Is Bird ’88 Flying Again?
Back in the 1980s, Peter Bird considered how the Rockies might have been formed through a pair of papers. The first laid out the physical pieces of how a subducting plate could affect the overriding continent if it was in contact; the second combined all those pieces into a numerical model to see what would happen if enough stress was transmitted into the crust to create the Rockies in Colorado and Wyoming. One clear answer at the time was that the mantle lithosphere in the western U.S. would basically go away; the very clear response from those studying volcanic rocks sourced in the mantle was no, that won’t work. Despite this, the flat-slab hypothesis remains the front runner with most geoscientists.
A new paper by Copeland et al. in Geology [paywalled] seems to return to the basic hypothesis Bird envisioned:
Following the hypothesis that Laramide shortening was a consequence of the traction between the base of the North America (NA) plate and the top of the Farallon plate (e.g., Yonkee and Weil, 2015; Heller and Liu, 2016), we suggest that the southwestward migration of the inboard deformational edge (Fig. 1B) was a consequence of a narrowing of the zone of FA–NA lithospheric interaction by progressive rollback of the Farallon plate from northeast to southwest beginning at ca. 55 Ma and continuing into the Oligocene.
Now a lot of the paper is dealing with that rollback, which is actually an investigation of the post-Laramide landscape, but it is some of the material dealing with the start of the Laramide that caught GG’s eye. So we’re going to unpack in detail one figure in order to see if the data is what it seems to be–and to see if this is different that what GG outlined in a 2011 paper. (And hopefully today GG won’t anger yet another unsuspecting author who never expected their work to be examined in public). So hang on if you are coming for the ride….
OK, so let’s start with the figure in question: (it is big so you can see everything; it is tiny in the publication)
Caption: Figure 1. A: Map of western North America indicating locations of magmatic rocks (red circles between 120 and 20 Ma; green circles between 120 and 40 Ma), sites at which esti- mates have been made for age of youngest marine deposits (lavender diamonds), sites of estimates for timing of initiation and cessation of Laramide shortening (blue circles and yellow squares, respectively), sites at which estimates have been made for timing of attainment of maximum surface elevation (black stars), and projected trajectory of thickest part of CSR beneath North America (blue triangles). All points have been restored to their positions at ca. 25 Ma. White line indi- cates extent of Laramide province of DeCelles (2004). Black arrows in lower left indicate approximate relative convergence direction between Farallon and North America plates at times noted. Yellow line indicates approxi- mate extent of Siletzia terrane (Schmandt and Humphreys, 2011; Wells et al., 2014). Pink dashed line marks discontinuity in mantle tomography (Schmandt and Humphreys, 2010) and Laramide magmatism. B: Projection of data south of pink dashed line in A, using same symbols as in A, showing temporal distribution of magmatism, uplift, and deformation along line A-A′. For deformation cessation estimates, yellow rectangles indicate stratigraphic uncertainty. Gray line shows evolution of easternmost interaction of Farallon and North America plates from our thermal-mechanical model (Fig. 2).
OK, it was part B that really caught GG’s eye, because he had a somewhat similar figure years ago:
Now both figures are an eyeful and so we’ll take them apart piece by piece and compare. But first we have to deal with an oddity of Copeland et al.’s Figure 1: it shows modern geography but plots the data points in “their positions at ca. 25 Ma.” This makes it a bit hard to see just where some of these data points are from, so as a public service, GG has taken the discussion in the supplemental material and moved the state boundaries and some topography more or less in line with what was done with the data points:
Although this still isn’t quite right (western California needs to move more to the east or southeast), it at least gives an idea of the source of a lot of points. Note that they used a conservatively small restoration of both the San Andreas system and Basin and Range extension (can compare with McQuarrie and Oskin, JGR 2011 if desired). If you are curious, this is what GG gets from NAVDAT for points from 120-20 Ma with U-Pb and Ar-Ar dates:
OK, with that out of the way, on to the main event. We’ll be plotting parts of Jones et al.’s Fig. 7 alongside Copeland et al.’s Fig. 1B; they should be pretty similar…
Obviously a lot more points on the left, which is a bit surprising as Copeland et al. only use Ar-Ar or U-Pb ages; although this is in part because more data was added to NAVDAT after the figure on the right was made, more significantly, the figure on the right was a swath only 500 km wide just east of the Sevier frontal thrusts while the Copeland et al. figure includes everything south of the Snake River Plain. Only the larger points in the figure on the right are from U-Pb or Ar-Ar. Presumably the ages near 1400 km in Copeland’s figure are the Colorado Mineral Belt (COMB) rocks near 1200 km at right [the difference is presumably a combination of differing trench positions and the Jones et al. use of the larger Basin and Range displacements derived from the McQuarrie and Wernicke restoration].
