: ($\theta>110^{\circ}$) where the broad heliosheath absorption is
272: potentially prominent. Absorption predictions from the models must lie
273: above these limits to be consistent with the data. The dashed lines
274: cannot be extended to low velocities close to line center for the
275: same reason that the shaded regions of the $\theta>160^{\circ}$ lines
276: of sight are limited to the wings of the line (see above).
277:
278: \section{THE INTERSTELLAR MAGNETIC FIELD'S EFFECTS ON Ly$\alpha$
279: ABSORPTION}
280:
281: Figures 3 and 4 compare the HST Ly$\alpha$ data with the
282: heliospheric absorption predicted by 3D kinetic MHD models of
283: the heliospheric interface \citep{vi05,vvi06}, assuming
284: various directions and magnitudes for the ISM field. The models used
285: here are of the type initially developed by \citet{vbb93,vbb95},
286: with a fully kinetic treatment of neutral hydrogen
287: within the heliosphere to provide the most precise computations of
288: the velocity distribution functions of the neutrals. \citet{vi05}
289: expanded the 2D axisymmetric Baranov \& Malama code to
290: a fully 3D geometry, and also added the capability of including an
291: interstellar magnetic field in the model. The code separates all
292: heliospheric H atoms into several populations: 1. original interstellar
293: atoms and other atoms originating outside of the bow shock, 2. secondary
294: interstellar atoms originating between the bow shock and heliopause,
295: 3. atoms originating between the heliopause and termination shock, and
296: 4. atoms originating in the supersonic solar wind. We calculate number
297: densities, temperatures, and bulk velocities for these populations along
298: the lines of sight toward the observed stars by taking moments of the
299: velocity distributions. The heliospheric absorption for each line of
300: sight is computed from these traces of density, temperature, and flow
301: velocity. With this methodology, we are making the approximation
302: that the velocity distribution functions of the individual populations
303: are locally Maxwellian.
304:
305: \subsection{Previous Constraints on the Local ISM Field}
306:
307: The nature of the interstellar magnetic field surrounding the
308: Sun is poorly known, though some observational constraints exist.
309: The global Galactic field has a magnitude of $1.6\pm 0.2$~$\mu$G and
310: is directed towards a Galactic longitude of $l=96\pm 4^{\circ}$, but
311: there is substantial local variability \citep{rjr89}, meaning
312: that the actual local field could be significantly higher or lower and
313: could be in a completely different direction.
314:
315: A $\sim 4^{\circ}$ discrepancy exists between the flow vectors of
316: interstellar He and H within the solar system, and the most promising
317: explanation for this is that the LIC's magnetic field is skewed with
318: respect to the ISM flow seen by the Sun, which can deflect the flow
319: of interstellar hydrogen atoms in the heliosphere \citep{rl05}.
320: Helium atoms are not affected in this manner, since their charge
321: exchange cross sections are much lower than hydrogen and they are
322: therefore effectively blind to the presence of the heliosphere.
323: If correct, this interpretation identifies a plane in which the ISM
324: field must lie, which happens to be inconsistent with the orientation of
325: the global Galactic field. \citet{mo06} have argued that
326: an ISM field that is $\alpha=30^{\circ}-60^{\circ}$ from the apparent
327: flow direction can potentially explain {\em Voyager~1} and {\em 2}
328: observations of energetic particles flowing inwards from the TS.
329: The satellites have both seen these particles, but flowing in opposite
330: directions. \citet{mo06} demonstrate that asymmetries in the heliospheric
331: structure induced by a skewed ISM magnetic field can potentially
332: cause this effect, thanks in part to the satellites' positions on opposite
333: sides of the ecliptic plane.
334:
335: A magnetic field much stronger than the global Galactic
336: field has been proposed to explain an apparent pressure imbalance
337: between the hot, ionized plasma that dominates the Local Bubble and the
338: warm, partially neutral clouds that lie within it. The Local Bubble,
339: the cavity in which the Sun resides \citep{rl03}, is believed to account
340: for much of the soft X-ray background radiation \citep[see also][]{rks06}.
341: These X-rays seem to suggest thermal pressures of
342: $P/k \sim 15,000$ cm$^{-3}$~K \citep{sls98}. In contrast, the LIC and
343: other warm clouds within the Local Bubble appear to have much lower
344: pressures of $P/k \sim 2280$ cm$^{-3}$~K \citep{ebj02,sr04}.
345: Such a large pressure imbalance within the local ISM seems unlikely.
