MINOR PLANET 2002 EX₁₂ (=169P/NEAT) AND THE ALPHA CAPRICORNID SHOWER

Activity from the constellation of Capricorn was first seen in 1871 by Prof. N. de Konkoly from O'Gyalla in Hungary (in present day Slovak Republic, Tass 1921) and coworkers at Agram, Schemnitz, Hodmezövásárhely, and Szathmárnémithi (de Konkoly 1880). From observations on July 28 and 29, de Konkoly determined a radiant at R.A. = 3072, decl. = −30 (J2000). In 1873, he saw six meteors from a radiant at R.A. = 3143, decl. = −76 on July 25 and 26, and five meteors from R.A. = 3183, decl. = −97 on July 28. Later that century, Denning (1893) knew the Capricornids as an annual shower of slow and often bright meteors from the antihelion source direction. Visual observers continued to map the radiant and its daily movement (McIntosh 1935; Shigeno & Shigeno 2004), but the ecliptic low-inclination stream was established only with the publication of 12 multi-station photographic orbits from the Harvard Meteor Survey (Wright et al. 1956). Several of the Capricornids detected between July 16 and August 1 had similar orbits with inclination ∼8°, and this stream was called the alpha Capricornids.

A dearth of data has led to some confusion in recent years, with several authors finding other Capricornid showers in the same region and time frame. A second more diverse group in the August 8–22 time frame (August Capricornids, IAU no. 199) had lower inclinations, while a third group with a smaller semimajor axis was found active in late July. In the latest analysis of meteoroid orbits, Hasegawa (2001) identified 10 different showers, tentatively linked to seven corresponding parent bodies. These latter streams are more spurious. The alpha Capricornids, on the other hand, can be recognized by visual observers from mid-July until mid-August.

Jenniskens (1994) derived a peak zenith hourly rate of ZHR = 2.2 ± 0.3 at solar longitude 1217 ± 09, and an activity curve slope factor of B = 0.041/° ± 0.007/° from NAPO-MS visual observations (Table 2). This translates to a duration FWHM = 151 ± 26. The shower was found active from July 19 to August 18. On closer inspection, however, more weight should have been given to the Dutch Meteor Society (DMS) observations reported in the same paper, derived by using plotted meteor tracks to verify association, from which we have a peak ZHR = 9 ± 2 hr⁻¹ at λ_o = 127° ± 1° and a duration FWHM = 10° ± 1° (fitted Lorentzian profile), which agrees with the more abundant NAPO-MS data prior to λ_o = 121°, but not with the more scattered data from later days. This new result is in better agreement with Dubietis & Arlt (2004) in time of the peak and peak rate, but implies a slightly narrower profile (Table 2).

Dubietis & Arlt (2004) concluded from an independent set of visual observations that the shower was active from about July 4 to August 14, peaking on July 31 (1277 ± 02 solar longitude, J2000) at a peak zenith hourly rate of ZHR = 5.0 ± 0.2 meteors hr⁻¹ of duration FWHM = 154. They found a magnitude size distribution index χ = 2.31 ± 0.03, not unlike that of other annual showers. Others, too, have reported such peak rates (Bellot 1990; Trigo 1991; Zvolánková 1993).

The model (Vaubaillon 2004; Vaubaillon et al. 2005; Jenniskens & Vaubaillon 2007) evolved a cloud of dust grains under planetary perturbations, which were ejected from the comet at perihelion in 3000 b.c., 2500 b.c., ... etc. in 500 year intervals. The meteoroids were ejected according to the Crifo model for comet dust ejection (Crifo & Rodinov 1997), which is a slight adaptation of Whipple's comet ejection model (Whipple 1951), in a uniform manner in and out of the plane of the comet orbit on the sun-lit hemisphere.

The planetary perturbations are calculated for 10,000 particles each in three size bins (0.1–1 mm, 1–10 mm, and 10–100 mm). Assuming a density of 1 g cm⁻³, the corresponding mean masses for each interval are 0.017 g, 17 g, and 17 kg and the corresponding visual magnitudes +6.7, −0.8, and −8.3, respectively (Jenniskens 2006, p. 594). The relevant size range for comparison to the observations is the middle 1–10 mm bin, while the other two bins represent the extremes of the known size distribution of alpha Capricornids and demonstrate the effects of grain size on the orbital evolution.

