1 Introduction

Long-period comets, such as C/1995 O1 (Hale-Bopp), have long been thought to be responsible for some of the largest impact craters on Earth (e.g., Zimbelman 1984; Weissman 1990, 2007). They have a mean impact probability of 2.2 × 10−9 per perihelion passage if the perihelion distribution is uniform and the inclination distribution is random, accounting for a small but important fraction of the potential impacts on Earth. They can be big and tend to approach the Earth from all possible directions at random times, typically at a high velocity (56.4 km/s most probable value according to Weissman (2007)). They offer little advanced warning, except for a trail of dust particles in their orbit released during the previous return to the Sun. The Earth encounters these dust trails on occasion, causing brief 1–2 h meteor showers (Jenniskens et al. 1997).

Now, the imminent encounter with the dust trail of comet C/1911 N1 (Kiess) may teach us how long-period comets lose large dust grains, how to translate the observed dust trail crossings into physical data of the parent comet, and even to find more evidence for the hypothized “pristine crust” of a comet.

Since the confirmed detection of the predicted (Jenniskens 1995) outburst of 1995 alpha-Monocerotids from an unknown long-period comet and the subsequent Leonid storms, the basic physical principles behind these transient showers are understood (Kondrat’eva and Reznikov 1985; Jenniskens 1997; Jenniskens et al. 1997, McNaught and Asher 1999; Lyytinen 1999). Dust ejected from the parent comet is dispersed due to small differences in orbital period from ejection speed and radiation pressure, causing some particles to return earlier than others. Upon return, the thin stream of dust wanders in and out of Earth’s path due to planetary perturbations by the major planets, which work slightly differently on particles at different positions along the dust trail. A meteor shower outburst is observed only when the trail is in the Earth’s path at the very moment when Earth passes the node of the particles (for a review see Jenniskens 2006).

In the case of so called intermediate long-period comets such as Kiess, with orbital periods of 200–10,000 years, the trail is so much perturbed that the second revolution dust trail is dispersed beyond recognition. As a result, the outburst meteors observed in prior encounters of the dust trail in 1935, 1986, and 1994 all date from the last time (approximately 2,000 years ago) that the comet was near the Sun (Lyytinen and Jenniskens 2003). In 2007, that same dust trail will shower Earth again.

Until now, the alpha-Monocerotid shower is the only encounter with the dust trail of a long-period comet observed by modern instrumental techniques (Spurný et al. 1995; Rendtel 1995; Znojil and Hornoch 1995; Borovička and Spurný 1995; Rendtel et al. 1996; Langbroek 1996; Šimek 1996; Jenniskens and Docters van Leeuwen 1997; Jenniskens et al. 1997). Interestingly, these meteoroids were very unusual. They were found to be almost completely lacking in sodium (Borovička et al. 2002, 2005) and penetrated relatively deep in Earth’s atmosphere (Jenniskens et al. 1997). Presumably because material was sampled that came from a “pristine crust” caused by exposure to cosmic rays at the time of cold storage in the Oort cloud. Short period comets such as the parent of the Leonid shower have long lost this pristine crust.

We do not know if comet Kiess still had its pristine crust and lost some of it 2,000 years ago, perhaps now causing similar unusual meteors in 2007. It is interesting, however, that George Zay and Bob Lunsford, the only two visual meteor observers to witness the 1994 Aurigid outburst, described the outburst Aurigids as having a greenish or bluish look to them, while being more white outside this interval (Zay and Lunsford 1994). That suggests that the meteoroids produced unusually strong iron and magnesium atom line emissions from ablating metal atoms, relative to the air plasma emissions in the orange and red. This could point towards a different particle morphology of outburst Aurigids than those of other Aurigids seen outside the outburst, because the ratio of metal atom to air plasma emissions is a function of how the meteoroid matter is ablated.

2 Predictions for the 2007 Aurigid Shower

We investigated the distribution of dust from comet C/1911 N1 (Kiess) using a comet ejection model developed by Crifo and Rodionov (1997), based on the model by Whipple (1951), and calculated rigorously the planetary perturbations on the particles from the point of ejection until intersection with Earth’s orbit (for a full review of the method see Vaubaillon et al., 2005a, b). Earlier results were published in Jenniskens and Vaubaillon (2006, 2007a, b), each publication aimed at a different audience.

