Keith James replies to Iain Neill (for the Pacific Ocean Paradigm) and Karsten Storetvedt
Geoscientist Online 2 February 2010
Now that the Origin and Evolution of the Caribbean Plate,
GSL Special Publication 328, is available it is timely to respond to contributions to the Geoscientist on-line Caribbean debate.
The fundamental tools we have are data. Chapter 3, SP 328, synthesises and comprehensively references an abundance of these. Chapter 4 presents my current understanding of Caribbean geology, carefully separating data from interpretation and suggesting questions and tests. Readers should compare these with Chapter 1, which makes no acknowledgement of other models or challenges to the model it describes.
I reply first to Iain's article in Geoscientist, November 2009.
The Pacific idea is 40 years old. It is indeed well established. It is not well tested. Arguments supporting it
1, 2, 3 (20 arguments in total),
4 (8 "controversies resolved") were challenged in detail by James, 2006. However, Pacific champions never address arguments against their ideas or discuss arguments favouring an
in-situ origin
5,6. Instead they plea for censorship: "The literature is simply too mature for the proposal of
ad hoc regional or sub-regional models based on small, limited or local data sets to mean anything significant" and: "the peer-review process needs to be tightened up: several recent attempts to build "new" Caribbean models violate so much basic geology, ignore so much geophysical and geochemical data, and/or show a lack of understanding of tectonic synthesis, that they cannot be taken seriously"
4. These proclamations contain no references and detail/disprove none of the "violated" geology or "ignored" geophysical or geochemical data. This is not science.
Now we read Iain's psittacine “breaks so many fundamental geological and especially geochemical concepts."
Let's look at his four pro-POP fundamental concepts.
1) Accommodation of the Caribbean Plate was not possible in the early Cretaceous. The reconstructions Iain refers to are based upon fracture zones in the Central and Southern Atlantic oceans, estimations of Pacific crust that no longer exist and models of hotspot motions (SP 328 Chapter 1). Fig. 3 of Chapter 1 shows only a minor part of Atlantic reconstruction in the Equatorial Atlantic and an earlier version
4 (fig. 2) even shows the Caribbean area blanked out. Whatever sense can be extracted from Chapter 1 on derivation of plate movements, I note that even GPS data do not rigidly define present day motions in the Caribbean area
7,8 (65% change in estimate of Caribbean/N. America motion from earlier model) and
Geoscientist 13/5, p.16-17, noted that hotspots move relative to each other at centimetres per year.
Separation of N and S America began when N America started to move NW and it continues today. Magnetic anomaly and fracture data from the Central and Equatorial Atlantic record this separation and reconstructions based on these allow the presence of the Caribbean in the Cretaceous (SP 328 Chapter 3, fig 4). As the curved (convex-to-west) Mid-Atlantic Ridge moves away from Africa it has to extend. It does so over Caribbean latitudes (Vema Wedge, Ch. 3, fig. 2). Simple observation, that the westerly offset of the Mid-Atlantic Ridge over latitudes 0 – 15°N equals the offset of the Caribbean "plateau" (Ch. 3, fig. 3; magnetic lineations, western Venezuela Basin) from the eastern seaboard of N America, supports the suggestion (Ch. 3, 4) that the two areas share the same geology.
2) Iain, like SP 328 Chapter 1, claims a "general lack of arc-derived tuffs in Proto-Caribbean passive margins until the Maastrichtian-Cenozoic". Chapter 1 also notes that this is a "primary argument" for the Pacific origin of the Caribbean arcs. Chapter 3 (Stratigraphy, Cretaceous) summarises data on Lower Cretaceous sedimentary/metasedimentary sections of northern S. America. They show that rocks grade/continue from continental platform to deep marine environments. They all contain tuffs, exhibit northward increasing silica and volcaniclastic content, and change to metatuffs, pillow lavas and basalts in distal areas. Tuff occurs in Cuban upper Jurassic limestones9 and in metamorphic rocks above Aptian-Albian limestones of the Dominican Republic. The latter contain Orbitolina concava texana showing correlation with Texas, Cuba, northern Central America, Colombia and Venezuela
10.
