Vasiliki Mouslopoulou | Helmholtz Zentrum Berlin (original) (raw)

Papers by Vasiliki Mouslopoulou

Coastal uplift is common in continental fore-arc systems, with elevated paleoshorelines indicatin... more Coastal uplift is common in continental fore-arc systems, with elevated paleoshorelines indicating that uplift rates can vary dramatically over time on individual margins. The origins of these changes in uplift rates are examined using a global data set of paleoshorelines together with 2-D numerical models of subduction systems. Empirical paleoshoreline data (N = 282) from eight subduction margins indicate that uplift rates are generally not steady state and varied by up to a factor of 20 during the late Quaternary (≤125 ka). On many subduction margins uplift rates increase to the present day, a finding which we attribute to sampling bias toward those locations where Holocene uplift rates have been highest (with respect to other global margins which have undergone fast subsidence or no vertical motion—e.g., a property akin to the so-called Sadler effect). Paleoshorelines and 2-D models suggest that transient uplift rates at subduction margins are mainly a short-term (<20 ka) phenomenon that cannot be accounted for by plate-boundary scale processes such as changes in the rates of plate convergence, sediment underplating or isostatic unloading. Instead, time-variable uplift rates are ascribed to temporal clustering of large-magnitude earthquakes on upper plate faults and, to a lesser extent, the subduction thrust. The potential for future damaging earthquakes and tsunamis may have been underestimated at active subduction margins with no measureable Holocene uplift, and in such cases, late Quaternary paleoshorelines could provide an important constraint for hazard analysis.

[1] Historical earthquakes are often strongly clustered in space and time. This clustering has be... more [1] Historical earthquakes are often strongly clustered in space and time. This clustering has been attributed to static stress triggering associated with tectonic fault interactions and/or to fluid migration. Discrimination between these two models requires detailed information on the timing, location and size of earthquakes. The Matata earthquake sequence in the Taupo Rift, New Zealand, provides a unique opportunity to chart spatial and temporal patterns of earthquakes along individual faults and their relations to other faults in the system over timescales of days to years. This is possible because there are 2563 accurately relocated (within <± 100 m) earthquakes (1 < Mw < 4.7) that ruptured the rift's crust during 49 months of earthquake activity. These earthquakes define subparallel faults which were active over discrete periods of time, with earthquake activity migrating initially along and subsequently across the rift. Although the process of microearthquake generation on individual faults is highly complex, we show by means of semivariogram analysis that system‐wide correlations extend over at least eight successive earthquakes. The system‐wide coherence requires the interaction of earthquakes located on neighboring faults over timescales that span tens of days, and it is achieved when about 76% of the total faults in the system are included. This pattern is consistent with tectonic (as opposed to fluid‐triggered) fault interactions that have been established for larger magnitude (e.g., Mw > 5.5) earthquakes that occur over thousand year timescales (e.g., <60 kyr).

The catastrophic earthquakes that recently (September 4th, 2010 and February 22nd, 2011) hit Chri... more The catastrophic earthquakes that recently (September 4th, 2010 and February 22nd, 2011) hit Christchurch, New Zealand, show that active faults, capable of generating large-magnitude earthquakes , can be hidden beneath the Earth's surface. In this article we combine near-surface paleoseismic data with deep (<5 km) onshore seismic-reflection lines to explore the growth of normal faults over short (<27 kyr) and long (>1 Ma) timescales in the Taranaki Rift, New Zealand. Our analysis shows that the integration of different timescale datasets provides a basis for identifying active faults not observed at the ground surface, estimating maximum fault-rupture lengths, inferring maximum short-term displacement rates and improving earthquake hazard assessment. We find that fault displacement rates become increasingly irregular (both faster and slower) on shorter timescales, leading to incomplete sampling of the active-fault population. Surface traces have been recognised for <50% of the active faults and along 50% of their lengths. The similarity of along-strike displacement profiles for short and long time intervals suggests that fault lengths and maximum single-event displacements have not changed over the last 3.6 Ma. Therefore, rate changes are likely to reflect temporal adjustments in earthquake recurrence intervals due to fault interactions and associated migration of earthquake activity within the rift.