One million meteoroids in five bins of mass were ejected from the comet orbit in 83 BC, which is the perihelion time of the nominal comet orbit (Minor Planet Center comet orbit database; Marsden et al. 1978) when integrated backward in time. Forward integration, from 83 BC until the current perihelion return, confirms that planetary perturbations occur only on the inward leg. As a result, the overall motion of the dust trail is not sensitive to the adopted perihelion time of the comet in that previous return, although the precise position is.

All particles that are at the descending node ±2 months before Earth encountered this point are included in Fig. 1,which shows the point where each particle crossed the ecliptic plane. Also shown is the position of Earth in 1-h intervals (on the dates listed in Table 1 below). Relatively bright meteors responsible for visual meteors are shown as small dots. Large dots show the distribution of faint +6 magnitude meteors detected by video cameras and some radar systems.

Fig. 1
figure 1

Position of the node of the model 1-revolution Aurigid stream particles that are within 2 months from passing the ecliptic plane at the time of past Aurigid outbursts (grain diameter: • = 0.1–0.2 cm (faint ∼ +6 magnitude meteors); . = 0.2–2 cm (bright +3 to −3 magnitude meteors)

Full size image
Table 1 Calculated circumstances for the encounter with the 1-revolution (83 BC) trail of C/1911 N1 (Kiess) at the time of Aurigid outbursts
Full size table

Much of the shape of the distribution in Fig. 1 is the result of motion of the dust trail over a 4-month period. In 1935 and 1994, planetary perturbations caused the trail to move rapidly from outside to inside Earth orbit over the months around the outburst. In 1986, the trail moved from inside to outside Earth orbit. In contrast, the trail will be nearly stationary in 2007. The daily motion of the dust trail relative to Earth orbit can be removed by fitting a first or second order polynomial to the X and Y positions as a function of perihelion time and then interpolate each position back to the perihelion time corresponding to the particles encountered by Earth. This is justified, because the dust trail shows no gaps or strong density variations in the periods ±2 months from the encounter times.

The result is shown in Fig. 2. Table 1 summarizes the statistical data of the calculated encounters in their usual meaning (Jenniskens 2006), including the expected time of the peak, the width of the shower, the mean miss-distance Δ(E − D), the initial difference in semi-major axis of the meteoroid orbit relative to that of the comet (Δa), and the dilution factor f M. Observed parameters of past Aurigid showers are summarized in Table 2.

Fig. 2
figure 2

Position of the node of the model 1-revolution Aurigid stream particles after correcting for motion of the trail

Full size image
Table 2 Observed parameters of past Aurigid outbursts
Full size table
Table 3 Observed magnitude distributions (Teichgraeber 1935; Tepliczky 1987; Zay and Lunsford 1994)
Full size table

After correcting for the trail motion (Fig. 2), we find that the model predicts a trail position that is, within a fraction of the width of the trail, at the same distance from Earth orbit Δ(E − D) in all years. Because of that, we are confident that the shower will return based on it having been seen in 1935, 1986, and 1994.

One objective of future work is to understand why the model puts the trails always just inside Earth orbit, as noticed before in the early results by Lyytinen and Jenniskens (2003). In the absence of previous sightings of a shower, these effects make it difficult to predict meteor showers from other long-period comets. Some of that discrepancy could come on account of an uncertainty in the orbital period of the comet orbit, or it could be due to the specifics of comet dust ejection. The dust trails of Leonid parent comet 55P/Tempel-Tuttle, for example, were calculated with the same ejection model +0.00077 AU too far outward than observed (Jenniskens 2006, Fig. 15.33). Tempel-Tuttle’s orbit is well known. Hence, that discrepancy is thought to be due to ejection conditions being slightly different than those in the Crifo model. No effort has yet been made to improve the dust ejection model accordingly.

The observed peak time in past Aurigid encounters was off by −1 min in 1935, −16 min in 1986, and −7 min in 1994. Therefore, our best estimate for the peak time, 11:36 UT, has an uncertainty of about ±20 min. The predicted encounter time makes the shower favorable for viewing from the western states and provinces of the US and Canada, and western Mexico, where the radiant will be high in the sky just before dawn in the early morning of September 1 (Fig. 3).