3) Tethyan – Caribbean connections since the late Cretaceous only? Brachiopods migrated from Tethys to Mexico in the Middle-Late Jurassic
11. Early Aptian rudists extended from Mexico and Trinidad to France and Tethyan margins
12. Aptian – Cenomanian gastropods of Mexico, the Gulf of Mexico and the Caribbean show strong affinity with Tethyan faunas
13, 14.
4) Tomography. Van der Hilst
15 summarized that his tomographic studies imaged the Lesser Antilles subduction zone to a depth of 600 km. Seismicity indicates dips of 60° (vertical south of Grenada), so only 700 km of crust could have been subducted below some 340km of the eastern Caribbean. The data support neither "tomographic suggestion of > 1500 km of subducted crust" nor overriding of more than 2,500 km of "proto-Caribbean" crust by a migrating plate
4, fig. 6.
In his thesis Van der Hilst wrote several cautionary notes. While regional seismic data are desirable for tomographic analysis, Caribbean data are restricted to narrow zones along plate boundaries. He stated " --- the results necessarily have a preliminary character and discussions and conclusions should be considered tentative." He described his interpretation “of inclined, slab-like velocity anomalies as transections through the blurred image of the Atlantic lithosphere subducted below the eastern Caribbean” as "a working hypothesis". Iain joins a long list of authors who (mis) quote Hilst's findings as fact without noting these reservations.
Dziewonski
16 observed that the most famous "example" of subduction is the Farallon slab, which seems to penetrate the lower mantle from the 650 km discontinuity in both P- and S-velocity models. There is a mirror-image high velocity body beneath the Pacific. It lies in an area where there has been no subduction in the last 200 m.y.
I now turn to Iain's understanding of the Pacific paradigm.
1) Migration of the Caribbean “Great Arc” between the Americas. He should take a ruler and measure the total length of the Cretaceous Caribbean "Great Arc" components. By this I mean the actual size of the Greater Antillean blocks, not just the smaller, subaerial parts commonly illustrated by Pacific models such as SP 328 Chapter 1, fig. 7 with Jamaica, Cuba, Hispaniola represented by blobs. While Figs. 9 and 24 of Chapter 1 show definitive outlines of the islands these are subaerial only and they show none of the rotations mentioned by Iain. These figures also show the Caribbean Plate in the Pacific already carrying the Lower Nicaragua Rise, part of the continental Chortis Block supposedly accreted to the Caribbean after it entered between the Americas.
Iain should consider why these Antillean blocks, supposedly extinct Cretaceous volcanic arcs, are many times larger than the islands of extant arcs of the Lesser Antilles, Scotia, the western Pacific and Aleutians. He should quantitatively illustrate how the linear arc entered the Caribbean and became highly curved (decapitated volcanoes dragged in like string of beads, extending, colliding, obducting and rotating?).
2) There is no NE tectonic fabric of the western Venezuela Basin? Donnelly
17 identified linear magnetic anomalies over thick crust (the Caribbean "oceanic plateau") in the western Venezuela Basin where Edgar
et al.18 described a corresponding structural grain of buried scarps and seismic isopachs. Diebold et al.
19 combined these in the same map and they appear on the Exxon world geological map. Chapter 3 of SP 328 discusses these in detail and in regional context. They are also referred to in the Chapter 1, which in addition claims prediction and identification of E-W shears across Costa Rica and seeks to name the "Panama Block". The E-W shears were already published in 2007
20, fig. 11.2b and the "Panama Microplate" has been in the literature since the early 1990s.
3) Half the literature ignored? Iain states that many recent models do not necessarily believe that the oceanic plateau collided with an arc, quoting Pindell et al. 4,
21. That is one model, not half the literature. Moreover, it is a model that includes a "nearly inescapable slab-gap" origin for the plateau, which Iain himself rejects in the same paragraph.