[1] The amount of extension accommodated in active rifts by earthquakes that do not rupture the g... more [1] The amount of extension accommodated in active rifts by earthquakes that do not rupture the ground surface (e.g., <Mw 5.5) is often poorly constrained. The Matata Earthquake Sequence (MES), a high-quality dataset of 2563 relocated microearthquakes (1 < Mw < 4.7) that ruptured the Taupo Rift in New Zealand over a period of 49 months, has been used to quantify the proportion of extension produced by small to moderate sized earthquakes. Analysis shows that the extension rate across the rift due to the MES is 2.4AE0.7 mm/yr (at 1s standard deviation), an average extension rate for small to moderate magnitude earthquakes which also prevailed during the preceding 28 years (1977–2004) and represents up to ~30% of the total contemporary deformation recorded across the north Taupo Rift by GPS (15 AE 5 mm/yr). The bulk of the MES (94%) occurred at depths of 1.5 to 6.5 km and may not be observed in geological datasets (e.g., as displacements of the ground-or near-surface horizons). Small-scale faulting associated with microseismicity may thus record strains not measured by geological datasets and constitute an important component of the ~3–10 mm/yr disparity between geological and GPS rates of extension across the Taupo Rift. Citation: Mouslopoulou, V., D. T. Hristopulos, A. Nicol, J. J. Walsh, and S. Bannister (2013), The importance of microearthquakes in crustal extension of an active rift: A case study from New Zealand,

This article appeared in a journal published by Elsevier. The attached copy is furnished to the a... more This article appeared in a journal published by Elsevier. The attached copy is furnished to the author for internal non-commercial research and education use, including for instruction at the authors institution and sharing with colleagues. Other uses, including reproduction and distribution, or selling or licensing copies, or posting to personal, institutional or third party websites are prohibited. In most cases authors are permitted to post their version of the article (e.g. in Word or Tex form) to their personal website or institutional repository. Authors requiring further information regarding Elsevier's archiving and manuscript policies are encouraged to visit: http://www.elsevier.com/copyright High-quality Light Detection and Ranging (lidar) data collected across the Rangitaiki Plains, the fastest extending section of the onshore Taupo Rift, New Zealand, reveal 122 active fault traces and provide new constraints on displacements, displacement rates and paleoearthquakes of the normal faults in the rift. The identified lineaments are scarps that record vertical offset of geomorphic surfaces (e.g., beach ridges, meander channel floors, river terraces, etc.) and trend parallel or sub-parallel to other active faults in the rift; these lineaments are interpreted to be active faults. Active fault traces trend ∼ 060° and their lengths range from 0.25 to 6 km. They mainly traverse and displace a diachronous landscape of b 6.5 kyr age, with throws that vary from 0.05 to 7 m and form a graben. Historic, geometric and kinematic constraints have been used to aggregate individual traces into eight fault zones that range in length from 5 to 40 km. Displacement rates in the rift beneath the Rangitaiki Plains vary between and along individual faults by more than one and three orders of magnitude, respectively, over the last 0.64 and/or 1.72 kyr. Variability of displacement rates on individual faults arises from episodic slip accumulation during a minimum of 15 paleoearthquakes of variable slip and recurrence interval. Repeated fault movements have produced about 3 mm/yr of subsidence over the last ∼ 2 kyr within the rift (i.e. between the Edgecumbe and Matata faults) while relatively stable conditions have persisted (uplift/subsidence of 0–0.6 mm/yr) on the rift shoulders for the last ∼ 3.3 to ∼ 6.5 kyr. These plain-wide signals of vertical movement were interrupted by short-lived episodes of rapid uplift (0–1.72 kyr) and subsidence (1.72 to ∼ 2.1 kyr) at the western and eastern margins, respectively, which we infer to result from prehistoric earthquakes.