Fig. 3
figure 3

Earth as seen from the perspective of the approaching dust grains at the peak of the predicted meteor outburst on September 1, 2007

Full size image

The density of particles in the stream in 2007 will be the nearly the same as in the 1986 and 1994 returns. Unfortunately, data of past outbursts were hampered by bad observing circumstances. Rates continued to rise when twilight interfered in Germany and the Czech Republic during the 1935 outburst (Teichgraeber 1935; Guth 1936), from which we have ZHR > 100/h. The single eyewitness of the 1986 outburst, Istvan Tepliczky of Hungary, derived an average ZHR = 39.6 ± 8.1 from the period between the first and last Aurigid (00:47–02:12 UT), during which 24 Aurigids were seen (Tepliczky 1987), which corresponds to a peak ZHR = 200 ± 25/h based on 10-min intervals. The rate measurement in 1994 was hampered by a low radiant elevation. For the hour between 7:22 and 8:22 UT, with the radiant being at 13° elevation at 7:49 UT, Zay and Lunsford (1994) calculated a ZHR = 55/h (Zay) and 37/h (Lunsford), respectively. Again, the rate varied strongly during that interval. In small 10-min intervals, we calculate a peak ZHR of again about 200 ± 25 per hour. Hence, meteor rates in 2007 are expected to increase to ZHR = 200/h in a short time interval at the center of the outburst depending on the exact position of the trail crossing.

Based on past Leonid storm observations, the width of the trail is expected to be wider if we pass further from the trail center (Jenniskens 2006). Given that the trail will be at the same location as in past returns, we can use the sparse data from past observations to predict what to expect from the 2007 encounter. Each of the past observed showers lasted about 1.5 h, with a Full Width at Half Maximum ∼28 min. For the center of the trail, our model predicts FWHM = 27.3 min for 1986 and 32.9 min for 1994, in good agreement. If Earth will pass through the trail center, then the 2007 return would have a FWHM duration of about 25 min.

In 2007, we will be only 15% further from the comet than in 1994. Our model (Fig. 4) shows that large particles can make it out to this position in the trail, and the model predicts that meteoroids as big as 20 cm in diameter (−10.6 magnitude, according to Eq. C.12 in Jenniskens 2006) may be observed and meteors down to magnitude +4 should be abundant. The model also predicts that a lack of relatively faint meteors (1 mm, or +6.6 magnitude meteors) be present in the stream, mainly because the smaller meteoroids are ejected at higher speed. Past magnitude estimates of the Aurigids are summarized in Table 3 and show that most were in the range −3 to +3 magnitude. We expect that to be so again in 2007. A Moon 4 days past full will not dampen the display much.

Fig. 4
figure 4

Particle distribution along the comet dust trail for different particle diameters. The location where Earth crossed the trail in 1935, 1986, 1994, and 2007 is indicated in the last panel. The vertical axis gives the number of particles in the model, which is a relative measure of the particle density along the dust trail

Full size image

3 Future Work

The short duration and the abundance of zero and +1 magnitude meteors will make for a very impressive shower. Accurate measurements of the trail width as a function of particle mass will help validate predictions in the model. Note how faint meteors are expected to have a wider stream cross section. The particle size distribution will calibrate the dispersion of meteoroids along the comet dust trail. The mass of the comet nucleus is a free parameter in the model and may be derived from these measurements. This mass can then be compared to mass estimates based on the brightness of comet Kiess in 1911.

Of interest too is the annual shower activity associated with comet Kiess, because the intensity of the shower and duration hold clues about how many revolutions the comet has completed since being captured. This Aurigid shower (IAU #206) is known from only three meteoroid orbits (Jenniskens 2006). Dubietis and Arlt (2002a, b) calculated a Zenith Hourly Rate curve with a peak of about 7 meteors/hour, but it is not clear whether the visual observers were sufficiently capable of discriminating the annual Aurigids from other apex source meteors at that time.

It will also be interesting to study the light curves, the penetration depth, and the spectra of the Aurigid meteors for clues about the presence of pristine comet crust material. It is not certain that the dust of Kiess contains such unusual meteoroids. If such meteors lacking in sodium are found, the material properties of this dust should be investigated to derive the density of the crust material and its main element composition. Unusual meteors should be looked for in both the outburst and the annual Aurigid component.

The fast meteors are an impact hazard to satellites in orbit (Beech and Brown 1993). At the peak of the Aurigid shower, the influx rate of fast meteoroids of several mm size will increase briefly by a factor of ∼100, more so for large particles and for spacecraft surfaces oriented towards the Earth’s apex, but the flux will not rise so high that an impact is certain, even considering the whole surface area of active satellites.

