MODELING THE DIURNAL EVOLUTION OF A DUST FEATURE IN COMET HALE-BOPP (1995 O1) Zdenek Sekanina

Transcription

1 The Astrophysical Journal, 494:L121 L124, 1998 February The American Astronomical Society. All rights reserved. Printed in U.S.A. MODELING THE DIURNAL EVOLUTION OF A DUST FEATURE IN COMET HALE-BOPP (1995 O1) Zdenek Sekanina Jet Propulsion Laboratory, California Institute of Technology, 4800 Oak Grove Drive, Pasadena, CA Received 1997 May 9; accepted 1997 November 25; published 1998 January 30 ABSTRACT A Monte Carlo computer code for simulating dust features in comets is applied to comet Hale-Bopp in order to model the diurnal evolution of a bright jet observed, in late 1997 February, to point in the southwesterly direction from the nucleus. The jet s morphology is closely matched by a model, which is based on the assumption that the ejecta were released nearly continuously from an isolated source on the nucleus surface. The fitted characteristics include the initial direction of the ejecta s motion, the feature s overall dimensions and shape, and its temporal variations during the rotation cycle. In particular, it is found that the relatively faint appendage some directly to the south from the nucleus can be identified with the residual ejecta from the previous rotation cycle. No attempt has been made to fit the observed brightness changes, which apparently are affected by variable seeing, especially in daylight. It is suggested that if the system of concentric halos to the south from the nucleus consists of two independent branches, its southwesterly branch can be identified with the ejecta from the same jet that are several rotations old. Modeling of a jet s diurnal evolution presents a powerful tool in our quest for understanding the emission of dust from discrete active regions on cometary nuclei. Subject heading: comets: general comets: individual (Hale-Bopp 1995 O1) 1. INTRODUCTION In early 1997, the impressive list of morphological dust features displayed by comet Hale-Bopp (1995 O1) grew even longer, documenting the complexity of the comet s activity. In this Letter, I demonstrate the available capabilities of computer modeling by simulating the diurnal evolution of a widely observed, short, bright jet. On images taken between 1997 January 12 and February 10, Lecacheux, Jorda, & Colas (1997) detected a measurable change in the jet s orientation on a timescale of less than 2 hr with a recurrence period of hr (see also Sarmecanic et al. 1997). A system of nearly concentric halos (sometimes called rings, ripples, shells, waves, envelopes, hoods, etc.) was also observed in comet Hale-Bopp in early First detected in comet Donati 1858 L1 (Bond 1862) and other comets since, the halos have been identified with the sharp outer boundaries of conical sheets of dust ejected from discrete sources on rapidly rotating cometary nuclei (Sekanina & Larson 1984; Sekanina 1987, 1991). Although an investigation of halos is not a prime objective of this Letter, a likely association of one of their branches with the modeled jet is proposed. 2. MONTE CARLO SIMULATION OF DUST MORPHOLOGY Jet morphology is modeled in this study by applying a Monte Carlo computer simulation code, which generates synthetic images of dust ejecta released from discrete sources on a rotating nucleus. Brief descriptions of this technique can be found elsewhere (Sekanina 1996a, 1996b and references therein). A dust-morphology model is defined by its rotation and dustemission parameters. If significant changes in the rotation vector occur over a period of 20 days (e.g., Sekanina 1996a; Jorda, Lecacheux, & Colas 1997), the assumed pure spin should satisfactorily approximate the comet s state of rotation during a fraction of a day. The spin-axis orientation is described by two Eulerian angles, the argument F and the obliquity I (e.g., Sekanina 1981a), or by the coordinates of the north rotation pole (from which the nucleus is seen to spin counterclockwise). L121 The spin period for comet Hale-Bopp is from Jorda et al. (1997). The dust ejecta in the observed features are assumed to have been released from one or more discrete sources activated from local sunrise to sunset. The source locations, the parameters of their activity, and the rotation constants are derived by fitting the features on the observed images. Since photometric information is not considered at this time, some of the dust-ejecta parameters are not tightly constrained. For this reason, a constant rate of dust production between sunrise and sunset is being assumed, an adequate first approximation. Limits on the particle acceleration