Global Earthquake Forecast Testing Experiment on Multi-Resolution Quadtree Grids
Creators
- 1. GFZ German Research Centre for Geosciences
- 2. Southern California Earthquake Center
- 3. University of Bristol
Description
The RISE (Real-time earthquake rIsk reduction for a reSilient Europe, (www.rise-eu.org/) research group in collaboration with CSEP (Collaboratory for the Study of Earthquake Predictability, (https://cseptesting.org/) is conducting a global earthquake forecast experiments using the newly defined multi-resolution grid technique called Quadtree.
The participating forecast models include:
- WHEEL (Bayona et al., 2021)
- TEAM (Bayona et al., 2021)
- KJSS (Kagan and Jackson 2011)
- SHIFT GSRM2f (Kreemer et al., 2014)
- GEAR1 (Bird et al., 2015)
The forecasts for this experiment have been submitted by their respective authors but were aggregated from the initial high-resolution grid with 0.1° x 0.1° cells to various data-driven multi-resolution Quadtree grids to investigate the effects of these grids on the evaluation results. Quadtree grids are generated based on earthquake catalog data and strain data points, the same data which have been used during the creation of forecast models (before 2014). The forecast file naming, after aggregating on a particular Quadtree grid, is derived from the criteria used to generate the grid. For example, 'N' stands for number earthquakes, 'SN' stands for Number of earthquakes and Strain data points, and 'L' stands for maximum zoom-level allowed for the grid.
The forecasts (number of earthquakes) for each bin are defined by its Quadkey which represents a pre-defined latitude and longitude cell extension and location, and by its magnitude bin in 0.1 magnitude units, starting from magnitude M5.95, represented as:
Tile | depth_min | depth_max | 5.95 | 6.05 | 6.15 | ... |
'0010' | 0 | 70 | 0.00235 | 0.00183 | 0.00054 | ... |
In this repository, we provide Global Earthquake Forecast Experiment, as floating software (also available in GitLab repository, https://git.gfz-potsdam.de/csep-group/gefe-quadtree), thereby having all the codes and data registered under public version control. With the git version control facility, the experiment can keep on evolving over time in terms of new tests, data, and grids, however, the provided forecast models shall remain the same. Every interested user can download the full repository and run the experiment on their hardware. They are allowed to change the codes, add their models to the experiment and run evaluations on their own. Every user is invited to contribute to the repository and the CSEP project. The RISE Floating Community Experiments constitute the new CSEP approach to earthquake forecast testing experiments.
Files
gefe-quadtree_results_2023-03.zip
Files
(68.6 MB)
Name | Size | Download all |
---|---|---|
md5:3dbdd34d9222726043c0fd2ae031d92b
|
65.7 MB | Preview Download |
md5:94aca4174ac9f9a2c9079f3144078199
|
2.9 MB | Preview Download |
Additional details
Funding
References
- Bayona, J.A., Savran, W., Strader, A., Hainzl, S., Cotton, F. and Schorlemmer, D., 2021. Two global ensemble seismicity models obtained from the combination of interseismic strain measurements and earthquake-catalogue information. Geophysical Journal International, 224(3), pp.1945-1955.
- Kagan, Y.Y. and Jackson, D.D., 2011. Global earthquake forecasts. Geophysical Journal International, 184(2), pp.759-776
- Kreemer, C., Blewitt, G. and Klein, E.C., 2014. A geodetic plate motion and Global Strain Rate Model. Geochemistry, Geophysics, Geosystems, 15(10), pp.3849-3889.
- Bird, P., Jackson, D.D., Kagan, Y.Y., Kreemer, C. and Stein, R.S., 2015. GEAR1: A global earthquake activity rate model constructed from geodetic strain rates and smoothed seismicity. Bulletin of the Seismological Society of America, 105(5), pp.2538-2554