A Fireball and Potentially Hazardous Binary Near-Earth Asteroid (164121) 2003 YT₁

The worldwide meteor survey networks have established the procedure for identifying meteoroid orbits in streams and associated parent bodies, asteroids, and comets, mostly known as near-Earth objects (NEOs; SonotaCo 2009; Rudawska & Jenniskens 2014; Ye et al. 2016; Jenniskens 2017). Some NEOs, meteorite falls, and fireballs have been linked with potentially hazardous asteroids (PHAs; Madiedo et al. 2013, 2014; Svetsov et al. 2019) of which the Taurids are studied in many cases (Brown et al. 2013; Olech et al. 2017; Spurný et al. 2017; Clark et al. 2019). Physical disintegration of NEOs result in producing orbit-hugging dust (streams) which may cross Earth's orbit. Suggested mechanisms, especially for those of asteroids, include rotational instability, thermal stress, collisions (impacts), and so on (Jewitt 2012; Jewitt et al. 2015). Asteroidal stream parents should be, or used to be, losing mass while among the few mass-loss activities other than activity driven by sublimation of ice are identified (Kasuga & Jewitt 2019).

A relatively slow, bright fireball was detected in the sky over Kyoto, Japan on UT 2017 April 28 at 15^h 58^m 19^s through the SonotaCo Network (SonotaCo 2009). The small semimajor axis (a = 1.111 au) and high inclination (i = 439) present its peculiar orbit. The dynamical properties, as given by orbit-linking D-criterions (see Southworth & Hawkins 1963), find a close association with the near-Earth asteroid (NEA) (164121) 2003 YT₁ (hereafter, 2003 YT₁; see details in Section 3). The short distance from the asteroid orbit to Earth's orbit (see 0.0026 au at the descending node) is compatible with those of meteoroid streams for showers (≲0.01 au; Vaubaillon et al. 2019), suggesting that both the fireball and 2003 YT₁ practically cross Earth's orbit. This asteroid–meteor pair is likely to be secured, giving a rare opportunity for understanding of meteoroid production.

The NEA 2003 YT₁ was discovered on UT 2003 December 18 in the course of the Catalina Sky Survey (Tichy et al. 2003). Based upon the absolute magnitude of H = 16.2 and the low minimum orbit intersection distance (MOID) of ∼0.003 au (NASA/Jet Propulsion Laboratory (JPL) Small-Body Database), the object is a PHA (Larson et al. 2004; Hicks et al. 2009). The impact probability to Earth is calculated as ∼6% per 10⁷ yr (Galiazzo et al. 2017). The Arecibo radar delay-Doppler and optical photometric observations independently identified 2003 YT₁ has a binary system (Nolan et al. 2004a). The suggested formation process is a rotational instability, a breakup/fission driven by Yarkovsky–O'Keefe–Radzievskii–Paddack (YORP) torques (Pravec & Harris 2007). The primary has 1.1 ± 0.2 km in diameter (D_p) and the secondary with a diameter of 0.21 ± 0.06 km (D_s), having a distance of 2.7 km (Nolan et al. 2004b). The primary's rotation period is 2.343 ± 0.001 hr, and the light-curve amplitude of ∼0.16 mag exhibits its nearly spheroidal shape (Galád et al. 2004; Larson et al. 2004; Warner et al. 2018). The secondary's rotation period of ≲6 hr and its orbital period of ∼30 hr (eccentric orbit) suggest the asynchronous state (Nolan et al. 2004a, 2004b). Geometric albedos (in visual and infrared) are measured by thermal infrared observations, p_v = 0.24 ± 0.16 from the ground-based (Delbo et al. 2011), and p_v = 0.20 ± 0.10 and p_IR = 0.33 ± 0.14 from the space (Wide-field Infrared Survey Explorer (WISE)/NEOWISE; Mainzer et al. 2012). Near-infrared spectra (0.7–2.5μm) reveal the surface assemblage dominated by orthopyroxene with any lack of olivine content on 2003 YT₁, implying a taxonomically V-type asteroid (Abell et al. 2004, 2005; Sanchez et al. 2013). The regolith breccia (<25μm in size mostly; Ieva & Dotto 2016) could be originated in a larger, extensive-igneous processed precursor body (howardites, eucrites, and diogenites-assemblage, HED). The V-type NEAs (Cruikshank et al. 1991) remain an open question for their origin (from (4) Vesta?; Cochran et al. 2004; Burbine et al. 2009).

In this paper we present the orbital and physical properties of the Kyoto fireball taken by the SonotaCo Network, including the trajectory, radiant point, geocentric velocity, orbit, and meteoroid mass (size) and further discuss the possible relation to the parental binary NEA 2003 YT₁ by examining its potential dust production mechanisms.

The fireball studied here is from the SonotaCo Network database. Automated multi-station video observations use more than 100 cameras at 27 sites in Japan (SonotaCo 2009).⁴ The database is advantaged in the similar type of camera setup of all the network sites. The CCD cameras are mostly WATEC series with f = 3.8–12 mm lens having a field of view (FOV) ≈ 30°–90°. The video format is digitized in 720 × 480 or 640 × 480 pixels audio video interleave (AVI) from the National Television System Committee (NTSC) signal (29.97 frames per second, interlaced), and the video field with a time resolution of ≃0.017 s (1/59.94 s) is used for measurement. Meteors are detected by UFOCaptureHD2 software, and the data reductions and orbit determinations are conducted by UFOAnalyzerV2 and UFOOrbitV2, respectively. Limiting magnitude for multi-station observations is estimated to be apparent magnitude <+3 and absolute magnitude <+2 for each (SonotaCo 2009).