b by solar radiation pressure (a dimensionless quantity in units of the Sun s attraction) are only crudely estimated from the recognized properties of the specific features ( 3 and 4). From experience with other comets, the peak b-value is probably near 2.5 (e.g., Sekanina & Farrell 1980, 1982; Akabane 1983; Lamy 1986; Beisser & Boehnhardt 1987; Nishioka & Watanabe 1990). However, on short timescales, particle motions are determined primarily by the ejection velocity v eject, which is assumed to be related to the acceleration b by 1 1/2 veject A Bb, (1) where A and B are constants whose meaning is extensively discussed in the studies of comets 109P/Swift-Tuttle (Sekanina 1981b) and 1P/Halley (Sekanina & Larson 1984, 1986). The values of A and B, which depend on the nuclear size and the gas and dust emission parameters, can be calculated from the adopted limits on b and v eject. Since all observed images of dust features contain noise (both of circumstantial nature, such as due to imperfect seeing, and of physical nature, such as due to dispersion effects in the vector field of particle velocities) that cannot be removed, noise is added to the synthetic images (Sekanina 1991) in order to simulate better the appearance of the observed images. 3. DIURNAL EVOLUTION OF THE BRIGHT JET Monitoring the bright jet on 1997 February 28 ( 1), Jorda et al. (1997) used the 105 cm telescope of the Pic-du-Midi

2 L122 SEKANINA Vol. 494 TABLE 1 Rotation and Dust-Emission Parameters Used to Model Evolution of Bright Jet Parameter Rotation Value Argument F of subsolar meridian at perihelion Obliquity I of orbital plane to equator Right ascension of north rotation pole a Declination of north rotation pole a Angle between spin vector and Earth s direction during Feb. 28 observation Adopted sidereal period of rotation b h 15 m Dust Source Cometocentric latitude Cometocentric longitude of subsolar meridian at perihelion... Local hour angle of Sun at onset/end of dust emission (sunrise/sunset)... Sun s elevation at local noon Duration of emission during one rotation... Dust Ejecta h 50 m Power index of reciprocal differential particle size distribution law Particle acceleration b (units of Sun s attraction): Lower limit b min Upper limit b max Particle ejection velocity v (m s 1 eject ): Lower limit ( v eject ) min Upper limit ( v eject ) max a Equinox b Based on information from Jorda et al Observatory with a CCD detector and a broadband Gunn z filter (whose peak transmission is at 1.02 mm; cf. Schneider, Gunn, & Hoessel 1983). Their observing run, from 3:50 until 15:35 UTC (i.e., mostly in broad daylight), covered an entire rotation cycle. 1 A set of selected images from the animation sequence, spaced at approximately equal steps of 30 minutes in time, is presented in the left panel of Figure 1 (Plate L8). At the level of detail seen on the animation sequence (consistent with Sarmecanic et al. s 1997 description), the jet first appeared in a position angle of 240, growing gradually longer, rotating clockwise toward the south, and terminating near a position angle of 180. A new jet then began to evolve in 240 after the onset of the next cycle of activity. The jet s rotational motion, its gradual development from an initial short rectilinear wisp pointing to the west-southwest (in rows 1 and 2 of the left panel of Fig. 1) into a check-sign like feature (in rows 3 and 4), and the persisting nebulous appendage to the south, prominently displayed in rows 2 4, unquestionably represent the intrinsic attributes of the ejecta that should serve as constraints and must be satisfied by any viable model. On the other hand, judging from the observers comment on the varying signal-to-noise ratio in the course of the observing run, the rapid brightening of the feature in the second row from the foot and the subsequent sudden fading in the last row may primarily be artifacts of the Earth atmosphere s uneven transparency; hence, their modeling has not been attempted. Experimentation with the simulation code has shown that the constraints established from the jet s evolving appearance are restrictive to the extent that changes in the model s angular parameters as small as 5 10 yield perceptible differences in the quality of fit. Since activity, including dust ejection, from 1 At the time of publication, an animation of the jet s diurnal evolution is presented at a Web site URL halebopp.html#fevriere, and a copy of it is also at comet/anim/picanim.gif. discrete emission areas is known to be confined to the sunlit side of the nucleus (Keller et al. 1987), the activation of the source has been assumed