Defining tectonic zones

From BemlarGroup

Defining tectonic zones (v.3)
by Peter Bird, 5 September 2007

I. Objective

In our NSF proposal, “Time-dependant earthquake forecasts with a tectonic basis” we promised to define various “tectonic zones” and then test whether ETES models optimized separately in each zone might show significantly different behaviors. While writing the proposal, we made some basic decisions:

• Zones will be defined as surface areas into which epicenters (and/or epicentroids) of shallow earthquakes (<70 km) may fall. Depth of the earthquake will not be a criterion because it is usually poorly known (within the 0~70 km depth range) unless there is local station control. Focal mechanism will not be a criterion because this is not usually available for the smaller earthquakes (<5.6) which make up large portions of the aftershock swarms we need to include in the study. Also, delays in obtaining focal mechanisms would be a problem for real-time forecasting.
• The number of zones should be relatively small (e.g., ~5) so that there are lots of earthquakes falling in each zone, allowing for reliable ETES model optimizations and quick turn-around in forecast tests. Also, this will keep our workload under control, since we have to do a separate analysis for each zone. Therefore, no zone will have a local geographic name like “Aleutian”; instead, zones will be of generic type like “Trench”. We will allow one tectonic zone to be the union of many non-contiguous patches.
• The preceding rules will make it impossible to separate strike-slip from normal-faulting earthquakes on mid-ocean spreading ridges, or in continental rift zones. (We assume that strike-slip earthquakes will tend to dominate these zone subcatalogs.) Also, along many trenches it will be impossible to separate subduction-related earthquakes from back-arc-spreading earthquakes. (We assume that subduction earthquakes will completely dominate these zone subcatalogs.) Therefore, the tectonic zones will not be equivalent to the 7 plate-boundary classes of Bird [2003], and they need new names that will not be confused with plate-boundary classes.

Based on these considerations, I propose the following short list of tectonic zones (with identifying integers for compact representation in computer files):

• (4) Trench (including SUBs & OCBs, & earthquakes in outer rise or overriding plate)
  
• (3) Fast-spreading ridge (oceanic crust, spreading rate > 40 mm/a; includes transforms)
  
• (2) Slow-spreading ridge (oceanic crust, spreading rate < 40 mm/a, includes transforms)
  
• (1) Active continent (including continental parts of all orogens of PB2002, plus designated continental plate boundaries of PB2002)
  
• (0) Plate-interior (the rest of the Earth’s surface)
  

Since previous Jackson & Kagan forecasts using ETES have emphasized trenches, and since trench seismicity tends to dominate on our planet, we expect that the Trench zone results will be similar to published results. There is more chance for surprises in the other 4 tectonic zones. Note that fast vs. slow ridges are divided at 40 mm/a because (a) Bird & Kagan [2004] found a change in OTF earthquakes at ~39 mm/a, and (b) there is a change in ridge morphology at this rate (from central-graben to peaked, as spreading rate increases). “Active continent” seems like a terribly general zone which might be profitably subdivided; however, much of continental seismicity is in “orogens” of PB2002 where plate boundaries (other than SUB) are not yet defined—this means that we could only subdivide by focal mechanism, which we have decided not to do.

II. General Method

To create Figure 5 in the proposal, I manually drew lines on a map in Adobe Illustrator. This is not a good method for serious work. The tectonic zones should be defined by objective rules, implemented in a computer program, because this is:

• more reproducible, and therefore more scientific;
• easier to explain in the papers we will write;
• easier to revise if we decide to change any rule(s); and
• less subject to procedural errors.

I believe it will be easier to write a program that assigns tectonic zone integers to grid points (or earthquake epicenters/epicentroids) than it would be to write a program that draws curves along boundaries of tectonic zones. It will also be slightly easier to use grid-of-points output to classify earthquakes from catalogs (or future earthquakes). If we need zone-boundary curves for some reason in the future, I think I could create them from a grid of points with another program. So, I propose to write a program which loops through latitudes (+90 to -90, in 0.1-degree steps) and also loops through longitudes (0 to 360, in 0.1-degree steps), and creates a grid of integers to identify the tectonic zone at each point. The output will be a relatively compact representation (1801 latitudes * 3601 longitudes * 2 bytes = 12 MB). The necessary datasets (e.g., elevation, lithospheric age, PB2002 plate boundaries and orogens) are already available in digital form.

III. Specific Method

At every point in the grid (or every epicenter), the procedure will be:

• Test whether the point lies in any “subduction lane” as defined by Bird & Kagan [2004] using SUB (SUBduction zone) plate-boundary steps and Euler poles from the PB2002 model of Bird [2003]. If so: (4) Trench.
• Otherwise, test whether the point is adjacent to any OCB (Oceanic Convergent Boundary) of PB2002. If so: (4) Trench.
• Otherwise, test whether the point lies within any of the 13 orogens of PB2002. If so:
  • Test whether crust is “continental” at that point? Yes: (1) Active continent.
  • Non-continental crust is “oceanic”. Assign tectonic zone according to orogen name:
    • “Alps” orogen: (4) Trench.
    • “Laptev Sea” orogen: (2) Slow-spreading ridge.
    • “Persia-Tibet-Burma” orogen: (4) Trench.
    • “Ninety East-Sumatra” orogen: (4) Trench.
    • “Philippines” orogen: (4) Trench.
    • “New Hebrides-Fiji” orogen: (3) Fast ridge.
    • “western Aleutians” orogen: (4) Trench.
    • “Alaska-Yukon” orogen: (4) Trench.
    • “Gorda-California-Nevada” orogen: (3) Fast ridge.
    • “Rivera-Cocos” orogen: (3) Fast-spreading ridge.
    • “Peru” orogen: (4) Trench.
    • “Puna-Sierras Pampeanas” orogen: (4) Trench.
    • “west central Atlantic” orogen: (2) Slow-spreading ridge.
• For points NOT within any orogen of PB2002, test whether crust is “continental” at that point?
  • Yes, continental: Test for proximity to any non-SUB/OCB plate boundary:
    • Yes: (1) Active continent.
    • No: (0) Plate interior.
  • No, not continental; therefore oceanic crust:
    • Adjacent to any slow OTF and/or OSR?: (2) Slow-spreading ridge.
    • Else, adjacent to any fast OTF and/or OSR?: (3) Fast-spreading ridge.
    • Else: (0) Plate interior.

Note that the terms “lane” and “adjacent” and “proximity” in the lines above refer to the finite-width seismicity zones around each of the 7 classes of plate boundaries whose widths were defined by Bird & Kagan [2004]. The definition of “continental” crust is based on elevation and/or lithosphere age, as specified in Bird [2003].