NSERC Research Proposal 1997

Introduction

The focus of our research is lithospheric deformation during orogenesis and the associated interactions of the solid earth, hydrosphere and atmosphere acting as a system (Fig. 1). This field of research is important because orogeny is the principal process that creates continents. Most of our research is collaborative (see form 100) and significant progress has been achieved by numerical model experiments combined with comparisons with observations. This proposal outlines the mechanical aspects of the research. For other aspects, reviewers are referred to the accompanying: PAPER 1- a summary of the surface process modelling and coupled surface process and tectonic models; PAPER 2- an example of insight gained from a simple model; PAPER 3- results of thermal-mechanical models and initial metamorphic application, and; PAPER 4- a collaborative project concerning a young orogen. The approach in each theme is to start from basic principles and to develop the underlying idea and its variations systematically using model experiments.

Principal Results of Projects - Mantle Subduction (MS) Models of Orogenesis

This research addresses a class of problems in which crustal deformation is forced by subduction of the underlying mantle lithosphere (Fig. 1). The proposition is that continental and oceanic mantle lithospheres behave in a similar way during convergence - they subduct. What are the styles of deformation of the overlying crust and to what degree are compressional orogens mechanical analogues of accretionary wedges?

Below, I list the results of specific projects designed to address this problem. Table 1 relates developments concerning Processes to the basic doubly-vergent model. Model types are:

Class 1: Crustal deformation driven by mantle subduction (MS)

• Plane strain vertical section (MS-PS), Thermal Mechanical (MS-TM)
• Thin crustal sheet (MS-TS)
• Fully 3D (MS-3D)

Class 2: Whole lithosphere deformation by lateral indentors (WL)

• Single thin sheet (WL-TS)

1) Basic Model (MS-PS). Willett, Beaumont and Fullsack (Geology 21, p.371, 1993) demonstrates underlying concepts (Fig. 1), the generality of the basic mantle subduction model and relates the model results to small orogens (Fig. 2a style), large orogens and active margins.

2) Thin Sheet Models Forced by Basal Boundary Conditions (MS-TS). Ellis, Fullsack and Beaumont (Geophys. J.I., 120, p.24, 1995). The mantle subduction model can also be investigated in planform. It is shown that there is a non-dimensional parameter, the Ampferer number (ratio of basal coupling to crustal strength) that characterizes the problem in addition to n(the creep power law exponent) and the Argand number (ratio of gravitational force to crustal strength). Two main results contrast with WL-TS results (England et al - several papers during the 1980's). In oblique convergence, the ratio of MS-TS horizontal deformation scales for compression and shear is 2:1, whereas the corresponding WL-TS result is 4:1. The corresponding absolute scales of deformation in MS-TS are determined by the Ampferer number and not the length scale of the bounding indentor as in the WL-TS model. Ellis (Geology, 24, p.699, 1996) discusses how subduction and indentor models of lithospheric deformation may be reconciled.

4) Strain Partitioning in Obliquely-Convergent Mantle Subduction Models. Braun and Beaumont (JGR, 100, p.18059, 1995) use a fully MS- 3D model (development by Braun) to demonstrate that, in weakly transpressional orogen models with a Griffith crust, strain is partitioned into thrust-sense pro- and retro-shears (as in basic PS-MS model, Fig. 2a, see also PAPER 1 for terminology) co-existing with and bounding strike-slip `flower' type structures. Under stronger transpression, the strike-slip `flowers' merge with the step-up shears to create non-partitioned oblique-slip zones. The partitioned mode corresponds closely with the Coast Ranges California, whereas the non-partitioned mode is like the continental collision zone, South Island, New Zealand.

5) Styles of Small Compressional Orogens (MS-PS). Beaumont, Fullsack and Hamilton (Tectonophysics, 232, p.119, 1994). Shows results of a range of numerical experiments for small convergence in relation to initial crustal thickness. Specifically, the effects of plasticity and power law creep, thermal structure of the crust, degree of crustal subduction and denudation are investigated. Predictions are made for basic seismic reflectivity patterns in orogens. Oppositely vergent primary structures in orogens do not require a change in polarity of subduction.

6) Large Orogens (MS-PS), Himalaya-Tibet (H-T). Willett and Beaumont (Nature, 369, p.642, 1994) use large orogen MS-PS models. The work combines these models with `rollback' (ie. retreat of the subducting mantle lithosphere) to propose a model of the H-T system. Comparison of the numerical experiments with the first order asymmetry of the H-T orogen and position of the Indus-Tsangpo suture indicates these observations are best satisfied by subduction of Asian lithospheric mantle. Accompanying rollback explains anomalous mantle beneath northern Tibet as a sub-crustal asthenospheric window. In general, both WL-TS and MS-PS models make similar predictions. The main difference is that no mantle thickening (with or without convective destabilization) occurs in the MS-PS models. The MS and WL models are therefore opposite end members of a spectrum of lithospheric deformation styles.