to coincide with local sunrise. The parameters for the best model, derived by trial and error, are listed in Table 1. The jet s evolution is found to be determined by ejection velocities (of up to almost 600 m s 1 ) of micronand submicron-sized particles. While having no measurable effect on the synthetic images, the lower limit to the particle ejection velocities has been taken to be 7 m s 1, the velocity of escape from the nucleus of an assumed bulk density of 0.2 g cm 3 and 40 km in diameter (Weaver et al. 1997). The jet s modeled diurnal evolution is displayed in the right panel of Figure 1. The individual synthetic images, arranged in a oneto-one correspondence with the frames in the left panel of Figure 1, are identified by the time elapsed since activation. My best guess for the activation time in the course of the monitoring period is 12:05 UTC (as observed at Earth). With a spin period of hr, the source was previously activated at 0:50 UTC, so that the monitoring began 3 hr into the cycle. The image in the upper left corner of either panel in Figure 1 then refers to the time 12:15 UTC (and also to 1:00 UTC one rotation earlier), whereas the image in the lower right corner refers to 13:00 UTC. Because of the jet s short length in the early phase of its development, no trace of it can in fact be detected for some time after its activation. The source is found to have passed through the subsolar meridian at 5:45 UTC, and its activity is calculated to have terminated at 10:40 UTC (the time of local sunset), or 9 h 50 m after activation. The source s activity is thus calculated to have extended over 87% of the rotation cycle. The right panel of Figure 1 demonstrates that the characteristic properties of the observed jet are indeed well reproduced by the model. A schematic diagram that illustrates the basic geometry of the modeled jet as seen from Earth is presented in Figure 2. In order to recover information on clean dustejecta trajectories, one needs to inspect the deterministic (noise-

3 No. 2, 1998 MODELING DIURNAL EVOLUTION OF DUST FEATURE L123 Fig. 2. A schematic diagram illustrating the proposed model for the diurnal evolution of the dust jet observed to emanate from the nucleus of comet Hale- Bopp on 1997 February 28. The nucleus (in reality, of unknown shape) is shown with the calculated position of its spin vector and with the jet s rotating source at a derived cometocentric latitude of 55 (dotted ellipse). The projection plane makes an angle of 16 with the spin vector and 49 with the direction toward the Sun, which is above the plane. At the time of observation, the comet had 32 days to reach its perihelion, and it was 1.50 AU from Earth and 1.08 AU from the Sun. less) images. For three particular times, the noiseless images and the images with noise added are shown side by side in Figure 3. It is noted, for example, that the jet to the south of the nucleus on the frame taken 1 h 10 m after activation is caused by clustering of particle loci from afternoon emissions during the previous rotation. On the other hand, morning emissions from that same rotation are responsible for the extension to the west-southwest of the nucleus. Emissions that are two rotations old are predicted to have made up a nebulous appendage farther to the south, but its surface brightness was apparently too low to show on the observed images. Fig. 3. Comparison of the computer-generated images without noise (right) and with noise added (left) for three times selected from the right panel of Fig. 1. The orientation and the frame size are the same as in Fig. 1. The morning (early postsunrise) ejecta are located near a position angle of 240, the afternoon ejecta are toward the south of the nucleus, and the evening (late presunset) ejecta are near a position angle of RELATION BETWEEN THE JET AND A SYSTEM OF NEARLY CONCENTRIC HALOS Extrapolation of the model for the bright jet ( 3) to a period of time spanning several rotation cycles reveals that before merging with other emissions in the dust tail, the particle ejecta (especially submicron, rapidly accelerating grains) from the discrete source make up a sequence of halos. Prominent dust halos were indeed observed to extend to both the southeast and the southwest from the nucleus, as shown, for example, on the Fig. 4. A computer-generated image of the halo system in the southwestern quadrant in late 1997 February, modeled as remnants of the bright jet in the left panel of Fig. 1, made up of residual ejecta from earlier rotations. The field is 180 by 90 (204,000 km by 102,000 km at the comet). North is up, east to the left.

