This question already has an answer here:

How can an ultra-deep hole or canyon form naturally on an earth like world?

By ultra-deep I’m thinking something like the Marianas Trench but on land and not filled with water. If it’s not possible why not and what would be a more realistic depth be? If it is possible how much deeper might it realistically become?

By earth like I mean a world the same as ours, but with any alternative configuration of oceans, continents, tectonic plates and mountains that might be required. The surface configuration can be as you wish but must be plausible geologically even if it might be a very rare occurrence.


marked as duplicate by Mołot, Frostfyre, anon, Anonymous, sphennings Nov 16 '17 at 14:15

This question has been asked before and already has an answer. If those answers do not fully address your question, please ask a new question.

  • 2
    $\begingroup$ Does a deep valley between two tall peaks count? $\endgroup$ – nzaman Nov 15 '17 at 17:17
  • $\begingroup$ @nzaman not realy, unless they are long ridges rather than peaks and the valley is very steep sided $\endgroup$ – Slarty Nov 15 '17 at 18:30
  • $\begingroup$ @Alexander very interesting - although I don't think its a duplicate, even though its similar. That question allowed magic and called for a chasm to the core of the Earth. Mine is science based and a mere 11km. useful link though. $\endgroup$ – Slarty Nov 15 '17 at 20:45
  • $\begingroup$ I believe the operative phrase is 'earth-like world', NOT 'the Earth'. I assume you mean that it does not have to be possible on EARTH, just has to be possible on an EARTH-LIKE world. IE. a world with a very different tectonic plate structure than earth, and totally different geology, is within the scope. The characteristic does not HAVE to be found on earth, or even possible on earth given the specific geology and tectonic structure. $\endgroup$ – Justin Thyme Nov 16 '17 at 1:19
  • $\begingroup$ @JustinThyme Yes you are correct it is an earth like worlld with a different plate set up and geology. $\endgroup$ – Slarty Nov 16 '17 at 9:40

The fundamental problem is that as the trench becomes deeper and deeper, the walls will tend to crumble because of the hydrostatic pressure pushing sideways. Underwater trenches can be deeper than trenches on land because the pressure of column of water in the trench serves to counter in part the pressure of the column of rock in wall.

Another difficult problem is how to keep the trench free of sediment.

That being said, the deepest canyon in the world is the Yarlung Tsangpo Grand Canyon in Tibet, with an average depth of about 2.3 km and a maximum depth of over 6 km.

As to the mechanism, this kinds of canyons are formed when a mountain range is risen quickly, for example because the Indian plate crashed into the Asian plate raising the Himalayas; the pre-existing rivers continue to erode their bed keeping it more or less at the elevation it had before the birth of the mountain range.

  • 1
    $\begingroup$ You could have a very deep but shortlived (geologically speaking) canyon form quickly by a catastrophic glacial melting event - these have happened in Iceland from volcanoes under glaciers and several big ones closed the last Ice Age in N. America. $\endgroup$ – Willk Nov 15 '17 at 18:03

Trenches, No

You basically can't make a dry tectonic trench. The tectonic forces that make plate trenches only work with oceanic plate being subducted. Oceanic plates are were oceans will be on an earth-like planet because they are wide spread lowlands. There is no way to keep it from filling with water on an earth like planet.

Canyons, possible but not that big

Canyons can be mostly dry because they are cut by rivers and the river does not need to be large, but completely dry canyons are rarely large since the the river needs to change course drastically, and if the river had an alternate course it probably would have taken it when the canyon started.

The maximum depth a dry canyon can reach is the from the height of the plateau they cut through to sea level, after that they become inland fjords instead. It is no coincidence the deepest canyon is on the highest plateau on earth. So your maximum is probably around 5000 meters, (highest plateau on earth) but if the river has been cutting for that long it will not resemble a canyon so much as a wide valley. Of course 5000 meters is nowhere near the depth of the marianas trench (11,000 meters).

  • $\begingroup$ Bonus for noting that water is an important problem. $\endgroup$ – JBH Nov 15 '17 at 22:24
  • $\begingroup$ Actually, I'd disagree. If we look at the en.wikipedia.org/wiki/Hellenic_Trench, this has a depth of over 5200 meters and as part of the Mediterranean sea, has dried out in recent geologic history. It takes special circumstances - subduction in an enclosed oceanic basin in a dry climatic zone - but it;s possible. $\endgroup$ – Andrew Dodds Feb 8 '18 at 8:42
  • $\begingroup$ the mediterranean was not dry ,it was drier, it was still one (or two) of the largest inland seas on earth. sedimentology support the hellenic trench being a lake/sea prior to the zanclean flood. $\endgroup$ – John Feb 8 '18 at 21:51

Drop the gravity

If your world has Mars-like gravity, it can have Mars-like mountains and canyons. Tharsis is a 7 km high plateau on Mars. Yarlung Tsangpo has an average depth of 2300 meters as it drops from a 2900 m plateau to a 600 m lowland. You could thus expect to get average canyon depths of 6000 m + with a highland as high and extensive as Tharsis (which is about the size of Russia). Consider that Tharsis is also a dead volcanic province on a dead planet. An active one might be even higher, and thus the canyon even deeper.