The database includes orbital and physical parameters of meteors, such as trajectory (apparent position on the sky plane), radiant point, geocentric velocity, orbital elements, brightness (magnitude), and height above sea level.⁵ Astrometry and photometric calibrations for meteors are conducted using field stars in the background and the SKY2000 Master Catalog, Version 4 (Myers et al. 2001) installed in the UFOAnalyzerV2. Single-station observation has some uncertainties of measurements but are negligibly small, as estimated by the position in the sky plane ∼003 (SonotaCo 2009), distance to meteor ≲200 m, and elevation angle ∼002–003. Lens distortion is corrected by background stars' positions fitted by a polynomial equation. The aperture radius used for the stars is 5 pixels in the image (∼05) and the sky background is determined within a concentric annulus having projected inner and outer radii of 5 pixels and 7.5 pixels (≈05 ∼ 07), respectively. For meteors on the other hand, the aperture sets a minimum rectangle that covers the total brightness of a meteor including its tail, and the sky background was subtracted by the field prior to the meteor appearance. More than five background stars are used to count the flux of the meteor. Then we obtain apparent magnitude of the meteor, m(obs). The photometric uncertainty (mag) is estimated from the typical uncertainty of comparison stars, ∼0.5 mag, and the correction for the saturated apparent magnitude of meteor is expressed as

$$\begin{eqnarray}&&\sqrt{{0.5}^{2}+{\left(m(\mathrm{obs})^{\prime} -m(\mathrm{obs})\right)}^{2}},\end{eqnarray} \tag{ 1 }$$

where m(obs)' is the corrected apparent magnitude. The m(obs)' is derived from m(obs)' = m(obs) + k(m(obs))², where m(obs) < 0 and k = −0.03. Details of the analysis procedure is described in the UFOAnalyzerV2 manual,⁶ and private communication with SonotaCo.

The fireball trajectory and observing sites (ID⁷ ) are shown in Figure 1. The images of the fireball are represented in Figure 2. This event was simultaneously detected at 11 sites with 12 cameras. The data sets taken at Tokyo (TK8_S7) and Osaka (Osaka03_3N) have imaged the most part of trajectory, from the beginning to the end. The numbers of video fields that have acquired the fireball position and brightness are 159 out of 173 in the Tokyo data and 194 out of 204 in the Osaka data, respectively. Therefore these two data sets are primarily used for orbit determinations and photometric measurements. Orbital results are listed in Tables 1–3. Photometric results are given in Tables 4 and 5.

Zoom In Zoom Out Reset image size

Zoom In Zoom Out Reset image size

3.1. D-criterions

3.2. Meteoroid Mass

For initial meteoroid mass, the classical meteor luminous model (Bronshten 1983; Ceplecha et al. 1998) has been used but with non-negligible uncertainty in the ablation coefficient. Instead, we have made a new meteor luminous model as described in Appendix B.

The total mass of the meteoroid (source of fireball), m(g), can be estimated from the light curves (Figure 3) using the new luminous model (Equation (18)), which is given by

$$\begin{eqnarray}&&m=\displaystyle \sum _{N}{\left[\displaystyle \frac{2I}{\tau {v}^{2}}{\left(\displaystyle \frac{2}{\sigma {v}^{2}}-1\right)}^{-1}{\int }_{0}^{t}{dt}\right]}_{N},\end{eqnarray} \tag{ 2 }$$

where N is the number sign of the video field (see Tables 4 and 5), I is the meteor luminosity, τ is the luminous efficiency, v is the meteor velocity (cm s⁻¹), σ is the ablation coefficient (s² km⁻²), and t is the time (s). We define the meteor luminosity in visual magnitude-based units as $I={10}^{-0.4{M}_{{\rm{v}}}}$, where M_v is the absolute magnitude (as seen from distance of 100 km).

Zoom In Zoom Out Reset image size

The luminous efficiency, τ, is the fraction of a meteoroid's instantaneous kinetic energy loss converted into light in a particular bandpass. The uncertainty within is substantial (0.05 ∼ 10 s%) as it depends on many factors, e.g., the speed, mass, composition of meteoroid and the height at which it ablates (different flow regimes), and the spectral sensitivity of the detector (see Weryk & Brown 2013; Subasinghe & Campbell-Brown 2018; see the review in Popova et al. 2019). For this study, we use the velocity dependence (Table 1 in Ceplecha & McCrosky 1976) considering the performance of CCD cameras (e.g., low resolution). Setting v = 23.7 km s⁻¹ finds τ = 5 × 10⁻¹³ erg⁻¹ s 0 mag. The τ-value corresponds to 0.75% efficiency. The conversion is given by multiplying 1.5 × 10¹⁰ erg s⁻¹ 0 mag⁻¹, i.e., the luminous energy equivalent to zero magnitude in visual (Table 6 in Ceplecha et al. 1998).

The critical bulk density, ρ, for the meteoroid and 2003 YT₁ is estimated. An asteroid shape is approximated as an ellipsoid with axes a ≥ b = c, in rotation about the c-axis. A limit to the ratio of the equatorial axes is f = a/b = 10^0.4Δm, where Δm is the light-curve amplitude. Rotation around the c-axis with period, P_rot, gives a condition that the gravitational acceleration is greater than the centripetal acceleration which is the largest at the top of the shape. The net acceleration toward the center of a rotating object is >0, giving the relation as (Equation (4) of Jewitt & Li 2010)

$$\begin{eqnarray}&&\rho \gt \left(\displaystyle \frac{3\pi }{{{GP}}_{\mathrm{rot}}^{2}}\right){\left(\displaystyle \frac{a}{b}\right)}^{2},\end{eqnarray} \tag{ 3 }$$

where G is the gravitational constant. We substitute G = 6.67 × 10⁻¹¹ m³ kg⁻¹ s⁻², P_rot = 2.343 hr, Δm = 0.16 mag (i.e., f = a/b = 1.16) into Equation (3), then obtain ρ ≳ 2700 kg m⁻³. This is consistent with the lower limit for rubble pile asteroids with diameters of 0.3–10 km (ρ = 2.7 g cm⁻³), as formulated by the observed light-curve amplitude versus spin rate (Pravec 2005). The proposed bulk density of 2010 ± 700 kg m⁻³ (Brooks 2006) may be uncertain due to the assumption of the circular orbit of the secondary. The orbit is actually eccentric (e ≃ 0.18; Fang & Margot 2012; see Pravec et al. 2016).

