Number of Basic Activities :

• Seismic Monitoring of Georgia
• Seismic Hazard Assessment of Georgia
• Case Study of Tbilisi earthquake of 25.04.2002
• Detailed Site Studies, Local effects and Risk Assessment for the Selected Sites in Tbilisi (Saburtalo District)
• Strong Motion Network of Enguri Arc Dam
• Seismic Hazard and Risk Assessment for the Enguri High Dam

Seismic Monitoring of Georgia

During the past two years construction of Georgian Seismic Network has started with the help of number of International organizations, Projects and Private companies. The works conducted involved scientific as well as organizational activities: Resolving technical problems concerning communication and data transmission. Due to complicated topography and non-operational telephone lines and power, Dial Up connection, leased telephone lines, or microwave transmission is used depending on local conditions. Providing Power supplies at remote locations using Solar batteries and/or Uninterruptible power supplies, with high capacity batteries. Special server room with two PC servers, internet connection and uninterruptible plover supply is operating. At the present period data from 4 seismic stations (Mtatsminda, Delisi, Gareji and Gori) is transmitted on online regime, processed and archived.  All the stations are already integrated into the database.

Modern, Digital Seismic Network of Georgia

Special attention is drawn to construction of Tbilisi local network. The recent earthquake of 2002 has highlighted the poor seismic resistance of buildings in Tbilisi. Detailed seismic risk assessment must now be done taking into account present conditions of buildings. Unfortunately we should state that after this earthquake Tbilisi has become much more vulnerable. The moderate earthquake located in Tbilisi requires detailed geophysical studies of adjacent area and continuous observations of local seismicity. At the moment two digital stations are operating in Tbilisi and in nearest future it is planned to increase the number up to five.

Map of Tbilisi Local Network

Modern software systems installed for data acquisition, interactive seismogram analysis and continuous data archiving. Catalogue of events recorded is started, already including up to 600 events for past two years and associated up to 2000 waveforms.

Now a minimum configuration for local earthquake survey is completed. These recent achievements allow Georgian seismologists to carry out following tasks:

·         Observation of earthquakes in Georgia necessary for an estimation of earthquake hazard. Especially should be mentioned two extremely weak earthquakes in Feb. 05, 2004 and Apr. 10, 2005 occurred in the same location where the destructive Earthquake of Apr. 25, 2002 has occurred, and providing us with invaluable information about the seismic source zone within Tbilisi. Monitoring of such a weak event became possible only using modern digital equipment.   

·         Implementation of an online earthquake alarm system for reliable short term information of official institutions. This system involves calculation of the epicenter for strong earthquake, assessment and mapping of the distribution of possible damage, within the number of minutes after the event occurred. Main feature in this system is online data availability, four online stations gives a sufficient coverage for Tbilisi area, though successful realization of this task for the whole territory of Georgia requires additional online stations.

·         It opens good perspectives for creation of regional earthquake analysis system planned in cooperation with Armenia and Azerbaijan, some activities in this direction are already started. Monitoring of Seismicity and assessment of seismic hazard in a regional scale is very important, because Caucasus represents a natural entity and separation by the artificial state borders its Seismicity field leads to loss of information and improper results.  

·         Last but not least contribution to global seismological network has to be also mentioned. During the past two years Georgia has started cooperation with number of international seismic data centers, such as ISC, CTBTO, USGS, EMSC.

Epicenters of earthquakes that happened in Tbilisi and near Tbilisi in 2004-2005 years

Seismic Hazard Assessment of Georgia

Caucasus is one of the most seismically active regions in Alpine-Himalayan collision belt. The main seismotectonic feature is the junction between the Arabian and Eurasian plates. The analysis of the historical and instrumental seismological shows, that this is the region of moderate seismicity. The strong earthquakes with magnitude up to 7 and macrosiesmic intensity 9 (MSK scale) occurred here. Caucasus is characterized with so-called moderate seismisity. Reoccurrence period for strong events is of order of 1000 years. In this case possibility to stretch catalogue of strong events (instrumental records) down to the beginning of 20^th century is very important for seismic study of the region

Seismicity map of the Caucasus

Documentary historical catalogue is available dated back to the beginning of the Christian era. The information about the earthquakes in this period has been extracted from ancient Georgian and Armenian annals, as well as from other sources. The parameters of these earthquakes were determined on the basis of the macroseismic data analysis, from contemporary documentary description of damage caused by earthquakes.

On the basis of multidisciplinary data (seismological, geological, geophysical etc) a set of probabilistic seismic hazard maps for Georgia has been compiled. The methodology used in most probabilistic seismic hazard analysis was first defined by Cornell: it consists of four steps: 1. Definition of earthquake source zones, 2. Seismicity (definition of recurrence characteristics) of source zones. 3. Estimation of earthquake effect at site. 4. Determination of hazard at site.

The seismic hazard maps defined as the probability of exceeding a fixed value of shaking during different times of exposure has been constructed. Computer code SEISRISK III after Bender and Perkins (1987), was used for calculations The set of maps for peak ground acceleration (PGA for a 50 year exposure time and 1%, 2%, 5% probability of exceeding has been constructed.

