> In 2005, the nuclear-powered USS San Francisco collided with an underwater volcano, or seamount, at top speed, killing a crew member and injuring most aboard. It happened again in 2021 when the USS Connecticut struck a seamount in the South China Sea, damaging its sonar array.
How in the world does a submarine going at < 40 mph not detect that it's running into a mountain? Why are light, radar, and sonar still not enough to detect obstacles that big?
> How in the world does a submarine going at < 40 mph not detect that it's running into a mountain?
Seamounts don’t give off active emissions (except maybe if they are erupting), and subs usually use passive hydrophones (which rely on the active emissions of targets), not active sensors when operating underwater to avoid detection.
Navigation is done by charts using inertial navigation alongside dead reckoning. Hitting a mountain is generally because the mountain wasn't on the chart being used, though in principle in could also be due to an error in positioning (with inertial navigation, that should be much rarer than with dead recknoning alone, but its not impossible.) (In the case of the San Francisco, for instance, other charts had a notation of "discolored water" which suggested a probable seamount, which the investigation noted should have been transferred to the chart in use, but was not.)
> Are light, radar, and sonar still not enough to detect obstacles that big?
Radar is useless underwater, light would also be useless for spotting anything in time to react at speed, and sonar isn’t routinely used because it compromises stealth.
Naive suggestion from someone who has zero knowledge about these things: can't an unmanned submarine drone be constantly moving ahead of the main submarine to detect these obstacles before they become dangerous to the submarine and crew?
Best solution is to pre-map the ocean to the degree possible (mountains don't tend to move, at least on the timescales we care about...); this wouldn't need to be done stealthily (so you could use active sensors), and using surface ships is one option, also drone surface ships and drone underwater vehicles now. I assume there are relatively highly classified high-detail government maps of certain parts of the ocean of particular interest.
We also have highly classified models of the exact shape and mass distribution of the Earth, as this is essential to accurate long-range ballistics (i.e. for ICBMs to hit their targets) -- it affects the gravity experienced over the course, and the precision/accuracy needed is higher than just uniform-density sphere or oblate spheroid. Also of magnetic and other anomalies since that can be used as a baseline when searching for submarines or ships using magnetic sensors, although I think this is less important now.
Its already been largely done, which is why when subs run into seamounts, you tend to see the leadership team relieved.
“Maybe it would help to have a map of the space we’re going to be driving very large, very expensive, strategically important pieces of equipment (that also have crews) through blind” is not, as it turns out, something that never occurred to the Navy.
Oceans cover 361M km^2. So if we built 13,333 drones with a failure rate of 25% and they map at a rate of 50 km^2/day at a higher column depth with active sonar, it would take approximately 2 years
Also speaking from utter naivety, this seems like a super good idea. Each submarine gets a little attendant constellation of drones. Control them with Bluetooth. Or usb cables.
Or whiskers! That's how cats and catfish do it. A dozen 300' plastic prongs protruding frontward.
Why does everyone assume you need wireless communication? You can have a cable running between the submarine and the drone, same channel would be used to power the sub.
That is a lot of heavy, droopy, cable (if it is going to be able to withstand the stresses and ‘unkindness’ of the ocean) which is very likely to get tangled in things like spinning propellers unless very carefully managed.
What about using a pressure sensor fixed to a long pole in front of the submarine? That way, if they hit anything, the pole would hit first, and could activate the stopping mechanism.
Imagine swimming with a long pole attached to the front of you. Even if it would be thin, it would still disturb you very much. Also the disturbances in the water created by the pole would make noise, that can be detected.
Submarines are optimised in everything for absolute stealth, up to the point that the crew may not make any noise, like dropping a wrench.
I recall looking at the various charts and from what I recall the "discolored water" notation was sourced from a chart several decades before and was offset by a few miles (incorrectly charted in the first place).
>> Are light, radar, and sonar still not enough to detect obstacles that big?
> Radar is useless underwater, light would also be useless for spotting anything in time to react at speed, and sonar isn’t routinely used because it compromises stealth.
Any people with knowledge in the field care to comment on whether it's feasible for a light (em?)-type, active, near-range detection system to work?
It seems to me the basic problem is:
- enough energy to get a signal at a range sufficient to move around any detected obstacles
- high enough dissipation in water to minimize range at which detection by external sensors is possible
Water absorbs photons at most frequencies outside the visible spectrum. What we deems as visible light is actually where water is most transparent. Since eyes evolved underwater, it makes sense they would use this section of the spectrum.
Your two requirements contradict each other. You can choose one or the other, but not both.
The world's militaries have invested a great deal of time and money into researching underwater sensors. The solution is not going to be discovered by random people on the Internet.
> Any people with knowledge in the field care to comment on whether it's feasible for a light (em?)-type, active, near-range detection system to work?
I’m not in the field, but I follow it enough to say: in theory, yes, (blue-green lidar), and there is a lot of work on moving this from theory to practice.
But approximately everyone is working on it (for bathymetry and for detecting subs, and similar blue-green laser tech for underwater comms), so its likely once its deployed so that a sub could in principle use it as a “headlight”, it’ll raise similar concerns to sonar, because if it is used actively, passive sensors looking for active use will also be common.
The issue is that the cloudiness of water, even open water, varies greatly. So there would be either lots and lots of false positives, or the system would be too insensitive.
To clarify: I mean that, generally speaking, the presence of 'cloudy' water would presumably signify that the seafloor was close. In the open ocean, maybe a sea mount is surrounded by essentially a cloud of particles.