Substituting v(=v_g) = 2.37 × 10⁶ cm s⁻¹, τ = 0.75%, σ = 0.0017 s² km⁻² (Appendix B) and t = 0.017 s into Equation (2), we obtain m (see Tables 4 and 5). The weighted mean of the total mass is m = 29 ± 1 g, corresponding to the meteoroid size of a_s = 2.7 ± 0.1 cm for ρ = 2700 kg m⁻³. For reference, the classical luminous model (Equation (11)) is applied too. The resulting masses are compared in Table 7.

Table 7.  Mass Comparison

Data	Classical Modela	New Modelb
Tokyo	11 ± 1 g	31 ± 1 g
Osaka	9 ± 1 g	26 ± 1 g
⋯	10 ± 1 g c	29 ± 1 gc

Notes.

^aEquation (11) (e.g., Ceplecha et al. 1998). ^bEquation (18) from this work. ^cThe weighted mean of measurements.

Download table as:  ASCIITypeset image

Here, we recapitulate the binary formation process of 2003 YT₁ and evaluate possible dust production mechanisms for millimeter–centimeter scale particles.

4.1. Binary Formation

The 2003 YT₁ binary system is presumed to be formed from a breakup/fission by rotational instability with YORP spin-up. The primary with D_p ≲ 10 km and the normalized total angular momentum of the binary system α_L⁹  = 1.13 suggest that the 2003 YT₁ binary system was formed from a precursor body spinning at the critical rate, resulting in fission and mass shedding (Group A in Table 1 and Figures 2 and 3 from Pravec & Harris 2007; reviewed in Margot et al. 2015; Walsh & Jacobson 2015). The 2003 YT₁ primary rotates (P_rot = 2.343 hr) closely to the spin barrier period of ∼2.2 hr (Warner et al. 2009; Chang et al. 2015). This can reasonably lead to a rotational breakup when centrifugal forces have exceeded the gravitational and cohesive forces (Pravec et al. 2008).

We calculate the YORP timescale of the spin, τ_Y, using the ratio of the rotational angular momentum, L, to the torque, T. The relation is given by Jewitt et al. (2015) as

$$\begin{eqnarray}&&{\tau }_{{\rm{Y}}}\sim K\,{D}_{{\rm{e}}}^{2}\,{R}_{{\rm{h}}}^{2},\end{eqnarray} \tag{ 4 }$$

where K is a constant, D_e is the asteroid diameter (kilometer), and R_h is the heliocentric distance (astronomical unit). The value of constant K is experimentally estimated from published measurements of YORP acceleration in seven well-characterized asteroids (Table 2 from Rozitis & Green 2013). Scaling K to the bulk density of primary ρ = 2700 kg m⁻³ and its rotation period P_rot = 2.343 hr, we find K ∼ 5 × 10¹³ s km⁻² au⁻². By Equation (4), the primary with D_e(=D_p) = 1.1 km orbiting at R_h ∼ 1.11 au takes τ_Y ∼ 2 Myr. This is consistent with the previous study (∼1 Myr; Vokrouhlický et al. 2015) and much shorter than the catastrophic collisional lifetime for 1 km NEAs (∼100 Myr, Bottke et al. 1994; see also Section 4.2.3). The YORP spin-up plays a contributory role.

The cohesive strength is a required parameter for asteroids rotating near or faster than the spin barrier to resist rotational forces (Scheeres et al. 2010). The strength at a rotational breakup of a body is estimated by ${S}_{{\rm{c}}}\sim \rho \left({D}_{{\rm{s}}}/{D}_{{\rm{p}}}\right){({\rm{\Delta }}v)}^{2}$ (Equation (5) of Jewitt et al. 2015), where D_p and D_s are the dispersed fragmental sizes of the primary and secondary, respectively, Δv is the excess velocity of escaping fragments, assumed comparable to the escape velocity (v_e) from the primary, and ρ is the bulk density. With the same value for ρ (see Section 3.2) and substituting (D_s/D_p) = 0.19 (the diameter ratio of the secondary to primary), and Δv (=v_e) = 0.68 m s⁻¹, we find S_c ∼ 240 N m⁻². This value is comparable to weak, van der Waals forces (∼10–100 N m⁻²) bounded in a modeled rubble pile asteroid (Scheeres & Sánchez 2018), while 10⁵× weaker than those of competent rocks (10⁷–10⁸ N m⁻²). Therefore, given a rubble pile structure, a rotational breakup/fission is a probable process for the 2003 YT₁ binary formation.