346:
347: One way out of this dilemma is to propose that the LIC is supported
348: by a strong magnetic field of order 7~$\mu$G \citep{dpc03,vf04}.
349: This would seem to be disallowed by heliospheric models,
350: which imply that such large magnetic pressures would force the termination
351: shock well inside the 94~AU distance measured by {\em Voyager~1}
352: \citep{gg97,ecs05}. The only way to avoid this problem
353: is for the field to be nearly parallel to the ISM flow. However,
354: this would contradict the evidence mentioned above that the LIC field is
355: {\em not} parallel to to the LIC flow (i.e., the discrepant H and He
356: flow vectors within the solar system, and the opposite flow directions of
357: termination shock particles observed by {\em Voyager~1} and {\em 2}).
358: Thus, large fields are still difficult to reconcile with heliospheric
359: observations and models. An alternative solution to the Local Bubble
360: pressure problem is that the pressures normally estimated from the soft
361: X-ray emission are too high for various reasons: contamination from
362: heliospheric foreground emission \citep{tec00,rl04a,dk06},
363: contamination from X-ray emission from the walls of the
364: Local Bubble \citep{rl04b}, and improper assumption of collisional
365: ionization equilibrium \citep{db01}.
366:
367: \subsection{Absorption Constraints on the ISM Field}
368:
369: Most of our models are computed assuming a modest magnetic field
370: of $B=2.5$~$\mu$G, which can compress the TS somewhat depending on
371: the field orientation, but not enough to be inconsistent with the
372: {\em Voyager~1} encounter distance of 94~AU (see Fig.~1). We assume
373: that the magnetic field is oriented within the plane suggested by
374: \citet{rl05}. Figure~3 shows the absorption predicted
375: by models with different field orientations within this plane,
376: for angles ranging from parallel to the ISM flow ($\alpha=0^{\circ}$)
377: to perpendicular to the ISM flow ($\alpha=90^{\circ}$). Table~1
378: lists the ecliptic and Galactic coordinates that correspond to
379: these field directions. Figure~4 shows the absorption predictions
380: for three models with $\alpha=45^{\circ}$,
381: but with different field strengths of 0, 1.25, and 2.5~$\mu$G.
382: The ISM hydrogen and proton densities assumed in these models are
383: $n_\infty(H~I)=0.18$ cm$^{-3}$ and $n_\infty(H^+)=0.06$ cm$^{-3}$, the
384: ISM flow speed is $V_\infty=26.4$ km~s$^{-1}$, and the temperature is
385: $T_\infty=6400$~K \citep{vi05}.
386:
387: Figure~3 shows that in upwind directions, where most of the
388: heliospheric detections lie, the absorption decreases
389: with increasing $\alpha$. To better
390: illustrate this behavior, Figure~5 plots the absorbed Ly$\alpha$
391: flux predicted by the Figure~3 models versus $\alpha$.
392: No one model clearly
393: fits the data better than all the others, presumably due to the
394: absorption dependence on $\alpha$ being only a modest one. This
395: also may be indicative of the systematic uncertainties in the
396: estimation of the ISM absorption. The 61~Vir line of sight seems
397: particularly discrepant. This is probably due to difficulties in
398: removing geocoronal emission blended with the red side of the
399: Ly$\alpha$ absorption line \citep[see Fig.~1 in][]{bew05b}, so
400: this line of sight should be regarded with caution.
401:
402: The $\alpha=60^{\circ}$ and $\alpha=90^{\circ}$ models underestimate
403: the absorption in all upwind directions (see Figs.~3 and 5), so perhaps
404: these models should be considered inconsistent with the data. If one
405: ignores the problematic 61~Vir line of sight, the $\alpha=0^{\circ}$
406: model is a worse fit to the data than the $\alpha=15^{\circ}-45^{\circ}$
407: for all but the 36~Oph line of sight. Considering the evidence
408: described above that the ISM field is skewed from the flow direction,
409: $\alpha=0^{\circ}$ seems unlikely anyway. Thus, we consider
410: $\alpha=15^{\circ}-45^{\circ}$ to represent the most likely field
411: orientations for the local ISM, which overlaps the
412: $\alpha=30^{\circ}-60^{\circ}$ range quoted by \citet{mo06}.
413:
414: These conclusions are based on $B=2.5$~$\mu$G models, but
415: Figure~4 shows that assuming $B=1.25$~$\mu$G does not change the
416: absorption very much. Thus, at least for these low-to-moderate
417: field strengths our conclusions concerning $\alpha$ are relatively
418: sound. However, Figure~4 shows that decreasing the field all the
419: way to zero does change the absorption, with the absorption being
420: somewhat higher in upwind directions.