4) Salt solution or spreading – the impossible diapir. Instead speculating on salt behaviour and cursorily dismissing sea-floor diapirs in the Caribbean because he thinks they could not exist, Iain should refer to data
22, 23, 24. Salt diapirs rise several hundred metres above the floor of the Gulf of Mexico, the Mediterranean and several locations in the Atlantic, protected by solution cap rock. Figure 8, SP 328 Chapter 3, shows the similarity between seismic data over the Gulf of Mexico and Caribbean features. It is also unscholarly, Iain, to imply that interpretation of a salt diapir is my only argument for continental crust in the Venezuela Basin.
Iain's further misunderstandings.
1) Thicknesses alone do not identify crustal type. Agreed, but combined with rocks, gravity and seismic data, chemistry and regional geology they offer good clues. Crust up to 40 km thick below the Aves Ridge and dredged granodiorites combined with seismic velocity data indicate continental granodiorites below the ridge and the adjacent, similarly layered Grenada and Venezuela basins
25, 26.
Some of the best, most recent data on "intra-oceanic" arcs come from the Mariana Arc where velocity data also indicate layered crust including thick continental, granitic or andesitic rocks
27. This is an arc whose northern extension, Izu-Ogasawara, next to continental rocks in Japan, is built of continental material.
Over 80% of volcanoes produce andesitic lava and clastics while melts generated in the mantle at subduction zones are mostly basaltic
28. Continental crust is andesitic in composition
29, 30. Tatsumi and Kosigo
31 offer a model of continental crust formation via magma mixing involving present arc basalt magma, present arc mantle, present amphibolite, ancient amphibolite, ancient arc basalt and felsic magma. So you can choose – nine-stage complex magma mixing or presence of original continental crust. Occam would have something to say here. Take away the back-arc basin behind the Mariana Arc and a much simpler explanation of the origin of Mariana continental rocks emerges.
2) Extension lasting 170 Ma seems geologically unfeasible? Pangean breakup was underway by 190 – 180 Ma
32, 33 and divergence between the Americas continues today (SP 328, Ch. 3).
3) I do not argue that the "Caribbean Plateau" (up to 20 km thick) formed by serpentinization of mantle (Iain needs to re-read my words). I suggest this mechanism for formation of Horizon Rough B" and thin (3 km) parts of Caribbean crust. We certainly do not know that the "Plateau" formed in a short space of time. The only samples of the original plateau
34 are a few cores just metres long that sampled Horizon B" - dolerite sills or basalt flows at the top of the "plateau". They do not calibrate a 20 km thick section. Iain's supervisor, Andrew Kerr et al.
35, Table 2 summarized occurrences and ages of plateau rocks in the Pacific and the Caribbean. Ages cluster at 124 - 112 Ma (Barremian-Aptian), 91-88 Ma (Turonian) and 78 - 59 Ma (Campanian-Danian). Kerr et al.35 also noted that basalts of marginal basins of island and continental arcs are potentially the most difficult to distinguish from oceanic plateaux.
According to Pacific models of the “Caribbean Plateau” rocks on Cuba, Jamaica, Hispaniola, Puerto Rico, Aruba, Curaçao, Trinidad and Tobago, in Costa Rica, Guatemala, Panama, Venezuela, Colombia and Gorgona Island are seen as uplifted parts of the plateau. Except for Hispaniola and Puerto Rico, all these are removed from the original "plateau"
34 of the Venezuelan Basin. Cuba lies on the North American Plate and is separated from the Caribbean by the Cayman Ridge and Nicaragua Rise, both with continental basement rocks
36, 37, 38. Trinidad and Tobago lie beyond the Lesser Antilles arc, not behind it, on the South American Plate. The North Coast Schist of Tobago exhibits continental input in the Albian
39, when Iain and his friends would have the island in the Pacific.
As for plumes below plateaux, Schaltegger
et al.
40 found continental zircon xenocrysts in basalts from Iceland and Mauritius while Amundsen et al.
41 noted that continental material has been dredged, drilled or reported as outcrops from a number of volcanic ocean islands and submarine ridges, suggesting that many are underlain by continental fragments. Continental crust is inferred to exist at Seychelles, Faeroes, Rockall Bank, Jan Mayan, Kerguelen, Ontong Java Plateau, Cape Verde and the Cameroon Line. Seismic data indicate that Caribbean "Plateau" architecture replicates that of the calibrated (drilled), extended, eastern continental margin of N America (SP 328, Chapter 3).
