Impact of topography on amplification of seismic response at regional scale focusing on 2005 Kashmir Earthquake Muhammad Shafique Faculty of Geo-Information Science and Earth National Centre of Excellence in Geology, Observation, University of Twente University of Peshawar, Enschede, The Netherlands Peshawar, Pakistan
Earthquakes Natural Hazards Floods Landslides Tsunamis Hurricanes Typhoons
Earthquakes in action
Indian plate movement in time Source: USGS
Topography and active tectonics in Pakistan 54 +/- 5 mm/yr
Seismicity in Pakistan and Surroundings since 1901
Location of past earthquakes and seismic gaps around Indian plate Source: Aydan, 2006
Earthquake secondary hazards Liquefaction Floods Earthquakes Ground Shaking Landslides Tsunami Fire
Earthquake ground shaking Topographic effects Regolith thickness effects Geologic effects SITE EFFECTS Surface Bed rock . . . . . MEDIUM EFFECTS . . . . . . Hypocenter SOURCE EFFECTS
I MPACT OF T OPOGRAPHY ON S EISMIC R ESPONSE INCIDENT SEISMIC WAVES
Date: 8 th October 2005 Kashmir earthquake Magnitude: 7.6 Depth: 10 Km Epicenter: 34.43ºN, 73.53ºE Death Toll: 80,000 Economic Loss: 5 billion US$ Pakistan
73°0'0"E 74°0'0"E Location Map Upper Dir of Study Area Diamir Kohistan ± Ghizer 35°0'0"N Shangla Batagram Neelum Mansehra _ ^ Buner Mardan Muzaffarabad 60°E 70°E 80°E Pakistan Abbottabad Swabi 34°0'0"N Haripur Bagh Attock Rawalpindi 30°N Poonch 73°0'0"E 74°0'0"E Kilometers 0 10 20 40 60 80 60°E 70°E 80°E
Framework of the study Topographic Parameters Seismic Modeling • Slope Numerical models • Height • Aspect • Curvature Seismic Parameters • Wavelength • Damping • Frequency • Angle of incidence Topographic seismic response at regional scale
Framework of the study Topographic Parameters Seismic Modeling • Slope Numerical models • Height • Aspect • Curvature Seismic Parameters • Wavelength • Damping • Frequency • Angle of incidence Topographic seismic response at regional scale
Data utilized • Shuttle Radar Topography Mission (SRTM) DEM (90m) • Advanced Spaceborne Thermal Emission and Reflection Radiometer ( ASTER) DEM (30m) • Instrumental ground shaking records • Drainage network and landslides map
Methodology SRTM & ASTER DEMs Literature on topographic seismic modeling Monte Carlo Simulations Slope Angle DEM disparity Height from local base Impact of resolution level Numerical models Aspect Kashmir earthquake induced seismic waves wavelength Horizontal and vertical TAF Sensitivity analysis TAF from ASTER & SRTM Topographical Seismic Uncertainty in TAF Zonation
Digital Terrain Representation Vector Representation Raster Representation DEM Point Contour TIN DTM DSM
DEM Uncertainties Uncertainities with Satellite Remote Sensing DEM Specific General Errors Errors Absolute Relative Absolute Relative Blunders Systematic Random Edge Slope Errors Horizontal Vertical Vertical Horizontal Errors Errors Matching Accuracy Accuracy Accuracy Accuracy Errors Areas of Constant Error void areas suspect areas water bodies
Disparity of SRTM and ASTER DEMs Correlation ( SRTM-ASTER) Standard Minimum Mean Maximum Deviation -1531 32.80 71.11 1621
Impact of DEM resolution on aspect computation Impact of DEM resolution on slope computation 90 m 8.04 2.15 0.61 3.39 6.35 90 m E E E N.W S 30 m 15.10 2.18 0.43 1.65 8.04 30 m N.E E N.W W S 15 m 20.24 10.55 0.62 1.74 7.04 15 m N.E E W N.W S 5 m 24.23 10.10 0.74 0.86 9.20 5 m N.E E N.W S N.W 1 m 31.30 10.20 1.22 0.86 10.82 1 m N.E E N.W S.W S
Impact of DEM uncertainty on topographic attributes using Monte Carlo Simulations Step1 Step 2 Step 3 Original DEM Random Filtered random numbers field error field + Filtered random error field Step 4 Aspect Slope Subtract Subtract Step 5 Original Aspect Slope DEM Residual aspect Residual slope Statistical summary Statistical summary
Step 1 Step 2 Step 3 Step 4 Step 5 Atb from RMSE original DEM Random Error Filtered L.P Filter Minus Disparity numbers Map E.M DEM Realisa- Error + Terrain Plus tion DEM Attributes INTERNATIONAL INSTITUTE FOR GEO-INFORMATION SCIENCE AND EARTH OBSERVATION
