Holmes, after having returned from his dry fly adventure: Take your seat Watson, and get ready for my lecture on subduction.
– I‘m all ears, Holmes, Dr. Watson replies.
Part 1: Sherlock Holmes and the Hippopotamus in the Basin
Part 2: Sherlock Holmes applies “meteorology” to geology
Part 3: Sherlock Holmes gets a big surprise from Dr. Watson
Part 4: Dr. Watson on the Coriolis Effect and Drifting Continental Plates
Holmes begins by stating that subducting plates may descend in many directions and at many locations on Earth.
Holmes: The plates may be narrow or broad, and have curved or straight edges when viewed by the shape of their subduction trench. I will not touch on everything that might occur around subducting plates. Complicating factors like corner flow and differences in buoyancy between plate segments will have to be explained by someone else. I will concentrate on the events related to the Coriolis Effect.
Holmes continues: In principle, plates moving along the surface and subducting into the interior of the Earth will experience the Coriolis Effect in both the horizontal- and the vertical plane. Advancing plates into a subduction zone will experience both categories, whereas, in retreating subduction zones where the surface plate is virtually stagnant, the Coriolis Effect will mainly act on the plate from the subduction trench and downwards.
– During our previous discussion on continental drifting, we agreed that plate position and direction of movement relative to the poles and Equator play an important role. I assume this is still true even if some of the effects are observed in the interior of the Earth, Dr. Watson interrupts.
Holmes:
Indeed, it is and they are, my friend. The relationship between the Coriolis Effect and subducting plates is relatively complex, as you shall soon come to realize. Hence, we have to sort out a few illustrative cases independently by separating different aspects.
As the subducting slab is often attached to a plate moving along the surface, one might expect the upper parts of the slab to behave in a manner dictated more by the surface plate than the deeper part. Suppose the deviations created by the lateral and vertical Coriolis Effects produce different results. In that case, one might expect to see a transitional depth where one will dominate the other upon further descent. This might occur somewhere in the subduction zone and not at the subduction trench as such. About this, the steepness of the slab and the transformation of the slab from brittle to ductile would of course be of importance.
Let‘s start by assuming that subducting plates descend at an angle of 45o from the horizontal. Slab angles are not necessarily 45o; however, the principles are easier to understand if we start with this assumption. Have a look at my side view drawing of north- or south-directed subduction in the northern hemisphere.
What you see here is two hypothetical subduction slabs at latitudes P – and two more at O. Two of them are subducting southwards and two subducting towards the north. Since we have chosen a slab angle of 45o, this angle defines the latitude where the southwards descending slab is moving neither closer, nor further away from the axis of rotation during descent. At higher latitudes, a slab descending southwards will move away from the axis, and at lower latitudes, it will move closer to the axis.
Anyway, Watson, in our example, southwards subduction at latitudes less than 45o N will enter regions rotating faster as they descend, and slower moving regions during descent above 45o N. To generalize, if the southwards subducting slab is dipping by “a” degree, the neutral latitude will be at “a” degree N.
However, northwards descending slabs behave differently. In the northern hemisphere, subducting plates always move closer to the rotational axis and enter regions rotating slower, regardless of latitude. We now have the tools to consider the Coriolis Effect for plate segments moving and subducting northwards or southwards in the northern hemisphere.
Let‘s have a look at some of the nitty-gritty details, Watson.
A subducting plate dipping by “a” degrees, moving southwards at latitudes above “a” o N, will move into regions having a higher rotational speed and will therefore deviate in a westerly direction as it descends. In my drawing, westwards deviation means moving into the paper. As discussed before, an advancing, connected plate at the surface would deviate westwards while rotating counter-clockwise. In other words, the plate at the surface and the subducted part would both deviate westwards.
A subducting plate dipping by “a” degrees, moving southwards at latitudes less than “a” o N, will move into regions having lower rotational velocity and therefore deviate eastwards during descent. At the surface, an advancing, connected plate would deviate westwards while rotating counter-clockwise. The implication of this is that the advancing plate is still on the surface and the subducted plate would deviate in opposite directions. This would increase the tendency for eastward tearing of the slab as the plate attached to it at the surface is deviating westwards.
Northwards subducting slabs at all latitudes above the Equator will be moving closer to the axis of rotation and therefore deviate eastwards during descent. Subduction slabs further north will deviate more to the east than subduction zones closer to the Equator. At the surface, an advancing, connected plate would deviate eastwards while rotating clockwise.
