Efficiency of a wide-area survey in achieving short- and long-term warning for small impactors

We consider a network of telescopes capable of scanning all the observable sky each night and targeting Near-Earth objects (NEOs) in the size range of the Tunguska-like asteroids, from 160 m down to 10 m. We measure the performance of this telescope network in terms of the time needed to discover at least 50% of the impactors in the considered population with a warning time large enough to undertake proper mitigation actions. The warning times are described by a trimodal distribution and the telescope network has a 50% probability of discovering an impactor of the Tunguska class with at least one week of advance already in the first 10 yr of operations of the survey. These results suggest that the studied survey would be a significant addition to the current NEO discovery efforts.

💡 Research Summary

The paper investigates the performance of a proposed wide‑area, all‑sky nightly survey designed to detect small Near‑Earth Objects (NEOs) in the 10 m–160 m size range, which are capable of causing regional devastation. The authors introduce two key concepts: “blind time,” the interval after a survey’s start during which a given impactor cannot be detected before impact (often because it approaches from the Sunward direction), and “lead time,” the interval between the first successful orbit determination and the actual impact. To quantify the minimum lead time required for effective mitigation, they define a simple exponential function of absolute magnitude H: t(d) = 30 e^{‑0.5(H‑22)} days, yielding 30 days for H = 22 (≈160 m) and about one week for H = 24.5 (the Tunguska‑size, ≈50 m).

The simulation uses the synthetic impactor set of Chesley & Spahr (2004), comprising 4 950 objects whose orbital elements (a, e, i) follow the Bottke et al. (2002) NEO model. Each impactor is assigned a fixed H value ranging from 22 to 28 (approximately 160 m down to 10 m) and the population is divided into ten 10‑year bins relative to the start of the survey, allowing the authors to track discovery performance as a function of elapsed survey time.

The envisioned optical network relies on a “fly‑eye” telescope design with a 1 m equivalent aperture, 45 deg² field of view, high‑efficiency CCDs (80‑90 % QE), and rapid 2 s readout. One such telescope is placed in each hemisphere, each covering roughly half the sky. In addition, a dedicated follow‑up telescope per hemisphere is positioned ≈30° west of the survey instrument. Each night (10 h of observing) a telescope can acquire 766 images, amounting to about 34 500 deg², i.e., the entire visible sky (≈88 % of the celestial sphere) is imaged twice per night, excluding regions within 40° of the Sun, 30° of the Moon, and 15° of the Galactic plane. The limiting magnitude for the survey mode is V = 21.5 mag (≈45 s exposure), while the follow‑up mode reaches V = 23 mag.

Data processing proceeds in real time: images are calibrated, detections are linked into “tracklets” (2–5 detections within 15 min–2 h), and tracklets are combined to produce preliminary orbits using the methods of Milani et al. (2004, 2010). A detection is considered a true discovery only when an orbit based on at least four tracklets (six orbital parameters) is obtained. Gaussian noise of 0.3″ is added to astrometry, and photometric errors of 0.2 mag (correlated) plus 0.2 mag (random) are applied. The simulation does not include background main‑belt asteroids, but prior work shows that >99.5 % of cases with ≥4 tracklets lead to successful impact probability computation (CLOMON2).

Results are presented as differential completeness curves (Figure 4). For objects with H = 23 (≈100 m), about 70 % are discovered within the first decade, reaching 90 % after roughly 30 years. For H = 25 (≈30 m), 60 % completeness is achieved after two decades, but the 90 % level is not reached even after 100 years. The smallest objects (H = 28, ≈10 m) start below 40 % and increase slowly. When the minimum lead‑time requirement is imposed, the effective completeness drops modestly: H = 23 reaches ≈60 % in the first decade and 90 % after 30 years; H = 25 reaches ≈50 % after two decades, never attaining 90 % within the simulation span.

These findings demonstrate that a wide‑area nightly survey can provide meaningful early warning for small impactors, especially in the 10–30 m range where current deep surveys are ineffective due to their longer revisit times and limited sky coverage. The ability to scan most of the sky twice per night reduces blind time, even for objects approaching from near the Sun, because earlier apparitions can be captured. However, the authors note that real‑world factors—weather, cloud cover, lunar brightness, and the need for multiple telescopes to ensure longitudinal coverage—will likely increase the required number of units to 5–6 per hemisphere.

In conclusion, the proposed survey would complement existing US‑based deep‑survey programs, offering a substantial improvement in the probability of detecting Tunguska‑scale (≈50 m) impactors with at least a week’s warning within the first ten years of operation. This capability could enable timely mitigation actions such as orbit alteration or evacuation, thereby reducing the potential human and economic losses from small but hazardous NEO impacts.