Subduction zone backarcs, mobile belts, and orogenic heat Roy D. Hyndman, Pacific Geoscience Centre, Geological Survey of Canada, P.O. Box 6000, Sidney, British Columbia V8L4B2, Canada, and School of Earth and Ocean Sciences, University of Victoria, Victoria, British Columbia V8W3P6, Canada, [email protected]; Claire A. Currie, School of Earth and Ocean Sciences, University of Victoria, Victoria, British Columbia V8W3P6, Canada; and Stephane P. Mazzotti, Pacific Geoscience Centre, Geological Survey of Canada, P.O. Box 6000, Sidney, British Columbia V8L4B2, Canada ABSTRACT hot backarc lithosphere, not from the Two important problems of continental orogenic deformation process itself. tectonics may be resolved by recogniz- INTRODUCTION ing that most subduction zone backarcs The model of plate tectonics with have hot, thin, and weak lithospheres narrow plate boundaries provides an Figure 1. The North America Cordillera mobile over considerable widths. These are (1) excellent first-order description of global the origin of long-lived active “mobile belt. The high elevation and complex current tectonics. Plate tectonics also provides tectonics illustrate the hot, weak, backarc belts” contrasted to the stability of cra- an elegant explanation for orogenic lithosphere deformed by variable margin forces. tons and platforms, and (2) the origin crustal shortening and thickening in of the heat of continental collision orog- terms of continental or terrane colli- eny. At many continental margin plate sion. However, a number of large-scale deformation over long geological peri- boundaries, there are broad belts with tectonic problems are not explained by ods. They have some characteristic that a long history of distributed deforma- the simple rigid plate and continental allows them to maintain an especially tion. These regions are mobile because collision models. In this article, we pre- thick, strong lithosphere, such as a more the lithosphere is sufficiently weak to sent explanations for two such tectonic refractory mantle composition (Jordan, be deformed by the forces developed problems: the origin of long-lived active 1978; Forte and Perry, 2000). The rea- at plate boundaries. We conclude that “mobile belts” that lie along a number son for the long histories of tectonic mobile belts are weak because they are of continental margin plate boundaries, activity in the mobile belts has not been hot, and they are hot as a consequence contrasted to the stability of cratons and clear. Also, most continental mobile of their position in present or recent platforms, and the origin of the heat of belts have high elevation; they are backarcs. Most continental backarcs continental collision orogeny. Current mountain belts. Surprisingly, however, have thin and hot lithospheres, not just mobile belts are up to 1000 km wide there is commonly no crustal thicken- those with active extension or rifting. and cover nearly a quarter of the conti- ing. In this article, we emphasize the Moho temperatures are 800–900 °C and nents (e.g., Stein and Freymueller, 2002; important distinction between long-lived lithosphere thicknesses are 50–60 km, Thatcher, 2003). For example, the North mobile mountain belts where the eleva- compared to 400–500 °C and 150–300 American Cordillera has been tectoni- tion is mainly thermally supported and km for cratons and other regions with a cally active for >180 m.y., with evidence short-lived continent or terrane collision thermotectonic age greater than ca. 300 for older tectonic events extending this orogenies where there is major crustal Ma. The temperature differences result history to >350 m.y. (e.g., Burchfiel et shortening and thickening. in backarc lithospheres that are weaker al., 1992) (Fig. 1). Rates of deformation Plate tectonics provides an elegant than cratons by more than a factor of in current mobile belts are commonly explanation for orogenic crustal short- 10. Backarcs may be hot because shal- 5–15 mm/yr; i.e., 10%–20% of the main ening and thickening in terms of con- low asthenosphere convection results plate boundary rates as indicated by tinental or terrane collision when an from the reduction in viscosity due distributed seismicity, precision global ocean closes. However, an important to water rising from the underlying positioning system (GPS) measure- question remains unresolved that we subducting plate. Hot, weak, former ments, and geological studies. attempt to address: What is the origin backarcs are expected to be the locus The long geological histories of defor- of the heat of orogeny? Significant heat of most deformation during continent mation suggest that these mobile belts from frictional deformation processes or terrane collision orogeny. The heat exhibit long-term lithosphere weak- has been discounted through studies by indicated by orogenic plutonism, high ness compared to cratons. In contrast, a number of authors, and most other grade metamorphism, and