[Tutorial] Using GPlates to model continental drift

[Charerg]

Right, so now we have our continents moving, oceanic crust forming, and we can even check their velocity! However, like I mentioned, the velocity can vary at different points of a plate. So, it can be useful to actually generate a velocity field to see how the velocities differ. A velocity field in GPlates looks like this:

VField Example.png

Let’s fire up our model then. I’ve created a new continent for the purpose of this tutorial, but you could also use one of the old ones (they’re just real slow moving in my version, as you noticed in the previous section). Anyway, for velocity fields to work, we need to first create some actual plates rather than continents. Since plates deform and change shape over time and we’re only interested in creating a velocity field for the present day, we only need to do so for 0 Mya though.

So, go under the Digitise tool and choose “Digitise new polyline geometry (hotkey L)”. We will use this to draw the different boundaries. To start off, draw the mid-oceanic ridge, Create feature and choose MidOceanicRidge as the feature type. Save the line in its own feature collection called “Plate Boundaries” (and remember to give it the right Plate ID!). Here are the settings I used (note that I only set this to exist at 0 Mya):

E1.PNG

I also gave the “Plate Boundaries” feature collection a unique colour to make it easy to see the lines. Next, if you envision your plate has a subduction boundary like mine has here, draw another line, Create feature again, and choose SubductionZone as the feature type. Again, save the feature in “Plate Boundaries”.

E2.PNG

Then we need to define the rest of the plate boundaries. Draw another line to connect the ridge and the subduction zone. Make sure that this line intersects both of the previous lines! I decided that this boundary is a transform fault, and chose Fault as the feature type (depending on the tectonics, could be a different type as well).

E3.PNG

And finally draw the 4th line to finish the plate, again making sure it intersects. I picked MidOceanicRidge as the feature type, but it could be something else as well. Your boundaries should now look something like this:

E4.PNG

Note that we could have also drawn just two or three lines (or five or nine, you get the idea). The important thing is that the lines intersect each other, not the number of lines drawn.

Now we can actually define the plate itself. Go under the “Topology“ tools (hotkey 4) and choose “Build new boundary topology (hotkey B)”. Remember to check you don’t have any feature selected, otherwise the option will be greyed out. After choosing “Build new boundary topology”, click one of your lines. The “Topology Tools” window to the right of your screen should now have some options:

E5.PNG

Now click “Add”. Notice how the “Topology boundary selections” window at the bottom of your screen shows your newly added feature:

E6.PNG

Now go ahead and add the rest of the lines into your topology, going around the plate in either clockwise or counter clockwise order. Once done, your screen should look something like this:

E7.PNG

Then click Create and now you can choose ClosedPlateBoundary as the feature type:

E8.PNG

Then give your plate the right Plate ID and the appropriate time of appearance. I used the following settings:

E9.PNG

Save your newly created plate in a new feature collection called “Plates”. Congratulations, you’ve now defined your plate! I chose to give the “Plates” feature collection a unique colour, infill at 0.5 opacity, and hid the previously created “Plate Boundaries”. So, my plate looks like this:

E10.PNG

Now that we have our plate, we can generate the velocity field. I should warn you that doing so will clutter our “Layers” with something like 20+ extra layers. And as far as I’m aware, there’s no way to organise the layers into several layer groups in GPlates. So, you might want to either create a duplicate model for the sake of generating velocity fields, or only generate them after your tectonics model is finished. But, to generate the velocity field Features->Generate Velocity Domain Points->CitcomS. The following window will pop up:

E11.PNG

Set the node density at 9x9 and I also recommend making a separate folder (name it Velocity Field or something) for the output directory, since this will generate quite a few files. Once done, click OK, and the velocity field should now appear.

E12.PNG

You’ll note that the “Layers” are now very cluttered up. You can define the rest of the plates in exactly the same way, and their movement directions will also be shown on the field (no need to recreate the field). I also recommend hiding the nodes (the “mesh” layers). I created the plates A and B as well, and here’s how my planet looks like in Mollweide:

E13.PNG

As you can see, the arrows denote absolute movement direction (movement relative to the planet’s spin axis), and the length of the arrows display the velocity. We can see here that Plate B is moving very slowly (0.7 cm/year), whereas Plate C is moving very swiftly relative to the others (about 9.3 cm/year). Generating a velocity field can be a useful tool if you’re using your GPlates model as a basis for creating a map of your present-day tectonic plates.

