[Tutorial] Using GPlates to model continental drift

[Charerg]

a. CREATING A FLOWLINE

Okay, so now that we have our continents separating, it’s time to look at modelling the new oceanic crust being created as part of the process. We’re going to use the flowlines feature of GPlates to achieve that, but before we start with the actual modelling it’s worthwhile to take a closer look at the behaviour of mid-oceanic ridges.

First off, a mid-oceanic ridge generates new oceanic crust at equal rate on both sides of the ridge. This has some implications for the movement of the mid-oceanic ridge that I’ll now try to demonstrate with a few pics taken from the model used in the GPlates tutorials (based on Seton et al. 2012).

First off, I’ve drawn the initial location of the India-Antarctica rift (130 Mya):


And then look at where the present-day ridge is located. You can still see ”the original location” close to where the Kerguelen plateau is.


As you probably noted, there is a rather substantial discrepancy between the location of the initial rift and the present-day location of the mid-oceanic ridge. That is because the distance between a point in India and Antarctica that were once connected is about 9000 km. Of this distance Antarctica has covered 1500 km, and India 7500 km (very rough Google Maps measurements btw, nothing official). However, the ridge has created 4500 km of new crust both on the Antarctic and Indian plates. What this means in practice, is that the ridge has migrated 3000 km northwards due to the difference in relative movement between India and Antarctica. Well, actually the reality is a bit more complicated than that in this case, since the ridge in question became extinct around 40 Mya (the extinct mid-oceanic ridge between Indian and Australian plates today is the one initially located between India and Antarctica). Also, a good chunk of that ridge has now been subducted by Eurasia (so maybe not the best example). But you catch my (continental) drift, eh?

Since I think it’s somewhat intuitive to think of mid-oceanic ridges as static, fixed features I feel the above point needed to be made. The ridges are in fact not static at all, they move depending on the relative motions of the plates. So, we need some method to the track the movement of these ridges, and that’s where flowlines come in.

Here, I’ve created a flowline between South America and Africa:


Again, note the “white outline” showing the location of the initial rift and note the discrepancy between its location and the modern ridge.

Here’s a second image, this time showing the actual present-day ridge as well:


As you’ll note, the flowline isn’t an exact match (I wasn’t too careful with point placement), but it does approximately match the shape and location of the mid-oceanic ridge. Also, the flowlines should depict the orientation of the “stretch marks” on the ocean floor, which is an additional useful feature of them.

The previous image with a global gravity anomaly raster on top:


Right, so after that rather long-winded introduction, let’s jump pack to our previous fictional model, and actually create one of these vaunted flowlines. First off, set the Time back to the point when the supercontinent starts breaking. Before you start, be aware that creating flowlines can cause crashes if you pick the wrong option or something goes wrong. So always save everything before attempting to create a flowline.

After you’ve made sure everything is saved, go under the Digitise tools and choose “Digitise multi-point geometry (Hotkey M)”. Go ahead and place some points at the seam between the two continents. Once you’re happy, click Create feature and choose Flowline from the list.



Then, click Next and you will be presented with some options. First off, you have the choice between spreading centres, left-plate end points and right-plate end points. Now, this is the critical option for those crashes I mentioned. I believe left-plate end points and right-plate end points are intended to work with plates. And plates in GPlates are not defined using the Digitisation tools, instead they are defined with the Topology tools (which we haven’t covered yet). This is why in the previous section we chose either ClosedContinentalBoundary or ContinentalCrust as feature type for our continents. Note that in Astrographer’s tutorial one suggestion was to define the continents as ClosedPlateBoundary. I wouldn’t recommend this, and it may cause problems with the flowlines we’re attempting to create here.

So, in case you didn’t guess it yet, choose spreading centres, otherwise I can almost guarantee you’ll get a crash. Then, type the appropriate Plate IDs in the “Left Plate ID” and “Right Plate ID” slots, set the flowline to exist from 150 Mya to present, and give it an appropriate name.



Next, you’ll have to insert an array of time instants. The flowline calculates the “half-stage rotations” between the two continents at specific intervals, and now you’ll have to define what those intervals are. Click “Add” and then under “Insert multiple times”, set it to insert times from 150 to 0 Mya at 10 Mya intervals. Then click “Insert” and “Ok”.



Click next and create a separate feature collection called “Flowlines” for storing flowline features. If your GPlates is going to crash, it will do so now (okay, it shouldn’t crash as long as you remember to pick spreading centres). Assuming everything went fine, you should now have a nice new feature collection. Go ahead and play the animation, to see the flowline in action. For me, this is the end result:



As you can see by looking at the flowline, something is dreadfully wrong with my rotations here. Apparently, I have a “hinge type rotation” with the hinge in the middle of the ocean! You’ll note that flowlines are a good way to check if your movements are valid, or if they’re total nonsense like my animation here. One way to fix this (in this case), is to redraw Continent A at present:



Then I can fiddle with the rotations a bit, and again redraw the continent at 150 Mya to fit the supercontinent. In this case, I originally redrew the continent at 0 Mya instead of rotating it to avoid giving it a “non-zero” present-day rotation (which will partially mess up flowlines 100% of the time). Flowlines do sort of “auto-update” to changes in rotation, btw. That said, you might have to re-create the flowline if you change continent shapes and location substantially or add a lot of new rotations. After some further redrawing and changing rotations (and recreating the flowline, several times in this case), here’s the result:



Note that I added a few pieces of extra crust to both continents here. You can do that by simply digitising a new polygon, just as we did when drawing the continent outlines. Then simply give it the same Plate ID as the parent continent, and it will stick along. Also, in this case save it in the same feature collection as its parent to give it the same texture. Now, looking at the flowline we see that the movements actually make sense! Note that since those “arrows” along the flowline are generated at 10 Mya intervals, you could directly estimate from this the ages of the oceanic crust, as well as the location of the mid-oceanic ridge!