The breakup/fission period is inferred from the spin asynchronous state of the 2003 YT₁ binary system in the present day. The timescale from asynchronous to synchronous state, τ_sync, limits to the age of the binary system. Two models are applied for 2003 YT₁ using the data of known synchronous binary asteroid systems (Table 3 in Fang & Margot 2012). One estimates τ_sync = 10^7–8 yr by the tidal Love number proportional to the radius (Goldreich & Sari 2009), another estimates τ_sync = 10^4–5 yr by the tidal Love number inversely proportional to the radius (Jacobson & Scheeres 2011). The former just agrees with the large-sized binaries having a primary with D_e ∼ 4 km; on the other hand, the latter fits well for smaller-sized objects too (down to D_e ∼ 0.4 km). For 2003 YT₁ we thus take τ_sync = 10^4–5 yr (Fang & Margot 2012). The interpretation is that this binary is of an age of <10⁴ yr, comparable with the upper limit of the meteoroid stream lifetime, <10⁴ yr (Jenniskens & Lyytinen 2005).

Another example is proposed by the small-sized V-type NEA pair (D_e ∼ 25–50 m) also having the young age of separation <10⁴ yr (D' = 0.0035 for 2017 SN₁₆ and 2018 RY₇; Moskovitz et al. 2019). The YORP-driven breakups for the (sub)kilometer-sized bodies may suggest moderately recent events.

4.2. Dust Production Mechanisms

We look into possible dust production mechanisms from 2003 YT₁. The consequences of YORP-driven breakups are reported from the (sub)kilometer-sized main-belt asteroids, as exampled by P/2010 A2 (Jewitt et al. 2010, 2013; Agarwal et al. 2013), P/2013 R3 (Jewitt et al. 2014a, 2017; Hirabayashi et al. 2014), and (6478) Gault (Chandler et al. 2019; Hui et al. 2019; Jewitt et al. 2019b; Kleyna et al. 2019; Moreno et al. 2019; Ye et al. 2019). Additionally, other different mechanisms may work together, e.g., impact for P/2010 A2 and outgassing torques from sublimated ice for P/2013 R3 (Jewitt et al. 2015). Here, we estimate breakup/fission (rotational instability), resurfacing, impact, thermal fracture, photoionization, radiation pressure sweeping, and sublimation of water ice.

4.2.1. Breakup/Fission (Rotational Instability)

Binary NEAs show a trend of having large values of thermal inertia, Γ ≳ 400 J m⁻² s^−0.5 K⁻¹, typically twice those of non-binary NEAs, suggesting that the fine regoliths were swept away during the YORP-induced binary formation (Walsh et al. 2008; Delbo et al. 2011). For 2003 YT₁, it would be difficult to determine the sizes and speeds of released dust particles at the presumed breakup time, whereas the measured values of the recent precedents infer the large particles (millimeter–centimeter scale) with nearly the gravitational escape speeds ≲1 m s⁻¹ (see a_s = 6 mm–40 cm from P/2010 A2, ∼1 cm from P/2013 R3, and ≲1 cm from Gault; Jewitt et al. 2013, 2014a, 2019b). On the process, resurfacing could be partly involved (Gault; Marsset et al. 2019). A similar situation might be expected for 2003 YT₁. The dust particles are, if released, supposed to reach the Earth within the typical stream lifetime (10⁴ yr). The short distance to Earth's orbit, e.g., Δr = 0.0026–0.0279 au (Table 2), may help. Accordingly, the rotational breakup/fission ejecting the millimeter–centimeter scale dust particles is considered as a likely cause.

4.2.2. Resurfacing

Planetary encounters, space weathering, and thermal processes could induce resurfacing, which might lose dust particles on the surfaces to some extent. For example, the timescale for Q-type NEAs to be refreshed into S-type (at 1 and q ≲ 0.9 au) is estimated to be 10^5–7 yr by planetary encounters (Binzel et al. 2010; Nesvorný et al. 2010), space weathering (Graves et al. 2018), and thermal processes (Graves et al. 2019). For V-type NEAs, the aftermath of those processes are unclear (space weathering; Pieters et al. 2012; Fulvio et al. 2016), while the timescale of resurfacing itself seems to be 10–1000 times longer than the typical stream lifetime. The resurfacing is thus unlikely to be responsible for releasing the source of meteors.

4.2.3. Impact

Impacts can cause catastrophic disruption of asteroids and/or dust production. The catastrophic disruption is defined as the impact resulting in losing a half of the target's mass. The specific impact energy threshold is expressed as ${Q}_{{\rm{D}}}^{* }$ = (1/2)(D_i/D_t)³${\rm{\Delta }}{V}_{\mathrm{NEA}}^{2}$, where D_i, D_t, and ΔV_NEA are the size of the impactor and the target (an assumed precursor body) and the relative velocity among NEAs, respectively (Jutzi et al. 2010). With Q_D* ∼ 1400 J kg⁻¹ for catastrophic disruptions of stone meteorites (Flynn & Durda 2004; Flynn et al. 2018), D_t ≈ D_p = 1100m (assuming the primary size occupying >80% of the precursor body) and ΔV_NEA = 17–20 km s⁻¹ (Bottke et al. 1994; Jeffers et al. 2001), we find D_i ∼ 20 m. This catastrophic event is inferred from the interval between impacts, τ_col (Davis et al. 2002), as

$$\begin{eqnarray}&&{\tau }_{\mathrm{col}}\simeq \displaystyle \frac{4}{\pi {({D}_{{\rm{t}}}+{D}_{{\rm{i}}})}^{2}\,{P}_{\mathrm{NEA}}\,{N}_{{\rm{i}}}(\geqslant {D}_{{\rm{i}}})},\end{eqnarray} \tag{ 5 }$$

where P_NEA is the collision frequency per unit area in the near-Earth region (km⁻² yr⁻¹), and ${N}_{{\rm{i}}}(\geqslant {D}_{{\rm{i}}})$ is the cumulative number of impactor larger than D_i. The NEA cumulative size distribution is measured by WISE/NEOWISE, N_i(D_i ≥ 140 m) ≃ 13,200 × ${\left(140{\rm{m}}/{D}_{{\rm{i}}}\right)}^{1.32}$ (Mainzer et al. 2011), and we presumably extend the equation down to 20 m in diameter. With D_i ∼ 20 m, N_i(≳20 m) ∼ 1.7 × 10⁵, D_t ≈ D_p = 1.1 km (see above) and P_NEA = 1.5 × 10⁻¹⁷ km⁻² yr⁻¹ (Bottke et al. 1994), we find τ_col ≳ 10¹¹ yr. This is much longer than τ_sync = 10^4–5 yr (Fang & Margot 2012) and the mean dynamical lifetime of NEAs ∼ 10⁶ yr (Bottke et al. 2002; Morbidelli et al. 2002), suggesting the absence of a catastrophic event (see Section 4.1).

