GEM Global Active Faults
The GEM Foundation's Global Active Faults project (GEM-GAF) is building a comprehensive, global dataset of active fault traces of seismogenic concern. The dataset comprises GIS files hosted here of fault traces and small amount of relevant attributes or metadata (fault geometry, kinematics, slip rate, etc.) useful for seismic hazard modeling and other tectonic applications. The dataset is being assembled primarily as a part of GEM's global Probabilistic Seismic Hazard Modeling efforts, although we hope that the data find wide use in research, education and general interest among many users.
The dataset is freely and publicly available here, under a Creative Commons attribution license.
The dataset currently covers most of the deforming continental regions on earth with the exceptions of, the Malay Archipelago, Madagascar, Canada, and a few other regions. These are to be added progressively through 2018.
The data is viewable in an interactive map here.
|downthrown_side_id||string||direction of downthrown side||NE|
|average_rake||tuple||Slip rake of fault||(45,25,55)|
|strike_slip_rate||tuple||Strike slip rate on fault||(1.5,0.5,2.5)|
|dip_slip_rate||tuple||Dip slip rate||(1.5,0.5,2.5)|
|vert_slip_rate||tuple||Vertial slip rate||(1.5,0.5,2.5)|
|shortening_rate||tuple||Horizontal shortening rate||(1.5,0.5,2.5)|
|accuracy||integer||Denominator of map scale||40000|
|activity_confidence||integer||Certainty of neotectonic activity||1|
|exposure_quality||integer||How well exposed (visible) fault is||2|
|epistemic_quality||integer||Certainty that fault exists here||1|
|last_movement||string||Date of last earthquake||1865|
|name||string||Name of fault zone||Polochic|
|fz_name||string||Name of fault zone||Motagua-Polochic|
|reference||string||Paper used||Rogers and Mann, 2007|
|notes||string||Any relevant info||May be creeping|
|ogc_fid||integer||ID used by GIS||8|
There are three main data types used in the GAF attribute table,
tuple is a 3-tuple of real (floating-point or integer) numbers
representing continuous random variables such as slip rate. The tuple has the
(most-likely, min, max). In some instances where there is no estimated
uncertainty in the parameter of interest, the tuple may be simply given as
(most-likely,,); this is most common for the dip of purely strike-slip
faults. In typed databases it is actually represented by a string, so the
parentheses and commas will be preserved.
Rake is in Aki-Richards
convention. All slip rate fields except
shortening_rate describe the slip
rate or component on the actual fault, and are in magnitudes, i.e. are always
shortening_rate describes the horizontal contraction rate (heave)
of a fault (such as a GPS measurement); this is not the dip slip rate.
Extension across a fault is negative.
Integer is used as a categorical variable in this database, typically to
denote the relative epistemic uncertainty in a parameter.
1 is most certain,
2 is moderately uncertain, and
3 is highly uncertain. The other uses of
Integer types are for table indices in many constituent datasets, and in the
accuracy attribute which denotes the denominator of the map scale during
fault mapping and digitization; for example, a fault that was mapped in GIS at
a 1/40,000 (or 1:40,000) scale will have an accuracy of
Strings for fields with words.
The database is currently available in 4 formats, GeoJSON, GeoPackage, KML,
and ESRI ShapeFile. Which file format is most appropriate depends on the
software package that is being used. QGIS users and anyone making webmaps or an
API will find the GeoJSON format most useful. This is also the version of
record as it is tracked best with version control. However, ESRI does not seem
to provide acceptable GeoJSON support; ArcGIS users should be able to use the
GeoPackage format (note that we have no access to ArcGIS and are unable to test
these files for compatibility). ESRI's legacy ShapeFile format is also provided
but this is not a good choice as that format truncates both column names and
longer text fields such as
If you are downloading individual files instead of cloning the repository: Please note that GitHub may display a version of the data that is formatted for viewing within its built-in webmap, and not display the actual data. If you press the 'Download' button, however, you may download the real file.
Additional file formats will be provided once the Version 1 of the database is complete; if you have specific needs, please contact richard dot styron at globalquakemodel.org.
Contributions to the GEM GAF-DB are highly encouraged. Please see the page on [contributing] for more details.