Asteroid 2026 JH2 Close Approach Visualizer
Interactive Earth-centered orbital visualization of the safe close approach of near-Earth asteroid 2026 JH2 on 18 May 2026 at 21:23 UTC. The asteroid passes at a closest approach distance of about 91,500 km (0.238 lunar distances), well inside the Moon orbit and outside the geostationary ring. Trajectory propagated with a geocentric two-body plus lunar perturbation model using a Dormand-Prince adaptive integrator. Data attribution: JPL CNEOS and the Minor Planet Center. Produced by Astrophyzix.
Asteroid 2026 JH2
Close Approach Visualizer
18 May 2026 at 21:23 UTC - 22:23 BST United Kingdom
Live Telemetry
Data: CNEOS / MPC - Astrophyzix
Inside Moon orbit, outside GEO ring
Safe flyby -- no impact risk
Asteroid 2026 JH2 will pass well inside the Moon orbit at 0.24 lunar distances, threading between Earth and the GEO satellite belt. At 9-20 meters diameter, it is far below the 140 m threshold for Potentially Hazardous Asteroids. Condition code 7 indicates orbit uncertainty that will improve with further observations.
Reference frame and model
The animation shows an Earth-centered, ecliptic-plane projection. The asteroid is propagated as a test particle under the gravity of Earth (point mass) with an additive lunar third-body perturbation. The Moon follows a circular sidereal orbit at the mean Earth-Moon distance. The state vector is held as a 64-bit float array [x, y, z, vx, vy, vz] in km and km/s.
Physical constants (provenance)
- GM Earth = 398600.4418 km^3/s^2 (JPL DE-series geocentric value).
- GM Moon = 4902.800118 km^3/s^2 (JPL DE-series value).
- Earth equatorial radius = 6378.137 km (IAU/WGS-84).
- Geostationary radius = 42,164 km.
- 1 lunar distance (LD) = 384,400 km (mean), also Moon orbit radius.
- Moon sidereal period = 27.321661 days.
- Close approach epoch = 2026-05-18 21:23 UTC.
Initial conditions (reconstruction)
A geocentric hyperbolic encounter is constructed to reproduce the catalogued close-approach geometry: perigee radius r_p = 91,500 km (0.238 LD) with a hyperbolic excess speed chosen so the speed at 130,285 km equals 9.21 km/s. This yields semi-major axis a approx -5,065 km and eccentricity e approx 19.1 (a fast, nearly rectilinear flyby typical of a small near-Earth object). Because the catalogued close-approach distance already folds in all perturbations, the inbound aim (impact parameter) is then fitted by a bisection shooting solve so that the full Earth-plus-Moon force model reproduces 91,500 km, rather than the bare two-body value. This is standard close-approach reconstruction practice. The start state is placed on the inbound asymptote at about 30 million km (around 39 days before perigee).
Equations of motion
acceleration = -GM_E * r / |r|^3 + GM_M * ( (r_m - r) / |r_m - r|^3 - r_m / |r_m|^3 ), where r is the geocentric asteroid position and r_m the geocentric Moon position. Specific orbital energy used for the conservation check is E = |v|^2 / 2 - GM_E / |r|.
Numerical integrators
- Dormand-Prince 5(4) with embedded error estimate and adaptive step control (default; used for the reference ephemeris).
- Runge-Kutta-Fehlberg 4(5) with adaptive step control.
- Classical fixed-step Runge-Kutta 4.
- Symplectic kick-drift-kick leapfrog (Verlet) for the energy-conservation demonstration.
Adaptive steps use a per-step relative tolerance of 1e-10 with standard PI-free step-size scaling and a safety factor of 0.9.
Validation against benchmarks
Running validation...
The catalogued close-approach distance is 91,500 km (0.238 LD). After the shooting fit, the full Earth-plus-Moon integrated perigee matches this value to the reported sub-percent tolerance. The symplectic leapfrog energy drift is bounded and oscillatory, as expected for a symplectic scheme, and is reported alongside the non-symplectic Runge-Kutta result as an independent measure of integrator accuracy.
Computational approach (scope statement)
Propagation uses a deterministic state machine and a single contiguous 64-bit float ephemeris buffer; the render loop takes zero-copy typed-array views into that buffer and, when available, caches the computed ephemeris in IndexedDB keyed by parameters and integrator. For a single planar test-particle propagation of this size the double-precision path resolves in well under a millisecond, so GPU compute, WebAssembly SIMD, SharedArrayBuffer threading and worker offload were deliberately not added: they would increase embedding fragility on a hosted blog page with no accuracy or performance benefit. Relativistic time-scale terms (TT vs TDB) and BCRS framing are below 2 milliseconds and far below the kilometre visualization tolerance over this arc, so a uniform dynamical time base is used and this approximation is disclosed here for full parameter provenance.
Limitations
This is an illustrative visualization, not an operational ephemeris. Distances are to scale relative to 1 LD. Solar radiation pressure, Earth oblateness (J2), planetary perturbations and the formal orbit-uncertainty covariance (condition code 7) are not modelled. Not for navigation.