Edit: I think your reply made the thread reach maximum depth, so I'll end it here.. I should do more reading on the topic.
a) whales don't tend to go where volcanos are. Whales dive to find food, and their food sources aren't near the seabed. I think the deepest diving whale species are sperm whales, and they go after squid, which are not typically found near volcanos.
b) it's not a question of matching frequency with the whales, it's a question of noise emissions and the amount of energy you're putting into the water. Modern active sonar systems are incredibly energy intensive, and the technology to modify the waveform and limit the power of the waveform to match something biologic doesn't really exist, and would have to justify all the added complexity
c) USN and US geological survey charts are really really good, and regularly updated, to the point where "It wasn't on the charts!" is a claim met with extreme skepticism. 9 times out of 10 some sort of navigation or systemic error was involved.
There's no way you could have gotten a gravity sensor to work reliably enough in that environment. Active sonar would defeat the point of a stealth submarine.
Gravity detectors work from moving aircraft (it's a common method of deployment), so motion's no problem and can be made incredibly sensitive—sensitive enough to suit the job in hand (place one on a table and then on a book on the table and you'll easily read the difference in gravity). Moreover, seamounts are incredibly large objects and would have huge signatures and thus easy to detect.
I'd suggest the real issue around this matter is that it'll be classified.
What's the technical term for this type of gravity detector? Do you mean a gravity gradiometer? It seems hard to steer a modern pigheaded search engine away from LIGO when you search for "gravity detector".
We'd have difficulty in fitting LIGO in a sub methinks and seamounts are static and thus no G waves. You're right though, I should have been more specific, perhaps a more understandable description would be a mass detector or sensor.
I reckon the reason we don't hear much about this subject and the reason it's likely classified is that's its primary purpose would be the detection of other subs, detecting seamounts a secondary benefit.
Sonar is a brilliant beacon identifying the subs location; other than in combat where they are already detected or desperately need a hard fix on a target now, or someplace they don’t need to be stealthy, subs aren’t using it.
AFAIK, the first sub-deployed gravity gradiometers were developed for the Ohio-class ballistic missile subs and Los Angeles-class attack subs like San Francisco don’t, AFAIK, have them.
No, the ocean is not fully mapped. It's written in the article: "With only one-quarter of the sea floor mapped with sonar, it is impossible to know how many seamounts exist."
In the specific case of the San Francisco, because notations indicating a probable seamount were not transferred from other charts when the chart being used was compiled; there is some lack of clarity on how much the failures involved are the responsibility of the command crew of the sub and how much of higher command, but its definitely a failure to appropriately use existing information, not a “the information did not exist” problem.
While the area was apparently not well charted, the responsible crew not using the information at hand (including information as to the the quality of charting data for the area) was also apparently at play in the Connecticut grounding.
The USS San Francisco crash-type problem was discussed some while ago on HN. At the time I questioned how this could happen when one would expect a submarine of that sophistication to be equipped with both sonar and gravity detection.
Detecting changes in gravity is common practice in mining surveys (ores often have higher densities) and geological work, so for me why the sub collided with the seamount remains a mystery.
I cannot believe subs of this class don't have such detection on board so could someone more knowledgeable about the subject than me provide an explanation.
Easy way to not be found is to not emit energy be it sound, heat, light, radio frequency, etc... until we discover a passive sensor that works underwater that can see static objects that may not emit energy, seamounts and other things that change the map are occupational hazards for submariners.
One of the main features of military submarines is being undetectable by others, to various degrees. This include visual, audio and movement detection. Drones would probably make all of those three factors worse and make the submarine more detectable, than without a drone travel in front.
> Do these volcanoes really have steep/craggy dramatic slopes like the image suggests?
Probably not many. They will have formed slowly at plate boundaries such as the mid-Atlantic ridge, then moved apart over millions of years, then very, very, slow erosion.
> I remember being disappointed to find that the Mariana Trench didn’t have looming near-vertical canyon walls, but rather had a relatively tame slope.
It has always boggled my mind that military subs run around under the water largely blind, due to the fact that using active sonar would give their position away, and passives are going to miss many things like this.
“From the ocean surface to depths of about 500 meters CO2 would exist as a gas and would tend to rise in the water column, since it is less dense than seawater. Between 500 meters and 2,700 meters CO2 would exist as a liquid, but one that is still lighter than seawater and would thus also rise in the water column. At depths up to 2,700 meters CO2 will create a rising plume and dissolve into the surrounding seawater relatively quickly.
At depths deeper than 3,000 meters the weight of the water column compresses the liquid CO2 and it becomes denser than seawater and sinks slowly to the seafloor.” [0]
So yes, even at normal temperatures for that depth C02 would be a gas; definitely so at volcanic temperatures!
The San Francisco struck it at 160m depth, I am also curious to this question. 1km, sure I doubt we see bubbles - but 160m is not as deep (and the bubbles would get bigger in volume as they rise). Great question, does anyone know?
That is quite many, if you start to think spacing them, 19000km if they were 1km apart on average... Then again, there is lot more fault line distance on Earth.
Land moves around, raises and lowers, plates crash lifting mountains and subduct continents under each other so stuff moves, but stuff goes up and down at the same time. Sometimes the land sinks, sometimes it's pushed up, but these geological processes happen on geological time. Sea level rise is a concern within a few human generations.
These are just much different time and volume scales.
Yes? That is exactly the commenters concern. Fill a bowl to the brim with water, throw a ball in. It spills over the edges of the bowl by "just displacing" the water.
Probably not. I would think that similar to above water formation, there is also at least three ways mountains happen. Volcanic, folding(two plates collide) and Block.
How in the world does a submarine going at < 40 mph not detect that it's running into a mountain? Why are light, radar, and sonar still not enough to detect obstacles that big?