The meteoroids from each epoch of ejection were integrated forward in time until the current epoch (taken as 2000 a.d.). Subsequently, all those meteoroids were selected that passed to within 0.05 AU from Earth's orbit at the descending node, irrespective of the position in their orbit. The number of particles that did so in the 1–10 mm size bin was 120 (3000 b.c.), 86 (2500 b.c.), 14 (2000 b.c.), 19 (1500 b.c.), 24 (1000 b.c.), 20 (500 b.c.), and 11 (0 a.d.), respectively.

Only dust ejected around 0 a.d. or earlier has nodes at Earth's orbit in 2000 a.d. Any dust ejected more recently would not be seen as a meteor shower. Hence, if the alpha Capricornids were caused by a single fragmentation event, that event would have had to be prior to 500 a.d., in agreement with the (minimum) age estimated by Jenniskens (2006).

In Figure 3, all model meteors are compared to those of observed shower meteors. The chosen meteoroid orbits were passing so close to Earth's orbit (<0.05 AU) that the calculated radiant of each meteoroid orbit was nearly independent of the method used. We used the method "H" of Neslusan et al. (1998), which involves changing the argument of perihelion to make the orbit intersect Earth's orbit. The dispersion of radiants calculated for ejection between 0 a.d. and 3000 b.c. does not vary significantly and all results are combined.

The dispersion of radiants is in excellent agreement with the observed distribution (Figure 3). In contrast, Figure 4 includes the distribution of radiants for other proposed parent bodies in Table 1. Again, we calculated back in time to 3000 b.c. the orbits of 14P/Wolf, 72P/Denning-Fujikawa, 45P/Honda-Mrkos-Pajdušáková, 141P/Machholz 2, 2101 Adonis, and 9162 Kwiila, then integrated the orbit of dust particles forward to 2000 a.d. We plotted the radiant of those meteoroids that passed in 2000 a.d. to within 0.05 AU from Earth's orbit at solar longitudes between 115° and 135°. None of these proposed parent bodies are in better agreement with the observed alpha Capricornids.

The best results were derived from 3000 b.c. ejecta by 141P/Machholz 2. Asher & Steel (1996) looked into the likelihood of meteors from this comet, but found that recent orbits stayed too far from Earth's orbit. Instead, we find that 3000 b.c. ejecta do arrive from the correct direction R.A. = 3130, decl. = −78 with Vg ∼ 22.9 km s⁻¹, but are more dispersed and peak later (centered at solar longitude 1354). Jenniskens (2006) pointed out that it is perhaps possible that 2002 EX₁₂ and 141P/Machholz 2 were split long ago.

No meteoroids met the required encounter conditions for 45P/Honda-Mrkos-Pajdušáková, 2101 Adonis, and 9162 Kwiila. 45P/Honda-Mrkos-Pajdušáková, like 72P/Denning-Fujikawa and 14P/Wolf, did not generate a compact annual shower from ejecta in 3000 b.c. Instead, all three comets could have contributed to the August Capricornids. Meteoroids ejected by 2101 Adonis did stay together (Figure 4), but arrive at Earth from R.A. = 2935, decl. = −220 at Vg = 24.8 km s⁻¹ during solar longitudes 929–1093, before the alpha Capricornid activity period, while ejecta from 9162 Kwiila arrived later from R.A. = 3226, decl. = −29 at Vg = 18.0 km s⁻¹ during solar longitudes 1350–1450 (peaking at 1416).

Much of the spread in the right ascension of the observed alpha Capricornids is on account of the daily radiant drift. Figure 5 shows that drift for the direction (geocentric right ascension and declination) and magnitude (speed) of the approach vector. The observed drifts are in good agreement with the model predictions for ejecta from 2002 EX₁₂.

Zoom In Zoom Out Reset image size

From the graph of R.A. and speed, it is clear that there are two strands to the model stream, one leading and one trailing strand. The leading strand has higher R.A. at a given solar longitude, the same decl., and a higher geocentric speed. The number ratio of the two strands is not a function of epoch of duration, but does vary with the size of the meteoroids. The leading strand is absent in the smallest size bin.