On the other hand, micrometeorite impacts may result in ejecting the meteoroid-sized particles. The velocity distribution for micrometeorites near Earth, U, is 12 ∼ 70 km s⁻¹ (see radar observations; Nesvorný et al. 2010; Janches et al. 2014; Carrillo-Sánchez et al. 2015). For equal target and impactor densities, the ratio of the ejecta mass, m_e, traveling faster than the escape velocity, v_e, to impactor mass, m_i, is related by

$$\begin{eqnarray}&&{m}_{{\rm{e}}}/{m}_{{\rm{i}}}=A{({v}_{{\rm{e}}}/U)}^{\alpha },\end{eqnarray} \tag{ 6 }$$

where A ∼ 0.01, α ∼ −1.5 (Housen & Holsapple 2011). Substituting m_e ∼ 30 g (fireball mass), v_e = 0.68 m s⁻¹, and U = 12–70 km s⁻¹ into Equation (6), we find that micrometeorite impactors in the size range of 0.4 mm ≤ a_i ≤ 1 mm (m_i = (0.1–1.3) × 10⁻³ g with ρ = 2.7 g cm⁻³) can eject the centimeter-sized dust particles. The perpendicular impact strength of >10¹¹ N m⁻² is estimated from the equation of impact force per unit area given by m_iU/δt × 4/π ${a}_{{\rm{i}}}^{2}$, where δt = a_i/U is the assumed extend impact time (s). The value is by orders stronger than the compressive strengths of stone meteorites, ∼10⁸ N m⁻² (Flynn et al. 2018), suggesting that micrometeorites are certainly smashing the surface. In this case, many of unknown relevant physical parameters (e.g., impact frequency, population of micrometeorite near 2003 YT₁) prevent exact estimation, however, offer probable insight for dust production.

4.2.4. Thermal Fracture

Thermal fracture and fatigue of the asteroid surfaces can be caused by desiccation stress, with the release of dust particles (Jewitt & Li 2010). For 2003 YT₁, the peak perihelion temperature, T_q ∼ 440 K, is about half or less of those of near-Sun asteroids (see Phaethon; Jewitt 2013), while the thermal stress ≲50 MPa is somewhat responsible for the breakdown of the rocky surfaces of most asteroids in the inner solar system (Figure 9(b) in Molaro et al. 2015). The characteristic speeds of dust particles produced by thermal disintegration can be computed by conversion from thermal strain energy into kinetic energy of ejected dust particles. We use the required conversion efficiency, η, given by (see Equation (3) of Jewitt & Li 2010)

$$\begin{eqnarray}&&\eta \sim {\left(\displaystyle \frac{{v}_{{\rm{e}}}}{\alpha \delta T}\right)}^{2}\left(\displaystyle \frac{\rho }{Y}\right),\end{eqnarray} \tag{ 7 }$$

where again v_e = 0.68 m s⁻¹, α ∼ 10⁻⁵ K⁻¹ is the characteristic thermal expansivity of rock (Lauriello 1974; Richter & Simmons 1974), δT ∼ 80 K is the temperature variation between the q and aphelion, and Y = (1–10) × 10¹⁰ N m⁻² are Young's moduli for rock in general (Pariseau 2006). With ρ as above we find η ≳ 2%–20% is needed for the velocities of ejected dust particles to surpass the escape velocity. The value of conversion efficiency is small enough for most dust particles to be launched into interplanetary space.

Note that micron-sized particles are observed from the Phaethon tails at perihelion, possibly produced by a combination of thermal fracture and radiation pressure (Jewitt et al. 2013; Hui & Li 2017). Such tiny particles are distinct from the millimeter–centimeter scale dust. Larger, mass-dominant particles could be launched, but the acquisition of more and better data for estimation is needed (Jewitt et al. 2018, 2019a). This mechanism is hence pending.

4.2.5. Photoionization

Photoionization by solar UV induces electrostatic forces to eject very small particles. For a 1 km diameter asteroid, 2003 YT₁, the critical size is estimated to be a_e ≲ 4 μm (Equation (12) of Jewitt et al. 2015). Therefore, millimeter–centimeter scale particles cannot be launched. We conclude that this process is improbable.

