Shake maps have become a valuable tool for real-time hazard and risk management worldwide.
They are commonly used for rapid evaluation of potential damage following a major earthquake in order to inform emergency response, loss estimation and public information.
Shake maps provide an estimate of ground motion amplitudes and shaking intensity created by a seismic event across a region of interest. Shake maps demonstrate the effects of radiated seismic energy on the ground surface, and are complementary to event magnitude and location data. Shake maps are also used to predict the level of ground shaking that may be created by future events, in order to assess the structural integrity of critical facilities, and to prepare emergency action plans.
Shake maps are produced in near real-time. They include heat maps showing the spatial distribution of peak ground motions (PGA and PGV), response spectral amplitudes at various periods of engineering interest (e.g. 0.1, 0.3, 0.5 and 1.0 seconds) and estimated Modified Mercalli Intensity (MMI).
Shake maps are developed using two key inputs:
- A ground motion prediction equation (GMPE) that models the regional source and attenuation effects, defining how ground motions scale with earthquake magnitude and distance, and
- A site amplification map that shows the spatial variation of the effects of local site conditions.
Ground Motion Prediction Equations (GMPEs)
A regional GMPE for the area of interest will be derived from existing high-quality data using the following workflow:
- We review phase arrival picks and locations of selected events, characterizing the events on a quality and geographical distribution scale.
- We determine moment magnitudes (Mw) and source mechanisms of events with sufficient signal-to-noise ratio (SNR) by moment tensor inversion. For relatively low-SNR events, we estimate Mw values using displacement spectral fitting approach and compute source mechanism using a template matching technique.
- We calculate source-to-site distances, peak ground motions and response spectral amplitudes in order to generate an empirical data set for GMPE inversion.
- We examine the decay of ground motion amplitudes with distance to understand regional attenuation attributes.
- We determine horizontal-to-vertical (H/V) spectral ratios for each recording station. We group stations with similar site characteristics based on H/V ratios and other available geological/geotechnical information available.
- We supplement the empirical ground motion data with simulations of large events in order to ensure accuracy of predictions at large magnitudes (M>6) where empirical data is sparse.
- We construct a GMPE as a function of magnitude, source mechanism, distance and local site condition. We then perform regression analysis to determine the GMPE coefficients, relating the model parameters and ground motion amplitudes.
- Finally, we validate the derived GMPE by means of residual analysis and statistical resampling methods.
Typically there are a small number of recordings obtained from large events, which introduces relatively poor empirical constraints on the GMPE for large magnitudes of engineering interest. This is a common problem among regional data sets in different parts of the world due to long return periods of large events.
In order to overcome this limitation, we can use stochastic simulations to develop a magnitude scaling function (FM) that defines how ground motions scale with magnitude over a wide magnitude range (up to Mw=8). The regional source and attenuation attributes (e.g., stress drop, geometrical spreading and anelastic attenuation) computed from available empirical data will be used in these simulations.
The derived FM function will then be incorporated into the regional GMPE in order to ensure accurate ground motion predictions for large events (Mw>6). The final GMPE will be tested against observed ground motions for potential sub-regional variations in source and attenuation effects for further improved accuracy.
Local site conditions can significantly impact the level of ground shaking experienced at a particular location. This means that the accuracy of shake maps is highly dependent on the reliability of estimated local site effects. Site effects at a given location can be predicted from local surface geology. These estimations can be further refined for mission critical locations via direct measurements taken during site characterization surveys.
Accurate assessment of local site effects at locations of warning recipients is a prerequisite for reliable estimation of ground shaking in EEW applications.
For instrumented locations, local site effects are determined empirically by decoupling site effects from source and attenuation effects via regression analysis.
For mission-critical locations with no instrumentation, site effects can be determined either by the empirical analysis of new ground motion data collected from temporarily deployed stations, or by the theoretical site response analysis of a sub-surface model constructed based on physical characteristics determined from geotechnical surveys.
We can work with local institutions to perform geotechnical site surveys such as borehole sampling, standard penetration test, multichannel analysis of surface waves, at mission-critical locations where high accuracy of ground motion prediction is required.