421: The greater amount of absorption upwind is somewhat
422: surprising considering that the hydrogen wall is narrower for
423: $B=0$ (see Fig.~1). However, the inclusion of even a modest ISM
424: field weakens the bow shock and lowers the H~I density in the
425: hydrogen wall, more than offsetting the broader width of the wall.
426:
427: It should be stated that definitive conclusions are difficult
428: to make at this point since the Ly$\alpha$ absorption
429: is at least somewhat sensitive to other input parameters, such
430: as the assumed ISM densities and temperature that also have observational
431: uncertainties \citep{vvi02}, though not nearly as large as
432: those involving the magnetic field. A time-consuming
433: thorough exploration of parameter space would be necessary to fully
434: characterize the constraints on ISM properties provided by the
435: Ly$\alpha$ data.
436:
437: We have focused so far on comparing the models with data in
438: upwind directions, where absorption from the hydrogen wall is
439: dominant. Any conclusions drawn from downwind directions, where
440: heliosheath absorption is more prominent, will be far more
441: tentative. The primary reason for this is that the grid used for
442: our current models only extends 500~AU from the Sun,
443: which is not nearly far enough to capture all of the heliosheath
444: absorption for $\theta\gtrsim 120^{\circ}$.
445: Thus, the absorption predictions shown in
446: Figures~3 and 4 for these directions will
447: underestimate the amount of absorption that the models would
448: really predict if the grid were extended further downwind.
449:
450: Many of the models seem to predict too much absorption downwind
451: even with the limited grid extent, particularly in the velocity
452: range of $80-120$ km~s$^{-1}$. But there is another potential
453: difficulty with downwind absorption that concerns the treatment of
454: the plasma in the models. Although a fully kinetic treatment is
455: applied to the neutrals, the plasma is assumed to be a single Maxwellian
456: fluid throughout the heliosphere. However, this is clearly a poor
457: approximation, since pickup ions, for example, have non-Maxwellian
458: velocity distributions and are not thermalized with the solar wind
459: inside the termination shock \citep[e.g.,][]{gg04}.
460: \citet{ygm06} have replaced the simple single-fluid plasma
461: treatment in the 2D Baranov \& Malama code with a complex multi-component
462: representation of the plasma. This more sophisticated plasma treatment
463: does not result in significantly different hydrogen wall absorption in
464: upwind directions, but we have found that is {\em does} result
465: in a significant reduction in heliosheath absorption which can
466: potentially alleviate the problems these models have in predicting
467: too much downwind absorption \citep{bew07}. The most meaningful
468: comparison with the data in downwind directions would therefore
469: require that our 3D models also utilize such a multi-component
470: plasma treatment, as well as having grids that extend far enough
471: downwind to capture all the heliosheath absorption. We leave such
472: computationally intensive modeling for a future paper.
473:
474: We note that asymmetries in the heliospheric structure induced by
475: the ISM field are evident in the downwind absorption predicted by
476: the models. For example, there is significant model dependence in the
477: absorption towards DK~UMa ($\theta=112^{\circ}$), but no significant
478: model dependence towards $\tau$~Cet ($\theta=123^{\circ}$). Since
479: $\theta$ is similar for these stars but their azimuthal angles quite
480: different (see Fig.~2), this difference in behavior must be due to
481: azimuthal variability, which can only be due to magnetic field
482: induced asymmetries (see \S3.3).
483:
484: Finally, we note that \citet{nvp06} have also developed
485: 3D MHD heliospheric models that treat neutrals in a self-consistent
486: manner with the plasma. This code uses a less sophisticated 2-fluid
487: treatment for the neutral H velocity distributions, but it includes the
488: effects of the interplanetary magnetic field on heliospheric structure
489: as well as the ISM field. Absorption has been computed for a limited
490: number of these models, yielding results qualitatively similar to
491: those reported here, with absorption decreasing with both $\alpha$ and
492: the magnetic field strength \citep{bew06}. Like \citet{vi05},
493: \citet{nvp06} demonstrate that the 3D MHD
494: models can potentially reproduce the shift between the H and He flows
495: observed within the solar system \citep{rl05}, though they
496: emphasize that the magnitude of the shift depends not only on the
497: strength and orientation of the magnetic field, but also on the ISM
498: neutral hydrogen density.