The Hellenic subduction margin in the Eastern Mediterranean has generated devastating historical ... more The Hellenic subduction margin in the Eastern Mediterranean has generated devastating historical earthquakes and tsunamis with poorly known recurrence intervals. Here stranded paleoshorelines indicate strong uplift transients (0–7 mm/yr) along the island of Crete during the last ~50 kyr due to earthquake clustering. We identify the highest uplift rates in western Crete since the demise of the Minoan civilization and along the entire island between ~10 and 20 kyr B.P., with the absence of uplifted Late Holocene paleoshorelines in the east being due to seismic quiescence. Numerical models show that uplift along the Hellenic margin is primarily achieved by great earthquakes on major reverse faults in the upper plate with little contribution from plate-interface slip. These earthquakes were strongly clustered with recurrence intervals ranging from hundreds to thousands of years and primarily being achieved by fault interactions. Future great earthquakes will rupture seismically quiet areas in eastern Crete, elevating both seismic and tsunami hazards.

Keywords: paleoshorelines beachrock eustatic sea-level rise rock uplift Eastern Mediterranean Cre... more Keywords: paleoshorelines beachrock eustatic sea-level rise rock uplift Eastern Mediterranean Crete Paleoshorelines of Late Quaternary age in western Crete do not exclusively increase in age with rising altitude as is generally observed worldwide. At numerous sites, for example, Late-Holocene paleoshorelines decrease in age with increasing altitude while in other cases paleoshorelines at similar altitude vary in age by tens of thousands of years. We propose that the observed paleoshoreline altitude–age relationships can be accounted for by eustatic sea-level changes and tectonic rock uplift without requiring substantial errors on radiocarbon ages or tectonic subsidence, as has been previously proposed. To test this model we use a dataset consisting of altitude and age data for 71 individual paleoshorelines sampled from 21 sites distributed along the entire Cretan coastline. These data include radiocarbon ages of marine biota (40 new dates) within beachrock resting on paleoshorelines ranging up to 48 kyr BP in age and ≤20 m above present sea-level. We find that paleoshoreline formation reflects Late Holocene tectonic rock uplift in western Crete, preceded by eustatic sea-level rise and by >10 kyr BP rock uplift along the entire island. Our observations contravene existing models as they suggest that some paleoshorelines, and their associated lithified beachrock, survived passage through the wave-zone multiple times and formed throughout the sea-level cycle (i.e., preservation is not restricted to highstand deposits). These results may have application globally in regions where erosion-resistant carbonate beachrock mantles paleoshorelines.

Journal of Structural Geology, 2014

The late-Cenozoic kinematic and late-Pleistocene paleoearthquake history of the Spili Fault is ex... more The late-Cenozoic kinematic and late-Pleistocene paleoearthquake history of the Spili Fault is examined using slip-vector measurements and in situ cosmogenic ( 36 Cl) dating, respectively. The Spili Fault appears to have undergone at least three successive but distinct phases of extension since Messinian (~7 Ma), with the most recent faulting resulting in the exhumation of its carbonate plane for a fault-length of 20 km. Earthquake-slip and age data show that the lower 9 m of the Spili Fault plane were exhumed during the last~16,500 years through a minimum of five large-magnitude (Mw > 6) earthquakes. The timing between successive paleoearthquakes varied by more than one order of magnitude (from 800 to 9000 years), suggesting a highly variable earthquake recurrence interval during late Pleistocene (CV ¼ 1). This variability resulted to significant fluctuations in the displacement rate of the Spili Fault, with the millennium rate (3.5 mm/yr) being about six times faster than its late-Pleistocene rate (0.6 mm/yr). The observed variability in the slip-size of the paleoearthquakes is, however, significantly smaller (CV ¼ 0.3). These data collectively suggest that the Spili Fault is one of the fastest moving faults in the forearc of the Hellenic subduction margin.

