February 11, 2024

The Small Carrier Problem

The USN's supercarriers are extremely capable, but also very expensive, and even the USN can only afford a few. As a result, every few years, the suggestion of building smaller, cheaper carriers comes up. In recent years, these have gotten louder, driven in large part by the development of the F-35B, which can operate without the need for catapults or arresting gear and offers capability unmatched outside the F-35 family. There have even been tests with USS America, normally an amphibious helicopter carrier, serving as a "Lightning Carrier". Unfortunately, there are serious practical problems with all of this. The simple fact is that there are major economies of scale in carrier aviation, and STOVL isn't going to remove them, nor is something like America, lovely though she is, a good substitute for the CVNs.

But where do these economies of scale come from? The first and simplest is the fact that steel is cheap and air is free. What really costs is not size, but capability. A full-capability carrier is going to need radar, big engines, defensive weapons, mission-planning systems and fancy communications gear, all of which costs a lot more than the bare hull, regardless of the size of the carrier. Not to mention things like catapults, arresting gear, and heavy maintenance facilities, all of which need to be onboard to make a conventional carrier regardless of the size of the air group. Nor is the air group going to scale linearly. Sure, you can cut the fighter complement from four dozen to two dozen, but there's going to need to be some minimum number of fighters kept for self-defense, and since the threat doesn't halve for a smaller carrier, that might not be something you can cut too much, leaving fewer than half of the previous complement for strike missions. Things are even worse for the rest of the air wing. The complement of E-2 Hawkeyes (typically 4-5) is set by the need to keep one airborne at all times while allowing maintenance, and the size of the MH-60R detachment similarly comes from the need to keep a helicopter or two active to hunt subs. The EA-18G and MH-60S detachments could be cut, but you're still looking at needing more than half of the carrier wing for less than half the capability.

Of course, some of this can be compromised if you're willing to abandon CATOBAR1 carrier operations and rely on STOBAR2 or STOVL3 instead. These let you discard the catapults and, in the case of STOVL, the arresting gear, allowing aircraft to fit onto much smaller ships. But there are several major drawbacks. One is that a smaller ship inherently carries fewer aircraft, which limits striking power. Another is that ski jumps limit the use of fixed-wing aircraft to high-performance fighters, which most notably takes the E-2 Hawkeye off the table. Yes, it is possible to use a helicopter for AEW&C,4 but because a helicopter has to beat the air into submission to fly, it will lack the range and endurance of a fixed-wing aircraft. It also closes off the possibility of other low-performance aircraft for missions like sea surveillance or ASW, as performed by the retired S-3 Viking or the upcoming MQ-25 Stingray.

STOBAR and STOVL operations also impose significant performance penalties on the aircraft involved. Quantifying aircraft performance is quite difficult, so I'm not going to touch STOBAR right now, but for STOVL, I can turn to the excellent Command: Modern Operations. I loaded up two aircraft of each variant, one with two internal GBU-32s (Mk 83/1,000 lb bombs) and one with the maximum load (4 in the F-35B, 6 in the other two) of GBU-31s (Mk 84s/2,000 lb bombs).5

AircraftWeaponRange (nm)
F-35BMk 84 x 4163.6
F-35AMk 84 x 6253.8
F-35CMk 84 x 6280.8
F-35BMk 83 x 2321.0
F-35AMk 83 x 2789.5
F-35CMk 83 x 2816.1

I also investigated station time over base at 36,000' with a load of 4 internal AMRAAMS.

AircraftLoiter Time (h:mm)

Basically, when in full bomb truck mode, the conventional versions can carry 50% more payload than the F-35B 50% further, or the same light payload over twice as far. On CAP, they have close to three times the loiter time, which translates into a lot less time spent operating aircraft. Although the F-35B may be by far the best STOVL fighter ever built, the limitations of STOVL remain abundantly clear. Particularly in a world where the range of US carrier wings comes in for frequent criticism, advocacy for "Lighting Carriers" is a rather strange thing to do.

The USN is well aware of everything discussed above, and has a long history of fighting off attempts by various groups to downsize the carriers. In some ways, this makes the saga of the Americas surprising. The first two units of the class, America and Tripoli, were built without a well deck, to allow for a bigger hangar and more aviation supplies.6 This made them uncomfortably close to light carriers for the USN's taste, and it is rumored that making sure that Congress didn't get confused is the reason that the third unit of the class, Bougainville, got a well deck. Although tests with 20 or so F-35Bs have been carried out aboard Tripoli in 2022, the Americas are not particularly good carriers. Their design is definitely specialized for amphibious assault, with large areas for vehicle stowage and wide corridors for loaded Marines, which compromise survivability. Their magazines are set up to provide weapons for half a dozen F-35s operating in support of amphibious operations, less than 5% of the volume available on a Nimitz or Ford. Oh, and they're a lot slower than the proper carriers, capable of 22 kts instead of the 30 or so that the CVNs can make. This also explains the lack of a ski-jump. It would make them too similar to a real carrier, and take up valuable helicopter space during an amphibious assault.

