2024 YR4: Identification of Possible Precoveries in 2016 IPTF Data

Abstract

2024 YR4 is a 40-100 meter-diameter asteroid and former Torino Scale 3 object which currently has a roughly 4% chance of impacting the Moon on 2032 December 22, an event which recent studies suggest could pose a hazard on Earth due to impact ejecta. We present a search for, and identification of, potential precovery observations of the virtual lunar impactor in Intermediate Palomar Transient Facility (IPTF) survey data, as well as other publicly accessible surveys, dating from 2016. These candidate detections, not accounting for any currently-undetected Yarkovsky forces, predict a perilune of 22001 +/- 49 km and a perigee of 277534 +/- 46 km (relative to the center of each respective body) representing an improvement of > 300 times in the approach distance uncertainty above the existing orbit solution and, if confirmed, decisively ruling out a lunar impact in 2032. Using a matched filter tuned to 2024 YR4's predicted appearance in each image, we find the detection to be significant at Pnull = 5x10-9. The resultant possible orbit solution should be easy to confirm during 2024 YR4's 2028 approach to Earth, potentially greatly reducing the effort required by the planetary defense community at large to characterize 2024 YR4 before its potential lunar impact.

Figures (8)

• Figure 1: (Left to right, top to bottom) Representative frames of the datasets searched for 2024 YR$_4$ as described in Table \ref{['tab:possible images']}, in chronological order. The 3$\sigma$ line of variations is indicated with a line with the associated 2032 perilune distances marked with ticks at 10,000 km intervals, and positions corresponding to a 2032 lunar impact indicated with parallel lines -- displayed in greater detail on the upper right inset for each image. North is up and East is left; the true thickness of the LOV is less than a pixel in each image, including the insets.
• Figure 2: (Left of each column) Detections of a potential candidate for 2024 YR$_4$ in each individual IPTF image, with (right of each column) ZTF deep reference images in the same filter for comparison, centered on the expected location on each date using the nominal orbit produced by these measurements. The direction and length of 2024 YR$_4$'s motion streaking over the course of each exposure is indicated with a line. The expected SNR as described in Table \ref{['tab:IPTF images']} is included as well.
• Figure 3: Automated detection and significance testing methodology. $N$ samples were gathered in (a) along the LOV, then corrected with a synthetic flat. Reference images from ZTF were aligned, matched to the IPTF seeing, gathered along the LOV, flat-corrected and subtracted from the IPTF samples. Synthetic images of 2024 YR$_4$ were generated in (b) matching the streak length, seeing and expected magnitude of 2024 YR$_4$ at the time each IPTF frame was taken. The compensated image samples were then correlated against these reference images in (b), and null correlations were performed against the rest of the image (b) to estimate P-value $p_{i,k}$ for the data correlation coefficients $r_{i,k}$. Null and true detection distributions were also calculated from these same images through simulation of overlaid streaks in non-LOV areas. These null and true distributions were used to implement a maximum likelihood test against the P-values calculated in (b) to extract the most likely detection candidate[s], as shown in \ref{['fig:detection-testing']}
• Figure 4: Log-P-values $-{\log{(p_{i,x})}}$ plotted against frame index ($i$, vertical) and LOV sample coordinate ($x$, horizontal). Brighter regions are more significant (larger $-{\log{(p)}}$). The candidate detection is highlighted by a red arrow in the inset.
• Figure 5: Boxplots of predicted correlation P-value for each IPTF frame from simulated 2024 YR$_4$ detections. The actual correlations from the most significant detection at $x=0.4172$ are indicated by red points.