499:
500: \subsection{Quantifying Expected Absorption Asymmetries}
501:
502: Figure~1 illustrates that even a modest ISM field can result
503: in a heliospheric structure that is significantly asymmetric,
504: consistent with other models that also predict asymmetries of this
505: sort \citep{rr98,nvp06}.
506: In the bottom half of Figure~1, the heliosheath (between the TS and HP)
507: is narrower than in the upper half, but the hydrogen wall (between the HP
508: and BS) is wider. One might imagine that this would result in
509: corresponding Ly$\alpha$ absorption asymmetries. In other words, the
510: Ly$\alpha$ absorption should be azimuthally dependent as well as $\theta$
511: dependent.
512:
513: Some evidence that the models do indeed predict azimuthally
514: dependent absorption in downwind directions is mentioned in \S3.2.
515: However, in order to properly quantify the degree of absorption asymmetry
516: expected based on the models, it is necessary to compare absorption
517: predictions for lines of sight with identical $\theta$ values but different
518: azimuthal angles ($\phi$). This cannot be easily done with the set of
519: observed directions in Figures~3--4, which are scattered randomly
520: about the sky.
521:
522: Thus, in Figure~6 we show the heliospheric Ly$\alpha$
523: absorption predicted by the $B=2.5$~$\mu$G, $\alpha=45^{\circ}$
524: model for various $\phi$ angles, with $\theta$ fixed in the five
525: panels of the figure. The azimuthal angle is defined such that
526: $\phi=0^{\circ}$ and $\phi=180^{\circ}$ are in the plane of the
527: ISM magnetic field. This is the plane in which the
528: heliospheric structure is portrayed in Figure~1. The $\phi=0^{\circ}$
529: direction would be associated with the upper half of Figure~1,
530: and $\phi=180^{\circ}$ would be associated with the lower half.
531:
532: In general, the hydrogen wall will be responsible
533: for the steep, saturated absorption edges of the absorption
534: profiles in Figure~6, which are particularly prominent in upwind
535: directions (e.g., located at 20--30 km~s$^{-1}$ in the
536: $\theta=30^{\circ}$ panel), while the heliosheath is responsible
537: for the broad, unsaturated absorption wings that extend to high
538: velocities, which become more prominent in downwind directions.
539: Very little $\phi$ dependence is apparent in upwind directions.
540: This is a bit surprising given the hydrogen wall asymmetries
541: apparent in Figure~1. However, it turns out that azimuthal
542: density variations in the hydrogen wall offset the azimuthal width
543: dependence. For example, although the hydrogen wall is narrower
544: for $\phi=0^{\circ}$ (corresponding to the upper half of Fig.~1)
545: than for $\phi=180^{\circ}$ (corresponding to the lower half of
546: Fig.~1), this is offset by higher hydrogen wall densities in the
547: $\phi=0^{\circ}$ direction, so integrated H~I column densities are
548: actually not very different.
549:
550: Figure~6 shows that for $\theta=60-90^{\circ}$, the particularly
551: narrow hydrogen wall at $\phi=0^{\circ}$ results in somewhat less
552: hydrogen wall absorption than other directions, but the difference is so
553: small it would be very difficult to detect in practice. In contrast, the
554: heliosheath is thicker at $\phi=0^{\circ}$ and Figure~6 shows that
555: this does in fact lead to the broad heliosheath
556: absorption being significantly stronger in this direction than others,
557: though it is only in downwind directions ($\theta>90^{\circ}$) where
558: this azimuthal dependence becomes potentially detectable. As discussed
559: in \S3.2, actual comparisons with the data are currently problematic
560: in these directions. The asymmetries seen for the $\theta=90^{\circ}$
561: and $\theta=120^{\circ}$ surely extend to $\theta=150^{\circ}$ as well,
562: but Figure~6 does not show this due to the limited grid extent of the
563: models. According to Figure~6, the
564: heliosheath absorption is at a minimum in directions normal to the plane
565: of the ISM field ($\phi=90^{\circ}$ and $\phi=270^{\circ}$), indicative
566: of magnetic compression of the heliosphere in those directions,
567: leading to shorter distances through the heliosheath for those lines
568: of sight.