The trailing strand is more compact and centered at the densest part of the observed alpha Capricornid shower. This is also where most theoretical radiants are for meteoroids that pass longer than 0.05 AU from Earth's orbit; hence we call this the "main strand." The leading strand is perhaps observed at the upper part where the density of solutions is highest.

We also report a strong drift in geocentric speed with solar longitude (Figure 5). This dependence on speed versus solar longitude is also in good agreement with the model prediction. The observed variation is −0.16 km s⁻¹ per degree of solar longitude, while the model predicts an identical dependence. The absolute value of the speed is systematically underestimated by −0.8 ± 0.3 km s⁻¹, as would be expected if no account is made for deceleration of the meteoroid in the atmosphere.

To verify the model, results are compared to the precise photographic data obtained by Harvard in the 1950s and more recent data by DMS (Figure 6). These data are corrected for deceleration in the atmosphere. There is good agreement now between the measured and calculated radiant position and speed.

Zoom In Zoom Out Reset image size

The model stream is more concentrated than the observed shower. The leading strand is also not as prominent (like the results for our [0.1–1] mm size bin, or the results for 0 a.d. in the [10–100] mm bin). The main strand has FWHM = 3° ± 1°, the leading strand has FWHM = 5° ± 1° (peaking 2° later).

Now the distribution of the stream is known in three dimensions, we can calculate the total mass of the shower by including our measurement what fraction of meteoroids pass to within <0.05 AU. Earth only passes the outskirts of the stream. If the stream dates from 0 a.d., then this ratio is 11/10,000 = 0.0011 ± 0.0003. If the stream dates back to 3000 b.c., then the ratio is 120/10,000 = 0.012 ± 0.001.

For χ = 2.1, the total mass is determined by the mass of the largest meteoroids. A zero-magnitude alpha Capricornid is about 8 g (Jenniskens 2006). Five meteors of −5 and two of −10 were reported in the SonotaCo data. DMS reports one −10 mag alpha Capricornid. We will set the limit at −10 mag, corresponding to 80 kg (Jenniskens 2006). Within 0.05 AU from Earth's orbit, the dust density does not significantly fall off. Assuming that the orbital period of the meteoroids is close to that of the parent body (P = 4.20 years), χ = 2.1, ZHR = 5 hr⁻¹, 〈Vg〉 = 22.2 km s⁻¹, and a width of 10° in Earth's path, the total mass of meteoroids <80 kg that come to within 0.05 AU from Earth's orbit amounts to 1.07 × 10¹² kg.

Combined with the previously determined fraction of meteoroids near Earth's orbit, we find that the total mass in the stream is 8.9 × 10¹³ kg if the meteoroids were ejected in 3000 b.c. On the other hand, if the meteoroids were ejected in 0 a.d., less would have evolved into orbits with a node near Earth's orbit and the total mass in the stream would have to be about 10 times higher (9.7 × 10¹⁴ kg).

Now it has been demonstrated that the predicted stream is in good agreement with the observed shower, the origin and nature of the stream can be investigated. Figure 7 shows the distribution of meteoroids ejected in 0 a.d., as seen in 2000 a.d. The meteoroids have distributed all along the comet orbit and are found mainly at aphelion at any given moment of time. In three dimensions, the distribution of dust is warped by rotation around the line of apsides. The dispersion perpendicular to, and in, the ecliptic plane is smallest at perihelion. Near aphelion the radial dispersion is largest.

Zoom In Zoom Out Reset image size

Figure 8 shows the descending node of the orbit of these meteoroids. The nodes are close to that of the comet itself, passing just inside Earth's orbit. Some orbits have evolved by rotating the node. The nodal line rotates in the direction shown by an arrow. Hence, some dispersion of the nodes has occurred in the forward direction of the nutation cycle, representing dust particles that nutate at a faster rate than the comet itself. Those particles that are encountered by Earth are about 300 years ahead of the orbital evolution of 2002 EX₁₂ itself (Figure 2).

