3.1.2. Implications for the natural hazards, in particular the Vrancea seismicity
The societal impact of active tectonics in the SE Carpathians is one of the largest in Europe. Strong earthquakes within the lithospheric slab that gravitationally sinks into the mantle have a recurrence interval of 10 years for earthquakes with Mw>6.5, 25 years for Mw>7 and 50 years for Mw>7.5 (Oncescu and Trifu, 1987). Within the seismogenic volume (80x40x120km, Fig. 31) the five strong earthquakes of the last century (Oncescu and Bonjer, 1997; Bala et al., 2003) exhibit the largest present-day strain concentration in continental Europe of 2x10-7year-1 (Wenzel et al., 1999) and had a significant impact on densely populated areas, such as the city of Bucharest (Sokolov et al., 2004).
|
|
| Fig. 31. Background seismicity in the Vrancea region in the period 1990-2002, and >M6.0 earthquakes in the last century. Yellow triangles and green square show seismic stations. Inset shows the intermediate depth earthquakes along NW-SE and SW-NE profiles. Note the spatially limited occurrence of earthquakes in the slab and the seismic gap below the Moho at ~40 km depth. (Source: Collaborative Research Centre, CRC 461 Strong Earthquakes, University of Karlsruhe). |
|
|
| Fig. 32. 3D image of the high-velocity mantle body beneath the Vrancea area. Blue and red indicate the +2.5 % and -2.5 % Vp tomographic velocity anomaly, respectively (Martin et al., 2005). The red ellipse approximately shows the area of the seismogenic volume, as given in the inset of Fig. 31. The green arrow indicates that the lower part of the slab is probably laterally torn off (after Martin et al., 2005; see also Martin et al., 2006). |
In order to better assess seismic hazard in the Vrancea region, processes controlling the stress and strain evolution need to be understood. Questions of why seismicity is concentrated in a sub-volume of the high-velocity body (Fig. 32), whether there is a pattern, and what is the role of stress diffusion in the mantle around the high-velocity body, are highly relevant to seismic hazard assessment. A key question is whether the sequence of M > 7.0 earthquakes is random in depth and time, or if there is a causal relationship between them. A simple general hypothesis states that the co-seismic static stress changes, the Coulomb Failure Stress Changes (CFS) control the location of the succeeding earthquake (King et al., 1994). This CFS triggering hypothesis has been tested successfully on the 20th century earthquake sequence along the North Anatolian fault (Stein et al., 1997). However, at greater depth, such as in the seismogenic volume of the Vrancea slab, stress transfer also involves transient processes (Pollitz, 2003).
To clarify whether there is a predictable triggering process through static and transient stress transfer that controls the spatial succession of strong earthquakes in the Vrancea slab, a deterministic 4D numerical model is required. An independent control of such a model is the estimation of vertical movements with GPS and the geologically estimated subsidence and uplift rates on longer time-scales. These observations are assumed to be the surface expression of geodynamic processes that act at depth and control the stress evolution. Processes that play a role in the stress evolution and active tectonics include subduction, re-equilibration of crust and mantle, mass re-distribution (erosion, sedimentation), and post-seismic relaxation.
Several models have already been advanced for the geodynamics and tectonics of the Vrancea region (Bertotti et al., 2003; Cloetingh et al., 2004; Sperner et al., 2004; Dirkzwager et al., 2007). Each of these models addressed a specific question on its inherent time scale and delivered fundamental understanding of (a) the processes that contribute to the topography evolution and its changes in time and space or (b) the stress concentrations and state of stress at depth that are responsible for seismic events of intermediate strength.
 |
| Fig. 30. Numerical model for surface transport in the Romanian Carpathians. (a) Present-day observed topography and predicted drainage network using the historical mean runoff distribution. Numbers indicate water discharge (white, in m3/s) and sediment load (red, in kg/s) at selected locations of the Danube river and its tributaries. River width is plotted proportional to the predicted water discharge. (b) Predicted erosion/deposition (shade) and isostatic vertical velocity of the crust related to the surface mass transport (isolines labeled in mm/y; dashed lines correspond to uplift).U, S and O indicate present-day uplift, subsidence and stable topography, respectively, as inferred from geodetic leveling measurements (after Cloetingh et al., 2003b). |
The next generation of numerical models that address contemporary stress and strain accumulation needs to integrate these different approaches, and to physically link the 3D structural complexity of the crust with mantle processes as well as surface processes such as erosion and sedimentation and their mechanical response. Such an integrated model approach can for the first time address the feedback between processes that act in the crust and lithospheric mantle with sub-lithospheric mantle processes. In order to set up such an integrated model the necessary algorithms and tools are currently developed within the CRC 461 Strong Earthquakes (Wenzel et al., 1998) of the Karlsruhe University. The importance of such an approach has been revealed by Dirkzwager et al. (2007), where a 3D semi-analytical discretization method is used to construct a model geometry representing the present day lithospheric structure of Vrancea, in order to model and ascertain the driving forces behind the vertical GPS field (Pollitz, 1997; Pollitz, 2003). Model results show that the maximum post-seismic deformation due to post-seismic relaxation after the five intermediate strength earthquakes in the Vrancea region between 1940-1990 are approximately 2mm/y, and suggest that post-seismic relaxation is contributing to but not driving the observed vertical deformation pattern. The discrepancy between the modelled and observed vertical velocity rates may be due to inaccuracies in the visco-elastic model, the length of GPS monitoring or alternatively, the post-seismic signal may be superimposed on a positive background signal arising from isostatic rebound of the lithosphere after, for example, ‘break-off’ of the subducted lithospheric slab.
 |
| Fig. 33. 3D geometry of Quaternary deposits and post-orogenic relationship with topography development in the foreland of the SE Carpathians (after Cloetingh et al., 2005a). |
Apart from seismicity, landslides and flooding events that also pose large hazard risks have so far not been assessed in terms of neotectonic activity (Matenco et al., 2007). Displacement of the basement along active normal fault systems can cause large scale collapse of its poorly consolidated Quaternary cover, giving rise to long linear landslide alignments often threatening large inhabited areas in entire SE Moldavia. Due to active uplift of the external SE Carpathians units and subsidence of their foreland, a largely uncompensated and unstable river network developed (Rădoane et al., 2003). Whilst braided rivers rapidly incise in the rising Carpathians, meandering systems deposit thick alluvial sediments in the subsiding foreland (Fig. 30). Furthermore, the present-day subsidence axis is used by the drainage collector of the East Carpathians, the Siret River, which by meander shifts and flooding actively deposits alluvial material in the Focşani basin, an area close to sea level (minimum 2.5m above MSL; Fig. 33). Active subsidence increases the disequilibrium of the system and associated natural risk, as observed in recent years through increasing flooding damages in the Focşani area (e.g., ~1.5bn € of direct damages in 2005). Mitigation of this type of natural risk must take into account the pattern of Quaternary and active tectonics.