8) Effect of Subduction Zone Retreat on PS-MS Models. Waschbusch and Beaumont (JGR, 101, p.28133, 1996) investigate the consequences of the motion of the subducting slab away from the overriding (retro-) plate during convergence. Retreat of the subduction zone (ie. hinge point of the subducting plate) leads to tectonic underplating in the space created between the pro- and retro- lithospheres. The thickness and amount of deformation of the tectonic underplate are simple functions of the subduction zone retreat rate and fraction of the crustal layer that subducts. Model predictions agree with the low elevation, limited exhumation and single-vergence style of many orogens formed at retreating plate boundaries.

9)Thermal-Mechanical MS-TM Coupled Models. Coupling between mechanics and heat is a first-order process in orogens. Application of MS-TM models to Barrovian metamorphism, specifically heating by Tectonic Accretion of Radioactive Material (TARM), is discussed in Jamieson, Beaumont et al. (in press) PAPER 3 (see, in particular, discussion/Fig. 22).

Tectonic Model Applications

Application modelling is to determine: 1) if the mantle-subduction (MS) model agrees with observations from specific orogens; 2) what causes the particular characteristics of these orogens, and; 3) whether a general classification of orogens can be based on this approach. In each project (Table 2), we start from the basic MS model and determine the minimum variations in the model properties (ie. surface processes, crustal properties, basal boundary conditions) required to explain the primary observations. In the model experiments we test the proposition that the diversity of observed styles stems from interactions among relatively few, more simple components and recognize that interactions are key to the behaviour of the systems as a whole. Although this research is not complete, some preliminary conclusions can be drawn.

10) European Alps (<50Ma phase). Extremely well studied orogen. First-order features compatible with subduction-subduction/collision-collision model (Fig. 2d). Model results suggest retro-charriage related to slab-breakoff; transpression may be like non-partitioned Braun and Beaumont (1995) style; nappes formed from mechanically decoupled high-level layers; Lepontine dome Barrovian metamorphism may be caused by TARM.

11) Pyrenees (Alpine Phase) (Fig. 2c). Compatible with MS model with Iberian lower crust/mantle subducted. Large-scale structure explained by compressional reactivation of mid-crustal detachment formed during Cretaceous extension. Denudation not asymmetric but important, as is linkage between crustal deformation and detachment above Triassic salt (blue) in region of Tremp piggyback (Pg) basin.

12) Southern Alps, New Zealand (<10Ma). End-member collisional orogen with oblique convergence; Pacific mantle and lower crust (?) subduct; collision style controlled by asymmetric subduction and extreme asymmetric surface denudation (see PAPER 4). These results are compatible with recent observations (Stern et al., EOS, 1977). Non-strain partitioned tectonic style gives oblique slip on Alpine fault (retro-step-up shear). Strike variations in tectonics compatible with subduction zone retreat at southern end of orogen.

13) Central Coast Ranges, California (<3.5Ma). Tectonic style like strain partitioned, weakly transpressional MS-3D model.

14) Newfoundland Appalachians (Silurian orogenesis) are compatible with `Vise model' style (Fig. 2b) in which crust in the orogen core is weakly coupled to the mantle (low Ampferer number). External crust is strong and squeezes weak core like a vise. Tectonic styles contrast strongly with basic model (Fig. 2a). Deformation is distributed throughout the weak core with little evidence of S-position in the deformation (Type 2 Orogen behaviour, see proposed research and Fig. 3b).

Proposed Research - Orogen Tectonics and Surface Processes

An assessment of our research, considered both alone and in the context of CIAR-ESEP and other collaborations, indicates that several facets of the research warrant or require further attention. I outline these themes below and link them to specific projects, noting collaborations and hqp training.

A) Toward a Unified Model of Orogenesis

Our main focus has been the mantle subduction (MS) model of orogenesis. An opposite end-member hypothesis holds that the crust and mantle lithosphere initially deform together followed by the sinking of the dense mantle lithosphere (Fig. 3d). We propose to develop a `Unified Model' that embraces the end-member hypotheses, intermediates, other tectonic styles, and evolutionary differences from the Archean to present. A prototype UM (Fig. 3) for modern orogenesis proposes that type and evolutionary style be related to basal boundary conditions, internal properties and surface processes. Our medium term goal (4 yrs) is to develop modelling capabilities for Type 1-4 orogens (Fig. 3) (in 3D, if possible), use model experiments to determine characteristics of each Type, and compare these with observations.

