S31C-01
Interpretations on the Geologic Setting of Yogyakarta Earthquakes 2006 (Central Java, Indonesia) Based on Integration of Aftershock Monitoring and Existing Geologic, Geophysical and Remote Sensing Data
The unprecedented 26 May 2006 Yogyakarta earthquake (central Java, Indonesia) that took victims of 5,700 lives was generally accepted to have a depth of about 10 km and moment magnitude of 6.4. However, the definition of location of active fault is still under debate as the epicenter of mainshock was reported quite differently by several institutions. Many researchers believe that the Opak fault which is located at the eastern boundary of Yogyakarta low-land area (or Yogyakarta Basin) and the high-land region of Southern Mountains was the source of year 2006 earthquakes. However, our result of aftershocks observation suggests that the ruptured zone was not located along the Opak fault but from an unknown fault located about 10 km to the east from it and within the Southern Mountains domain. Unfortunately, surface geologic manifestations are scarce as this area is now largely covered by limestone. Therefore the suspected active fault system must be studied through interpretations of the subsurface geology and evaluation of the Cenozoic geo-history of the region utilizing existing geologic, geophysical and remote sensing data. This work suggests that the Yogyakarta Basin is a volcano-tectonic depression formed gradually since the early Tertiary period (Oligo-Miocene or older). Geological and geophysical evidence suggest that structural trends changed from the Oligocene NE-SW towards the Oligo-Miocene NNE-SSW and the Plio-Pleistocene NW-SE and E-W directions. The ruptured "X" fault during the Yogyakarta earthquakes 2006 is likely to be a NNE-SSW trending fault which is parallel to the Opak fault and both were firstly active in the Oligo-Miocene as sinistral strike-slip faults. However, while the Opak fault had changed into a normal faulting after the Pliocene, the evidence from Kali Ngalang and Kali Widoro suggests that the "X" fault system was still reactivated as a strike-slip one during the Plio-Pleistocene orogeny. As this new interpretation of active fault causes spatial discrepancy between locations of earthquakes epicenters and highly damaged regions, other geo-engineering factors must be considerably important in determining the final scale of seismic hazards. The most vulnerable areas for seismic hazards are those located nearest to the ruptured fault and are underlain by thick Quaternary unconsolidated deposits. In case of regions along the fault line, seismic hazards seem to reach more distance region, such as the case of Gantiwarno region, as the seismic waves can travel more easily along the fault line.
S31C-02
Geological Influence on the Site Response of Bantul Earthquake at Yogyakarta Special Province, Indonesia
On May 27, 2006 a magnitude 6.2 earthquake struck Yogyakarta Special Province in Central Java. The earthquake designated the Bantul earthquake resulted in the deaths of 5,800 and left 200,000 families homeless. Over 280,000 buildings were damaged or destroyed. Housing quality was generally poor with no accounting for earthquake-resistant design. Housing and building materials were also of poor quality. In spite of the poor design and construction materials, local geological conditions were determined be a significant influence on the site response which affected the intensity of building damage. In addition, strong ground shaking was unusually long at 57 seconds. In response to the findings of the initial field reconnaissance a series of micro-tremor surveys supported by aerial and satellite image interpretations, ground penetration radar and magneto telluric surveys, as well as engineering geological site investigations were conducted. The objectives of the surveys and investigations were: to investigate various factors controlling the levels of site response which induced damage to homes and buildings; and, to produce a seismic hazard micro-zonation map. Provision of this micro-zonation map is crucial to support the enhancement of building code and landuse management in the earthquake prone area at Bantul. The Bantul area is located in a valley formed a graben. This valley was due to two major normal faults extending towards North East - South West. The east border of the valley is bounded major normal fault formed the Progo River, whilst at the west border another major normal fault formed the Opak-Oya Rivers. Drilling correlations of the bedrock (of andesitic breccia) in the valley, was found at a depth of about 40 m below the existing ground surface. Bedrock was covered by a layer of clay produced from paleo-swamps with various thicknesses from 1 to 2 m. After the clay deposition, the basin then was filled by loose fluvial sediments consisting of gravelly sand and clayey sand, with intercalation of laharic deposits consisting of bouldery-gravelly sand. Micro-tremor surveys conducted across the Bantul Basin, indicated that amplifications occurred were limited to zones of loose gravelly sand - clayey sand within the first of 30 m from the ground surface. The surveys identified four lateral zones where varying degrees of site response was identified. The zones were differentiated by levels of very high, high, moderate or low amplification. The surveys and investigations concluded that the most significant control of the site response is the local stratigraphic conditions, especially related to the type, density and thickness of sediments. Some other controls on the amplifications of site response are the existence of faults, the depth to bedrock, the shallow depth of the groundwater table, and the distance of the site to the earthquake epicenter.