4 Added in Proof

These predictions were presented at the Meteoroids 2007 conference in June. Results were published in a paper in EOS, Transactions of the AGU on August 7 (Jenniskens and Vaubaillon 2007b). Subsequently, JPL issued a new orbit for comet Kiess derived from a more restricted dataset of observations (JPL-3), calculated by Jon Giorgini of NASA/JPL, which resulted in a perihelion time of 4 A.D. ± 40 years. We repeated the calculations and found that a 4 A.D. ejection date put the dust trails in Earth’s path in 2007, implying that this was a better solution for the comet’s perihelion time (Jenniskens et al. 2007). The peak time now was 11:33 ± 20 min UT.

The Meteoroids 2007 conference helped coordinate the airborne and ground-based observing campaign. NASA Ames facilitated the deployment of two privately owned Gulfstream V aircraft, which provided a team of 24 researchers an opportunity to observe the Aurigid shower from 47,000 ft altitude, where extinction near the horizon is low and a large surface area can be monitored (Fig. 5). Participating researchers in this Aurigid Multi-Instrument Aircraft Campaign (Aurigid MAC) were from the USA, UK, France, Germany, and The Netherlands, many with past experience in Leonid MAC and spacecraft reentry missions. The main purpose of the mission was to measure the duration, peak time, and particle size distribution in the dust trail accurately by imaging as many meteors as possible with cameras sensitive to meteors of different brightness. And to measure meteor light curves, penetration depth, and optical spectra, in search of evidence that some of these meteors may be pieces of the original cosmic-ray-produced crust of the comet.

Fig. 5
figure 5

Composite image of Aurigid shower with 15 Aurigids observed from one of the two aircraft by Jason Hatton (ESA/ESTEC). The meteors span the period 11:04:44–11:50:54 UT 1st September 2007 (a 46-min period covering the peak)

Full size image

The shower peaked at 04:15 ± 5 min PDT, earlier than our predicted 04:33 PDT ± 20 min (Jenniskens and Vaubaillon 2007b), but in line with the shower in 1935, 1986 and 1994, which also appeared to be slightly earlier than predicted. The shower may have been wider than expected, about FWHM = 0.68 h instead of the predicted 0.42 h, which could imply that we passed slightly further from the trail center than expected.

At the peak, meteors were detected at a Zenith Hourly Rate of about 130/h, within a factor of two from the anticipated rate. This number is a small improvement on the rates reported in near-real time, which peaked at 100/h. In an effort to inform satellite operators about the shower’s activity in near-real time, a team of four amateur meteor observers kept a tally of Aurigid meteors by means of a video headset display hooked to intensified cameras positioned at the windows and by using an automated counting tool. The Zenith Hourly Rates were phoned in every 10–15 min and immediately posted on our mission website (http://aurigid.seti.org). Figure 6 shows the rates calculated from those reports after a first re-evaluation of the calibration. The rates were simply scaled to the response from the Perseid shower 2 weeks earlier. We flew a test flight at that time, involving one aircraft, during which the Perseid shower was observed under no Moon conditions. The Aurigid shower rates are not expected to scale precisely in the same way, due to the different magnitude size distribution index, and so the rates are expected to slightly change again in the final result.

Fig. 6
figure 6

Zenith Hourly Rates of the Aurigid shower in 5-min intervals as derived from near-real time counts by the flux team of Aurigid MAC

Full size image

At the time of writing, in late September, the tally of Aurigid optical spectra (400–800 nm) was 44 individual Aurigids, and that number is still rising. Many Aurigids have a strong forbidden line of oxygen at 577 nm, which may account for some meteors appearing greenish, rather than being due to the production of metal atom emissions as we thought before. The Aurigid meteoroids appear to contain more sodium than the alpha Monocerotids. It is too early to tell if the sodium content is anomalous compared to that of Leonids and Perseids, as potential evidence of a comet’s pristine crust. There is some indication, perhaps, that the morphology of the grains was more sintered, because, at first glance, we do not see the early release of sodium relative to magnesium and iron that was common among 1–4 revolution dust trail Leonids, and thought to be a sign of sodium-containing minerals being efficiently exposed to heat by a finer fragmentation process.

Few meteors were observed from both planes in a favorable geometry for calculating penetration depth accurately, due to a misalignment of the aircraft trajectories at the time of the peak. The stereoscopic alignment was restored only during the declining tail of the shower. Fortunately, many Aurigids were imaged by ground teams at Lick Observatory and Fremont Peak, in an effort led by Tolis Apostolos (Armagh Observatory). These observations were supported by scattered ground-based observers in the wider Bay Area, where amateur astronomers photographed many Aurigids using digital cameras and low-light-level cameras. At the time of writing, at least 24 Aurigids are known to have been recorded from two or more sites simultaneously, sufficient for a detailed analysis of penetration depths.