4.2.6. Radiation Pressure Sweeping

Small dust particles on the surface of asteroids, if they briefly lose contact forces, can be stripped away by radiation pressure sweeping. By equating the net surface acceleration (gravitational and centripetal) with the acceleration due to radiation pressure, we estimate the critical size to be swept away, a_rad (μm), with Equation (6) of Jewitt & Li (2010),

$$\begin{eqnarray}&&{a}_{\mathrm{rad}}\sim \displaystyle \frac{3\,{g}_{\odot }}{2\,\pi \,{R}_{\mathrm{au}}^{2}\,{f}^{1/2}\,{D}_{{\rm{e}}}}{\left[\displaystyle \frac{G\rho }{{f}^{2}}-\displaystyle \frac{3\pi }{{P}_{\mathrm{rot}}^{2}}\right]}^{-1},\end{eqnarray} \tag{ 8 }$$

where g_⊙ is the gravitational acceleration to the Sun at 1 au, R_au is the heliocentric distance expressed in astronomical units, f is the limit to the axis ratio (=a/b), and G is the gravitational constant. We substitute g_⊙ = 0.006 m s⁻², R_au = 0.786 (non-dimensional), and adopt the same values of f, G, D_e(=D_p), ρ, and P_rot (as applied so far) into Equation (8), then we obtain a_rad ∼ 2900 μm ≈ 3 mm. The millimeter-sized dust particles can be swept by radiation pressure from 2003 YT₁, which could be the source of meteors. Even if they arrived at Earth, the relatively small size and slow velocity would produce meteors of m_obs ∼ +5 mag (Table 1 in Lindblad 1987), being too faint for most optical surveys. By contrast, the centimeter-sized dust particles (source of fireballs) are unlikely to be released.

4.2.7. Sublimation of Water Ice

Sublimation of water ice may be an improbable dust production mechanism for differentiated (V-type) or thermally metamorphosed (S-type) asteroids. On the contrary, the presence of aqueously altered minerals on those of surfaces have been reported (Rivkin et al. 2015, 2018), as well as further evidence such as the weakly active S-type Oort Cloud object driven by the sublimation of water ice (Meech et al. 2016) and the native water inclusion in Itokawa samples (Jin & Bose 2019). Asteroid (4) Vesta's current surface texture, fracture, and roughness (1 cm to 10 cm scale) could be caused by (carbonaceous) impactors (Hasegawa et al. 2003; De Sanctis et al. 2012; Russell et al. 2013, 2015). By contrast, the recent Dawn bistatic radar observation indicates subsurface volatile (water ice) involvement processes (Palmer et al. 2017).

Can buried water ice exist and survive even in V-type asteroids? The differentiation process would occur for the most part of the body, but partially may not. The Vesta's smoother terrain area (heightened hydrogen > 0.015%), on which subsurface ice might contribute to, occupies only ≳0.01% of the total surface area (Palmer et al. 2017). The extreme partiality might lead to the localized subsurface ice existence.

How deep can water ice survive in the 2003 YT₁ primary, if it were therein? Megaregolith-like materials (large, rubble, brecciated bedrock), with similar structure as found in Vesta (Denevi et al. 2012; Hoffmann et al. 2012), have low thermal diffusivity of $\kappa \sim {10}^{-7}-{10}^{-8}\,{{\rm{m}}}^{2}\,{{\rm{s}}}^{-1}$ (Haack et al. 1990; Fu et al. 2014). The diurnal thermal skin depth (at which the temperature is reduced to be a factor of 1/e), d_s, is estimated by ∼$\sqrt{\kappa {P}_{\mathrm{rot}}}$. Setting κ = 10⁻⁷–10⁻⁸ ${{\rm{m}}}^{2}\,{{\rm{s}}}^{-1}$ and P_rot = 2.343 hr we find d_s ∼ 0.9–3 cm. The blackbody temperature at the thermal skin depth is ∼120 K even at perihelion, below the sublimation temperature of water ice of 150 K (Yamamoto 1985). Conceivably, water ice might be preserved in the very shallow subsurface within a few centimeters.

To estimate the size of ejected dust particles coupled to the outflowing gas driven by the sublimation of water ice, the small source approximation (SSA) model is applied (Jewitt et al. 2014b). We assume a small patch of surface water ice on 2003 YT₁, and also assume that subsurface water ice acts in a similar way to the exposed ice. Spacecraft visits to comets find too small of ice exposure on the nuclei (67P/C-G; Hu et al. 2017) to explain the measured activities driven by sublimation on which a few ∼10s of percent of surface ice coverages are presumed to replenish (Tancredi et al. 2006). Alternatively, shallow subsurface water ice is proposed as having the most contribution (67P and Ceres; Agarwal et al. 2017; Küppers 2019). A non-rotating, spherical object is assumed for the physical essence of gas dynamics. This prevents complicated gas flows caused by the inhomogeneous distribution of gas release from the non-spherical object (Fulle et al. 2015; Agarwal et al. 2016). Then, ice sublimation from an exposed (≈subsurface) ice patch located at the subsolar point is examined. We solved the energy balance equation of a completely absorbing (sub)surface ice at the subsolar point, with 2003 YT₁ located at perihelion q = 0.786 au. The subsolar temperature at the d_s is ∼160 K, warm enough for water ice to sublimate. The flux energy completely absorbed from the Sun and energy lost from the asteroid surface by radiation and latent heat of ice sublimation are calculated. The resulting maximum specific mass-loss rate is (dm/dt)_ice = 8 × 10⁻⁴ kg m⁻² s⁻¹ at the subsolar point (at the highest temperature of 206 K of the non-rotating body). The terminal velocity in the SSA by gas drag is very small compared to the gravitational escape speed from the asteroid, but certainly assists in launching dust particles from the surface into interplanetary space. The radius of the ice sublimating area (patch), r_ice, is related with the critical size of dust particles to be ejected, a_c, as expressed by Equation (A6) of Jewitt et al. (2014b),

$$\begin{eqnarray}&&{r}_{\mathrm{ice}}\gt \displaystyle \frac{8\pi G{\rho }^{2}{D}_{{\rm{p}}}^{2}{a}_{{\rm{c}}}}{9{C}_{{\rm{D}}}{v}_{\mathrm{gas}}}{\left(\displaystyle \frac{{dm}}{{dt}}\right)}_{\mathrm{ice}}^{-1},\end{eqnarray} \tag{ 9 }$$

