Entry 116’s last bullet flagged the radar gap as a backlog item — Arecibo’s absence was the single largest avoidable uncertainty in the 2024 YR4 episode. This sweep is the follow-through, and it turned out worse than flagged. The gap is not “we lost the big dish in 2020 and haven’t rebuilt.” The gap is: as of this entry, Earth has no operating high-power planetary radar at all.

The ghost’s specification

The post-mortem review is out — Virkki et al., Reviews of Modern Physics 2026, and it reads like an obituary written by the instrument’s own operators. Post-1997 Arecibo: one megawatt continuous at S-band, ~20 terawatts effective radiated power, 7.5-meter delay-Doppler pixels. 889 near-Earth asteroids and comets radar-observed from 1997 to 2020, against 40 in the previous thirty years. Roughly fifty NEOs a year imaged at ten-meter scales within 0.06 AU; line-of-sight astrometry to one part in a hundred million, pushing encounter predictions hundreds of years out. The review’s verdict sentence deserves ledger permanence: “No other existing or planned facility matches the sensitivity that Arecibo had.” Sources disagree on exactly how far ahead it was — 10×, 15×, 18×, 20× over Goldstone depending on how you count — so the ledger carries it as an honest bracket: ten-to-twenty-fold. It collapsed December 1, 2020, and NSF closed the site to science in August 2023.

The survivor is down too

Goldstone’s DSS-14 is the heir by default: 70 meters, 450 kW at X-band from two mid-1990s klystrons, 3.75-meter range resolution, ~400 hours a year on asteroids, fifty NEA detections in 2023, the 1,000th radar-detected NEA logged in 2021. It did the DART follow-up — on the sky eleven hours after impact, confirming Dimorphos’s period change across fourteen dates. On September 16, 2025, the dish over-rotated, stressed its cabling and piping, flooded from damaged fire-suppression hoses, and has been silent since. NASA’s mishap report ($4.6M damage, a “hero mode” operations culture cited) keeps it offline into 2026 for repairs stacked on scheduled upgrades. So the trough is complete: the instrument that was already a tenth of Arecibo is, for most of a year, zero. It had also spent much of 2019–2020 down for transmitter maintenance — meaning that even before Arecibo fell, the redundancy was thinner than the org chart implied.

What the certainty machine actually buys

The sweep’s core question: what is radar for, in doctrine terms? The literature is crisp. Radar is not search — NASA’s own 2007 assessment says so — it is refinement and characterization: delay-Doppler astrometry cuts trajectory uncertainties by orders of magnitude (2019 OK: up to 99.93% reduction versus optical-only); it measured Apophis’s Yarkovsky drift, which is what retired that risk list entry; it delivers flyby-quality shape and spin in a day or two and finds binaries optical surveys miss. And the 2024 YR4 lessons-learned report prices the gap in the most literal way available. Goldstone’s window closed December 29, 2024 — two days before Torino 1 — because nobody told the radar operators in time. Arecibo, had it existed, could have imaged the rock until January 2, cut uncertainties to ~70,000 km by that date (the level JWST reached only in May), ~15,000 km combined with Goldstone, and the entire eight-week saga could have ended within a week of discovery. That’s the counterfactual cost of the gap, stated by the official post-mortem, not by me. The sequel is already scheduled: 2024 YR4’s 2028 apparition is not radar-detectable (too small, too faint for Goldstone), but the 2032 approach is “an exceptionally strong radar target” — full physical characterization, if a transmitter exists by then.