Maps for peak ground acceleration

Another task to be paid attention is spectral characteristics of ground motion. This factor  is very important for real assessment of seismic hazard of cities. The preliminary investigation  of spectral acceleration  (SA) attenuation for 5% damping has been carried out for Caucasus Smit et al. Calculated Maps shows that maximal acceleration is connected with frequency ~ 3Hz (0.3 sec). The probabilistic seismic hazard maps for Spectral  Accelerations with different frequencies has been constructed for the territory of  Georgia. Resulted SA hazard maps of 2% probability of exceedance during 50 years, for different frequencies (1, 3, 5 Hz) are shown.     

Maps for peak ground acceleration

Maps for Spectral acceleration for 1, 3, 5 Hz Natural frequencies

Case Studies: Tbilisi earthquake of 25.04.2002

On April 25, 2002 at 22:41 local time (17:41 Greenwich Time) the strongest earthquake (at least since 1900) rocked Tbilisi. The magnitude of event was M=4.5, the epicentre was located in the central part of the city. The earthquake was very shallow ~ 4.5- 5 km depth. Despite the moderate magnitude, the earthquake had quite strong effect. The maximum horizontal acceleration was 0.11g, recorded in Tbilisi at distance of 7 kilometres from epicentre in Tbilisi seismic station, on the rock site. Earth shaking was also recorded at the 4th flour of The Institute of Geophysics at a distance of about 2.5 km from epicentre, where maximum horizontal acceleration has shown 0.22 g. The earthquake was preceded by small sequence of foreshocks M=3.5 on April 11th, two events with M=2.0 on April 21st and followed by hundreds of aftershocks, with magnitudes less than 2.5, several of them with intensity 3.

The analysis of strong motion record of the main shock shows, that the effect of the earthquake was unexpectedly high. According to the attenuation model the pha in the epicenter should exceeded 0.2 g. Another peculiarity of the earthquake was high frequency – SA of the record shows the maximum at a frequency of 5 Hz.

Macroseismic intensity was studied by means of questionnaire used for intensity estimation in former SU, distributed via Internet. For compilation of intensity map the method described in USGS, 2000 was used.  This was the preliminary estimation. For the full-scale macroseismic analysis building stoke data and local geological conditions were investigated. Significant part of residential 2-3 storey brick and stone buildings constructed without antiseismic measures were damaged. They correspond to class B (EMS-98, MSK-64). Due to high frequency of earthquake this group of buildings received the most important damages.

Due to high vulnerability of buildings the earthquake caused severe damages and human losses. More then 10,000 buildings were damaged. Direct loss exceeded 160 million US Dollars. Results on Building Stoke damage obtained by the Ministry of Urbanization and Construction were integrated in to the GIS database, showing the moderate and high damage rates in epicen-tral districts of Nadzaladevi and Chughuireti. These data is extremely important for future seismic risk assessment tasks, as well as for proper planning of rehabilitation and reconstruction measures. 

Map of Damaged buldings from Nadzaladevi District and summary table of total damage in Tbilisi

Existence of seismic source within the city, study its nature and physical parameters, together with peculiarities of seismic waves distribution, forms the  most important tasks for detailed analysis of seismic hazards in Tbilisi. . In case of 25.04.2002 Tbilisi earthquake, digital strong motion records were made at Delisi seismic observatory and the same kind of record at the Institute of Geophysics (on fourth floor) were analyzed, by means of Wavelet transform and Polarization analysis. In spite of leek of data, it let us make hypotheses about complex source mechanism of the event and local geological structures. Based on spatial calculations made in ArvView (GIS software), possible 3D model of earthquakes source was constructed.

3D source model of Tbilisi Earthquake of April 25, 2002

Detailed Site Studies, Local effects and Risk Assessment for the Selected Sites in Tbilisi (Saburtalo District)

Selected site in Tbilisi, namely the Saburtalo district was studied. Based on Engeneering-geological maps and soil borehole data 3D local Soil structure under selected area of TBILISI was compiled.

According to aerial photography mapping (done in 2003 year), in GIS was created detailed map of the selected Urban District of the Tbilisi. Including: Houses (each  house number was indicated), Linear street scheme (within addresses), So called red zones, Define territory on the social and private  sectors, Green zones, Lifeline system network

GIS model of the building stock in the selected district

Sub-site in the selected area was studied more precisely. Local soil structure and P wave velocity model was investigated using the borehole samples and seismic prospecting survey. Providing us with very exact estimations for local site effects caused by seismic waves.

Seismic wave recordings made at site and the Seistronix 12 Channel, 24 bit equipment used

P wave velocity model derived for the selected profiles using seismic prospecting data

Strong Motion Network of Enguri Arc Dam

During the 10-17^th July, 2000 year at Enguri high dam strong motion network was installed, composd of 10 stations. Accelerometers were manufactured by GEOSIG. 5 stations were installed on crest of  the dam, others at the different levels (see attached schema), network was equipped by GPS connected to station No1, which acts as a master in local network and provides rest of the stations with the correct time. 