So, this is the last of the sections planned so far. You’ve essentially now learned all of my tricks! So, I’ll have to learn some new ones before the next update, which might be a while off. Best of luck with building up your models and happy (continental) drifting!

[Charerg]

Well, I ended up writing the “next update” a bit faster than planned. That’s because I finally bothered to learn how to use conjugate plates. Since this can be a useful feature for more complex reconstructions, I feel it’s a sort of necessary addition to the tutorial before it can be considered complete.

So, as we recall from writing that rotation file, all rotations in GPlates are defined as relative to a conjugate plate, which does not necessarily have to be 000 (the planet’s spin axis). Rotations can also be defined relative to other plates. This is especially useful if you want two plates to move as one piece.

Let’s fire up our fictional model and go back to 150 Mya, which is the point when Continents A and B start breaking apart. What if we wanted to model the movements further back in time? In that case, we’d want A and B to “stick together”, and that’s where conjugate plates come in. Let’s say that we want to set up Plate B as conjugated to Plate A. To do that, we need to first find out what is the 150 Mya Euler rotation for Plate B relative to Plate A (not relative to the spin axis). To find that out, we need to temporarily change the Plate ID of the anchored plate. Click Reconstruction->Specify Anchored Plate ID… and a window will pop up. In my case, I’ll pick Plate ID 100 (Plate A) as the anchored plate.

F1.PNG

Click OK and you’ll note the continents change location. That’s because Plate 100 is now anchored (stationary) instead of Plate 000 (spin axis). So, the spin axis rotates relative to Plate 100 now instead of the other way around. But we don’t need to worry about that, we’ll change this back later. Right now, we’re interested in finding out the Euler rotation of Plate B (Plate ID 200 in my model) relative to the anchored plate. To do that, click Reconstruction->View Total Reconstruction Poles. Once again, a window will pop up.

F2.PNG

As you’ll note, the window has just a single entry, that of Plate 000 relative to Plate 100. That’s because all the other rotations are described relative to Plate 0, so it is the “parent plate” of all the others. Click the arrow next to “Plate ID” in the window to display the rotations of Plate 0’s children:

F3.PNG

Now we can read from the Equivalent rotation rel. anchored plate columns what the rotation of B (200) is relative to A (100) at 150 Mya. Open your rotation file. For me, the relevant entries look like this right now:

200 0.0 90.0 0.0 0.0 000 !Plate B
200 150.0 45.8413 100.6152 -20.1685 000 !
200 2000.0 90.0 0.0 0.0 000 !Plate B

What we need to do now is to add a step over at 150 Mya from movement relative to Plate 000 to movement relative to 100. So, fill in the coordinates and the angle that we just picked up (remember to change conjugate plate to 100 too!). Mine looks like the following:

Plate ID	Time	Latitude	Longitude	Angle	Conj. Plate	Comment
200	0.0	90.0	0.0	0.0	000	!Plate B
200	150.0	45.8413	100.6152	-20.1685	000	!
200	150.0	33.62	25.67	-23.77	100	!Step over to follow A
200	250.0	33.62	25.67	-23.77	100	!Detach from A
200	2000.0	90.0	0.0	0.0	000	!Plate B

Note that you need to add two entries. In this case, I’ve specified that B (200) does not move relative to A from 250 to 150 Mya. Now save your rotation file and reload it into GPlates. Returning to our model, first set the "Anchored Plate ID" back to 0. You'll notice that now we can give A any rotation between 150 to 250 Mya, and B will automatically follow its movements! Also, any rotation we give B between 150 to 250 Mya will be defined relative to Plate ID 100 (rather than 000). To see the outlines of both continents while rotating them, check the "Highlight children" box (outlined in red in the following image).

F4.PNG

And with this addition, I think I’ve well and truly covered my extant knowledge of GPlates. Feel free to ask questions if something was unclear. Comments and ideas are also welcome. Happy drifting!