However, if you want your GPlates model to not only be functional, but also to look pretty, you might want to actually depict the creation and aging of the oceanic crust. The further sub-sections will cover this process (and will fortunately be much shorter as well).



b. CREATING OCEANIC CRUST

Right, moving on to drawing the continental crust that is created at the mid-oceanic ridge. This is actually a very simple process, but somewhat time-consuming. Basically, the crust needs to be drawn manually. Now, don’t panic, this is actually fairly quick with the help of that flowline we created earlier. The “time-consuming” part comes in because it needs to be done separately for different ages. Which is why you should only consider doing this once your rotations are finished (at least for the two particular continents breaking apart)!

So, first off you need to decide at which intervals you’re going to draw the “new slabs” of oceanic crust. To keep this quick and simple, I’m going to draw them at 50 Mya intervals in this case. So, I need to draw one at 100 Mya (depicting the crust 150-100 old), the next at 50 Mya and a final slab at 0 Mya. So, first set the Time to 100 Mya. Then simply “Digitise new polygon geometry”, just like you did when drawing the continents, and draw the slab of oceanic crust using the flowline as an indicator. If you want to do this quickly, just click from point-to-point along the flowline. If fancy, you can think about where the strike-slip faults should be located based on the shape of the ridge, and put some “zig-zag” there. Those strike-slip faults (the “stretch marks” we noted earlier) should have the same orientation as the flowline, btw. Once you’re happy, “Create feature” and pick OceanicCrust as the type.



Then give it the appropriate Plate ID so it follows the right continent, name it and specify that this slab appears at 100 Mya:



Next, go ahead and click through the final windows and save this in a new feature collection called “Oceanic Crust”. This is important because we want to colour this separately from the continents. Once done, you need to do what you just did for the slab connected to the opposite continent as well, of course. Just remember to give that the Plate ID of the other continent (200 in this case) and save in the feature collection we just created. This is how my continent looks like afterwards:



That done, we can move on 50 Mya. Since this is a tutorial, I’m now going to show you the “quick and dirty” way to draw those slabs faster. Normally, I would of course never consider such underhanded shenanigans, and would draw each slab down to the smallest strike-slip fault. But, you can directly click the flowline (using the “Choose feature” tool), and click “Copy geometry to digitise tool” to automatically copy-paste the points of the mid-oceanic ridge. Outlined in red in the following image:



Then you can just switch to “Digitise new polygon geometry” and finish the outline. Go ahead and create the 50 Mya slabs in this fashion. Again, be careful to give them the right Plate ID and specify that the new slabs appear at 50 Mya. Also, remember to save them in the Oceanic Crust feature collection. This is how the result looks like for me:



And finally, switch to 0 Mya and draw the final slabs. Again, remember the right Plate IDs and specify the time of appearance as 0 Mya. Having done this in the “quick and dirty” way, this is how my crust looks like:



As you can see, it doesn’t look too pretty yet, just the outlines. And if we wanted just the outlines we could have used the “quick and dirty” method to copy the points of the mid-oceanic ridge at various intervals and saved those as lines using “Digitise new polyline geometry (hotkey L)”. So, let’s go ahead and colour it, shall we?



c. COLOURING THE OCEANIC CRUST BASED ON AGE

To colour the oceanic crust we just created, open your Layers manager, and find the Oceanic Crust feature collection. Under “Reconstruction options” check “Fill polygons”. You might want to also set the “Fill opacity” at 0.5, so the flowlines are still visible.



Next, choose Features->Manage Colouring and a new window pops up which lets you to manage the colouring of your various layers. This is pretty useful for other features besides oceanic crust as well, btw. From “Select layer”, choose “Oceanic Crust” and then pick FeatureAge and Default. The result:



If you play the animation, you can now view new oceanic crust popping up, and the colour changes as it ages. Congratulations, you’ve managed to complete this section of the tutorial!

I should note that this method works best for Atlantic-type situations, where an ocean has two passive margins. For other situations, it’s more tricky, but the method is the same. For example, if you have an ocean with an active and passive margin (meaning the crust is subducted on the other side of the ocean), you’d need to create a sort of “ghost plate” to depict the movements of the subducted crust. Say, create a piece of oceanic crust, give it a unique Plate ID and then use it as an “anchor” for the flowline. Those plates will have to be “set up” in the rotation file, of course. Since it’s going to be tricky to specify the present-day location of these kinds of essentially random slabs, it’s probably best just to terminate the flowline at say, 1.0 Mya (or before) to avoid the need to fiddle with “non-zero” present-day rotations. Here’s a quick example I just made:

150 Mya:


100 Mya:


50 Mya:


Remember that flowlines auto-update to rotation changes, so they’re very useful when attempting something like this. In fact, a good “workflow” is to first set up just initial basic rotations, then make use of flowlines to get them right, inserting new rotations and what-not. However, while with these methods we can already get pretty believable movements, we don’t have any clue yet if the velocities of our plates are plausible at all. Next up, I’m going to quickly cover the use of motion paths to create hotspot tracks (they’re very similar to flowlines), and then we’ll look at some methods to find out the velocity of your plates.