Picket ships to defend against Pearl Harbor surprise

[psteinx]

IIUC, navies used picket ships - minor ships or submarines on station to watch for the enemy, as a sort of extended guard around their home waters. In fact, the Doolittle raid was partially spoiled when a Japanese picket ship spotted the strike force heading towards Japan (forcing the planes to launch prematurely).

So, what if Admiral Kimmel had stationed a line of submarines and small-ish ships along the lines of approach to Pearl, especially to the Northwest, where there was less naval traffic and thus the logical (and actual) path of approach of any attackers?

How practical would this have been? There's a correlation between degree of coverage/protection and number of ships needed, and of course, the Navy was scrambling to get to wartime readiness. Still, it seems plausible, but I haven't really seen it discussed...

 

[vizzer]

As picket ships could only see out the horizon ~20 miles, much less from a submarine, they would need large numbers to cover possible lines of approach & to provide sufficient rotation of vessels.
As, at the time, PH was believed to be out of range of any Japanese attack, as it very nearly was, why would they use such vessels here?
A more vulnerable & suitable area for such would be the Philippines.

 

[I want to learn]

When I think of picket ships, I think of smaller ships (destroyers) equipped with radar. Not many US ships had radar in Dec 1941 had radar, and I think very few (if any?) of those ships were destroyers. Destroyers without radar made poor scouts. Asking for destroyers to fulfill a picket role without radar is probably something like asking for a ship to take on a battlecruiser role without big guns, or a battleship role without armor.

 

[LegionoftheUnitedStates]

Impractical, Kimmel was short on smaller vessels, as a lot had been sent to the Atlantic. The war was expected to breakout in the Western Pacific, with attacks on Wake, Guam, and the Philippines to support attacks on the Dutch East Indies and Malaya.

Both the Army and Navy were short on resources in December 1941 and Hawaii was not the most vulnerable target.

 

[psteinx]

What if it's, say, 3-5 ships, not at a single position, but moving (independently) in pre-defined patterns (known to the US but presumably not to the Japanese) to the NW of Pearl?

I'm not sure the size of the Japanese strike force (how much physical space it occupied), but presumably it was pretty big/spread out? Several moving ships can cover a lot of ground. You may not get 100% protection across the patrol area, but even if you got 60-80%, that's good odds...

 

[I want to learn]

What if it's, say, 3-5 ships, not at a single position, but moving (independently) in pre-defined patterns (known to the US but presumably not to the Japanese) to the NW of Pearl?

I'm not sure the size of the Japanese strike force (how much physical space it occupied), but presumably it was pretty big/spread out? Several moving ships can cover a lot of ground. You may not get 100% protection across the patrol area, but even if you got 60-80%, that's good odds...

The Japanese approached within 230 miles of Pearl Harbor. Back of the envelope calculations say that's an area of >16,000 square miles that has to be patrolled. Each patrol ship can see, optimistically, 300 of those square miles at a time. Once you factor in the fact that the last part of the Japanese approach was done under the cover of darkness and it's too late if they're detected at dawn after they're already launching the planes, it becomes very obvious the math doesn't work.

 

[YYJ]

IIUC, navies used picket ships - minor ships or submarines on station to watch for the enemy, as a sort of extended guard around their home waters. In fact, the Doolittle raid was partially spoiled when a Japanese picket ship spotted the strike force heading towards Japan (forcing the planes to launch prematurely).

So, what if Admiral Kimmel had stationed a line of submarines and small-ish ships along the lines of approach to Pearl, especially to the Northwest, where there was less naval traffic and thus the logical (and actual) path of approach of any attackers?

How practical would this have been? There's a correlation between degree of coverage/protection and number of ships needed, and of course, the Navy was scrambling to get to wartime readiness. Still, it seems plausible, but I haven't really seen it discussed...

Long range patrol aircraft like the PBY could cover a much larger area far more efficiently.

 

[psteinx]

Long range patrol aircraft like the PBY could cover a much larger area far more efficiently.

And yet they didn't cover the key area at the key time.

 

[psteinx]

The Japanese approached within 230 miles of Pearl Harbor. Back of the envelope calculations say that's an area of >16,000 square miles that has to be patrolled. Each patrol ship can see, optimistically, 300 of those square miles at a time. Once you factor in the fact that the last part of the Japanese approach was done under the cover of darkness and it's too late if they're detected at dawn after they're already launching the planes, it becomes very obvious the math doesn't work.