Make it dry

A large part of why the deepest canyons are underwater is that the Earth has a lot of oceans that tend to get low places wet. However, if there is less water, you can get some real deep places on land. A good example is the Zanclean period, when the Mediterranean was sealed off from the Atlantic at the Straits of Gibraltar and dried up almost completely. This would have exposed places up to 5000 m deep or more to the atmosphere.

For another example, what if the Red Sea was not open at the Bab al Mandeb? It is right in the middle of the desert, so it would be almost devoid of water; like a bigger version of the Dead Sea valley (already the deepest place on land). The Red Sea is over 3000 meters deep, and it is easy to imagine it being deeper, if the tectonic plates moved just so. Given how wide the Red Sea is, the grade wouldn't be steep enough to cause the walls to immediately collapse, even if it was 10000 meters deep. The island of Lombok is around 4 km high with a radius of 40 km and isn't collapsing; at that ratio you could get to 35 km deep where the Red Sea is 350 km wide.


It can't,

Anything below sea level naturally accrues water be it from the active water cycle (weather) or subterranean water pockets.

And depth only increases its chance of flooding.

Even man made quarries, which can be some of the deepest holes capable of fitting a human, often need pumps in order to not flood.

All water goes down hill


If a feature went as deep as the Marianas trench and didn't have the pressure of water baring down on it. I would be surprised if it didn't explode with volcanic eruptions.

  • $\begingroup$ So lower the sea level! $\endgroup$ – JeffUK Nov 15 '17 at 17:20
  • $\begingroup$ @JeffUK then its no longer Earth like $\endgroup$ – anon Nov 15 '17 at 17:21
  • $\begingroup$ So raise the ground level! Make a high plateau with a relatively deep canyon that extends down to sea level. The question says "with any alternative configuration of oceans, continents, tectonic plates and mountains that might be required"... If you don't allow lowering the sea level, then raise the ground level $\endgroup$ – JeffUK Nov 15 '17 at 17:23
  • $\begingroup$ @JeffUK Im not even sure such a geological structure could naturally exist on a planet. it would be like an egg shaped planet. $\endgroup$ – anon Nov 15 '17 at 17:34
  • 1
    $\begingroup$ @JBH: Sure, there is (or was) water on the bottom of all those places, but my point is that there are large areas that are below sea level, yet not filled with water. $\endgroup$ – jamesqf Nov 16 '17 at 5:43

I would suggest you consider tectonic plates splitting apart. The plates would perhaps have first pushed against each other, and now they are moving away from each other. Of course, the area would be very seismically active.

  • $\begingroup$ Yes – something that had actually crossed my mind. A rift valley in the middle of the Himalayas perhaps? Maybe it was initially flooded then becoming land locked and the water evaporated relatively quickly on a geological time scale? $\endgroup$ – Slarty Nov 15 '17 at 18:41
  • $\begingroup$ We have a number of rifts on Earth, but they all get filled with water very quickly. $\endgroup$ – Alexander Nov 15 '17 at 21:15
  • $\begingroup$ @Slarty, as a rule-of-thumb, "water begets water." It creates a natural cool zone that brings about condensation (aka, rain). Lake Bonneville was a great inland lake. Evaporation had a little effect, but it's gone because of an earthquake that opened a gash and drained it. A deep hole, canyon, or trench needs a drain or it will have water with the possible exception of a desert over basalt bedrock. Maybe... $\endgroup$ – JBH Nov 15 '17 at 22:52

It depends on what you mean by "ultra deep". If 3-5 km depth would satisfy you, there's an actual precedent in Earth's not-too-distant geologic past. The Straits of Gibraltar closed, stopping inflow from the Atlantic. There's not sufficient river flow into the basin to match evaporation, so the entire Mediterranean Sea pretty much dried up. See e.g. https://en.wikipedia.org/wiki/Messinian_salinity_crisis for more details.

But you aren't going to get much deeper than existing ocean basins, due to the strength of the rock - same basic reason there are limits to how high a mountain can be.


Make an ocean world that lost most of it's water and in one of it's many oceans have a trench as big as you want... but be warned it would break down fairly quickly(few millenia).

If you want it to be maintained ... you could add a ton of trees that grow on it's sides and with their roots maintain the structural integrity. Ofc this trees would have to be an scify-esk plant type that could have the necessary strength.


Not the answer you're looking for? Browse other questions tagged or ask your own question.