[cliftonprince]

Thanks for all this info. It's running right up my alley, and will help me out some, though I'm not sure I can accomplish my longer-term project (of getting an accurate simulation of Earth's real tectonic plates but moved in fictional directions). Will give it a go.

[shadixdarkkon]

Hey Charerg, I'm having a problem with GPlates flowline function. No matter what I do my flowlines are showing up like the attached image. I've even done a complete reinstall of GPlates. Do you have any idea what I'm doing wrong?
GPlates 5_22_2018 11_09_46 AM.png

[Charerg]

Hey Charerg, I'm having a problem with GPlates flowline function. No matter what I do my flowlines are showing up like the attached image. I've even done a complete reinstall of GPlates. Do you have any idea what I'm doing wrong?
GPlates 5_22_2018 11_09_46 AM.png

Looks like one (or both) of the continents have a non-zero present day (0 Mya) rotation. Those tend to mess up the flowlines.

[shadixdarkkon]

Ah, gotcha. So that makes it a bit harder to *start* with a supercontinent and work forward. Thank you though, I appreciate the help!

[Charerg]

Ah, gotcha. So that makes it a bit harder to *start* with a supercontinent and work forward. Thank you though, I appreciate the help!

Not necessarily, you can kind of work around it. For example, treat 1.0 Mya as "present day" (in other words change your current 0 mya poles to be 1 mya instead), that way you can keep the 0 Mya reconstruction poles as zero.

[kacey]

Wow, this has come along way since I last checked and I'm just about to dive in to you're tutorial but I wanted to ask a question before I start, I hope that's ok.

I've decided against my better judgement that I'm going to start with a Pangea type continent and break it up. Right now I have the basic land mass setup and the locations of major cratons and weak spots where old mountains ranges used to be, I don't know how they got there or why but that's where I'm going to start. I also have some key areas set up where subduction is going to start to pull it all apart, an ancient ring of fire so to speak, I just have to draw up the plates as they are now a little more clearly then I'll be ready to go but all the G.Plates tutorials I've found start with present day and I really don't want to go backwards...

So how to get my land mass into G.Plates at an earlier date? I would like to start back in time and see where the continents go so I have no idea what the present day situation would look like. Is there an easy way to do this? Or do I need to know where the continents will eventually end up before hand?

Sorry if this has already been asked I haven't gone over the whole thread yet just skimmed through it a little.

[Charerg]

Wow, this has come along way since I last checked and I'm just about to dive in to you're tutorial but I wanted to ask a question before I start, I hope that's ok.

I've decided against my better judgement that I'm going to start with a Pangea type continent and break it up. Right now I have the basic land mass setup and the locations of major cratons and weak spots where old mountains ranges used to be, I don't know how they got there or why but that's where I'm going to start. I also have some key areas set up where subduction is going to start to pull it all apart, an ancient ring of fire so to speak, I just have to draw up the plates as they are now a little more clearly then I'll be ready to go but all the G.Plates tutorials I've found start with present day and I really don't want to go backwards...

So how to get my land mass into G.Plates at an earlier date? I would like to start back in time and see where the continents go so I have no idea what the present day situation would look like. Is there an easy way to do this? Or do I need to know where the continents will eventually end up before hand?

Sorry if this has already been asked I haven't gone over the whole thread yet just skimmed through it a little.

You can just follow the normal procedure, it shouldn't be too problematic. You can import the raster image of your Pangaea, then draw the continents based on that just like you would when starting with the present date arrangement. The only major problem is that some features won't work properly with non-zero 0 Mya pole. Like I mentioned, it should be possible to circumvent that limitation by treating another date (say, 10 Mya) as the "present day".

As an example, if you were to start from 200 Mya, your rotation file for one plate could look something like this:

0 Mya: Zero
10 Mya: Final position
50 Mya: Intermediate position
100 Mya: Intermediate position
150 Mya: Intermediate position
200 Mya: Zero (Starting position)

Let me know how it goes (I might consider adding an extra section to the tutorial about starting from Pangaea if it's too difficult based on the instructions so far).

[kacey]

Thanks Charerg, I was also curiouse, could I not just keep the Pangea at present day 0 mya and have the whole time line go from like 200 mya to 200 million years into the future and figure out the before and the after? Is this possible?