4) There are no regional Caribbean unconformities? Mattson
42 catalogued episodic unconformities and hiatuses around the Caribbean in the Aptian-Albian, Santonian, Palaeocene, Middle-Late Eocene and Oligocene. To these we can add Campanian and Miocene unconformities (SP 328, Chapter 3)
5) Why the appearance of continental input in volcanic arc rocks? Uplift and erosion during development of the regional unconformities would be obvious candidates. Note here that the Aptian-Santonian Sans Souci volcanic rocks of Trinidad include quartz sandstones and conglomerates with continental provenance
43, yet Iain et al. would have these forming, like Tobago, in the Pacific ocean.
6) If Iain has proof that the Aves Ridge is an extinct island arc and not a back arc let him present it. Otherwise, Iain, the moon is made of green cheese - I have the proof. At a September 2009, conference in Cardiff Neill et al. wrote: Aves Ridge rocks were weathered on the sea floor and pervasive hydrothermal alteration has occurred. So what confidence do we have in their chemistry (see below)? Geophysical studies do indeed indicate thick crust below the Ridge (up to 40 km, Holcombe et al., 1990). So, no, this is not a mistake on the authors' part.
Geochemists are examining their data? Is the following
44 good, statistical science?
“Although the Mesozoic igneous rocks are generally poorly exposed and badly altered, we have collected and geochemically analyzed 25 samples that place new constraints on plate tectonic models of the Caribbean region. From our recognizance (sic) sampling, six main lava types have been identified within the Mesozoic igneous rocks of Cuba." From this small sample the paper identifies rift basalts, oceanic tholeiites, backarc basin lavas, boninites, island arc tholeiites and calc-alkaline lavas, a short-lived, aborted arc that was mostly eroded away and proposes a new tectonic model of the Caribbean. Table 2 of this paper shows just six samples related to the Cretaceous volcanic arc (see below, Iain's second article).
Similarly, Hastie
et al. 200945, studied just 25 samples from a 1000 m section of lower Cretaceous rocks on Jamaica. They used immobile elements Th and Co as proxies for K
2O and SiO
2, mobilised during hydrothermal alteration, to determine their volcanic arc origins. How do we know the (absent) mobile elements were there in the first place?
I have no wish to, and I do not, trash geochemistry. I genuinely admire efforts of those such as Iain's colleagues
45 and I sincerely wish them success in studying the weathered rocks of the Caribbean. However, like many similar studies, these begin with the premise that the rocks are from a volcanic arc. For all I know they are and I am willing to accept this given proof. That still does not mean they formed in the Pacific. However, I would like to see all rocks analysed in the same way and the data subjected to the same numerical analyses, without presumption of origins - multivariate statistics, impersonal analyses, repeatable, significant/confident discriminants - used consistently. Not “this rock is green - that one is heavy; this island has oranges - that one has apples."
I am sceptical when geochemists themselves cannot agree on origins of the wide range of intrusive igneous rocks on the small (8x3 km) island of Gorgona, offshore Colombia. Early work attributed the rocks to a mid-ocean ridge. Now "mantle melt modelling suggests that enriched magmas are the product of deeper, small degree melting of a pervasively heterogeneous plume comprising a refractory matrix with enriched streaks/blobs, whereas shallower, more extensive melting, results in the formation of relatively depleted magmas"
46. Chemical and radiogenic isotopic heterogeneity of the rocks requires "at least three isotopically distinctive mantle sources (two depleted one enriched) derived in part from recycled oceanic crust and lithosphere, with some transfer of material from the outer core to the plume source region"
47. Wow!
Some attribute Gorgona and the Caribbean plateaux to the Sala y Gomez and Galapagos hotspots, respectively
48. Others conclude that Gorgona is part of the Caribbean large igneous province associated with opening of a slab window within an intraoceanic subduction system
49, 50. Ar-Ar data and palaeontology of overlying sediments document Gorgona activity over the period ~92 - >56 Ma, inconsistent with formation above a stationary hotspot
51.
