Micro Seismic Hazard Zonation Workshop/Training on Earthquake - - PowerPoint PPT Presentation

INTERNATIONAL INSTITUTE FOR GEO-INFORMATION SCIENCE AND EARTH OBSERVATION

Micro Seismic Hazard Zonation

Workshop/Training on Earthquake Vulnerability and Multi-Hazard Risk Assessment: Geospatial Tools for Rehabilitation and Reconstruction Efforts Siefko Slob

2

Approach to use GRA in GIS for seismic microzonation purposes

Seismic microzonation based on spatial variation

• f ground ground response

Ground response is mainly dependent on:

Soil model or overburden thickness – LATERAL variation Maximum shear strength of soil units (Gmax) or Seismic velocity of the soil units (Vs) – VERTICAL variation

Therefore, if the spatial variation of ground properties can be modeled, the spatial variation in ground response can also be calculated

3

How to model the spatial variation in ground properties?

Create a model of the subsurface

True 3D soil model Boundary layer model Overburden thickness model

Acquire and model soil properties

Lateral – may assume more or less homogeneous properties for same soil unit Vertical – Generally large difference between different units

4

What parameter do we use for the microzonation?

We want to spatially model the spatial hazard for different building types Therefore, model spatial variation of

Amplification (qualitative), or Spectral acceleration (quantitative)

For different frequency ranges (i.e. corresponding to natural frequencies of specific building categories)

5

Seismic microzonation: scope

Seismic hazard analysis first step towards earthquake risk reduction strategy in earthquake-prone areas Particularly for “smaller” cities in developing countries risk reduction is important Challenge: to develop simple method for microzonation on basis of limited information

6

How do we do microzonation?

Calculation of spatial distribution of seismic response

Identification of areas where amplification of seismic signal results in unacceptably high acceleration levels

Knowledge of subsurface characteristics essential. e.g.:

Overburden thickness Soil profile Material properties of soil layers

7

How can we do microzonation?

Standard approach:

Classify into areas with homogeneous subsurface conditions Calculate seismic response (with e.g. SHAKE) for typical soil profile

Disadvantage:

Large generalisation beforehand Discrepancies at borders of “homogeneous” areas

8

Improved method (1)

• 1. Create a continuous (3D) ground model in

GIS using any available surface and subsurface information:

• Boreholes, SPT’s
• Geophysical profiles (VES, refraction)
• (Engineering) Geological maps
• Any other a-priori geological knowledge of the

“model builder”

9

Improved method (2)

2. Calculate seismic response for every surface point on the continuous layer model (using SHAKE) 3. Visualise spatial variation of seismic response

• ver the entire modeled area using GIS

4. Classify seismic response analyses into areas with different hazard levels (e.g. exceeding design acceleration levels)

10

Spectral acceleration as classification criterion

Before, Peak Ground Acceleration (PGA) was considered as main criterion More important to know the acceleration or amplification for different frequencies

Energy-content of seismic signal is frequency- dependent.

This approach allows for:

Visualisation of spatial variation of the acceleration for different frequencies, i.e. corresponding to natural frequencies of typical buildings. In this way, hazard maps can be made specific for different building types, e.g. high-rise buildings v.s. low-rise buildings.

11

• Earthquake 28 Jan.

1999, M 6.1

• 1,100 persons killed,

4,800 injured

• 45,000 houses

damaged and destroyed

• Approach: boundary

layer model

Case study 1 – Armenia, Colombia

12

Modelling of the seismic response of this earthquake

• 1. Creation of a (semi-) 3D model of the

subsurface using:

• Geotechnical maps
• Borehole information
• Seismic surveys (refraction and resistivity)
• 2. Calculation of the seismic response of the

ground at regular grid intervals (15x15 m)

13

Step 1: Creation of subsurface boundary layer model in GIS

Import data

Overburden thickness map Borehole data Geophysical profiles (resistivity) Geological surface map Laboratory tests for soil properties

Interpolation of input data

14

(Semi-) 3D grondmodel

How can we create a 3D model in a 2D GIS?

By modelling the boundaries of the different geotechnical units as surfaces (done in Ilwis) I.a.w: interpolating the surfaces as DEM’s This only works well with relatively simple geology (no faults, folds, etc.)