Zoom In Zoom Out Reset image size

One nutation cycle for 2002 EX₁₂, shown in Figure 2, takes 4500 years. The dashed line is the inclination, while solid lines show the heliocentric distance of the node. During one cycle, the inclination varies from 5° to 20°. The high-inclination phase corresponds to when one of the nodes is near Jupiter, as expected for an evolved orbit. The rotation of the nodal line is slow when the inclination is high and fast when the inclination is low, at which time the nodes move by Earth's orbit.

To speed up the nutation cycle, the orbit of the particles has to be slightly wider. Indeed, the model particles that pass within 0.05 AU from Earth's orbit have a slightly higher geocentric speed (Figure 6). The dust responsible for the alpha Capricornids represents dust that was either initially ejected in a wider orbit (due to ejection speed), or dust that evolved into slightly wider orbits due to planetary perturbations.

The initial ejection conditions are not likely responsible for putting the particles in those wider orbits. Radiation pressure and ejection speeds could do so, but then we would have expected smaller particles to evolve more rapidly than larger ones. Instead, the percentage of particles passing <0.05 AU from Earth's orbit does not change with size. It varied from 0.88% ± 0.09% to 1.21% ± 0.11%, and to 1.00% ± 0.10% for bin sizes [0.1–1], [1–10], and [10–100] mm, respectively. Hence, direct planetary perturbations are likely responsible for putting the particles in a wider orbit.

The mass of 2002 EX₁₂ is about 9.7 × 10¹³ kg (for density 1 g cm⁻³). If the breakup occurred around 3000 b.c. or slightly earlier, then the remaining fragment is ≤52% of the original mass. We consider this a reasonable number, because it is in line with other streams thought to have been created by comet fragmentation (Jenniskens 2008). In those cases, too, we find that about as much mass is present in the stream than is left in the remaining comet.

If, instead, the comet broke as recently as 0 a.d., then the remaining fragment is only about 9% of the original mass. Although such a small remaining mass cannot be fully excluded, it is not in line with other known showers.

In other words, only after 4000 years does enough dust evolve into Earth's orbit to account for the observed alpha Capricornid rates.

This dust was not released in normal comet activity from water vapor drag (Whipple 1951), because normal active Jupiter Family comets would eject dust at a rate of only about 1000 kg s⁻¹ for a period of 0.3 years around perihelion, which amounts to 1 × 10¹⁰ kg per orbit (Jenniskens 2006). In that case, enough dust would be released only after 8900 orbits, or about 37,000 years, rather than <5000 years suggested by the observations. Hence, a massive breakup is implied that occurred about 4000–5000 years ago.

According to our model, the alpha Capricornids have only gradually increased in activity since they first became visible in the late 1800s (Figure 9). In the next 300 years, however, the alpha Capricornids are likely to grow into a major annual shower when the nodes of the bulk of dust grains evolve into Earth's orbit. The predicted alpha Capricornid activity between 1800 and 2500 a.d. is shown in Figure 9. Rates are expected to stay around the current peak rate of ZHR = 5–9 hr⁻¹ until about 2140 a.d., but will increase dramatically in the 23rd and 24th centuries to a peak of ZHR = 2200 hr⁻¹ on an annual basis (assuming that the width of the shower does not change), half the visible shower peak rate during the 1999 Leonid storm.

Zoom In Zoom Out Reset image size

At this rate, the shower can become a concern for satellite operators. ZHR = 2200 hr⁻¹ translates to a peak particle flux of about 1.5 × 10⁻⁸ meteoroids larger than 10⁻⁶ g m⁻² s⁻¹.

Much sooner, the meteoroid stream of 2002 EX₁₂ will become a potential danger for space missions to Venus. At this moment, the nodes of most meteoroids cross the ecliptic plane near heliocentric distance R_h = 0.65 AU, just inside the orbit of Venus, and this heliocentric distance will gradually increase.

At least 64 meteor showers are listed as established by the IAU Meteor Data Center (Jopek 2009), the first shower being the alpha Capricornids (IAU no. 1). Only 16 of those have active parent comets. Those that have are mostly of Halley and long-period type (Jenniskens 2006). The majority of short-period showers do not have an identified active Jupiter Family comet parent body.