In his second article ("… a tale……full of sound and fury, signifying nothing") Iain gives us all a short course in plate tectonics (thank you) and the summary that continental crust forms by accretion of oceanic rocks (really?). Re: Caribbean geology he is certainly right to highlight problems of poor exposure, limited access, weathering and incomplete, misused and often inaccurate data. His use of "tweaking" is indeed appropriate for the Pacific paradigm. The "updates" in SP 328, Chapter 1, once again fail to reference arguments against Pacific origins while massaging the ever more complex model to accommodate them.
He twice mentions the Aptian-Albian commencement of N America's westward movement - there is Jurassic – lower Cretaceous crust in the Central Atlantic. Iain, these are data.
Paring away the empty verbiage (cursory glances, fervent debates, albeit in my opinion, eminently resolvable, gamut of facts, hot under the collar, grossly summarising, pristine plateaux, unpublished data, vaguely close to, in general terms, transiently form) leaves the notion that study of a west-dipping, formerly east-dipping, Great Arc will resolve Caribbean geology. Then second and third arcs appear (perhaps even several arcs
52, while the conclusion appeals to "tectonics".
The conference Iain co-chaired in September 2009 included a contribution that denied the Cretaceous Great Arc of the Caribbean
53. Mesozoic, Palaeozoic and Precambrian zircons and undated coarse detrital muscovite, "presumably derived from the South American continental margin" occur in Cretaceous turbidites on Curaçao arc rocks
54. Precambrian and Palaeozoic zircons occur in Cretaceous calc-alkaline volcanic arc rocks in central and eastern Cuba
54. As for the record of long-lived subduction in Cuban Cretaceous HP/LT rocks, buried to 80km or more and resurrected, he should know that such rocks occur intermingled with and grade into sedimentary equivalents.
Caribbean evolution will not be resolved by geochemistry of numerically small samples of hydrothermally- and/or sea-floor- and/or tropically-weathered rocks, geophysics, any other specialization or even "tectonics". It will be resolved by integration all available data and testing of interpretations. Data synthesised in SP 328 Chapter 3 converge to indicate simple extension of exposed and/or drilled geology of nearby continents. Chapter 4 suggests tests (a core hole to check salt diapir vs. seamount would be pretty useful). There is no need to invent complex migrations and rotations of plates, major blocks and arcs, arc polarity flips, complex derivation of silicic andesites from basalt, slab gaps opening in various directions, enriched streaks and blobs in heterogeneous plumes and hotspots or oceanic plateaux. It is needless to invoke Cretaceous subduction burial to/resurrection from 80km when modern basins, such as the Trinidadian Colombus Basin, can produce conditions necessary for blueschist metamorphism and perhaps senseless to do so when no such rocks occur along present day subduction zones.
I welcome new talent to Caribbean geology and I am glad Iain approves of Special Publication 328. I laud his appeal to address data; that is exactly what he can find in Chapter 3 and several other articles. As originator of the Sigüenza conference and senior editor of the resulting volume, SP 328, I accepted a variety of talks and papers in the spirit of collaboration and debate. I disagree with several of these, but readers should have the opportunity to form their own opinions. I hope Iain will open his mind to multiple working hypotheses. He has learned his lessons well but now I believe he should think for himself and consider the problems of plate tectonics raised by, for example, Pratt, 2000
56, Anderson, 2007
57, and would do well to absorb the message of Anderson & Hamilton, 2007
58.
This
Geoscientist on-line debate offers the opportunity for constructive interchange of ideas/understanding. I am sorry that Pacific champions remain hidden behind a research student's skirts. Jim Pindell has told me he has written a long reply to my 2006 arguments paper and would contribute it to this debate. That would be constructive. John Dewey wrote that he and Jim are ready to contribute to the debate if called upon. I'm sure we all look forward to hearing from him. It would also be good to hear from the father of the Caribbean Great Arc, Kevin Burke, and many others with Caribbean opinions (and data).