On the whole, the simple fact is that the small carrier is a bad idea if you have other options. The F-35B is a formidable aircraft, and it's far better than nothing if it's the only way for your nation (or service) to get its own sea-based airpower. But it is an inferior substitute for a big CATOBAR carrier for a nation with the resources and interests of the United States.

1 Catapult-Assisted Takeoff But Arrested Recovery

2 Short Takeoff But Assisted Recovery

3 Short Takeoff Vertical Landing

4 Airborne Early Warning and Control. Sometimes also called AWACS.

5 The one caveat to these is that CMO only has one load condition for each aircraft, and the maximum takeoff weight for the F-35B is probably going to vary a lot depending on the operating environment. That said, I suspect they picked a value for typical shipboard STOVL operations, so this should be broadly accurate, even though I was operating from a land base for the test.

6 A purist would argue that this means they should have been designated LPH-13 and 14.


  1. February 12, 2024Ian Argent said...

    Nuclear-powered CATOBAR carriers are the ultimate in "eschew foreign entanglements, but still wish to project power globally for long periods of time."

    Since they're basically large airbases that can go anywhere there's sufficient water under their keels

  2. February 12, 2024ack-acking said...

    What if... just spitballing here... What if we built a modern version of the IJN I-400 submarine aircraft carrier (https://en.wikipedia.org/wiki/I-400-class_submarine)

    Optimized for STOVL, of course. Imagine something like the Soviet Typhoon-class, but for launching planes instead of missiles.

    You wouldn't need a catapult, big engines, or big radar. You could also skip most of the defensive weapons and planes, relying on stealth instead to keep it safe. And it could get closer to its target than a carrier, so range would be less of an issue. Just pop up close to the target, launch enough planes for the strike mission, and then submerge again until its time to recover them.

  3. February 12, 2024John Schilling said...

    ack-acking: Volume is going to be a killer for this app. Air is free, and maybe steel is cheap but it takes a lot of steel, the good stuff, to enclose a hangar's worth of air and keep it leak-free at depth. Plus, you've got to push that entire volume through the water, not just the submerged displacement. The largest submarines ever built probably have about the useful volume of Thailand's "Chakri Narubet", and nobody ever expected Narubet to fight anything more than pirates.

    But even if you could make it work, carriers are about much more than strike missions. Providing fighter cover for other friendly forces, surveillance across a broad area, and ASW are all important and poorly suited for something that can only occasionally launch small groups of aircraft. Plus, of course, the "presence" mission, that submarines almost by definition don't do.

    And for the strike mission, you've got a critical vulnerability during the recovery portion. The enemy knows he's been struck, so he's got a pretty good idea where your strike aircraft are, you don't have the fuel to do much beyond go straight home, and your planes aren't all that stealthy from behind. So you recover your strike aircraft, and twenty minutes later the enemy's equivalent of a P-8 is circling overhead dropping sonobouys. And then torpedoes.

    Or, you notice that we can pack a hundred and fifty expendable strike aircraft into an Ohio hull, if we don't insist on recovering them. You can do a lot of striking with 150+ sorties.

  4. February 12, 2024redRover said...

    @ack-acking and @John Schilling

    You also have to submerge said volume of air. (I don't think the F-35B would take kindly to being treated as a submersible weapons system...)


    How much do the loiter time numbers change based on the takeoff mode of the F-35B? Looking at the world's most reliable open source aircraft spec site (Wikipedia) the delta in combat radius is 160nm, which is probably 45 minutes flying time at their cruise speeds.

    For some of the other stuff, it seems like tilt-rotor is an inferior but workable compromise as an engineering matter for the support aircraft, though by the time you do all of the engineering to make an EV-22 and a CV-22 and a KV-22 it's probably cheaper to just buy a bigger carrier.

    Also, going the other way, why not make the current CVNs bigger? I assume you start to run up against port limitations and the like, as well as a big infrastructure hurdle in terms of dry docks, but the largest container ships are about 3x the displacement and another 50m longer. They're admittedly slower at ~20kts, but with nuclear propulsion that doesn't seem like a big deal.