Now, first, do you see a west to east migration of magmatism? From the Copland et al. dates, it isn’t clear there is any migration in this timeframe. The Sierran arc is continuing to c. 83 Ma; the Peninsular Ranges aren’t terribly different. Yet the predicted edge of Farallon-North American interaction (gray line in Copeland’s Fig. 1B) has these arc rocks being emplaced several million years after the asthenospheric wedge is gone. Worse, it would seem, are the volcanics between 70 and 85 Ma in the Copeland et al. compilation: these are probably ages from southwestern Montana likely tangled up in the Idaho batholith and the igneous activity in the Siletzia embayment, so perhaps this isn’t a real problem (but it does beg the question of why they are included).
But wait, there is more. If we take the Copeland et al. dates at face value–and they are U-Pb and Ar-Ar, so hopefully relatively free of the problems of many older K-Ar dates in the western interior–considerable magmatic activity continued tens of millions of years above the flat slab, where no volcanism should be. Now to be fair, this is mostly because the width of their swath picks up the arc to the south, which simply migrated to the east some over this time. Drop those and probably the picture comes closer to the Jones et al. figure–but there are still these annoying 70-75 Ma plutons in the Mojave Desert that are being created some 15 million years after the asthenospheric wedge is gone in the Copeland et al. model, and some are calc-alkaline (presumably arc) rocks. These are right on the main profile line. Somebody needs to explain the origin of these plutons….[they are a big problem for everybody, really].
Anyways, it seems that by using this wide swath that Copeland et al. actually diminish the strength of their argument–though that argument isn’t helped a lot by looking at the details of the igneous ages. How about the start of Laramide deformation?
Here we are dealing with two very distinct datasets. On the right, Jones et al. used Peter Bird’s 1998 compilation of fault ages and plotted the age when faults were first active from that compilation. The blue line added here would correspond to the initiation of Laramide deformation at a given distance from the trench. On the left, Copeland et al. used a number of sources with varying criteria, many of which postdate the Bird 1998 compilation (many of these presumably should have a finite width at this scale). Unlike the age data, the difference in swath width is only adding in the very old points at left and a few points in Montana, so these are quite comparable.
Despite the different sources, the same picture emerges of Laramide uplifts emerging at about the same time from the Colorado Plateau out to the Wyoming ranges; the newer compilation pushes the ages back a bit towards 80 Ma from the 75 Ma in Jones et al./Bird. There is only a west-to-east gradation if you take the average of the initiation times, which would seem odd: after all, the first structures at a given distance should reflect the creation of the stress field capable of driving deformation at that distance; any later uplifts is just reflecting some evolution of the strain field and just tells you that deformation continues at that distance. The three older points nearer the trench in the Copeland et al. compilation are a bit suspect: two are in Mexico within the active arc and the third is tangled up with interpretations of some complex structures in the Maria fold and thrust belt.
We won’t get into the marine regression data Copeland et al. show: broadly speaking, the marine facies retreat eastward over this time. This seems at odds with the geodynamic model used in the paper where thick oceanic plateau crust has gone to eclogite; you might expect subsidence [the figures show eclogite, but the text claims that the plateau crust stays basaltic until ~58 Ma; this paper does not show any elevation predictions of the model. GG feels he is missing something here].
So what’s new here? Although the datasets are somewhat different, both the Jones et al. (2011) and Copeland et al. (2017) igneous ages and initiation of deformation ages tend to be pretty similar. So it seems the big difference for this aspect of the Laramide is the assertion of Copeland et al. that these age reveal a southwest-to-northeast sweep consistent with migration of the edge of a flat slab moving at 5-7 cm/yr; Jones et al. argued that events were too synchronous for a migrating slab edge to be responsible. Arguably the Copeland compilation makes that case a bit more strongly than the data used by Jones et al. [Again, the Copeland paper also deals with the return of magmatism to the west, which Jones et al. didn’t examine, and presents a geodynamic model that we shall perhaps discuss another day]. So who is right?
Well, you can look at the plots and decide for yourself. GG’s take is that you would not draw a simple sweep of magmatism on the Copeland et al. compilation; near-trench magmatism continues far too long and starts far too early in the distant foreland (arguably at 75 Ma, there is a roughly 1200 km wide swath of continent with some igneous activity). Similarly, the start of Laramide deformation is pretty synchronous over more than 1000 km distance; if there was a migrating deformation front, we would expect more than ten million years of difference from west to east. A model that relies on a migrating zone of igneous activity and deformation does not do well at the scale considered here; given that you want the edge of that migrating zone to precede or track the earliest magmatism and the earliest deformation, the appearance of the line bisecting the majority of points is not confirming a migration but in fact arguing against it.