4 L124 SEKANINA Vol. 494 three images at the foot of page 32 in the article by Aguirre (1997). A computer-generated image of a system of halos that I display in Figure 4 suggests that the halos to the southwest are indeed mere remnants of old jets observed a few rotations earlier. If so, however, the southeastern branch of the halos must have a different origin, presumably deriving from another jet formation. This interpretation thus tacitly assumes that the two halo branches are separate, independent features. To conclude, the applied Monte Carlo image simulation technique offers very encouraging results; it provides a highly satisfactory representation of the jet s diurnal evolution and suggests that remnants from jet activity over several rotations make up a halo system that is observed a few cycles later. I thank L. Jorda, J. Lecacheux, and F. Colas for permission to reproduce selected images from the animation sequence of their observations of the comet on 1997 February 28. I also thank H. Boehnhardt and S. M. Larson for their comments on a draft of this Letter. This research has been carried out at the Jet Propulsion Laboratory, California Institute of Technology, under contract with the National Aeronautics and Space Administration. REFERENCES Aguirre, E. L. 1997, S&T, 93(5), 28 Akabane, T. 1983, PASJ, 35, 565 Beisser, K., & Boehnhardt, H. 1987, Ap&SS, 139, 5 Bond, G. P. 1862, Ann. Astron. Obs. Harvard Coll., 3, 1 Jorda, L., Lecacheux, J., & Colas, F. 1997, IAU Circ Keller, H. U., et al. 1987, A&A, 187, 807 Lamy, P. 1986, Adv. Space Res., 5(12), 317 Lecacheux, J., Jorda, L., & Colas, F. 1997, IAU Circ Nishioka, K., & Watanabe, J.-i. 1990, Icarus, 87, 403 Sarmecanic, J., Osip, D. J., Fomenkova, M., & Jones, B. 1997, IAU Circ Schneider, D. P., Gunn, J. E., & Hoessel, J. G. 1983, ApJ, 264, 337 Sekanina, Z. 1981a, Annu. Rev. Earth Planet. Sci., 9, 113 Sekanina, Z. 1981b, AJ, 86, , in Diversity and Similarity of Comets, ed. E. J. Rolfe & B. Battrick (Noordwijk: ESTEC), , AJ, 102, a, A&A, 314, b, in ASP Conf. Ser. 104, Physics, Chemistry, and Dynamics of Interplanetary Dust, ed. B. Å. S. Gustafson & M. S. Hanner (San Francisco: ASP), 377 Sekanina, Z., & Farrell, J. A. 1980, AJ, 85, , AJ, 87, 1836 Sekanina, Z., & Larson, S. M. 1984, AJ, 89, , AJ, 92, 462 Weaver, H. A., et al. 1997, Science, 275, 1900

5 Fig. 1. Left panel: selected frames from the animation sequence of images of a bright jet in comet Hale-Bopp, taken by Jorda et al. (1997) between 3:50 and 15:35 UTC on 1997 February 28 with the 105 cm telescope of the Pic-du-Midi Observatory through a Gunn z filter. The observations cover a complete rotation period of the comet. The beginning of the sequence of images has been moved: the image in the upper left corner corresponds approximately to a time 10 minutes after jet activation, or to 12:15 UTC (as observed at Earth); one rotation earlier, it would have corresponded to 1:00 UTC. The temporal separation between the neighboring frames is about 30 minutes, so that the approximate UTC times of the leftmost images are 14:45 in the second row, 6:00 in the third, 8:30 in the fourth, and 11:00 in the last row. The image in the lower right corner refers to 13:00 UTC (and would again refer to 0:15 UTC on 1997 March 1), whereas the observing run actually terminated between the second and the third images in the second row. Each frame is 29.7 on a side, corresponding to 32,400 km at the comet. North is up, east to the left (for the direction to the Sun, see the right panel). Right panel: computer-generated images simulating the diurnal evolution of the bright jet from the left panel. There is a one-to-one correspondence between the frames in the two panels, with the times, the orientations, and the frame sizes being all identical. The neighboring frames again show changes in the jet in 30 minute intervals. The dot density is proportional to the integrated geometric cross sectional area of dust ejecta. It is assumed that dust is produced at a constant rate between sunrise and sunset and that no emission occurs between sunset and sunrise. The expansion velocities of the dominant microscopic-sized ejecta are up to almost 600 m s 1 (Table 1). The indicated times are reckoned from the estimated activation time of the emission source, or 0:50 UTC (as observed at Earth) on 1997 February 28. The projected directions of the spin vector (q) and the Sun (circled dot) are shown in the first frame. Sekanina (see 494, L122) PLATE L8