S31C-03
Fine-Tuning the Seismic Potential of the Nicoya Gap in NW Costa Rica
The Nicoya seismic gap is a subduction segment along the Middle American Trench where the Cocos plate subducts under the Caribbean plate. This seismic gap, located under the Nicoya Peninsula in northwestern Costa Rica, has ruptured with large earthquakes in 1853, 1900 and 1950. Its strong coupling contrasts with the immediately adjacent weak segments: Nicaragua to the NW and central Costa Rica to the SE. The Nicaragua segment has a very high level of background seismicity with several earthquakes per year with magnitudes above 5.0; its most recent large event was an abnormally slow earthquake (Mw=7.6) that generated a destructive tsunami in 1992. The central Costa Rica segment also has a high level of background seismicity with very frequent events with magnitudes above 4.5; its last large event (Mw=7.0) occurred in 1990 just at the boundary with the Nicoya seismic gap. The aftershock areas of the Costa Rica, 1990 and Nicaragua, 1992 earthquakes allowed the geographic extent of the Nicoya seismic gap to be clearly defined. Evidences that support a strong coupling for the Nicoya segment are: a) very low background seismicity; b) the sudden end of aftershocks of the 1990 and 1992 earthquakes at its boundaries; and c) the fast NE motion (parallel to convergence) of the Nicoya Peninsula (nearly 30mm/yr.) observed with GPS. Recent seismological and geodetic studies (Protti et al., 2001; Newman et al., 2002; Iinuma et al., 2004; Norabuena et al., 2004; DeShone et al., 2006) restrict the seismic coupling to 50 +/- 5 percent and the potential rupture area to 8000 +/- 1500 km2. These values, together with a convergence rate around 88mm/yr and no significant seismic slip since 1950, give the Nicoya seismic gap a potential, for the following 5 years, to generate a earthquake with Mw=7.8 +/- 0.1 magnitude.
S31C-04
Assessing Seismic Hazards - Algorithms, Maps, and Emergency Scenarios
Public officials in charge of building codes, land use planning, and emergency response need sound estimates of seismic hazards. Sources may be well defined (e.g., active faults that have a surface trace) or diffuse (e.g., a subduction zone or a blind-thrust belt), but in both cases one can use a deterministic or worst-case scenario approach. For each scenario, a design earthquake is selected based on historic data or the known length of Holocene ruptures (as determined by geologic mapping). Horizontal ground accelerations (HGAs) can then be estimated at different distances from the earthquake epicenter using published attenuation relations (e.g., Seismological Res. Letters, v. 68, 1997) and estimates of the elastic properties of the substrate materials. No good algorithms are available to take into account reflection of elastic waves across other fault planes (e.g., a common effect in California, where there are many strands of the San Andreas fault), or amplification of waves in water-saturated alluvial and lacustrine basins (e.g., the Mexico City basin), but empirical relations can be developed by correlating historic damage patterns with predicted HGAs. The ultimate result is a map of HGAs. With this map, and with additional data on depth to groundwater and geotechnical properties of local soils, a liquefaction susceptibility map can be prepared, using published algorithms (e.g., J. of Geotech. Geoenv. Eng., v. 127, p. 817-833, 2001; Eng. Geology Practice in N. California, p. 579-594, 2001). Finally, the HGA estimates, digital elevation models, geologic structural data, and geotechnical properties of local geologic units can be used to prepare a slope failure susceptibility map (e.g., Eng. Geology Practice in N. California, p. 77-94, 2001). Seismic hazard maps are used by: (1) Building officials to determine areas of the city where special construction codes have to be implemented, and where existing buildings may need to be retrofitted. (2) Planning officials to evaluate plans for new growth (though in most cities land use patterns are historically established). (3) Emergency response officials to plan emergency operations. (4) Insurance commissioners to estimate losses and insurance claims (e.g., with FEMA's software HAZUS).