Our current modelling capabilities address Types 1,2 and 3 orogens (Fig. 3) in cross section, in thermal-mechanical coupled form, and Type 1 orogens in 3D. We propose to use these techniques to complete research on relatively small Type 1-2 orogens, where available information indicates that the mantle lithosphere subducts. Do Type 1 or 2 styles (Fig. 3) apply to: i) the Swiss Alps and Convergent Margins (current research with Susan Ellis (hqp) and Adrian Pfiffner (Bern)); ii) Small Lithoprobe Orogens (current research with Susan Ellis (hqp) Bern); iii) the New England and Lachlan Orogens, Australia (current and proposed research with Paula Waschbusch (hqp) and Russell Korsch (AGSO, Canberra)); iv) the Urals? (Parallel research by other members of the CIAR Earth Systems Evolution Program poses equivalent questions in relation to Taiwan, Central Australia, and the Andes.)

By including subduction zone retreat in Type 1 models, we propose to investigate whether, v): the tectonic style of the Apennines/Tyrrhenian Sea and Aegean Sea can be explained; the associated crustal extension/thinning explains rapid exhumation of low temperature-high pressure (p<15kbar) metamorphic rocks, and; the mechanism is a candidate for the exhumation of ultra-high pressure (UHP) rocks (p~30kbar). (Research with Becky Jamieson and graduate student (hqp)).

B) The Role of the Mantle Lithosphere in Orogenesis

Deformation of mantle lithosphere is key to transitions among the Type 2-3-4 orogens (Fig. 3). Development of a Unified Model requires that our kinematic requirement, that mantle lithosphere subduct, be relaxed. We have noted this limitation in our research in earlier proposals. We have proposed a more dynamical approach that requires the modelled region of the Earth to expand to include the lithosphere and asthenosphere and in which subduction is modelled as a density driven flow (ie. generalized slab pull) into a viscous asthenosphere. Two main difficulties arise. 1) An efficient large-deformation numerical model is required. In our approach this is an Arbitrary Lagrangian-Eulerian implicit finite element thermal-mechanical model for creeping flows with viscous and plastic/frictional material properties. It must include total gravitational forces and a true upper free surface (not a perturbation analysis). Philippe Fullsack has developed a new code (MOZART) that meets these requirements and can efficiently solve 100 x 400 fe meshes, giving sufficient resolution for the enlarged model domain. We propose: vi) experiments to determine deformation style as functions of the model Argand and Ampferer numbers, buoyancy of the subducting slab (ie. including slab advance/retreat and breakoff). The significance is to link lithospheric deformation (eg. the crustal expression of orogenesis) to the larger scale mantle dynamics. 2) Suitable high performance computer resources are required - either through the proposed upgrade to the Dalhousie SP-2 (see accompanying equipment proposal) or buying time elsewhere. I regard this (B) theme as a priority. In order to maintain the momentum of our research, we must broaden the modelling capability beyond the crust. Significant resources (50/50 Lithoprobe/NSERC) have already been invested in the development of MOZART.

C) Internal Styles of Deformation in Orogens.

There is also the small-scale, in which the focus is part of an orogen, eg. the external thin-skinned fold-and-thrust belts (FTB's), regions with significant petroleum and mineral resources. Evolution of FTB's is understood from critical wedge (Coulomb and other) mechanics, laboratory scale modelling, geometrical analysis (eg. balanced cross sections), surface mapping and subsurface seismic and drilling. The extensive database makes FTB's prime candidates for numerical model experiments designed to understand dynamical processes involved in their evolution. We propose to: vii) develop a modelling capability, based on MOZART, with relatively high resolution (potentially 100m vertically x 250m horizontally), to investigate the importance of stacked lithologies (laminate nature), fluid pressures, flexure, surface processes (denudation and sediment deposition) on the cross-sectional evolution of FTB's. The coupled fluid flow computation is the only part that needs development. The initial focus will be the Southern Canadian Cordillera and Western Newfoundland and the collaboration will involve (Daniel Melanson, hqp, Glen Stockmal (GSC, Calgary, an expert on mapping and analysis of these FTB's) and Philippe Fullsack, hqp). The first target will be `triangle zones,' followed by `in sequence and out of sequence thrusts,ramp/flat structures, duplexes, imbricates' etc. The same approach can be applied at the medium scale and we will seek opportunities for collaborative research on specific orogens with projects like the Pyrenees and Alps (Fig. 2) in mind.