where C_D ∼ 1 is a dimensionless drag coefficient that depends on the shape and nature of the grain and v_gas is the thermal speed of gas molecules. We set a_c = 1 mm–1 cm using v_gas = 490 m s⁻¹ (Equation (10) of Graykowski & Jewitt 2019) and (dm/dt)_ice = 8 × 10⁻⁴ kg m⁻² s⁻¹ and again take the same values of G, ρ, and D_p. We then find r_ice > 3–25 m corresponding to the fractional area of (sub)surface ice of ∼0.001%–0.05%. This value is 10 times smaller than, or comparable with, those of Vesta (≳0.01%; Palmer et al. 2017) and the S-type Oort Cloud object (0.04%–0.1%; Meech et al. 2016). These give a crude but useful estimation, by showing that even a tiny (sub)surface ice coverage can release the meteoroid-sized particles. Yet note that no exposed water ice is observed on 2003 YT₁. Note also that it is difficult to detect subsurface ice by observations. Laboratory data find that even a few millimeter thick crust (organic mantle) perfectly attenuates the near-infrared absorption band depths of the subsurface water ice (Poch et al. 2016). Spacecraft missions for excavations like NASA's Deep Impact (A'Hearn et al. 2005; Kasuga et al. 2006, 2007) and JAXA's Hayabusa2 (Watanabe et al. 2017) would be advantageous for the detection in the kilometer-scale NEAs. Until then the sublimation of water ice, at least, remains as a potential dust production mechanism for 2003 YT₁.

Briefly we have examined a variety of processes capable of launching dust particles from 2003 YT₁. Rotational instability, impacts, and radiation pressure can produce millimeter to centimeter scale dust particles. By contrast, resurfacing and photoionization are implausible. Insufficient evidence exists in the thermal fracture and sublimation of ice, requiring future work.

We present the SonotaCo meteor survey of a fireball taken in Japan on UT 2017 April 28 at 15^h 58^m 19^s. The data is measured for orbit and physical properties. Specific detections give the following results.

• 1.  
  Radiant point, geocentric velocity, and orbital elements of the fireball are determined. The similarity to asteroid 2003 YT₁ with D-criterions (see D_SH = 0.0079) gives an order of smaller values than the significant threshold, indicating a parental association.
• 2.  
  The absolute visual magnitude is M_v = −4.10 ± 0.42 mag. Light curves give the meteoroid mass of m = 29 ± 1 g, which corresponds to the size a_s = 2.7 ± 0.1 cm with the density of 2700 kg m⁻³.
• 3.  
  The meteor luminous model comprising time derivative of momentum in the drag equation is suggested to employ a velocity-dependent ablation coefficient, as determined by σv² < 1.
• 4.  
  The 2003 YT₁ binary could be rotationally disrupted asteroids with mass shedding, consistent with Pravec & Harris (2007). The YORP spin-up timescale is τ_Y ∼ 2 Myr, which shortly induces rotational instability. The resulting end-state is a breakup/fission if it is the rubble-piled body held by a weak cohesive strength of S_c ∼ 240 N m⁻².
• 5.  
  Micrometeorite impactors with ≃1 mm in size sufficiently produce the centimeter-sized dust particles, given that they populated near the 2003 YT₁ orbit.
• 6.  
  Radiation pressure may sweep out the millimeter-sized particles from 2003 YT₁, which could be the source of faint meteors with an apparent magnitude of ∼+5 mag. The centimeter-sized particles are too large to be removed.
• 7.  
  The other dust production mechanisms are unprovable or pending.

We are grateful to SonotaCo for support. We appreciate Hideaki Muroishi, Hiroshi Yamakawa, Kazuhiko Yoneguchi, Naoya Saito, Hiroyuki Inoue, Chikara Shimoda, Toshio Kamimura, and Koji Okano for data contributions. We acknowledge Masahisa Yanagisawa, David Čapek, Takaya Okamoto, and David Jewitt for discussion, and David Asher for review. T.K. gives special thanks to Petr Pravec for the presentation source and to Daniel J. Scheeres as the scientific editor. Finally, we express a deep gratitude to Juraj Tóth and his local organizing committee (LOC) members of Meteoroids 2019 held in Bratislava, Slovakia for providing the opportunity to enhance this study.

Software: UFOCaptureHD2, UFOOrbitV2, UFOAnalyzerV2 (SonotaCo 2009, 2016, 2017; SonotaCo et al. 2014).

We used the SonotaCo and EDMOND databases to find other probable meteors that could be orbitally associated with asteroid 2003 YT₁ and the Kyoto fireball. Note that slow-speed meteors infer large uncertainties in the radiant points (Sato & Watanabe 2014; Tsuchiya et al. 2017), while the databases include little or nothing about estimations for errors on orbital information. Hence based on the asteroidal solar longitudes and radiant points in Table 2, we set the wide search ranges of λ_s−30° ≤ ${\lambda }_{{\rm{s}}}{\rm{^{\prime} }}$ ≤ λ_s+30°, α−30° ≤ α' ≤ α + 30°, δ−20° ≤δ' ≤ δ + 20° and v_g < 35 km s⁻¹ at the descending or ascending node, respectively. In which λ_s', α', and δ' are those of meteors in the databases. Among them we take any of the D-criterion for either the asteroid or the fireball that presents the value of <0.2. Selected meteors are shown in Tables 8 and 9. The scattered results may not be sufficient to be part of the association.