ductile defor- the Precambrian cratons and stable orogenic processes should absorb rather mation may come from the preexisting platforms have exhibited little internal than generate heat. GSA Today: v. 15, no. 2, doi: 10:1130/1052-5173(2005)015<4:SZBMBA>2.0.CO;2 4 FEBRUARY 2005, GSA TODAY In this study, we examine the hypoth- 2004). It is well recognized that exten- eses that (1) most mobile belts are sional backarcs are hot (e.g., Wiens located in backarcs or recent backarcs and Smith, 2003), including the Basin that are characterized by hot, thin, and and Range and extensional west Pacific weak lithospheres; and (2) most conti- backarc basins. In these regions, rifting nent or terrane collision orogenic belts and spreading are obvious sources of involve shortening of former hot back- heat. We therefore focus on nine back- arc mobile belts. We do not discuss less arcs where we conclude that there is no common weakening mechanisms that current or recent thermally significant subsequently may be reactivated in col- extension in the region, as indicated lision, such as localized extension and by GPS, seismicity, or geological stud- hotspots. ies (Fig. 3). Our areas are all more than BACKARC THERMAL STRUCTURE 100 km from such extension; Morgan (1983) showed that the thermal effects Why are mobile belt regions so weak Figure 2. Temperature vs. depth for the of rifts extend only a short distance. The and why are cratons so strong for long Cascadia backarc and adjacent craton (from geological periods? The primary reason Hyndman and Lewis, 1999). thermal effect of extension decays with appears to be that the mobile belts are a time constant of ~50 m.y., so we have hot, whereas cratons are cold (e.g., Q), effective elastic thickness, Te, ther- also excluded areas with significant Vitorello and Pollack, 1980; Chapman mally supported high elevations, seismic mid-late Cenozoic extension. and Furlong, 1992), and, below the estimates of lithosphere thickness, and We summarize the two principal indi- upper crust brittle zone, there is a strong upper mantle xenoliths. The high tem- cators of high temperatures in Figure decrease in strength with increasing peratures of current mobile belts are 4 (for details and references, see the 1 temperature. The systematic decrease also indicated by widespread sporadic GSA Data Repository ): (1) surface 2 in heat flow with age, or time since the Cenozoic basaltic volcanism, in con- heat flow greater than ~70 mW/m (for last “thermotectonic event,” has been trast to the almost complete absence of normal upper crust heat generation 3 well recognized. The greatest difference recent volcanism in cratons and stable of 1.0–1.5 μW/m ) (e.g., Lewis et al., is between the currently active mobile platforms. The same indicators of tem- 2003; Chapman and Furlong, 1992); and belts (70–90 mW/m2) compared to the perature show the lithospheres of cra- (2) low seismic velocities in the upper cratons (~40 mW/m2) (e.g., Chapman tons to be cold and >200 km thick. mantle (i.e., Pn velocities <7.9 km/s and Furlong, 1992; Vitorello and We have carried out a global survey compared to ~8.2 km/s for cratons) Pollack, 1980). Temperatures at the base of continental backarcs, and, in all (e.g., Black and Braile, 1982; Lewis et al., of the crust are ~400 °C lower for cold cases where there is sufficient data, we 2003) or seismic tomography veloci- cratons compared to young, hot mobile have found surprisingly high and uni- ties lower than the global average by at belts (i.e., 400–500 °C vs. 800–900 °C, form temperatures across wide zones, least 2% for P-waves or 4% for S-waves respectively). Although only the cool- indicated by heat flow and other deep (~4% and ~7% slower relative to cratons) est case of cratons is discussed here, temperature constraints (also see Currie, (e.g., Goes et al., 2000; Goes and van Paleozoic stable areas are only slightly warmer than Archean cratons and are estimated to be almost as strong. As an example, Figure 2 shows temperature estimates for the Canadian Cordillera mobile belt compared to the adjacent Canadian Shield (Hyndman and Lewis, 1999; Lewis et al., 2003). With these thermal estimates, the base of the mobile belt lithosphere (hot enough for vigorous convection, close to the solidus) is at a depth of only 50–60 km. The mobile belt high temperatures and thin lithospheres also are inferred from many other deep temperature indica- tors, including temperature-dependent uppermost mantle seismic properties (Moho refraction velocity, Pn, tomog- raphy compression and shear wave Figure 3. Examples of nonextensional backarcs with high heat flow and velocities, Vp and Vs, and attenuation, other high temperature indicators. 1GSA Data Repository item 2005030, Table DR1, notes, and references for Figure 4, is available on request from Documents Secretary, GSA, P.O. Box 9140, Boulder, CO 80301-9140, USA, [email protected], or at www.geosociety.org/pubs/ft2005.htm.