569:
570: \section{SUMMARY}
571:
572: We have compared H~I Ly$\alpha$ absorption profiles predicted
573: by 3D kinetic MHD models of the heliosphere with a large
574: selection of Ly$\alpha$ lines observed by HST, including many
575: lines of sight with detected heliospheric absorption. The primary
576: purpose of this comparison is to assess the sensitivity of the
577: predicted absorption to changes in the ISM magnetic field
578: properties assumed in the model. Our results are as follows:
579: \begin{description}
580: \item[1.] We find that the H~I Ly$\alpha$ absorption has a modest
581: sensitivity to both the strength and orientation of the ISM
582: magnetic field. Focusing on upwind directions where most of the
583: HST detections of heliospheric absorption reside, the models
584: presented here with $B=1.25-2.5$~$\mu$G and $\alpha=15-45^{\circ}$
585: appear to provide the best fits to the data, consistent with
586: constraints from other sources \citep{gg97,vi05,rl05,mo06}.
587: \item[2.] However, since the Ly$\alpha$ absorption is sensitive to other
588: model input parameters, such as the ISM H~I density, which have not
589: been varied here, the region of parameter space that yields acceptable
590: fits to the data will be complex. It will be very difficult,
591: perhaps impossible, for the Ly$\alpha$ absorption by itself to
592: yield a unique set of model input parameters that fit the data.
593: Nevertheless, the dependence of the absorption on many ISM
594: parameters means that the absorption does provide one constraint
595: on heliospheric models that is worthy of consideration in
596: assessing how precisely the models reproduce reality.
597: \item[3.] The models show that an ISM field that is skewed with respect
598: to the ISM flow vector results in substantial azimuthal asymmetries in
599: the heliospheric hydrogen wall. Surprisingly, these asymmetries do not
600: result in significant asymmetries in Ly$\alpha$ absorption from the
601: hydrogen wall, since densities within the wall vary in such a way as to
602: cancel out the effects of the spatial asymmetries on hydrogen wall
603: column densities.
604: \item[4.] The models also show that a skewed ISM field results in
605: significant azimuthal asymmetries in the heliosheath, and unlike
606: the hydrogen wall these asymmetries {\em do} yield significant azimuthal
607: absorption dependence, at least in downwind directions where the
608: heliosheath absorption is prominent. These directions are clearly the
609: best places to look for azimuthal dependences in Ly$\alpha$ absorption,
610: but there are problems with doing this in practice. One is simply
611: that we have few downwind detections of heliospheric Ly$\alpha$
612: absorption. Another is that the heliosheath absorption that dominates
613: in downwind directions should ideally be modeled using a complex
614: multi-fluid plasma treatment. And finally, the model grid must
615: be extended a much longer distance downwind than present models to
616: capture all the heliosheath absorption in these directions. We hope
617: to perform such computationally intensive modeling in the future.
618: \end{description}
619:
620: \acknowledgments
621:
622: This work was supported by
623: NASA grant NNG05GD69G to the University of Colorado. V. I. was also
624: supported by RFBR grant 04-02-16559, the ``Dynastia'' Foundation, and
625: the ``Foundation in Support of Russian Science''.
626:
627: \clearpage
628:
629: \begin{thebibliography}{}
630:
631: \bibitem[Baranov(1990)]{vbb90}
632: Baranov, V. B. 1990, Space Sci. Rev., 52, 89
633: \bibitem[Baranov et al.(1991)]{vbb91}
634: Baranov, V. B., Lebedev, M. G., \& Malama, Y. G. 1991, 375, 347
635: \bibitem[Baranov \& Malama(1993)]{vbb93}
636: Baranov, V. B., \& Malama, Y. G. 1993, J.~Geophys.~Res., 98, 15157
637: \bibitem[Baranov \& Malama(1995)]{vbb95}
638: Baranov, V. B., \& Malama, Y. G. 1995, J.~Geophys.~Res., 100, 14755
639: \bibitem[Baranov \& Izmodenov(2006)]{vbb06}
640: Baranov, V. B., \& Izmodenov, V. V. 2006, Fluid Dynamics, 41,
641: in press
642: \bibitem[Breitschwerdt(2001)]{db01}
643: Breitschwerdt, D. 2001, Ap\&SS, 276, 163
644: \bibitem[Cox \& Helenius(2003)]{dpc03}
645: Cox, D. P., \& Helenius, L. 2003, ApJ, 583, 205
646: \bibitem[Cravens(2000)]{tec00}
647: Cravens, T. E. 2000, ApJ, 532, L153
648: \bibitem[Florinski et al.(2004)]{vf04}
649: Florinski, V., Pogorelov, N. V., Zank, G. P., Wood, B. E., \& Cox, D. P.
650: 2004, ApJ, 604, 700
651: \bibitem[Gayley et al.(1997)]{kgg97}
652: Gayley, K. G., Zank, G. P., Pauls, H. L., Frisch, P. C., \& Welty, D. E.