Journal of Structural Geology, 2007

Interaction and displacement transfer between active intersecting strike-slip (or transform) and ... more Interaction and displacement transfer between active intersecting strike-slip (or transform) and extensional fault systems are examined. Outcrop data from a well-preserved strike-slip fault and rift intersection in New Zealand are compared to a global data set of 13 such intersections in both continental and oceanic crust. Displacement transfer between strike-slip and normal faults is typically accomplished by gradual changes of fault orientations and slip vectors close to the intersection zone. For two-and three-plate configurations, these changes result in sub-parallelism of the slip vectors of the component faults with their line of intersection. The dimensions of the area over which fault-strike and slip vectors change are principally controlled by the extent to which displacements on the dominant of the two intersecting fault systems are confined to a single slip surface or distributed across a zone. Where slip is spatially distributed, the region in which the two displacement fields are superimposed produces transtension and associated oblique slip. This distributed off-fault deformation facilitates the development of a quasi-stable configuration of the fault intersection region, maintaining both the regional geometry and kinematics of the intersection zone which, in many cases, would not be possible for rigid-block translations. The dimensions of the transition zone are larger for continental crust than for oceanic crust because oceanic crust is thinner, fault geometries in oceanic crust are simpler two-plate configurations and the slip vectors of the component intersecting fault systems are sub-parallel.

Geology, 2009

Displacement rates for normal and reverse faults (n = 57) are generally higher when averaged for ... more Displacement rates for normal and reverse faults (n = 57) are generally higher when averaged for the Holocene (~10 ka) than for the late Quaternary (~300 ka) and longer time scales. Holocene acceleration of displacement rates could be attributed to geological processes that produce increases of tectonic tempo. We propose an alternative model in which the observed rate changes arise from variability in earthquake slip and/or recurrence coupled with a sampling bias toward those faults that are best represented at the Earth's surface and that accrued displacement fastest during the Holocene. This model is consistent with displacement rates measured over time intervals of as long as ~300 ka for 129 faults from the Taupo Rift, New Zealand. Departures of earthquake parameters and associated displacement rates from their long-term (>300 ka) averages are attributed to fault interactions and occur on time intervals inversely related to these long-term displacement rates and to regional strain rates.

Geological Society, London, Special Publications, 2009

ABSTRACT The North Island Fault System (NIFS) is the longest and highest slip-rate active strike-... more ABSTRACT The North Island Fault System (NIFS) is the longest and highest slip-rate active strike-slip fault system within the Hikurangi subduction margin in New Zealand, accommodating up to 10 mm/a of the margin-parallel plate motion. Displacement of landforms over the lastc. 30 ka indicates a gradual northward change from right-lateral strike-slip to oblique-normal slip along the northern NIFS and within 60 km of its intersection with the active Taupo Rift. This change is expressed by a c. 608 increase in the pitch of the slip vectors. We use fault data from 20 trenches and displacements along active traces to explore whether changes in late Quaternary fault kin- ematics principally arise due to earthquake rupture arrest and/or variations in slip vector pitch during individual earthquakes that span the kinematic transition zone. Results show that earthquake rupture arrest occurs along the strike of the NIFS, with at least 60-80% of all events during the last 10-13 ka terminating across the zone of late Quaternary (c. 30 ka) transition from strike-slip to oblique-normal slip. The strike of the faults across the kinematic transition is unchanged, and we suggest that rupture was arrested there due to a 20- 308 northward shallowing of fault-dip across this zone. Rupture arrest limits earthquake lengths and magnitudes which, when combined with recurrence intervals from trenching, locally decreases the seismic hazard in the region of the faults. Simple kinematic earthquake slip models, which simulate the addition of slip vectors during individual earthquakes, suggest that rupture arrest was accompanied by a northward steepening of slip vectors during individual earthquakes. Changes in coseismic slip vectors may arise due to the northward decrease in fault dip and associated steepening of the principal com- pressive stress axis (s1) which, in turn, is due to fault interactions between the NIFS and the adjacent active Taupo Rift.