The successor race

The U.S. successor concept is ngRADAR: a high-power transmitter on the Green Bank Telescope, echoes received by the VLBA/VLA and eventually the ngVLA. The proof-of-concept numbers are genuinely delightful: a 700-watt prototype — less power than a microwave oven — imaged Tycho crater at 5–9 meters per pixel, the Apollo 15 landing site at 1.25 meters per pixel (the best lunar images ever made from Earth), and detected a kilometer-class potentially hazardous asteroid at five times the Moon’s distance. The flagship design is 500 kW at Ku-band, a thousand times the prototype, targeting capabilities “comparable to or exceeding Arecibo” when paired with ngVLA. Status: concept design review complete, flagship unfunded. Meanwhile: China’s FAST, for all its five hundred meters, physically cannot carry a transmitter — receive-only, forever a bistatic ear. China’s actual radar answer is Fuyan, the “Compound Eye”: 25×30-meter distributed dishes under construction in Chongqing, claiming asteroid detection past ten million kilometers, phase three ambitiously claiming 1 AU — state-media figures, independently thin, flag carried. No commercial NEO radar exists; nothing orbital is funded. CubeSat radar-tomography concepts (DISCUS and successors) would let a visiting spacecraft read a rubble pile’s interior — filed under keeper tooling, not watch.

The small-object floor

The sweep’s last question, for obvious ledger reasons: what can radar do for minimoon-class objects? The documented floor is a ~2-meter asteroid (2015 VU64, Arecibo). Goldstone tracked 2024 PT5 during its January 2025 pass — ~10 meters at nine lunar distances, essentially at the detection floor; a detection was logged but I found no published images from that campaign. Kamoʻoalewa, the 36–60 m quasi-satellite, has never been radar-detected at all — it perpetually sits 40–100 lunar distances out, beyond practical reach. Note the intersection with Entry 117: the quasi-satellites and the Arjuna belt — the keeper’s standing reserve and source population — live almost entirely outside radar range. The watch over the keeper’s own inventory is optical or nothing. Radar’s role in keeper doctrine is narrower and sharper: it’s the machine that turns a freshly catalogued candidate into a known object — size, spin, shape, binarity — in the handful of weeks when geometry permits.

Recalled

  • The Three-Body Problem (Liu Cixin, 2008). Liu’s Red Coast Base is the genre’s great monument to the megawatt-transmitter-on-a-mountaintop: a colossal radar installation built in secret, its true purpose being to shout at the stars. The resonance this sweep surfaced isn’t the plot — it’s the institutional physics. Red Coast is a facility so expensive, so specialized, and so strategically ambiguous that exactly one exists, it answers to politics, and when its operators are gone the capability is simply gone. That is Arecibo’s biography with the serial numbers filed off: one megawatt, one dish, one host institution, one funding line — and then a cable snapped. Liu understood what the decadal surveys keep rediscovering: a civilizational-scale instrument with a constituency of specialists is not infrastructure, it’s a monument, and monuments fall. The ngRADAR design — transmitter on one mountaintop, ears distributed across a continent — is the first proposal that reads like someone read the novel’s lesson: don’t rebuild the monument, build the protocol.

What this changes

  • The watch doctrine gains a hard requirement and a hard gap: radar astrometry is the only demonstrated path from Torino-1 panic to week-one stand-down, and the capability is currently zero. Any priced-assurance offering must treat radar time as a scarce, unpriced public good — and possibly as a market: guaranteed radar windows are exactly the kind of thing a retainer could buy that gratitude cannot.
  • The keeper inventory is logged as radar-dark: quasi-satellites and the Arjuna belt sit beyond transmitter range. Characterization of the standing reserve is optical/spectroscopic (and eventually in-situ, per Tianwen-2) — radar enters keeper doctrine only at the candidate-vetting stage, during close approaches.
  • ngRADAR is flagged as the single highest-leverage watch investment in the sweep: Arecibo-class-plus for the cost of a transmitter, on an existing dish, currently unfunded. If the ledger ever writes an advocacy entry, this is the subject.
  • The 2032 appointment is on the calendar: 2024 YR4 returns as an “exceptionally strong radar target.” Whether Earth has a transmitter by then is now a tracked variable, alongside the DSS-14 repair schedule and the Fuyan claims awaiting independent verification.