653: 1997, ApJ, 487, 259
654: \bibitem[Gloeckler et al.(1997)]{gg97}
655: Gloeckler, G., Fisk, L. A., \& Geiss, J. 1997, Nature, 386, 374
656: \bibitem[Gloeckler \& Geiss(2004)]{gg04}
657: Gloeckler, G., \& Geiss, J. 2004, Adv. Space Res., 34, 53
658: \bibitem[Gringauz et al.(1960)]{kig60}
659: Gringauz, K. I., Bezrukikh, V. V., Ozerov, V. D., Rybchinskii, R. E.
660: 1960, Sov.~Akad.~Doklady, 131, 1301
661: \bibitem[Holzer(1989)]{teh89}
662: Holzer, T. E. 1989, ARA\&A, 27, 199
663: \bibitem[Izmodenov et al.(2001)]{vi01}
664: Izmodenov, V.V., Gruntman, M., Malama, Y.G., J. Geophys. Res. 106 , 10681
665: \bibitem[Izmodenov et al.(2005)]{vi05}
666: Izmodenov, V., Alexashov, D., \& Myasnikov, A. 2005, A\&A, 437, L35
667: \bibitem[Izmodenov et al.(1999)]{vvi99}
668: Izmodenov, V. V., Lallement, R., \& Malama, Y. G. 1999, A\&A, 342, L13
669: \bibitem[Izmodenov et al.(2002)]{vvi02}
670: Izmodenov, V. V., Wood, B. E., \& Lallement, R. 2002, J.~Geophys.~Res.,
671: 107, 1308
672: \bibitem[Izmodenov \& Alexashov(2006)]{vvi06}
673: Izmodenov, V. V., \& Alexashov, D. B. 2006, in The Physics of
674: the Inner Heliosheath: Voyager Observations, Theory, and Future Prospects,
675: ed. J. Heerikhuisen, et al. (Melville, NY: AIP), 14
676: \bibitem[Jenkins(2002)]{ebj02}
677: Jenkins, E. B. 2002, ApJ, 580, 938
678: \bibitem[Koutroumpa et al.(2006)]{dk06}
679: Koutroumpa, D., Lallement, R., Kharchenko, V., Dalgarno, A., Pepino, R.,
680: Izmodenov, V., \& Qu\'{e}merais, E. 2006, A\&A, 460, 289
681: \bibitem[Lallement(2004a)]{rl04a}
682: Lallement, R. 2004a, A\&A, 418, 143
683: \bibitem[Lallement(2004b)]{rl04b}
684: Lallement, R. 2004b, A\&A, 422, 391
685: \bibitem[Lallement \& Bertin(1992)]{rl92}
686: Lallement, R., \& Bertin, P. 1992, A\&A, 266, 479
687: \bibitem[Lallement et al.(2005)]{rl05}
688: Lallement, R., Qu\'{e}merais, E., Bertaux, J. L., Ferron, S.,
689: Koutroumpa, D., \& Pellinen, R. 2005, Science, 307, 1447
690: \bibitem[Lallement et al.(2003)]{rl03}
691: Lallement, R., Welsh, B. Y., Vergely, J. L., Crifo, F., \& Sfeir, D. M.,
692: 2003, A\&A, 411, 447
693: \bibitem[Malama et al.(2006)]{ygm06}
694: Malama, Y. G., Izmodenov, V. V., \& Chalov, S. V. 2006, A\&A, 445, 693
695: \bibitem[M\"{o}bius et al.(2004)]{em04}
696: M\"{o}bius, E., et al. 2004, A\&A, 426, 897
697: \bibitem[Neugebauer \& Snyder(1962)]{mn62}
698: Neugebauer, M., \& Snyder, C. 1962, Science, 138, 1095
699: \bibitem[Opher et al.(2006)]{mo06}
700: Opher, M., Stone, E. C., \& Liewer, P. C. 2006, ApJ, 640, L71
701: \bibitem[Parker(1961)]{enp61}
702: Parker, E. N. 1961, ApJ, 134, 20
703: \bibitem[Parker(1963)]{enp63}
704: Parker, E. N. 1963, Interplanetary Dynamical Processes (New York:
705: Interscience)
706: \bibitem[Pogorelov \& Zank(2006)]{nvp06}
707: Pogorelov, N. P., \& Zank, G. P. 2006, ApJ, 636, L161
708: \bibitem[Rand \& Kulkarni(1989)]{rjr89}
709: Rand, R. J., \& Kulkarni, S. R. 1989, ApJ, 343, 760
710: \bibitem[Ratkiewicz et al.(1998)]{rr98}