Geological Society, London, Special Publications, 2007

The 500-km-long strike-slip North Island Fault System (NIFS) intersects and terminates against th... more The 500-km-long strike-slip North Island Fault System (NIFS) intersects and terminates against the Taupo Rift. Both fault systems are active, with strike-slip displacement transferred into the rift without displacing normal faults along the rift margin. Data from displaced landforms, fault-trenching, gravity and seismic-reflection profiles, and aerial photograph analysis suggest that within 150 km of the northern termination of the NIFS, the main faults in the strikeslip fault system bend through 258, splay into five principal strands and decrease their mean dip. These changes in fault geometry are accompanied by a gradual steepening of the pitch of the slip vectors, and by an anticlockwise swing (up to 508) in the azimuth of slip on the faults in the NIFS. As a consequence of the bending of the strike-slip faults and the changes in their slip vectors, near their intersection, the slip vectors on the two component fault systems become subparallel to each other and to their mutual line of intersection. This subparallelism facilitates the transfer of displacement from one fault system to the other, accounting for a significant amount of the NE increase of extension along the rift, whilst maintaining the overall coherence of the strike-slip termination. Changes in the slip vectors of the strike-slip faults arise from the superimposition of rift-orthogonal differential extension outside the rift margin, resulting in differential motion of the footwall and hanging-wall blocks of each fault in the NIFS. The combination of rift-orthogonal heterogeneous extension (dip-slip) and strike-slip, results in a steepening of the pitch of the slip vectors on the terminating fault system. Slip vectors on each splay close to their terminations are, therefore, the sum of strike-slip and dip-slip components, with the total angle through which the pitch of the slip vectors steepens being dependent on the relative values of both these two component vectors. In circumstances where interaction of the velocity fields for the intersecting fault systems cannot resolve to a slip vector that is boundary-coherent, either rotation about vertical axes of the terminating fault relative to the through-going fault system may take place to accommodate the termination of the strike-slip fault system, or the rift may be offset by the strike-slip fault system rather than terminating into it. At the termination of the NIFS, an earlier phase of such rotations may have produced the 258 anticlockwise bend in fault strike and contributed up to about one-third of the anticlockwise deflection in slip azimuth. On the terminating strike-slip NIFS, therefore, rotational and nonrotational termination mechanisms have both played a role, but at different times in its evolution, as the thermal structure, the rheology and the thickness of the crust in the rift intersection region have changed.

Earth and Planetary Science Letters, 2009

a b s t r a c t Editor: R.D. van der Hilst

Earth and Planetary Science Letters, 2011

Keywords: rare earth elements carbonate scarp soil paleoearthquake normal fault Crete Recent work... more Keywords: rare earth elements carbonate scarp soil paleoearthquake normal fault Crete Recent work has utilised a well-established earthquake record on a normal fault in Italy (the Magnola Fault) to successfully test a new method for identifying paleoearthquakes on carbonate rocks: that of chemical analysis of their exhumed fault planes. Here we take the next natural step, applying this novel method on a notionally active normal fault in Greece, the Spili Fault, for which no paleoearthquake record exists. Despite the 'blind' sampling, data reveal an outstanding record of systematic fluctuations in the concentrations of Rare Earth Elements (REE) and Yttrium (Y) upscarp, which closely resemble those recorded on the Magnola Fault. Chemical analysis of 35 core-samples extracted from a 10 m high section of the exhumed Spili Fault plane records upscarp depletion in the REE-Y concentrations at an average rate of ca. 9.3%/m. Depletion is overprinted by locally increased REE-Y concentrations upscarp. A minimum of four such concentration fluctuations, with wavelengths ranging from 0.5 to 3 m, are recorded. Each fluctuation is interpreted to be generated by at least one paleoearthquake that episodically exhumed a zone of the fault plane. Each zone consists of an upper domain that is enriched in REE-Y and a lower un-enriched domain. REE-Y enrichment is due to the prolonged (at least few 100's of years) contact of the limestone with the soil, whereas the un-enriched domain reflects instantaneous uplift from depths greater than the base of the soil, during the same earthquake. The REE-Y analytical method cannot resolve individual small-sized earthquakes (with slip less than the thickness of the soil-cover) and/or individual large-and small-sized earthquakes which are clustered in time (repeat timeb 100's yr). It may therefore yield better results when applied on large (≥20 km) carbonate faults that rupture the earth's crust at most once every ca. 0.5 kyr; nevertheless the number of identified earthquakes should always be treated as a minimum.

Over the last ~15 years we have excavated 73 trenches across active normal faults in the Taupo an... more Over the last ~15 years we have excavated 73 trenches across active normal faults in the Taupo and Hauraki Rifts, North Island, New Zealand. The stratigraphy in these trenches is quite similar because of the predominance of volcanic and volcanic-derived deposits, sourced from the active Taupo Volcanic Zone. These deposits, whether alluvial (reworked, mainly volcanics) or volcanic (tephra), are all