Uncertainties in Slope computation ASTER Min Mean Stand Dev Max Slope 0.005 2.853 0.595 6.001 (degrees) SRTM Min Mean Stand Dev Max Slope 0.03 0.93 0.17 1.67 (degrees)
Uncertainties in Aspect computation ASTER Min Mean Stand Dev Max Aspect 0 41.41 45.94 179.68 (Degrees) SRTM Min Mean Stand Dev Max Aspect 0.2 22.69 38.68 179.61 (Degrees)
TAF vs crest/base ratio comparison amplification
Models for horizontal and vertical TAF prediction Horizontal TAF Vertical TAF 0 . 4 2 6 0 . 8 H I 2 I H 0 . 225 0 . 5 5 0 . 75 I 1 . 5 I 3 I 0 . 02 A h 1 A , max v , max 1 0 . 9 0 . 5 1 0 . 15 Where H = Height from base level I = Slope angle = Wavelength ξ = Material damping Ah,,max = Horizontal TAF Av,max = Vertical TAF Source: Bouckovalas et al. (2006)
Seismic waves behavior in the assumed scenario Free Field Crest # " Crest Toe H # H " " Free Field Toe # " Incident SV seismic waves Homogenous soil and geological environment
Local minima for terrain height estimation Interpolate local Minima Base level DEM – Base level Height from local minima
Instrumental Ground Shaking Records Decay of Wavelength with Distance Epicenter Average Decay of Wavelength Per meter = 0.055m
Wavelength Estimation Abbotabad accelerograph Time of peak acceleration 0.8 s Sv velocity 3.2 km/s Dominant wavelength (using eq. 4.3) 0.8 x 3200 = 2560m Distance from epicenter 48.028 km Wavelength decay per meter 0.053302 Muree accelerograph Time of peak acceleration 1.2 s Sv velocity 3.2 km/s Dominant wavelength (using eq. 4.3) 1.2 x 3200 = 3840m Distance from epicenter 69.218 km Wavelength decay per meter 0.055477 Nilore accelerograph Time of peak acceleration 1.8 s Sv velocity 3.2 km/s Dominant wavelength (using eq. 4.3) 1.8 x 3200 = 5760m Distance from epicenter 99.519 km Wavelength decay per meter 0.057878 Average wavelength damping per meter 0.053302 + 0.055477 + 0.057878 = 0.166658 / 3 = 0.055553
Spatial distribution of wavelength
Horizontal aggravation of topographic seismic response Epicenter Epicenter ASTER Horizontal TAF SRTM Horizontal TAF
Sensitivity analysis Predicted results Sensitivity analysis Model Parameter TAF Base Slope Height Wavel Damping model 60º 400m 200m 10% Slope 30º A h ,max 1.10 1.17 1.14 1.08 1.06 Height 200 m Wavel 100 m A v ,max 0.76 2.85 1.33 0.44 0.69 Damp 5% Wavel = Wavelength A h,max = Horizontal TAF A v,max = Vertical TAF Slope angle is predicted to be sensitive terrain feature to topographic seismic response
Impact of DEM Inherent errors on TAF prediction ASTER Slope Slope Slope TAF Uncertainty (30º) (32.853º) (27.147º) A h,max 1.104 1.103 1.105 0.002 A v,max 0.765 0.831 0.708 0.123 SRTM Slope Slope Slope TAF Uncertainty (30º) (30.935º) (29.065º) A h,max 1.104 1.104 1.104 0.001 A v,max 0.765 0.786 0.746 0.040
Impact of DEM resolution on TAF prediction SRTM DEM ASTER DEM Disparity Min Max Min Max Min Mean Max A h,max 1 1.43 1 1.66 -0.260 0.003 0.227 A v,max 0 12.04 0 23.71 -14.992 0.030 6.521
Impact of direction of incident waves on seismic induced ground failures during Kashmir earthquake Direction Area (m 2 ) % of total Area 12800 Flat 0.09 576000 North 3.94 435200 2.98 Northeast 544000 3.72 East 1305600 8.94 Southeast 3462400 23.71 South 4409600 30.19 Southwest 3859200 26.42 West Total 14604800 100.00 80.32% of seismic induced Ground failures map landslides were facing away from courtesy of , epicenter HIC-Pakistan
Conclusions SRTM is more consistent in slope and aspect computation than ASTER, but not necessarily more accurate! Slope geometry is sensitive topographic parameter to the topographic seismic response DEM inherent errors and resolution have less impact on predicted TAF Significant impact of direction of incident seismic waves on ground amplification and seismic induced landslides
Recommendations for follow up Tracing of seismic waves in 3D topographic model Predict impact of direction of incident seismic waves on TAF prediction Predict role of soil type and depth on seismic response Predict impact of local lithology on seismic response Near real time seismic ground shaking maps
THA THANKS NKS
Backup slides
Recommend
More recommend