711: Ratkiewicz, R., Barnes, A., Molvik, G. A., Spreiter, J. R., Stahara, S. S.,
712: Vinokur, M., \& Venkateswaran, S. 1998, A\&A, 335, 363
713: \bibitem[Redfield \& Linsky(2004)]{sr04}
714: Redfield, S., \& Linsky, J. L. 2004, ApJ, 613, 1004
715: \bibitem[Smith et al.(2006)]{rks06}
716: Smith, R. K., et al. 2006, PASJ, in press
717: \bibitem[Snowden et al.(1998)]{sls98}
718: Snowden, S. L., Egger, R., Finkbeiner, D. P., Freyberg, M. J., \&
719: Plucinsky, P. P. 1998, ApJ, 493, 715
720: \bibitem[Stone et al.(2005)]{ecs05}
721: Stone, E. C., Cummings, A. C., McDonald, F. B., Heikkila, B. C., Lal, N.,
722: and Webber, W. R. 2005, Science, 309, 2017
723: \bibitem[Witte (2004)]{mw04}
724: Witte M. 2004, A\&A, 426, 835
725: \bibitem[Wood et al.(2007)]{bew07}
726: Wood, B. E., Izmodenov, V. V., Linsky, J. L., \& Malama, Y. G. 2007, ApJ,
727: in press
728: \bibitem[Wood et al.(2006)]{bew06}
729: Wood, B. E., Izmodenov, V. V., \& Pogorelov, N. V. 2006, in The Physics of
730: the Inner Heliosheath: Voyager Observations, Theory, and Future Prospects,
731: ed. J. Heerikhuisen, et al. (Melville, NY: AIP), 335
732: \bibitem[Wood et al.(2000)]{bew00}
733: Wood, B. E., M{\"u}ller, H. -R., \& Zank, G. P. 2000, ApJ, 542, 493
734: \bibitem[Wood et al.(2005a)]{bew05a}
735: Wood, B. E., M\"{u}ller, H. -R., Zank, G. P., Linsky, J. L., \&
736: Redfield, S. 2005a, ApJ, 628, L143
737: \bibitem[Wood et al.(2005b)]{bew05b}
738: Wood, B. E., Redfield, S., Linsky, J. L., M\"{u}ller, H. -R., \&
739: Zank, G. P. 2005b, ApJS, 159, 118
740: \bibitem[Zank(1999)]{gpz99}
741: Zank, G. P. 1999, Space Sci. Rev., 89, 413
742:
743: \end{thebibliography}
744:
745: \clearpage
746:
747: \begin{deluxetable}{ccccc}
748: \tabletypesize{\normalsize}
749: \tablecaption{Model ISM Field Orientations}
750: \tablecolumns{5}
751: \tablewidth{0pt}
752: \tablehead{
753: \colhead{$\alpha$ (deg)} & \multicolumn{2}{c}{Ecliptic Coord.} &
754: \multicolumn{2}{c}{Galactic Coord.} \\
755: \colhead{} & \colhead{l$_{e}$ (deg)} & \colhead{b$_{e}$ (deg)} &
756: \colhead{l (deg)} & \colhead{b (deg)}}
757: \startdata
758: %0 & 254.7 & 5.2 & 3.3 & 15.9 \\
759: %15 & 246.9 & 18.1 & 11.9 & 28.6 \\
760: %30 & 237.8 & 30.6 & 22.7 & 40.6 \\
761: %45 & 226.1 & 42.4 & 38.2 & 51.2 \\
762: %60 & 209.5 & 52.4 & 61.0 & 58.7 \\
763: %90 & 155.8 & 59.6 & 118.2 & 56.0 \\
764: 0 & 74.7 & $ -5.2$ & 183.3 & $-15.9$ \\
765: 15 & 66.9 & $-18.1$ & 191.9 & $-28.6$ \\
766: 30 & 57.8 & $-30.6$ & 202.7 & $-40.6$ \\
767: 45 & 46.1 & $-42.4$ & 218.2 & $-51.2$ \\
768: 60 & 29.5 & $-52.4$ & 241.0 & $-58.7$ \\
769: 90 & 335.8 & $-59.6$ & 298.2 & $-56.0$ \\
770: \enddata
771: \end{deluxetable}
772:
773: \clearpage
774:
775: \begin{figure}[p]
776: %\plotone{/home/origins/woodb/props/sh02/fig2.eps}
777: \plotone{f1.eps}
778: \caption{The locations of the termination shock (TS), heliopause (HP),
779: and bow shock (BS) for a model including a $B=2.5$ $\mu$G ISM field
780: (solid lines), and a model with no ISM magnetic field (dashed lines).