In a situation where two plates are advancing in opposite directions along the surface towards a southwards dipping subduction zone, the plate moving southwards would deviate westwards while rotating counter-clockwise. The northwards-moving plate would deviate eastwards while rotating clockwise. This would influence the direction and movement of the subduction trench itself. As the two plates at the surface move in opposite directions while also rotating in opposite directions, they would collide more intensely on the western edge and less on the eastern edge. At the same time, the subduction trench would experience some shear.
In a situation where two plates are advancing in opposite directions along the surface towards a northwards dipping subduction zone, the plates on the surface would behave in the same manner as described above whereas the subducted plate would always deviate eastwards regardless of latitude.
To discuss east- or westward subducting plates, we should probably look at two more drawings. Let‘s first look at this one, viewing the Earth from the North Pole:
As you have already observed Watson, in this configuration, the eastwards-oriented subduction is moving with the Earth‘s rotation, whereas the westwards-oriented subduction is moving against the direction of rotation.
In my next drawing, we may also see that the dimensions of the subducted plates are important. I have made a drawing of a green-colored plate subducting into the Earth‘s interior in the northern hemisphere.
As is clear from this drawing, in the northern part of the subducting plate, the direction of descent is almost aligned with the axis of rotation, whereas at the southern end, it is a lot more perpendicular to the axis of rotation. Thus, the southern part descends into regions having significantly lower rotational velocity, whereas the northern part experiences less change in velocity.
Therefore, in an eastwards-directed subduction zone, the southern part of the plate will deviate a lot more eastwards during descent than the northern part. This results in a relatively flat slab in the southern end and a gradual steepening of the subducting plate when moving northwards along the plate. According to this, eastwards subducting plates anywhere on Earth would be flatter near the Equator and steeper towards the poles.
Any westwards subducting plate would initially move in the opposite direction of the Earth‘s rotation. Thus, it will deviate eastwards, steepen, and pull the subduction trench along with it in an eastwards direction. A retreating subduction zone with slab rollback would emerge. It would be retreating closer to the Equator than nearer to the poles.
Every point on all subducting plates is being pulled by gravity towards the center of the Earth (albeit sometimes at an angle from truly vertical). This implies that the deeper corners of all plates move closer to each other during descent. To accommodate this crimping, the plates must stack up in the deeper sections, or corrugate to make a wavy surface. All oceanic crusts are created with weak zones in the form of transforms. The transforms might be reactivated to take part in the required size reduction during descent.
– I must say, Holmes, that was quite a mouthful. What else is there to tell, Dr. Watson asks.
Holmes continues:
So far, we have just discussed events in the northern hemisphere as well as some events that are independent of the hemisphere. The latter includes the east-or westward dipping subduction zones, of course. As you might have guessed, Watson, the hemisphere makes a bit of a difference for north- or southward subducting plates. I will sum it up very briefly.
North- or southwards subducting slabs on the southern hemisphere move opposite to their northern hemisphere counterparts.
That is nearly all I have to say concerning the behavior of different subduction zones when acting alone and not in combination with mantle upwelling. There is just one more thing I would like to mention before we call it a day.
Holmes is approaching his usual bedtime, but has one more subject to cover:
In some subduction zones, plates are moving in opposite directions, and colliding. One of the plates is usually attached to a subducting slab that is forced beneath the other plate, Holmes commences. Due to the opposite movement of the two plates, they therefore experience different Coriolis Effects.
My drawing illustrates two opposite situations south of 45 o N with slabs dipping at 45 o. To the left, I have drawn a northwards-moving plate subducting beneath a southwards-moving plate. To the right, – the opposite situation. Have a look at it, Watson.
The blue arrows indicate the deviations created by the Coriolis Effect due to the initial movements of the plates – on the surface, and within the subducting plate. For simplicity, I have omitted the slight rotation of the subducting plate itself. It also has a lateral component similar to the surface plates.
The dashed blue line in the right-hand figure indicates a hypothetical depth and location of a sideways slab tear due to the opposite deviation of the slab and the surface plate. In a real continental-continental collision zone, this tearing of the slab might probably be indicated at the surface by a shear zone in the middle of some mountains overriding the subducting slab.
Generally, the blue arrows in my drawing might be used to infer shear zones, zones experiencing rotation, more or less collision, etc. Remember, Watson, my reasoning and the blue arrows are also valid for individual slab segments subducting in parallel.
This is it, Watson. Bedtime. Tomorrow we will have a look at a few interesting cases where we combine different processes to see what might happen then. We might also look at something from the real world. Maybe we will celebrate by sharing a Whisky on the rocks – or maybe my excellent Gin and tectonics.
Good night, Watson.
– Thank you very much, Holmes, I am looking forward to that.
Dr. Watson turns off the lights and collects the drawings, and the two geology investigators leave the room.
HANS K JOHNSEN
Inspired by Arthur Conan Doyle