S31C-05
Seismic Hazard and Risk Assessment and Public Policy in the Central United States
Making a public policy on seismic hazard mitigation is not an easy task because it not only depends on seismic hazard itself, but also on seismic risk and other related social and economic issues. Seismic hazard and risk is the basis, however. Although seismic hazard and seismic risk are two fundamentally different concepts, they have been used interchangeably. Seismic hazard describes natural phenomena, such as surface rupture, ground motion, ground-motion amplification, liquefaction, and induced landslides, generated by earthquakes that have potential to cause harm. Seismic risk, on the other hand, describes the probability of experiencing a specified level of seismic hazard in a given time exposure. Seismic hazard occurs naturally and can be evaluated from instrumental, historical, and geological observations. Seismic risk depends not only on the hazard and exposure, but also on models (i.e., time-independent [Poisson] and time-dependent ones) used to describe the occurrence of earthquakes. High seismic hazard does not necessarily mean high seismic risk, and vice versa. In the central Unite States, large earthquakes (M7.0-8.0), similar in magnitude to the 1811-1812 New Madrid events, are of safety concern. Geologic evidences indicate that these large earthquakes have occurred several times in the past few thousand years with a recurrence interval of about 500 to 1,000 years. Consequently, the ground motions from these large earthquakes have the same recurrence interval (about 500 to 1,000 years). The risk posed by these large earthquakes is about 5 to 10 percent probability of exceedance in 50 years. Similarly, the risk posed by the ground motion generated by these earthquakes at a site is also about 5 to 10 percent in 50 years. The ground motion with 2,500-year return period (hazard) or ground motion with 2 percent probability of exceedance in 50 years (risk) has been generated from the same earthquakes and recommended for policy consideration in the central United States, however. This inconsistency is caused by the methodology being used for seismic hazard and risk assessment: probabilistic seismic hazard analysis (PSHA). Although PSHA is the most widely used method for seismic hazard and risk assessment, it contains a mathematical error in the formulation: incorrectly treating the ground-motion uncertainty as an independent random variable. The ground-motion uncertainty is an explicit or implicit dependent variable as it is modeled in the ground-motion attenuation relationship. The mathematical error results in difficulty in understanding and applying PSHA. PSHA mixes temporal measurement (occurrence of an earthquake and its consequence [ground motion] at a site) with spatial measurement (ground-motion variability due to the source, path, and site effects). Thus, use of PSHA in seismic hazard and risk assessment is problematic for policy consideration.
S31C-06
A new Kinematic Approach to Calculate Seismic Hazard Scenarios (Intensity and Peak Ground Displacement); an Example in SE Sicily
The part of the seismic hazard map of Italy (http:zonesismiche.mi.ingv.it/mappa_ps_apr04/italia.html) that
regards SE Sicily has recently been modified on the basis also of the new location of the source of one of the
strongest earthquakes that ever struck the Mediterranean basin (in 1693 in SE Sicily; 54,000 casualties; see
Sirovich and Pettenati, 2001 in BSSA). That source was located inland according to the KF geophysical inversion
of the regional damage patterns of the earthquake of Jan. 11, 1693 and of its destructive foreshock of Jan. 9 (see
Gentile et al., 2004 in BSSA, and Sirovich and Pettenati, 2004 in JGR, for the genetic KF inversion technique). The
damage patterns of Jan. 9 and 11, 1693 had been evaluated by three different groups of historians and
seismologists on three different intensity scales. (Detailed information was available from the reports of the
Officers of the "Regno delle Due Sicilie" of the time as well as from many other documents). Previously, both
earthquakes were traditionally ascribed to the well known Malta Escarpment, the most prominent physiographical
and structural feature of the area, which is found offshore, south-east of Sicily. However, given the inland damage
of 1693, an offshore epicenter would imply a mean radius of 45 km for the virtual area of degree XI (70 km for
degree X) and, thus, a magnitude of 8.3±0.2 with a fault length of approximately 280 km. This fault source
would cross the NE part of Sicily from the southern Jonian Sea to the Island of Lipari in the Tyrrhenian Sea; but
tectonically this seems unrealistic. All the inversion tests, however, pointed to a complex source inland which
could have hosted both earthquakes. Here, we used our kinematic KF model in the direct mode in a parametric,
deterministic-Montecarlo way to produce a seismic hazard scenario, in terms of maximum ground displacement.
