Essay

The FAA would let a booster take the lightning strike if nobody is in reach

Aviation has two answers to lightning: stay out of the cell, or build the airplane so a strike cannot bring it down. On October 5 the FAA proposed a third for rockets on the way down. Let the strike happen, as long as physics keeps every piece it could knock loose inside a box with nobody in it.

Rockets already had the first two. Section 450.163 of Part 450, the launch and reentry licensing rule in effect since March 10, 2021, says an operator “must” either set flight commit criteria that keep the vehicle away from a strike or “use a vehicle designed to protect safety-critical systems” through one. The FAA’s proposal adds an exception for descending phases of flight. Once the vehicle descends to an altitude where lightning is foreseeable, the operator can skip both if the vehicle’s “further flight path, including that of any possible resulting debris, is constrained by physics to remain within any area clear of the public and critical assets.” The worked example is a first stage coming back to a barge, through its coast and landing burn. That barge is no hypothetical. SpaceX and Blue Origin both land boosters on ships.

The FAA pitches this as burden reduction, one of five commercial-space proposals it printed that morning under Executive Order 14335, which told the Secretary of Transportation to “reevaluate, amend, or rescind” Part 450. The filing does two larger things. It widens a practice the agency says it already allowed, and it rests the whole lightning case on two legs, one engineered and one procedural.

Case by case, for reentries. During licensing, the FAA says, operators showed “that a reentry operation can mitigate some lightning hazards through clearance,” through an equivalent level of safety (ELOS) finding that the vehicle’s potential impacts “on an unguided trajectory” stayed inside a cleared area. An ELOS is not a waiver. Under § 450.37 it is a showing, made “clearly and convincingly,” that another route is as safe as the rule. The proposal exists “to allow more operators to take advantage of the rationale underlying these ELOS determinations,” and it names “first stage returns,” which fly under a launch license. The precedent the FAA describes is reentries. The rule reaches boosters.

The idea is older than the exception. The 2019 proposal for Part 450 carried a lightning option the FAA says was “intended to cover an operator’s use of physical containment.” The 2020 final rule cut it as unnecessary, since a contained vehicle, “by definition,” lacks the energy for its hazards to leave the hazard area. The text it kept still says “must,” and when operators could neither avoid nor harden, the new filing says, a contained, cleared area got them through by ELOS. Those grants are not in the Federal Register. The agency’s account is the only public one.

Two legs. The filing’s safety case is one sentence: the exception “preserves public safety because the hazards of a lightning strike are mitigated by physical containment and ensuring the area is clear of public and critical assets.” Containment is engineered. The applicant must describe a valid method for the limits of the flight path and its debris, held to the § 450.101(g) standard of accurate data, scientific principles, and statistical validity, with its assumptions stated and its validation shown. Clearing is a procedure. The applicant submits “a description of the methods used to ensure that the areas are cleared,” and where the box runs past the hazard areas an operator must already clear, the barge example fills the gap “by performing surveillance of the non-bounded area.”

So the trade is a weather call for a clearing call. Lightning near the landing zone drops off the lightning rule’s go/no-go list, and the empty box carries the load.

The money. Small: $3,131.16 per application, $715.39 for industry and $2,415.77 for the FAA. One sentence of the prose says $3,013 for the FAA and $3,729 in all. Each is about $597 above what the tables and two other sentences say, which looks like Table 2’s $597.40 supervisory-approval line added a second time. The FAA gives no annual total. Its same-day flight safety analysis proposal expects “two to five new applications per year” and counts 179 license modifications in five years. If every one used this exception, my arithmetic puts the savings at most around $125,000 a year.

Who wrote the numbers. A licensor with two jobs: act “consistent with public health and safety,” and “encourage, facilitate, and promote commercial space launches and reentries by the private sector” under 51 U.S.C. 50903. Like the four proposals printed beside it, this one expects to be “an E.O. 14192 deregulatory action.” None of that makes the numbers wrong, but one line is missing. The cost table books the ELOS review itself as saved, ten engineer-hours of staff review and discussion with the applicant, and prices nothing for checking the descriptions that replace it, though the FAA says it will use them “to evaluate whether the operator’s methods for clearance are appropriate.” That checking may always have lived inside the license evaluation, which the filing never priced. It does not say.

The case for it. An ELOS was never relief from safety, and neither is this. Section 450.37(b) keeps the core public-risk criteria of § 450.101, the (g) standard among them, beyond the reach of any ELOS, and the exception relieves § 450.163(a) alone. The preamble sums up the test as a flight path “clear of members of the public and critical assets.” The binding text is stricter, with debris included and the containment “constrained by physics.” Clearing is an old duty: § 450.161 already requires surveillance “sufficient to verify” the safety analysis. A struck booster can only land where physics lets it, the same as one that loses an engine. Part 450 protects the public, and an operator that reuses its boosters has its own reasons to keep them out of storms. And a published test beats a private grant, because every operator can read it.

Concede all of it. A bound set by what a vehicle can physically reach is the most engineered margin a regulator can write, and the floor under it does not move. Two things survive the concession. The first is reach. The analysis the FAA describes bounded reentries “on an unguided trajectory.” Its lead example is a booster under a landing burn, engines lit and a computer in the loop. Rockets set strikes off, too. In March 1987 Atlas-Centaur 67 triggered a flash that hit it about 48 seconds into flight, altered one word in its flight computer’s memory, and made the computer command the engines hard over. The vehicle broke up. That was on the way up, and a landing burn has the same ingredients. The FAA’s answer for the barge sits in its companion containment proposal, printed the same morning: in the final landing burn, “nearly depleted propellant and low velocity impose a physical limit on the reach of any hazard.” That describes a nominal burn. Section 450.119 already makes operators compute a malfunctioning vehicle’s “deviation capability,” so the tool exists, but nothing published shows it bounding a struck stage under thrust.

The second is the clearing. The duty to clear is old. What changes is that for this phase the clearing is the last layer between a strike and the public, and beyond the existing hazard areas the showing for it is a description of surveillance.

Comments close November 4. The useful ones will ask whether “constrained by physics” counts the thrust a struck stage still has, and how the surveilled part of the box gets verified empty.

Physics can draw the box. Somebody still has to look inside it.

Sources: