The Large Synoptic Survey Telescope as a Near-Earth Object discovery machine

Abstract

Using the most recent prototypes, design, and as-built system information, we test and quantify the capability of the Large Synoptic Survey Telescope (LSST) to discover Potentially Hazardous Asteroids (PHAs) and Near-Earth Objects (NEOs). We empirically estimate an expected upper limit to the false detection rate in LSST image differencing, using measurements on DECam data and prototype LSST software and find it to be about 450 deg −2 . We show that this rate is already tractable with current prototype of the LSST Moving Object Processing System (MOPS) by processing a 30-day simulation consistent with measured false detection rates. We proceed to evaluate the performance of the LSST baseline survey strategy for PHAs and NEOs using a high-fidelity simulated survey pointing history. We find that LSST alone, using its baseline survey strategy, will detect 66% of the PHA and 61% of the NEO population objects brighter than H = 22, with the uncertainty in the estimate of ±5 percentage points. By generating and examining variations on the baseline survey strategy, we show it is possible to further improve the discovery yields. In particular, we find that extending the LSST survey by two additional years and doubling the MOPS search window increases the completeness for PHAs to 86% (including those discovered by contemporaneous surveys) without jeopardizing other LSST science goals (77% for NEOs). This equates to reducing the undiscovered population of PHAs by additional 26% (15% for NEOs), relative to the baseline survey.-3the Ecliptic, at the faint fluxes probed by LSST, is a few arcminutes (object counts are dominated by main-belt asteroids). Typical asteroid motion during several days is larger (of the order a degree or more) and thus, without additional information, detections of individual objects are "scrambled". However, with two detections per night, the motion vector can be estimated. The motion vector makes the linking problem much easier because positions from one night can be approximately extrapolated to future (or past) nights. The predicted position's uncertainty is typically of the order of several arcminutes, rather than a degree, which effectively "de-scrambles" detections from different nights (for a detailed discussion of this algorithm, see Kubica et al. 2007 as well as Appendix A for a theoretical derivation of expected scalings).Early simulations of LSST performance presented by Ivezić et al. (2007) showed that the 10-year baseline cadence would result in 75% completeness for PHAs greater than 140 m (more precisely, for PHAs with H < 22). They also suggested that with additional optimizations of the observing cadence, LSST could achieve 90% completeness. An example of such an optimization was discussed by Ivezić et al. (2008) who reported that, to reach 90% completeness, about 15% of observing time would have to be dedicated to NEOs, and the survey would have to run for 12 years. More recently, estimates of LSST yields have been revisited by Grav et al. (2016) (predicted PHA completeness of 62% fo...

ATM: An open-source tool for asteroid thermal modeling and its application to NEOWISE data

“…When systematic errors are included, this small difference in accuracy between IR-based and optical size estimates is further diminished. This remarkable result bodes well for future optical asteroid surveys, such as the Large Synoptic Survey Telescope (Ivezić et al 2019), which might deliver such size estimates for over 5 million asteroids (Jones et al 2018, and references therein).…”

High-fidelity Simulations of the Near-Earth Object Search Performance of the Large Synoptic Survey Telescope

“…where a = 0.76, b = 1.16 and x is again the normalized trailing factor (Jones et al 2017). As shown in Figure 10, the detection losses are a factor of two worse than the SNR losses, which immediately implies that significantly trailed detections found by LSST will have SNR 5.…”

“…The first one is the the loss that happens when a Gaussian or a PSF-like filter is used to identify the sources in the image. The source function is traditionally modeled according to the static and well defined sources like unsaturated stars (Jones et al 2017), therefore, if the model finds a trail, it only captures a fraction of the flux in it ( Figure 9). We call this magnitude loss the "detection" loss (Ivezic et al 2009) and it is described by the function ∆m T rail = 1.25 log 10 1 + cx 2…”

Discovering Earth’s transient moons with the Large Synoptic Survey Telescope

“…These features are similar to those of smallest PHAs. In their analyses, Jones et al (2018) the performance analysis of MOPS by these authors is not directly applicable for our purposes.…”