  5. February 12, 2024bean said...

    It's worth pointing out that "steel is cheap and air is free" does not apply to submarines. Air in particular is very much not free (volume is very expensive on a sub), and submarine-grade steel costs a lot more than the normal stuff. Beyond that, what John said. Recovery kills submarine aircraft carriers, particularly today, and you can fit a lot more Tomahawks onboard.

    and a CV-22

    Cue screaming from the USN. (There's a reason that the Navy and Marines have gone very far to keep that designation at bay. I really wish they'd used UV-22 instead of CMV-22.)

    Also, going the other way, why not make the current CVNs bigger?

    The Fords are as big as they can possibly be without a lot of extremely expensive infrastructure work. More specifically, we'd have to rebuild all of the relevant drydocks, and that doesn't come cheap. There's also the fact that operating more than about 90 planes doesn't work too well, and that's about how many a Nimitz or Ford can theoretically carry, although they've been at more like 70 for the last couple decades. I have no idea how they drydock container ships, but those docks aren't qualified to work on nuclear-powered warships.

  6. February 13, 2024redRover said...

    I have no idea how they drydock container ships, but those docks aren’t qualified to work on nuclear-powered warships.

    My understanding is that they just use a normal (very large!) drydock. Looking at the Daewoo yard in South Korea, it appears that they have a mix of normal and floating drydocks. For cruise ships it also appears similar, with Grand Bahama (designed for cruise ships so that they don't lose revenue by moving out of the Carribean) able to take ships up to 980' on a floating drydock, and the construction yards in Finland and France appearing to use traditional drydocks.

    There’s also the fact that operating more than about 90 planes doesn’t work too well, and that’s about how many a Nimitz or Ford can theoretically carry, although they’ve been at more like 70 for the last couple decades.

    Sure, but a bigger carrier lets you have bigger (next generation) planes while still keeping the air wing at 70ish. Like, the F-15E has an MTOW of 81K lbs, vs. 70K for the F-35C and 66K for the F-18E. A 15% larger plane seems like it should net you some valuable improvement in range or armaments or sensors or whatever. (Plus you can make the ship more habitable and hopefully more attractive on the retention front, which I think is likely to be a bigger issue in the intermediate term)

  7. February 13, 2024muddywaters said...

    To put some rough numbers on size not being why carriers are expensive: the biggest oil tankers (about 500k tons loaded displacement) cost about $100M and container ships (300k tons) about $200M. Cruise ships (100k tons) are rather more expensive at $1.3B, but still well short of the carrier's $10B.

    The air wing costs about $4B-$7B.

  8. February 13, 2024Belushi TD said...

    Warning! Nitpick Ahead!

    Isn't CATOBAR defined as Catapult Assisted Take Off Barrier Arrested Recovery?

    Belushi TD

  9. February 13, 2024redRover said...


    That's not too surprising - a tanker (and to a lesser extent a container ship, but they're also generally faster) is basically the engine and propulsion section in the stern bolted to an arbitrarily long hunk of cargo, with a small crew accommodation stacked somewhere. A cruise ship at least has cabins and entertainment spaces throughout, even though they're lighter. (Though I bet if you look at the empty displacement, which is probably a better proxy for initial construction cost, the difference isn't quite as stark on a dollar per pound basis)

    It would be interesting to see the breakdown of the cost by system or function. Like, a Burke is $2B, and has weapons and sensor systems that are at least comparable to a carrier hull. So most of it is probably propulsion systems (plus nuclear surtax), and flight deck systems (EMALs, bomb lifts, etc.). I don't think that the crew accommodation or stores rooms and so forth would be that much more expensive on a normalized basis.

    Interesting bonus question - are the most modern carriers still nuclear capable, or has that been fully deprecated?

  10. February 13, 2024bean said...

    @Belushi TD

    That would require them to use the barrier to arrest airplanes on a regular basis, which they don't do. OK, that is what a lot of sources say, but I'm going to ignore them.

    Interesting bonus question - are the most modern carriers still nuclear capable, or has that been fully deprecated?

    I'm not totally sure, but I suspect that capability is effectively gone.

  11. February 14, 2024Commodore Perry said...

    Some of these economies of scale--- self-defense fighters, AEW&C, sub chasing--- seem like they belong to the battle group rather than the ship.

    The DeGaulle was about half the price of the contemporary Nimitz (CVN-77). A pair of DeGaulles would together carry a similar air wing to a Nimitz. The DeGaulle is nuclear, CATOBAR and carries E-2s, etc. The DeGaulle is slower but wouldn't 2 carriers have better survivability?

    The Queen Elizabeths are also about half the price of the contemporary US carrier, but the design choices where more different, so it seems less directly comparable.

  12. February 14, 2024ike said...

    Lets look at this from the other way. What sort of mission would you want an eighth of a CBG for and no more?