Accelerometers are equipped with the internal memory units of 2 MB capacity, based on row calculations this could keep the data gathered during 2-3 months. Each station could send and/or accept a network trigger, it also can trigger itself. For trigger two different parameters could be used: threshold value and STA/LTA ratio, in our case we use threshold value, because it could trigger recording system not only in case of earthquake but also for any kind of strong vibration. In both cases reaction of dam's construction on vibration is interesting and informative. Main goal of installation of Strong Motion Network is study of reaction of dam's body on vibration, this could help us to understand dam's reaction on earthquake of certain magnitude and on certain distance from dam's location.
Accelerometers are housed in water proofed boxes, protection boxes itself are covered by iron boxes for extra safety.
functioning.

Detailed Site Studies, Seismic Hazard and Risk Assessment for the Enguri High Dam

Probabilistic Seismic Hazard Analysis of Enguri Dam Area. In this study the probabilistic method of seismic hazard analysis has been used. The main steps of this method are well known and described in several studies (e. g.. Reiter 1990, Kramer 1996).

Maps of Peak Ground Accelerations and Spectral Accelerations for 1.5 Hz frequency, with different probabilities

Determination of hazard. The seismic hazard maps defined as the probability of exceeding a fixed value of shaking during different times of exposure has been constructed for the territory of Georgia. Seismic hazard was computed using the computer program SEISRISK III (Bender and Perkins 1987), using linear source setting. The final maps were done over 0.05^° ´0.05^° grid. The input file requires some control cards, the attenuation relationship in the form of table (ground motion parameter as a function of magnitude and distance) and separate blocks of description for each source zones. The source zone description consists of geometric characteristics of the source (only linear sources have been considered), maximum magnitudes and seismicity rates.

Seismicity rates are entered in terms of the number of earthquake occurrences at given magnitude levels, in a given number of years. Program doesn’t require rigid respect for Guttenberg-Richter formulation. However, this formulation was used for defining seismicity rates. This capability of SEISRISKIII of using individual seismicity rates was used to better fit the seismicity characteristics.

The set of maps for peak ground acceleration (PGA) Fig. 4 and spectral acceleration (SA) for 50 years exposure time and 1%, 2%, 5%, 10% probability of exceeding has been constructed. Set of maps of macroseismic intensity for Enguri region was analyzed in Chelidze at el 1998. The SA maps was calculated for frequencies 1.5 Hz (Fig. 5) and 2 Hz (fig. 6). These frequencies are close to the first modes of natural frequency of the construction and therefore they are most hazardous for the dam. The comparison of maps shows, that main hazard is connected with Gagra-Java fault zone. The effect of Ingirishi fault isn't seen on the maps. It means, that long term, high magnitude seismicity of the Gagra-Java fault is prevailing over the frequent, but moderate seismicity of local faults.

Estimation of natural frequency of Enguri arc dam

The knowledge of dynamical properties of high dams, constructed in seismically active regions, is crucial for the prediction of these structures behavior during the strong earthquakes or other geomorphological hazards. Even the rough estimation of natural frequency is very important for hazard analysis. This task can be carried out using sensitive seismographs. Installing sensor on the crest of the dam, we can estimate quite precisely the main mode (and in some cases fist modes as well) of natural frequency. In should be noted that in this case the dynamic characteristics that prevail at very low vibration levels will be investigated.

Seismic noise records analyzed. Time series, Fourier spectra and PSD are shown

The structure subjected to a near white-noise ambient excitation responds primarily in the vicinity of its resonant frequencies. These frequencies can be identified from the peaks in power spectral densities (PDW)  computed from the individual time histories.

The seismograph was installed in several places near the center of the crest. One of the horizontal components was oriented perpendicularly to crest.

Several records with duration up to 30 minutes has been done on the crest of the dam. The records was analyzed, amplitude and power spectra has been calculated using software package PITSA (fig.1) and MATLAB (fig. 2). Analysis of records show, that two mean frequencies can be identified ~1.6-1.7 and ~2.1-2.2. This result is in good agreement with theoretical studies (Losaberidze 1979) and is close to results, obtained on other dams with similar characteristics (De Smet at el 1998). That means that for relatively high frequencies the dam response is elastic, unlike its response to slow strains.

The results we have used in seismic hazard analysis.

reation of 3D Multifunctional Model of Enguri High Dam and Surrounding Territories Based on GIS and Remote Sensing Technologies. Using the modern technologies of GIS (Geographic Information Systems) and Remote Sensing (RS) 3D model of Enguri High Dam and Surrounding Territories was constructed. Model consists of: digital elevation map of Scale 1 : 50,000 and corresponding 3D surface model, water network (Rivers and Lakes), settlements, new map of anthropogenic landscape of territory down the stream. Work is done using topographic maps and Space Images (namely: NIMA  (C) CNES/SPOT Image 1992-1994 –10m resolution, LandSat TM – 30m resolution, ASTER 15m resolution). Another work concerning calculation of water stream and its parameters in case of dam failure is under progress. The main purpose of work already done is creation of catastrophe scenario and estimation of possible losses, based on calculation of water stream in 3D and modern database already compiled.