Solar versus fission on Mars: the dust storm decides

What the work claims

For two conceptual Mars missions, an uncrewed ISRU pathfinder at the equator and a four-crew 500-day surface mission at Jezero Crater, the team developed point-design power systems and compared solar options against a fission baseline.¹ For the crewed case, they found a distributed solar network of Orion-derived Ultraflex arrays plus lithium-ion batteries on every lander would mass about 9800 kg and could supply the 22 kW keep-alive load needed to survive a dust storm of up to 120 days at average optical depth 5, plus 35 kW peak power for normal clear-sky operations. Four 10 kW Kilopower fission reactors, at 7000 kg, provide full power regardless of dust.¹ The solar figure is less than half the mass estimated in a comparable 2007 NASA study, but still heavier than fission; landing-site selection, the authors conclude, is a key driver of the power decision. These are point designs from a trade study, not measured hardware: every mass figure is an engineering estimate, and the solar concepts carry a stated 30% mass growth allowance.

How it works

The comparison turns on one physical asymmetry. A fission reactor outputs the same power day, night, and through a dust storm, so its mass is roughly fixed by rated power plus shielding and radiator. A solar system's mass is set by two different operating points that must both be served: the clear-sky peak (35 kW here, sized by array area) and the storm survival point (22 kW keep-alive for up to 120 days, sized mostly by battery energy, since a global dust storm at optical depth 5 cuts sunlight to a small fraction of clear-sky levels).¹ Batteries, not arrays, dominate the storm case, and battery mass scales roughly with outage duration, which is why the solar mass is so sensitive to the assumed storm severity.

For the uncrewed pathfinder at the equator, the asymmetry nearly disappears. Three solar concepts were priced: daylight-only operation at roughly half the daily propellant production of a round-the-clock fission design, a battery-augmented system operating through the night at the fission production rate, and a daylight-only system at higher rate matching fission production. Including the 30% mass growth allowance, total masses across the three concepts ranged from 1,116 to 2,396 kg, against 2686 kg for the fission scheme.¹ Daylight-only solar wins at the equator precisely because there is no storm-survival requirement on a short demonstration and no crew to keep alive. The team flags the catch themselves: a daytime-only ISRU plant cycles on and off every sol, and the reliability of that cycling needs detailed analysis.

The latitude sensitivity check sharpens the geography conclusion. At Columbus Crater, 30 degrees south, solar array mass did not increase relative to Jezero, because dust storms do not occur in the southern hemisphere during winter; but longer nights demand more battery mass, and additional ISRU strings are needed for summer-winter cycling.¹ The trade therefore moves along different axes at different sites: storm statistics drive storage at low latitudes in the north, night length drives it in the south. For arithmetic the paper does not state: the solar system delivers 35 kW peak from 9800 kg, about 3.6 W/kg system level, while the fission system's 40 kW from 7000 kg is about 5.7 W/kg; both ratios are my calculations from the paper's own mass and power figures, and the fission figure's stated boundary is the four-reactor system mass as given, without a stated growth allowance comparable to the solar concepts' 30%.

Steelman: why the result matters

This is the study that almost overturned the field's prior. The 2007 NASA comparison had concluded fission was more reliable at much lower mass penalty; by 2016, better cells and better batteries had cut the solar estimate to less than half, close enough that site selection, not technology, tips the balance.¹ That is a genuinely useful engineering result: it converts a religious argument (solar versus nuclear) into a coupled optimization of landing site, load profile, and storage duration, which is how the decision should actually be made. The methodology is also appropriately humble for a trade study: explicit growth allowance, named designs derived from real hardware lineages (Orion Ultraflex arrays, Kilopower reactors), and an explicit sensitivity check at a second site rather than a single-point pronouncement.

Where a skeptic should push

The entire crewed-case mass gap is hostage to one requirement: survive 120 days at optical depth 5. The paper gives no derivation of that storm model in the two pages available; if the design storm were 150 days, or optical depth 6, battery mass would climb steeply and the solar case unravels, because storage scales with outage duration while fission mass does not. The solar figures benefit from a 30% growth allowance that is stated, but the fission 7000 kg figure carries no comparable allowance in the text, so the comparison's 2800 kg gap may be partly an accounting asymmetry rather than physics. Battery behavior at Mars temperatures, cycle life over hundreds of sols of storm-cycling, and degradation during the 120-day outage are not addressed in the available text. The equatorial pathfinder win for solar assumes the site is at the equator, which no serious candidate landing site is. And the reliability concern the authors raise themselves, frequent on-off cycling of a daylight-only ISRU plant, is exactly the kind of operational failure mode that point-design mass studies systematically underweight. Finally, these are conceptual point designs by a study team; no hardware in this paper has been built or tested.

What Martian dust storms mean for surface power

Implication: the load-bearing insight is that for solar-plus-storage architectures, the sizing event is not the average sol but the tail event, and the tail event is a seasonal, latitude-dependent climate phenomenon. That means Mars surface power cannot be designed before the landing site is chosen, and the site cannot be chosen without a power model: crewed-site selection becomes a coupled decision where storm statistics at 18 degrees north carry a 2800 kg price tag.¹ The non-obvious corollary is that southern-hemisphere sites buy storm immunity with night-length storage instead, moving the mass from storm batteries to night batteries; there is no site that escapes the storage bill, only different reasons to pay it.

Threat: the whole solar case rests on storage endurance that is unproven in this application. Lithium-ion batteries that must hold a crew alive through 120 days of optical depth 5, with degraded array input, cycling at low state of charge in a cold environment, are the single least-validated element of the 9800 kg system, and they are precisely the element that cannot fail. A storm model error is not a performance shortfall here; it is a crew-loss scenario, which is why the authors note a lengthy outage poses a safety hazard for crewed missions.¹

Opportunity: the study identifies the exact leverage points where a modest technology advance flips the answer. Storm-season solar at Jezero loses by 2800 kg; closing that gap needs either cheaper storage energy density (batteries, or a regenerative fuel cell handling the 22 kW keep-alive load) or a site at lower dust-storm risk with the mass moved to night storage, which is a battery problem with better-understood duty cycles. The daytime-only ISRU cycling question the authors flag is also a near-term experimental target: running an oxygen-production plant through hundreds of deliberate on-off cycles on Earth would retire the main reliability objection to the lightest solar concept at a bench-level cost.¹

The bottom line

Within its stated assumptions, the result is a well-bounded near-miss: solar nearly reached parity with fission for crewed Mars surface power and beats it outright for short equatorial missions without storm-survival requirements. Confidence is moderate: the framework is sound and the sensitivity to site is clearly demonstrated, but the decisive 120-day, optical-depth-5 storm requirement is taken as given, the growth-allowance accounting is asymmetric between the two systems, and nothing here is hardware. What would confirm the solar case: validated battery endurance data through months of simulated storm conditions at Mars temperatures, and a fission point design carrying the same stated growth allowance. What would break it: a design storm materially longer than 120 days, which would push storage mass up on a nearly linear slope while the fission number stays flat at 7000 kg.