But the Japanese have to move their (large) force THROUGH the patrolled area. All the picket ships have to do is ~randomly bounce into some element of the Japanese strike force, ONCE as they pass through. i.e. It's not static, so you can't just divide the total area to be patrolled with a ship's maximum visibility at one instant.

 

[CV12Hornet]

But the Japanese have to move their (large) force THROUGH the patrolled area. All the picket ships have to do is ~randomly bounce into some element of the Japanese strike force, ONCE as they pass through. i.e. It's not static, so you can't just divide the total area to be patrolled with a ship's maximum visibility at one instant.

Random chance is not something you can count on in military planning.

 

[John_MC]

And yet they didn't cover the key area at the key time.

Just remember in various war games, carrier aircraft came in the same direction as the Japanese did. Short and Kimmel were too focus on sabotage and not air strikes.

 

[psteinx]

Just remember in various war games, carrier aircraft came in the same direction as the Japanese did. Short and Kimmel were too focus on sabotage and not air strikes.

If you want to cast blame on Kimmel and Short, you'll get no objection from me. Perhaps the (primary) answer was straight up better radar usage/coverage. But in the event that that was too hard to do on short notice (weeks/months), picket ships seem like one possible, useful, supplement.

An advantage of picket ships is that if you can detect the Japanese fleet, say, 400-500 miles out, versus 150 miles out, you get MUCH more time, because the ships have to get close enough to launch, and ships go much slower than airplanes. So, 12-24 hours of advance notice opens up all kinds of possibilities...

 

[YYJ]

If you want to cast blame on Kimmel and Short, you'll get no objection from me. Perhaps the (primary) answer was straight up better radar usage/coverage. But in the event that that was too hard to do on short notice (weeks/months), picket ships seem like one possible, useful, supplement.

An advantage of picket ships is that if you can detect the Japanese fleet, say, 400-500 miles out, versus 150 miles out, you get MUCH more time, because the ships have to get close enough to launch, and ships go much slower than airplanes. So, 12-24 hours of advance notice opens up all kinds of possibilities...

The envelope math was started in Post #6, but I can continue.

16,000 square miles that has to be patrolled. Each patrol ship can see, optimistically, 300 of those square miles at a time.

I can't speak the the accuracy of the numbers, but if they are close, then you only need 16,000 / 300 = 54 picket ships.

Lets presume the battleships, cruisers, and aircraft carriers are not used as pickets. Stationed at Pearl at the time were 30 destroyers, 4 submarines, 3 Coast Guard Cutters, and 47 "Other Ships" including 22 mine warfare ships, 4 ocean going tugs, and 12 PT boats. The rest of the ships were large freighter sized auxiliaries, yard tugs, floating dry docks etc. So you almost have the numbers to do what you suggest, if the harbour was emptied right out and this strategy single-mindedly pursued. Some of the destroyers and patrol ships would want to be kept closer at hand to do the more normal patrol and submarine watch, since USS Ward and Monaghan actually stopped some of the mini-sub attack on Pearl that day.

https://www.history.navy.mil/resear...etically/s/ships-present-at-pearl-harbor.html

IIRC, one of the problems on Dec 7 was that the base had been on a high state of alert and a wartime patrol regime for a week, and they had exhausted themselves and were unable to sustain that level of readiness.

 

[I want to learn]

But the Japanese have to move their (large) force THROUGH the patrolled area. All the picket ships have to do is ~randomly bounce into some element of the Japanese strike force, ONCE as they pass through. i.e. It's not static, so you can't just divide the total area to be patrolled with a ship's maximum visibility at one instant.

Yeah I mean I'm sure you could do some really advanced math to figure out the exact statistics, but the fundamental problem is that the Pacific Ocean is a big place and even a task force is tiny in comparison.

 

[jlckansas]

The threat axis at the time was thought to be to the Southwest toward the Mandates like Truk and Kwajalein Atoll where the main IJN bases were. They knew that they had long range flying boats based there and were looking at an attack by them if it was an air attack and a raid by lighter forces (DD, CL, CA) on the shipping in the area from that area. The Northern approach was not looked at during that time of the year because the Northern Pacific is not a good route to take do to the Late fall early winter storms.

 

[I want to learn]

I can't speak the the accuracy of the numbers, but if they are close, then you only need 16,000 / 300 = 54 picket ships.