781: The directions of the LIC flow vector ($V_{LIC}$) and magnetic field
782: ($B_{LIC}$) are indicated. The distance scale is in AU. The region
783: between the HP and BS is sometimes called the ``hydrogen wall'' and
784: the region between the TS and HP is the ``heliosheath.''}
785: \end{figure}
786:
787: \begin{figure}[p]
788: %\plotone{/home/origins/woodb/props/sh02/fig2.eps}
789: \plotone{f2.eps}
790: \caption{Sky map in ecliptic coordinates of all HST-observed lines
791: of sight with useful Ly$\alpha$ spectra. The numbered symbols indicate
792: spectra that we will compare with model predictions of Ly$\alpha$
793: absorption (see Figs.~3 and 4). Boxes indicate lines of
794: sight with detected heliospheric absorption. The plus signs and
795: diamonds are both lines of sight with nondetections of heliospheric
796: absorption. The diamonds indicate lines of sight selected to
797: provide upper limits for absorption in those directions. The filled
798: and open circles indicate the upwind and downwind directions of the
799: local ISM flow vector, respectively.}
800: \end{figure}
801:
802: \begin{figure}
803: \plotone{f3.eps}
804: \caption{The red side of the H~I Ly$\alpha$ absorption line (histogram)
805: for the selected stars from Fig.~2, where the stars are placed in order
806: of increasing angle from the upwind direction of the ISM flow ($\theta$).
807: In each panel, the dotted green line is the ISM absorption alone.
808: Absorption predictions are shown for heliospheric models computed
809: assuming six different ISM field orientations, as quantified by
810: $\alpha$, the angle between the field and the ISM flow direction
811: (see the 36~Oph panel for line identifications).
812: For many downwind lines of sight ($\theta > 110^{\circ}$),
813: dashed lines show upper limits to the amount of absorption that can
814: be present --- absorption predictions from the models must lie above
815: these limits to be consistent with the data. For the three most
816: downwind lines of sight, the shaded regions indicate the amount of
817: absorption that the models {\em should} predict if the real stellar
818: Ly$\alpha$ profile is centered on the stellar rest frame rather
819: than blueshifted as suggested by the original recontructed profile.
820: For these lines of sight, the absorption predicted by the models
821: should not fit the data (which are dotted histograms in these cases)
822: but should instead fall within the shaded regions (see \S2).}
823: \end{figure}
824:
825: \begin{figure}
826: \plotone{f4.eps}
827: \caption{A figure analogous to Fig.~3, but in this case the absorption
828: predictions are for three $\alpha=45^{\circ}$ models that assume
829: different ISM magnetic field strengths (see the 36~Oph panel for line
830: identifications).}
831: \end{figure}
832:
833: \begin{figure}
834: \plotone{f5.eps}
835: \caption{For six upwind lines of sight, the predicted wavelength-integrated
836: Ly$\alpha$ flux absorbed by heliospheric H~I beyond that absorbed by the
837: ISM is computed for the six models from Fig.~3 and plotted versus
838: $\alpha$, the ISM field orientation relative to the ISM flow direction.
839: The fluxes are normalized to the observed amount of flux absorbed,
840: so in each panel a flux of 1 (dashed lines) corresponds to agreement
841: with the data.}
842: \end{figure}
843:
844: \begin{figure}
845: \plotone{f6.eps}
846: \caption{An illustration of the directional dependence of
847: H~I Ly$\alpha$ absorption predicted by a 3D MHD heliospheric model
848: assuming an ISM field strength and orientation of $B=2.5$~$\mu$G and
849: $\alpha=45^{\circ}$, respectively. Absorption is shown for five
850: values of the poloidal angle $\theta$ (the angle between the line
851: of sight and the upwind direction of the ISM flow), and eight
852: values of the azimuthal angle $\phi$ (where the plane of the ISM
853: magnetic field is in the $\phi=0^{\circ}$ and $\phi=180^{\circ}$
854: directions). The model grid does not extend far enough downwind
855: to properly search for azimuthal absorption variations at
856: $\theta=150^{\circ}$.}
857: \end{figure}
858:
859: \end{document}
860: