Rob Westaway

Uplift-driven valley incision and climate-controlled river terrace development in the Thames Valley, UK

Abstract The sequence of terraces of the River Thames in southeast England has previously been shown to span the period from the earliest Pleistocene to the present. This terrace sequence contains biostratigraphical and sedimentary evidence that testifies to the high-amplitude climatic changes of the Quaternary. Large-scale fluvial incision, resulting in basin-wide terrace formation, appears to have been concentrated at the warming limbs of the major climatic glacial–interglacial cycles, when sediment supply was greatly reduced. This incision and subsequent valley-floor widening created the accommodation space for the later aggradation of the terrace sediments during the following warm–cold transitions and during the cold stages, when high-sediment supply conditions prevailed. Although the timing of terrace aggradation may be controlled by climate change, the progressive valley incision recorded by terrace staircases cannot easily be explained in terms of Quaternary climatic change alone and recently developed models suggest that long-term incision by the Thames has been driven by uplift. This paper presents an overview of the available terrace data and tabulates incision amounts and rates between key stratigraphic horizons. Superimposed upon these broad changes, revealed by the complex internal sedimentary architecture of many terrace sediments, are the geomorphological system responses to both higher-frequency climate-driven changes and more localized intrinsic fluvial system adjustments.

The Quaternary evolution of the Gulf of Corinth, central Greece: coupling between surface processes and flow in the lower continental crust

Abstract The Gulf of Corinth in central Greece is an active normal fault zone with particularly clear evidence of isostatic footwall uplift, constrained by Quaternary marine terraces, and hanging-wall subsidence and sedimentation. It is bounded to the south by a Pliocene to Early Pleistocene sedimentary basin, which is now eroding into the Gulf. Previous work has suggested that the relief across this region has increased dramatically since the Early Pleistocene, due to the isostatic response to increased rates of footwall erosion and hanging-wall sedimentation. It is indeed assumed here that incision accompanying the draw-down of global sea-level at ∼0.9 Ma, during the first major Pleistocene glaciation, initiated the erosion of the basin south of the Gulf of Corinth and so abruptly increased the sedimentation rate in the Gulf. The resulting transient thermal and isostatic response to these changes is modelled, with the subsiding depocentre and eroding sediment source coupled by flow in the lower continental crust. The subsequent enhancement of relief, involving an increase in bathymetry from near zero to ∼900 m and ∼500 m of uplift of the eroding land surface in the sediment source, is shown to be a direct consequence of this change. The model is sensitive to the effective viscosity of the lower crust, and can thus resolve this parameter by matching observations. A value of ∼6×1019 Pa s is indicated, suggesting a viscosity at the Moho no greater than ∼1018 Pa s. Similar transient topographic effects caused by increased rates of sedimentation and erosion are likely to be widespread within the geological record, suggesting that this coupling process involving flow in the weak lower crust may be of major geological and geomorphological importance.

Present-day kinematics of the plate boundary zone between Africa and Europe, from the Azores to the Aegean

Anticlockwise rotation of Africa relative to Europe around an Euler pole in the eastern Atlantic causes plate convergence in the central Mediterranean towards ∼ N20°W at rate considered a priori to be ∼ 5–10 mm yr−1. However, previous summaries of the present-day deformation sense in the central Mediterranean are inconsistent with observations in southern Italy. Recent investigations of Italy and adjacent parts of North Africa enable local deformation sense observed to be reconciled with sense predicted from relative plate motions. Suggested angular velocity of the African plate relative to Europe is ∼ 0.07° Myr−1, around a pole near 21°N, 21°W. This predicts ∼ 2.6 mm yr−1 maximum extension rate towards N64°E near the Azores, consistent with the minimum local accretion of new oceanic crust in the past few millions of years, and ∼ 4.3 mm yr−1 plate convergence towards N20°W in the central Mediterranean at 13°E. Further east, the northern margin of the African plate can be regarded as the Gargano seismic zone in the central Adriatic Sea. The northern Adriatic Sea is a microplate rotating anticlockwise relative to Europe at ∼ 0.3° Myr−1 around a pole in northwestern Italy, causing northeastward shortening in Yugoslavia and up to ∼ 4 mm yr−1 extension in central and northern Italy. The African plate is deforming internally in Tunisia and northern Libya, extending obliquely in direction ∼ E10°S at rate up to ∼ 5.5 mm yr−1. The resultant velocity relative to Europe in the Ionian and southern Adriatic Seas, ∼ 5 mm yr−1 towards ∼ N50°E, explains the observed relative velocity across southern Italy. Kinematic consistency requires present-day shortening rate east of Calabria associated with subduction on the Tyrrhenian Sea Benioff zone and Tyrrhenian Sea extension rate to be both zero or equal to within ∼ 1 mm yr−1.

The mechanical feasibility of low-angle normal faulting

Abstract Low-angle normal faulting is precluded by elementary rock mechanics arguments. However, the existence of many low-angle normal faults in regions such as western North America has long suggested that this familiar theory is incomplete. Initiation of a low-angle normal fault requires the stress tensor within the brittle upper crust to have inclined principal axes, which necessitates a substantial shear stress within the vertical plane. Nonetheless, no-one has previously identified any specific extensional stress field which is suitably oriented to allow shear failure of previously unfractured rock at a low-angle dip of ∼30° in preference to the steeper dip of ∼50–60° predicted by standard theory. Despite previous claims to the contrary, this study shows for the first time that appropriately oriented stress fields can exist. However, they require the shear stress to locally reach ∼100 MPa near the base of a ∼10-km-thick brittle layer. This is only possible under extreme conditions, as it requires dramatic lateral variations in the state of stress across the extending region. It is suggested that a shear stress of this order can develop due to the combined effects of lower-crustal flow (which imparts a horizontal shear traction at the base of the brittle layer) and loading (which imparts shear traction in the vertical plane) associated with the isostatic response to changes in heat flow caused by changes to the geometry of subducting slabs beneath an extending region. Regional patterns of low-angle normal faulting in western North America are thus interpreted in terms of changes to the geometry of subduction of the Farallon plate.

Neogene evolution of the Denizli region of western Turkey

The Denizli region contains one of the easternmost Neogene sedimentary basins in the part of western Turkey that takes up SSW extension. This isolated active normal fault zone, which contains closely spaced en echelon normal fault branches, is investigated using field measurements of fault exposures and tilted sediments, and seismological observations, as a case study to address its style of extension. The Denizli basin is no more than ∼1 km thick and has accommodated up to 4 km of extension. Substantial (∼ 20°) sediment dips are readily explicable assuming extension is accompanied by distributed vertical simple shear, with initial and present-day dips of the main normal faults that control extension most likely 54–57° and 45–50°. Other aspects of the form of this basin require regional uplift at ∼ 0.1 mm yr−1, providing the first indication of major tectonic elevation changes in this actively-extending region that are not directly related to throw on normal faults.

Kinematics of the Malatya–Ovacik Fault Zone

Abstract We investigate the left-lateral slip on the 240-km-long, NE–SW-trending, Malatya–Ovacik fault zone in eastern Turkey. This fault zone splays southwestward from the North Anatolian fault zone near Erzincan, then follows the WSW-trending Ovacik valley between the Munzur and Yilan mountain ranges. It bends back to a SW orientation near Arapkir, from where we trace its main strand SSW beneath the Plio-Quaternary sediment of the Malatya basin. We propose that this fault zone was active during ∼5–3 Ma, when it took up 29 km of relative motion between the Turkish and Arabian plates; it ceased to be active when the East Anatolian fault zone formed at ∼3 Ma. The geometry of the former Erzincan triple junction, which differs from the modern Karliova triple junction, where the North and East Anatolian fault zones intersect, suggests a possible explanation for why slip on the Malatya–Ovacik fault zone was unable to continue. We interpret the SW- and SSW-trending segments of the Malatya–Ovacik fault zone as transform faults, which define an Euler pole ∼1 400 km to the southeast. Its central part along the Ovacik valley, which is ∼30° oblique to the adjoining transform faults, is interpreted as the internal fault of a stepover. The adjoining mountain ranges, which now rise up to ∼3 300 m, ∼2 000 m above the surrounding land surface, are largely the result of the surface uplift which accompanied the components of shortening and thickening of the upper crustal brittle layer that occurred around this stepover while the left-lateral faulting was active.

Chronology of Neogene and Quaternary uplift and magmatism in the Caucasus: constraints from K–Ar dating of volcanism in Armenia

Abstract The Greater Caucasus is one of Earths highest actively-uplifting mountain ranges; the adjoining Caspian Sea basin contains a substantial proportion of its hydrocarbon reserves. Like other parts of the former Soviet Union, the Neogene and Quaternary chronology of these important regions has not previously been well-defined. It has thus been impossible to obtain reliable estimates for rates of processes such as uplift of the Caucasus and sedimentation in the Caspian Sea. Previous studies have established the relative timings of events in the region, using correlation schemes between volcanism, glaciations, and the stratigraphy of the Caspian basin. However, a range of absolute chronologies has previously been proposed for these sediments and igneous rocks, based mainly on different interpretations of their magnetostratigraphic records. By K–Ar dating, we determine the ages of volcanism at three localities in Armenia as 1.1, 0.8 and 0.8 Ma. Using these data and other evidence, we propose a revision to the chronology of this region, in which a distinctive brief interval of normal magnetic polarity in the local sedimentary and volcanic magnetostratigraphic records is matched to the Cobb Mountain event in the global record rather than the Olduvai event or an earlier subchron as had previously been thought. We thus interpret a ∼1.5 Ma timing for the start of volcanism in the Lesser Caucasus, and also suggest a ∼1.2 Ma timing for the Late Akchagyl transgression of the Caspian Sea, a key event in the regional stratigraphy when this water body reached its greatest extent. We tentatively correlate this transgression with the melting event following glaciation during stage 36 of the oxygen isotope timescale, which was thus the first time during the Pleistocene when eastern Europe was covered by a lowland ice sheet. Time-averaged since ∼1 Ma, the flanks of the eastern Greater Caucasus mountains are shown to have uplifted at ∼0.6 mm a−1 and the Lesser Caucasus at ∼0.3 mm a−1. We show that the rate and spatial scale of this uplift are too great to be the result of plate convergence, and suggest instead that this uplift is caused by crustal thickening due to inward lower-crustal flow to beneath these mountain ranges. At the start of magmatism in both the Greater and Lesser Caucasus, the estimated crustal thickness was ∼45 km. We thus suggest that this magmatism has been caused by heating of the mantle lithosphere due to earlier crustal thickening, the temperature rise required to initiate magmatism being the same in both cases.

Rate of strike-slip motion on the Amanos Fault (Karasu Valley, southern Turkey) constrained by K–Ar dating and geochemical analysis of Quaternary basalts

The left-lateral Amanos Fault follows a ∼200-km-long and up to ∼2-km-high escarpment that bounds the eastern margin of the Amanos mountain range and the western margin of the Karasu Valley in southern Turkey, just east of the northeastern corner of the Mediterranean Sea. Regional kinematic models have reached diverse conclusions as to the role of this fault in accommodating relative motion between either the African and Arabian, Turkish and African, or Turkish and Arabian plates. Local studies have tried to estimate its slip rate by K–Ar dating Quaternary basalts that erupted within the Amanos Mountains, flowed across it into the Karasu Valley, and have since become offset. However, these studies have yielded a wide range of results, ranging from ∼0.3 to ∼15 mm a−1, which do not allow the overall role and significance of this fault in accommodating crustal deformation to be determined. We have used the Cassignol K–Ar method to date nine Quaternary basalt samples from the vicinity of the southern part of the Amanos Fault. These basalts exhibit a diverse chemistry, which we interpret as a consequence varying degrees of partial melting of their source combined with variable crustal contamination. This dating allows us to constrain the Quaternary slip rate on the Amanos fault to ∼1.0 to ∼1.6 mm a−1. The dramatic discrepancies between past estimates of this slip rate are partly due to technical difficulties in K–Ar dating of young basalts by isotope dilution. In addition, previous studies at the key locality of Hacilar have unwittingly dated different, chemically distinct, flow units of different ages that are juxtaposed. This low slip rate indicates that, at present, the Amanos Fault takes up a small proportion of the relative motion between the African and Arabian plates, which is transferred southward to the Dead Sea Fault Zone. It also provides strong evidence against the long-standing view that its slip continues offshore to the southwest along a hypothetical left-lateral fault zone located south of Cyprus.

Fault and bed ‘rotation’ during continental extension: block rotation or vertical shear?

Abstract For almost a century, the view has existed that the tilting of blocks between closely-spaced planar normal faults is rigid-body rotation. This interpretation requires only simple geometry, and has consequently found widespread application. However, it is not consistent with the deformation expected around normal faults given the present knowledge of stress fields and rheology in basement in the brittle upper crust, which is better regarded instead as distributed vertical simple shear. Rigid-body rotation and vertical shear involve different relations between fault and bed tilting, and thus predict different initial fault dips for particular present-day dips of faults and beds. These two schemes also predict different amounts of extension, and it is consequently important to establish which is correct. With this aim in mind, we examine normal faults associated with Neogene extension in western Turkey and the western United States, and with Mesozoic extension in the North Sea. Except where extension and the associated tilting are minimal, rigid-body rotation predicts unrealistically steep initial fault dips. Some extensional basins also exhibit reversals of normal fault polarity and tilt polarity of beds, which are incompatible with rigid-body rotation. We therefore conclude that the general cause of the tilting is vertical shear, not rigid-body rotation. This has three main observational consequences. First, the heave on any normal fault equals the amount of extension across it. Second, no feature near a normal fault can rotate through the vertical. A normal fault thus cannot rotate through the vertical and appear as a reverse fault. Third, any initially-vertical feature near a normal fault will remain vertical. A vertical dyke in the tilted surroundings of a normal fault is thus not necessarily younger than the extension that caused the tilting.

River terrace sequences in Turkey: sources of evidence for lateral variations in regional uplift

This paper reviews river terrace staircases in Turkey and examines their relation to regional uplift. Turkish fluvial records are shown to be similar to their counterparts in Europe, with aggradation concentrated in cold climatic stages, despite differences in present-day climate between these two regions. Furthermore, as in Europe, these Turkish river terrace sequences provide evidence for increases in regional uplift rates in the Late Pliocene and Middle Pleistocene. In both regions, this effect is unrelated to senses and rates of plate motion, being instead the result of crustal thickening caused by lower-crustal flow induced by surface processes. From the fluvial evidence, estimated amounts of regional uplift since the Miocene are typically c . 400 m in western Turkey and in the area of the border with Syria. However, they increase northward and eastward to c . 1 km or more in northeastern Turkey on this time-scale, reflecting the regional variations in mean altitude of the land surface. Estimated typical uplift rates during the Middle and Late Pleistocene have been c . 0.1 mm a −1 in the Arabian Platform and c . 0.2–0.3 mm a −1 in western and northern Turkey. These variations are interpreted as the isostatic response to lateral variations in erosion rates.