Following the identification of asteroidal-looking minor planet 2003 EH₁ as the source of the Quadrantid shower (Jenniskens 2004), we searched for other such minor planets as potential sources of our short-period meteor showers. Jenniskens (2006, 2008) identified numerous candidates, most of which are (mostly) dormant Jupiter Family comets not yet fully decoupled from Jupiter, with a Tisserand invariant T_J below 3 and an aphelion just inside of Jupiter's orbit. Based on bias-corrected discovery statistics, Stuart & Binzel (2004) estimated that 30% ± 5% of the entire NEO population resides in cometary orbits having T_J < 3 (still under Jupiter's influence).

The source of these streams is thought to be periodic fragmentation instead of gradual outgassing (Jenniskens 2006). The meteoroid streams provide an important historical record of fragmentation and decay of these potential Earth impactors. Combined, these massive disruptions are capable of producing the bulk of the zodiacal dust cloud (Jenniskens 2006).

Much work is ahead to identify with certainty the source of the established short-period meteor showers, in order to evaluate statistically the rate of decay of the population as a whole. The present work adds one such stream, the meteoroids ejected by 2002 EX₁₂.

During the imminent return of the comet in January of 2010, 2002 EX₁₂ is expected to brighten to a 10th magnitude object (15th magnitude if not active). This opportunity may be used to learn more about the physical properties of this object, in order to explain what drives the occasional massive disruptions.

The model predicts a brief daytime shower at the other node in January (Figure 1), with a radiant at R.A. = 3090, decl. = −294, Vg = 22.7 km s⁻¹, peaking at solar longitude 3035 ± 07 (J2000), with a width of 64. The predicted peak ZHR = 2–5 hr⁻¹.

The alpha Capricornids (observed at the descending node) have an ascending node that is at higher heliocentric distance for increasing ascending node (Figure 10). The model does reproduce this effect, when considering only the meteoroids that have the descending node in Earth's path (+).

Zoom In Zoom Out Reset image size

Until recently, the nearest identified shower from radar data are the Daytime Sagittariids-Capricronids (IAU no. 115), but those peak at solar longitude 3125 arriving at 26.8 km s⁻¹ from R.A. = 3150, decl. = −233. This shower is not caused by 2002 EX₁₂.

Now, Brown et al. (2009, 2010) report from a three-dimensional wavelet transform analysis of CMOR radar observations (Brown et al. 2008) that a shower of Daytime chi Capricornids (IAU no. 114) peaked at solar longitude 294 (ranging from 301° to 315°, or full width at half maximum about 6°), and arrived from R.A. = 304.7, decl. = −29.2, with Vg = 23.8 km s⁻¹, in good agreement. This implies that 2002 EX₁₂ (=169P/NEAT) is also the parent of the Daytime chi Capricornids.

Now the parent body has been identified, it is possible to study the alpha Capricornids in order to measure the comet's elemental composition. Borovicka & Weber (1999) have reported on a spectrum of an alpha Capricornid. They found that Na/Fe = 0.04, Mn/Fe = 0.003, Cr/Fe = 0.001, Mg/Fe = 3, and Ca/Fe = 0.001. This can be compared to the cosmochemical solar system abundances: Na/Fe = 0.07, Mn/Fe = 0.01, Cr/Fe = 0.015, Mg/Fe = 1.2, and Ca/Fe = 0.08 (Lodders 2003). The low calcium abundance was thought to be due to incomplete evaporation in the meteor, not a feature of the comet itself. Iron, too, may have been under represented by 30%, based on the volatile elements Na, Mn, and Cr. In that case, Mg may have been overabundant in 2002 EX₁₂ by a factor of 3–8. Before drawing conclusions, however, more work along these lines is needed.

It is of interest to note that the alpha Capricornid and kappa Cygnid showers, both known for meteors rich in flares, have in common both age and the fact that Earth passes through outskirts of the stream (Jenniskens & Vaubaillon 2008). If initial ejection conditions (and radiation pressure) are not responsible for the meteoroids ending up in the outskirts of the stream, then processes such as space weathering over the formation timescale of the streams can perhaps explain the peculiar physical properties of the meteoroids.