15

Geotechnical units mapped at the surface

16

Isopach map of the top Ash layer

17

Representation of the surface of the top of the basement unit

18

Profile through the 3D model (1)

19

Profile through the 3D model (2)

20

Step 2: Calculation of seismic response using SHAKE

SHAKE is 1D ground response analysis: for level or gently sloping sites with parallel material boundaries GIS raster data of 4-layer model; for every pixel a ground profile can be obtained Ground profile exported and seismic response calculated using SHAKE Calculated results subsequently visualised in the GIS

21

ON Soil profile - A nalysis No. 1 - Profile No. 1 Layer No. 4 Fourier Amplitude Spectrum Frequency (Hz) 0.000 0.005 0.010 0.015 0.020 0.025 0.030 0.035 0.040 0.045 0.050 2 4 6 8

Ground response is calculated for every pixel

Fourier amplitude spectrum: Frequency- dependency of amplification

• N. Brasilia - Layer 1 - CCA

LA EW Layer No. 1 Fourier Amplitude Spectrum Frequency (Hz) 0.000 0.005 0.010 0.015 0.020 0.025 0.030 0.035 0.040 0.045 0.050 2 4 6 8

Frequency Amplitude

22

Seismic response analysis using the external software ‘Shake’

Input for the program

Stratigraphy (in this case a simple 4 layer model)

Thickness of every layer

For every unit geotechnical parameters (assumed constant):

Density Shear wave velocity

Earthquake signal:

Accelerogram that was recorded with an accelerograph in rock 5 km SE of Armenia

23

Accelerogram of the Quindío quake used as input signal

Accelerogram CCALA EW

• 600
• 400
• 200

200 400 600 5 10 15 20 25 30 35 40 Time (sec) Acceleration (cm/sec2)

24

Important is the possible amplification of the input signal due to resonance of the soil layers

Peak Acceleration (g) Depth (ft)

CCALA EW - Profile N. Brasilia

• 20
• 40
• 60
• 80
• 100

0.0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1.0

Ash Ash Residual soil Residual soil Saprolite Saprolite Base Base lahar lahar/pf /pf

25

• N. Brasilia - Layer 1 - CCALA EW

Amlification (resonance) is dependent

• n frequency

Fourier spectrum of the

• utput signal at the surface

Fourier spectrum of the input signal in the basement Largest resonance

• ccurs at 3Hz

26

Respons spectrum: accelerations for different frequencies

CCALA EW - Profile N. Brasilia Spectral Acceleration (g) Period (sec)

1 2 3 4 5 6 0.01 10 0.1 1

Acceleration of 2g(!) at 3Hz Natural frequencies

• f low-rise buildings

27

Seismic hazard zonation

Map the differences in accelaration for a particular frequency, characteristic to a the natural frequency of certain buildings

Type of object or structure Natural frequency (Hz) One-story buildings 10 3-4 story buildings 2 Tall buildings 0.5 – 1.0 High-rise buildings 0.17

28

Practical realisation

Special software developed as interface between the 3D ground model of Ilwis and the seismic response program ‘SHAKE’

For GRID cells of 15x15m a stratigraphic column is created on the basis of the 3D GIS model which is exported as an ASCII point table:

(X,Y,thick1,thick2,thick3)

The output of the SHAKE analysis (the spectral acceleration) is subsequently added as an extra attribute to the ASCII table:

(X,Y,thick1,thick2,thick3,Accel10Hz,Accel5Hz)

29

30

31

32

33

Three-dimensional Effects

Three-dimensional modeling of topographic effects – Brasilia area Using DIANA: a real 3D finite element method Limited number of elements: only a small model of the topography could be build Surface accelerations are up to 2 to 4 times the base level (input) accelerations

34

Surface acceleration as result of 3D modeling: red is high; blue is low acceleration

35

Surface acceleration as result of 3D modeling: maximum acceleration ≈ 3 m/s2 which is about 3 to 4 times the maximum in the base level signal

36

Case study 2 – San Jose, Costa Rica

San Jose valley in earthquake-prone area Early 1990’s several large earthquakes in Costa Rica Seismic hazard zonation carried out, traditional approach Test case for improved method Approach: overburden thickness model

37

Approach

Only available data:

Overburden thickness map Depth-Vs relationships

Assumption 1: overburden thickness is the most important variable in computation of seismic response Assumption 2: Model Vs increase with depth

38

39

40

3D view of acceleration values

41

Other examples

Tessaloniki, Greece

42