We present a procedure to develop a new meteor luminous model based on the classical model (Bronshten 1983; Ceplecha et al. 1998; reviewed in Popova et al. 2019). The meteoroid kinetic energy is transformed into radiation during the meteor flight. The classical luminous model equating mass loss (ablation), luminosity, and deceleration is given by (Chapter 3.4 in Ceplecha et al. 1998)

$$\begin{eqnarray}\begin{array}{rcl}I & = & -\tau \displaystyle \frac{d}{{dt}}\left(\displaystyle \frac{{{mv}}^{2}}{2}\right)\\ & = & -\tau \left(\displaystyle \frac{{v}^{2}}{2}\displaystyle \frac{{dm}}{{dt}}+{mv}\displaystyle \frac{{dv}}{{dt}}\right)\end{array}\end{eqnarray} \tag{ 10 }$$

$$\begin{eqnarray}&&\,=\,-\tau \left(1+\displaystyle \frac{2}{\sigma {v}^{2}}\right)\displaystyle \frac{{v}^{2}}{2}\displaystyle \frac{{dm}}{{dt}},\end{eqnarray} \tag{ 11 }$$

where I is the meteor luminosity, τ is the luminous efficiency, m is the meteoroid mass (g), σ is the ablation coefficient (s² km⁻²) (=kg MJ⁻¹), v is the meteor velocity (cm s⁻¹), and t (s) is time. The meteor luminosity is defined as $I={10}^{-0.4{M}_{{\rm{v}}}}$ in magnitude-based units in the visual region, where M_v is the absolute magnitude in the 100 km distance. The ablation coefficient is generally defined as σ = Λ/2QΓ, where Λ is the heat transfer coefficient, Q is the energy necessary to ablate an unit mass of meteoroid (MJ kg⁻¹), and Γ is the drag coefficient. The motion and ablation of a single non-fragmenting body through the atmosphere has been traditionally represented by the drag and mass-loss equations as (Chapter 3.2 in Ceplecha et al. 1998)

$$\begin{eqnarray}&&m\displaystyle \frac{{dv}}{{dt}}=-{\rm{\Gamma }}S{\rho }_{{\rm{a}}}{v}^{2},\end{eqnarray} \tag{ 12 }$$

$$\begin{eqnarray}&&\displaystyle \frac{{dm}}{{dt}}=-\displaystyle \frac{{\rm{\Lambda }}S}{2Q}{\rho }_{{\rm{a}}}{v}^{3},\end{eqnarray} \tag{ 13 }$$

respectively. Here S is the cross-section of meteoroid and ρ_a is the atmospheric density. The Equations (12) and (13) are related as

$$\begin{eqnarray}&&\displaystyle \frac{{dv}}{{dt}}=\displaystyle \frac{1}{\sigma {mv}}\displaystyle \frac{{dm}}{{dt}}.\end{eqnarray} \tag{ 14 }$$

Substituting Equation (14) into (10), we obtain the classical luminous model in Equation (11).

A new luminous model is developed by refining the drag in Equation (12). Since the ablation process can lose the mass of meteoroid itself, the drag force should be expressed in the differential form of the momentum (Nagasawa 1981, in Japanese). The drag in Equation (12) thus can be rewritten as

$$\begin{eqnarray}&&\displaystyle \frac{d}{{dt}}({mv})=m\displaystyle \frac{{dv}}{{dt}}+\displaystyle \frac{{dm}}{{dt}}v=-{\rm{\Gamma }}S{\rho }_{{\rm{a}}}{v}^{2}.\end{eqnarray} \tag{ 15 }$$

Substituting Equation (13) into (15), the refined drag equation is obtained as

$$\begin{eqnarray}&&m\displaystyle \frac{{dv}}{{dt}}=-{\rm{\Gamma }}S{\rho }_{{\rm{a}}}{v}^{2}(1-\sigma {v}^{2}),\end{eqnarray} \tag{ 16 }$$

where σv² < 1 is required. The new relation between dv/dt and dm/dt using Equations (13) and (16) finds

$$\begin{eqnarray}&&\displaystyle \frac{{dv}}{{dt}}=\displaystyle \frac{1-\sigma {v}^{2}}{\sigma {mv}}\displaystyle \frac{{dm}}{{dt}}.\end{eqnarray} \tag{ 17 }$$

By substituting Equation (17) into (10), we obtain the new luminous model,

$$\begin{eqnarray}&&I=-\tau \left(\displaystyle \frac{2}{\sigma {v}^{2}}-1\right)\displaystyle \frac{{v}^{2}}{2}\displaystyle \frac{{dm}}{{dt}}.\end{eqnarray} \tag{ 18 }$$

The ablation coefficient, σ, characterizes the ability of a meteoroid to ablate. A larger value produces higher mass-loss, resulting in brighter luminosity. The estimated values in the published literature are, however, highly scattered and inconclusive. The distribution of a coefficient (single-body theory) showing 0.01 < σ < 0.6 s² km⁻² is used to classify the meteoroid type, such as ordinary (0.014 s² km⁻²), carbonaceous chondrites (0.042 s² km⁻²), or soft cometary materials (0.21 s² km⁻²) and so on (Ceplecha et al. 1993). Later, on the contrary, a fragmentation process is suggested to be dominant for mass loss, finding the low σ = 0.004–0.008 s² km⁻² in any type of meteorite (Ceplecha & Revelle 2005). But the process also depends on the assumed models. The works are reviewed in more detail by Popova et al. (2019).

Here, we propose an appropriate ablation coefficient for the new luminous model therein. It concisely depends on meteor velocity, as determined by σv² < 1 (see Equation (16)). Setting v = 23.7 km s⁻¹ (fireball) finds σ < 0.00178 s² km⁻². We thus use σ = 0.0017 s² km⁻² for this study (see Section 3.2).