Sorry, I was off by an order of magnitude, the area of ocean within a circle of 230 miles from Hawaii is actually more like 160,000 miles, not 16,000 square miles. I mean, if you're going to attempt some cartoonish saturation strategy, I think you can have all the ships steaming in a circle in a big circle around Hawaii 20 or 30 miles miles apart from each other, and it would accomplish the same thing, because you're patrolling the perimeter and not the entire area within the circle. That would take about fifty ships too. But like, the USN can't actually commit its entire pacific fleet to making 1400 mile circuits around Hawaii for obvious reasons. And that plan only works in daytime, in excellent visibility.

 

[YYJ]

OK, my own napkin math. If the OP is that a picket screen is set up 500 miles from Hawaii, and the ships are located such that their visibility is a set of barely overlapping circles with a view horizon-horizon, meaning the ships can not see each other but their fields of view barely overlap, and they are staking out the north west quadrant approach to the islands, then we are looking at the circumference of a circle with a radius of 500 miles. C=2pi r = 2 x 500x 3.14 = 3140 miles. The we divide that by 4 to get one quadrant = 784 miles. If the visibility in clear daylight is 20 miles to each horizon, then each ship can see 40 miles of this quadrant. So we need 20 ships to patrol the line. If the picket wants to patrol a semi circle, then they need 40 ships.

I have no idea what the practice of performing a search picket is, maybe less ships could patrol the area using lozenge shaped steaming patterns with a loop time so no one could sneak by. This interval would have to be tightened up for night and bad weather.

Edit: Extending the picket out to 1000 miles would require 40 ships at 40 mile intervals for a quadrant or 80 ships for a semi-circle.

 

[psteinx]

The geometry, if you're interested in defending a single point (Pearl) is a line, or an arc, not an area. The ships may have some lateral movement within whatever line/arc they're patrolling, but that's a fairly minor issue.

Some math:

Simple case - static pickets.

Defensive line is M miles
Sighting radius is R

Then you need M/(R*2) pickets at any given time.

The sighting radius would of course vary based on day/night, weather, etc. but you're not necessarily aiming for perfect coverage, but rather, say, 60-80% chance of visual contact.

But, if the pickets are in motion, roughly along the line of patrol, then you can potentially get away with fewer pickets. The key additional factors are the relative rates of speed of the pickets versus the presumed raider(s), and the presumed size of the raiding group (particularly it's length in motion, but to some extent it's breadth.

For relative speed, consider a picket that is infinitely faster than the raiders. It could ping-pong back and forth across the full picket line in an instant, and no raiders could make it by. 1 picket would suffice.

Of course, that's a little silly, but consider a picket that is 1.4X faster than the raiders. If it encountered any portion of the raiders (lead ship through tail ship) before the raiders fully passed by, it could report the group.

Now consider the size of the raiding group. If Kimmel is primarily concerned about a major raid, he's going to assume something like 15-60 ships or so, possibly broken into sub-groups (as I think occurred OTL). That takes up a fair amount of sea space. The longer and broader the group is, the more likely a fast-moving picket could spot a lead ship, a straggler, some steam or wake or debris or whatnot.

FWIW, the picket ship that spotted Doolittle's group appears to have been a small ship, manned by a crew of only 11. So you don't need a destroyer, and, while Wikipedia has few details, I believe the ship in question was intentionally being used as a picket, so the concept seems plausible.

 

[psteinx]

To play with the math of a patrol a little more:

Assume a ship with a speed of 20 mph is patrolling perpindicular to the path of a narrow, long raiding group, that is also moving at 20 mph. The raiding group stetches along a 40 mile long line. Visibility is assumed to be 20 miles.

So, if the picket passes within 20 miles of the lead or tail of the raiders, it sights them. The total "line of visibility" is thus ~80 miles (maybe a little more with some trigonometry for diagonals, but meh...)

Assume the patrol boat cycles back and forth across a 60 mile line, with a 20 mile gap to the next patrol boat's cycle path. (Thus, 1 patrol boat every 80 miles).

The patrol boat thus completes a one-directional sweep once every 3 hours. It (or one of it's neighbors on either side) likely spots any passing raid, except if the timing is bad at the far extreme of one of its sweeps. This probably gives you a ~60-80% sighting probability. 5 boats could patrol about a 400 mile arc/line. Plus some additional boats/crews for relief/rotation.

 

[CalBear]

And yet they didn't cover the key area at the key time.

There wer a limited number of aircraft assigned to patrol duties. Those that were concentrated to the west and south. No one expected any attack to come out of the Home Islands. Any enemy attack was expected to come from the Mandates.