This scenario will be compared with PSHA results for long recurrence times in the frame of a project of the Civil
Protection of Italy and the National Institute for Geophysics and Vulcanology (INGV). New tentative seismic rules
based on displacements are the goals of that project (S5, coordinated by E. Faccioli and A. Rovelli). In the present
paper, three linear sources were assumed, to accommodate all the area's guessed inland sources, those of
1693 included. The Montecarlo technique was applied to the eleven source parametres of the KF formula, with
the purpose of accounting for the large seismotectonic uncertainties we had. In so doing, 33,280 sources were
obtained, allowing us to calculate the mean plus-one-standard-deviation scenarios (in seismic intensity and in
horizontal displacement). Our mean scenario reaches 40-50 cm (80-90 cm summing one standard deviation),
which, as expected, is more than the maximum provisional PSHA values for the 475-year return period (S5
Project, Faccioli 2007); consider that the return period of an earthquake like that in 1693 is unknown, but it could
be of the order of thousands of years. The comparison with the damage in 1693 showed that our provisional
result is still not cautious enough, however. It is hypothesized that this is due to the asymmetric distribution of the
calculated values in each site also because of the presence of the XII intensity upper bound. In Acapulco, we plan
to show the parametric scenario obtained using the median values too.
http:zonesismiche.mi.ingv.it/mappa_ps_apr04/italia.html
S31C-07 INVITED
Public Education for Household Mitigation and Preparedness for Earthquakes in California: The Research Base and Program Innovations
This presentation summarizes the findings from previous research in the social sciences regarding the factors and processes that enhance the effectivenss of public education efforts for household mitigation and preparedness actions for earthquakes. The conclusions from this research base include that the most effective efforts are those that are designed as an ongoing process with multiple channels and types of public communications. Second, an anticipated survey to measure household mitigation and preparedness actions in the State of California is sumarized. This survey will measure actual household mitigation and preparedness actions taken, knowledge, perceived risk, and other factors that previous research suggests impact these actions and perceptions; each of these factors are reviewed. The presentation then illustrates how knowledge from previous research will be blended with the information obtained from the planned survey in order to desgin a state-of-the-art public education campaign in California that maximizes household mitigation and preparedness for earthquakes and mega-earthquakes. Among other things, this requires that government agencies, NGOs, and provate sector organizations cooperate to coordinate their efforts to maximize program effectivenss. Finally, how this program might be evaluated to inform program refinements over time is discussed.
S31C-08
An Overview Of Seismic Hazard Assessment Research Activities In California
California has undertaken a serious attempt to reduce loss of life and property by developing applied research products that are readily useful to engineers, planners, earth scientists and decision makers. The type and degree of complexity of the products vary according to application. This presentation will provide a overview of applied research activities including but limited to: an update on the Next Generation Attenuation (NGA) models overall program activities; potential for the application of NGA models outside of California: development update for the Uniform California Earthquake Rupture Forecast Model; a preview of the California Seismic Safety Commission's planned Earthquake Research Grant Program; and a update on the Tall Buildings Initiative with respect to potential contributions from the GeoSciences community. The presentation will conclude with a general discussion with a limited discussion on potential public policy issues related to the research.