EPISODES Platform Documentation

Application User Guides > Hazard and Risk Analysis

Seismic hazard forecasting user guide ​

TIP

This application enables seismic hazard modelling and forecasting based on an earthquake catalog. Users can model the Magnitude Distribution, Epicenter Location Distribution, and Seismic Activity Rate. Using these models — whether generated within the application or imported from external sources — the application computes probabilistic seismic hazard estimates of ground shaking intensity.
To obtain more general information about working with applications within the Platform, see Applications Quick Start Guide.

NOTE

CATEGORY Hazard and Risk Analysis

KEYWORDS Hazard and Risk Analysis, Ground Motion parameters, Hazard analysis, Induced seismicity, Magnitude, Probabilistic seismic hazard analysis, Statistical analysis, Statistical properties of seismicity, Statistical seismicity

CITATION If you use the results or visualizations retrieved from this application in a publication, then you must cite the data source as follows:

Orlecka-Sikora, B., Lasocki, S., Kocot, J. et al. (2020) An open data infrastructure for the study of anthropogenic hazards linked to georesource exploitation., Sci Data 7, 89, doi: 10.1038/s41597-020-0429-3.

Introduction ​

The application consists of four parts:

  1. Magnitude Distribution Modeling
  2. Epicenter Location Distribution Modeling
  3. Seismic Activity Rate Estimation
  4. Seismic Hazard Forecasting

The first three (1., 2., and 3.) can be run independently. The outputs of these are used as inputs into the last part (4. Seismic Hazard Forecasting).

Standard Workflow

Run parts 1., 2, and 3. separately and check their results. Adjust their settings as necessary to obtain satisfactory results. It is best to do this before activating part 4 because the Seismic Hazard Forecasting part of the application is very computationally time consuming compared to the others. Once parts 1 to 3 are optimized, run the application again with part 4 enabled.

Alternative workflows are also possible:

Alternative Workflow A

Instead of using 1. (Magnitude Distribution Modeling) as an input into 4. (Seismic Hazard Forecasting), it is possible to import the output files containing the results from the app Source size distribution functions/Stationary Hazard.

Alternative Workflow B

Instead of using 3. (Seismic Activity Rate Estimation) as an input into 4. (Seismic Hazard Forecasting), it is possible to manually specify a value for activity rate as events per time unit.

Input file specification

The application requires:

Other files are optional:

If using Alternative Workflow A, then provide the Magnitude file (m.mat) and PDF file (PDF.mat) files from the Source size distribution functions/Stationary Hazard application.

Filling form values ​

Each of the four parts of the app has settings that affect only those parts. There are also global settings used by all apps.

For each app there is checkbox controlling whether or not that part is active.

1. Magnitude Distribution Modelling

Estimate the probability density function and cumulative distribution function of magnitude using non-parametric kernel density estimation.

To activate this, use the following checkbox.

The settings for this module then become visible.

Magnitude column

  • a dropdown menu allows the user to select for the magnitude columns found in the catalog

Magnitude of completeness

  • the magnitude of completeness (Mc of the catalog. If the Mc was determined using the Completeness Magnitude Estimation application, then it’s possible to load it from the output file of that application.

Maximum magnitude

  • the maximum magnitude to consider in the distribution.

KDE method

  • the kernel density estimation method to use
  • recommended to first try AdaptiveKDE or DiffKDE

The PDF is estimated using the Magnitude of completeness and Maximum magnitude as endpoints.

2. Epicenter Location Distribution Modeling ​

Estimate the probability density function of earthquake epicenter locations using non-parametric kernel density estimation.

To activate this, use the following checkbox.

The settings for this module then become visible.

Grid cell size

  • The length (in metres) of the square cells of the 2D grid upon earthquake epicenter probability distribution function is estimated.
  • A smaller grid cell size will result in a higher resolution output in part 4 (Seismic Hazard Forecasting) at the cost of computation time.

Use time weighting function

  • When enabled, applies an exponential time weighting function to give more weight to recent events when estimating the location probability distribution function.

3. Seismic Activity Rate Estimation ​

To assist users in evaluating their time window selection, a Bayesian changepoint detection algorithm (Zhao et al., 2019) is applied to the smoothed activity rate time series. This algorithm identifies intervals that can be grouped under a common activity rate, providing visual feedback for further interpretation.

To activate this, use the following checkbox.

The settings for this module then become visible.

Time unit

  • The time unit to be used in parts 3 (Seismic Activity Rate Estimation) and 4 (Seismic Hazard Forecasting).
  • Available units are hours, days, weeks, and years.

Time window duration

  • The duration (in time units) of the window used in the Bayesian changepoint detection algorithm
  1. Seismic Hazard Forecasting

Forecast the exceedance probability of a ground motion product for a given forecast length and ground motion model. This operation is computationally extensive may take upwards of an hour depending mainly on the grid cell size selected in part 2 (Epicenter Location Distribution Modeling).

To activate this, use the following checkbox.

The settings for this module then become visible.

Forecast length

  • The point in time in the future for which to calculate the 95% exceedance probability
  • Calculated in time units past the last data point in the catalog

Time unit

  • The time unit to be used in parts 3 (Seismic Activity Rate Estimation) and 4 (Seismic Hazard Forecasting).
  • Available units are hours, days, weeks, and years.

Ground motion model

  • The model used to estimate actual ground motion values of acceleration and/or velocity
  • Each ground motion model specifies specific products that can be forecasted
  • Available ground motion models at this time are:
    • Lasocki 2013
      • developed specifically for the LGCD (Legnica-Glogow Copper District) episode
    • Atkinson 2015
      • from the OpenQuake library
      • Implements the Induced Seismicity GMPE of Atkinson (2015) Atkinson, G. A. (2015) Ground-Motion Prediction Equation for Small-to- Moderate Events at Short Hypocentral Distances, with Application to Induced-Seismicity Hazards. Bulletin of the Seismological Society of America. 105(2).
    • Convertito et al 2012
      • from the OpenQuake library
      • Implements the PGA GMPE for Induced Seismicity in the Geysers Geothermal field, published in Convertito, V., Maercklin, N., Sharma, N., and Zollo, A. (2012) From Induced Seismicity to Direct Time-Dependent Seismic Hazard. Bulletin of the Seismological Society of America, 102(6), 2563 - 2573
    • Akkar and Bommer 2010
      • from the OpenQuake library
      • Implements GMPE developed by Sinan Akkar and Julian J. Bommer and published as “Empirical Equations for the Prediction of PGA, PGV, and Spectral Accelerations in Europe, the Mediterranean Region, and the Middle East”, Seismological Research Letters, 81(2), 195-206.

Ground motion products

  • Typical products include PGA (peak ground acceleration, g), SA (spectral acceleration, g) at different frequencies (Hz), and PGV (peak ground velocity, cm/s)

Produced output ​

A grid of each selected ground motion product is displayed on top of a Google Maps basemap.

The units for acceleration products (PGA and SA) products are in terms of g, where 1 g is 9.81m/s2 . The units for velocity products (PGV) are cm/s.

Normally, the output grid displayed is upscaled by a factor of 4 and the edges trimmed. However, this is not done if the output grid consists of less than 10 points or the range of values is less than 0.1