  13. February 14, 2024bean said...

    Yes, in theory you could combine carriers to get a full wing's capability spread across two ships. This is a bad idea on several fronts. First, there's a good chance the ships get separated, and then you're back to not having full capability. Second, I am certain that the cost numbers don't actually work that way. Defense costing is notoriously terrible, but the Nimitz closest to CDG in time was Truman, and wiki gives her cost in 2007 dollars as $4.5 billion vs 3 billion Euros for the French carrier. I'm not going to simply say "see, it's only 1.5x the cost of the French ship" because there are so many things that could mess with reported costs. The US builds a lot more carriers than the French, which tends to push costs down. On the other hand, the French could have lower labor rates. Different nations use different cost accounting, so something that might get charged to the ship in the US is paid out of a different budget in France. Then there are questions like "what year do you do the currency conversions"? I stand behind the conclusion above. There's essentially no way that the smaller carrier is actually the better deal in terms of combat power per dollar. The engineering for that just doesn't work.

  14. February 15, 2024Basil Marte said...

    What are the practical difficulties that limit the number of operable aircraft to around 90?

  15. February 15, 2024bean said...

    Basically, it becomes hard to use all of the planes effectively. The controllers get saturated, and the deck gets too full. It becomes difficult to launch everything and then turn around and recover it without planes spending way too long in the air. This was a problem in the early careers of both the Lexingtons and Midways, in both cases solved by the growing size of airplanes.

  16. February 15, 2024Basil Marte said...

    Civilian airports tend to solve this problem by establishing multiple parallel runways. With the newest supercarriers having the island extremely aft, could they accommodate a short wire-assisted-recovery runway angled across the front of the deck, more or less parallel the "full-size" rear angled runway? I assume accommodating multiple catapults to not be the binding difficulty, since their placement is much less restricted.

    Failing that, an oil-tanker-sized and -shaped hull, with the island only slightly offset from the centerline, can have an "universal" set of facilities on the port side, and a smaller "fighters-only" set of runway&catapults on the starboard.

  17. February 15, 2024Basil Marte said...

    Basically an edit: the latter is not meant entirely literally. Take a page from the LCS' playbook and make it a trimaran, so as not to sacrifice speed.

  18. February 16, 2024bean said...

    Two problems with that. First, civilian airports are physically a lot bigger than carriers, which means that planes operating from multiple runways have a safe degree of separation. That's not possible on a carrier when landing. Second, while this bears some resemblance to plans in the late 40s and early 50s for separate bomber and fighter catapults (I think they were going to use the same arresting gear) these days, the fighters are the biggest things.

  19. February 18, 2024EdH said...

    I seem to recall reading, years ago, that the Fords were originally designed to be about 200' longer, but were cut down to fit into existing drydocks.

    It seems a poor decision, given what it will already cost to build a dozen or so Fords, but it's hard to say what that 200' would have brought to the table.

  20. February 18, 2024bean said...

    I didn't run across that claim when I researched the post on them, and I find it kind of dubious that that was ever the actual go-path plan. I'm sure that somewhere in the giant analysis that led to the Fords, they looked at "what could we do with an extra 200', and would it be worth it given that we'd have to lengthen the docks?", because good analysis looks at those kind of questions, but it clearly didn't pan out. And I could totally see someone seeing that and reading it as "this was the original design".

  21. February 18, 2024muddywaters said...

    Wiki says the drydock Nimitzes were built in is 662m, but they might be wrong about that (they note that it's unsourced), and the dates they list require usually 2 (1 under construction + 1 overhaul) and sometimes 3 drydocks.

  22. February 18, 2024EdH said...

    Beans, that could be, it was years ago and I can't source it (the extra 200' claim) now. I was actually hoping you could shed some light on it. There are indeed lots of twists and turns in development.

  23. February 23, 2024redRover said...

    Is the problem launch/recovery rate (i.e., number of catapults and arresting wires) or is it more about marshalling them on the deck? My guess is that the issue is not about 'runways' per se, so much as taxiway/parking/hangarage and avoiding gridlock.

  24. February 23, 2024bean said...

    I don't think that was the issue, given that the planes in question would physically fit. Avoiding gridlock might have been an issue, but largely in terms of just the sheer complexity of handling 100+ planes. I do think that the takeoff (and particularly landing) was a major part of it, given that you're very unlikely to get more than one set of arresting gear on a carrier.

Comments from SlateStarCodex:

Leave a comment

All comments are reviewed before being displayed.

Name (required):

E-mail (required, will not be published):


You can use Markdown in comments!

Enter value: Captcha