Over the past two years, I have been working for the Global Earthquake Model Foundation (GEM). My main task has been building the Global Active Faults Database (GEM GAF-DB), which is just what it sounds like—a database that contains all of the world's active faults. This work is nearing completion; I am doing a few small tweaks to the datasets that compose the GAF-DB, as well as the process that assembles the database from the underlying datasets, and we are waiting on permissions from some of our collaborators to publish the data with the open, permissive licensing that is important to us at GEM. These issues should be dealt with soon, and publications presenting the database are in progress.
In the mean time, here is an interactive webmap showing the most recent release of the data (which is available on GitHub here). You can zoom around, click on faults, etc.
The faults are colored by kinematics: blue faults are dextral, red faults are normal, purple faults are sinistral, black faults are reverse, and grey lines are contractional structures (i.e. folds) that are not mapped as faults. Green faults are of unknown kinematics.
The work itself has been quite enjoyable: a great blend of mapping (in GIS) and data management. I've made new active fault maps of Central America and the Caribbean, North Africa, Northeastern Asia, and modified existing mapping of South America and the Philippines (link coming soon, hopefully).
The data management stuff has been interesting as well, in a more limited sort of way (it's data management, after all). I've been given more or less complete latitude in deciding how project should be managed, what datasets should be included, how faults should be mapped, and how the data should be combined. It has been fun to assess the state of the data, make a plan, and then execute on that. I chose to keep the project as simple as possible, adding complexity only where needed, and to make as much of the project as possible automated, scripted, and therefore repeatable and reversible. This means that the many constituent datasets aren't modified unless absolutely necessary, so information that isn't used or should be used differently won't be lost; it means that newer versions of individual datasets may be used as they become available with as little effort as possible, and that datasets may be replaced with better datasets if those are made available. There is a final 'data harmonization' step where spatial conflicts between faults of different datasets are handled (for example, competing representations of a structure or of an area). This is also done algorithmically, based on a hierarchy of datasets, and is as mutable as anything else in the process. I've done a fair amount of programming in ways that are quite unlike the numerical simulations that I typically do.
Though I did get to map a few orogens, the vast majority of the work was done by many previous workers. The sources of the datasets we use are below. Please note that as a fluidly changing project, the references below may be out of date as new datasets are made available that can supersede those here.
Regardless, I extend my deepest gratitude to the workers in the references below (and the references used by those references) that have made this possible.
|Coverage Region||Reference||Dataset Name (if any)|
|New Zealand||Litchfield et al., 2014||none|
|East Africa||Macgregor, 2015||none|
|Middle East||Danciu et al., 2017||EMME|
|South America||SARA Faults||SARA|
|Europe||Woessner et al., 2015||SHARE|
|Northern Andes||Veloza et al., 2012||Active Tectonics of the Andes|
|Indo-Asian Collision Zone||Styron et al., 2010||HimaTibetMap|
|US mainland||Petersen et al., 2008||HazFaults|
|Taiwan||Shyu et al., 2016||none|
|Mexico||Villegas et al., 2017||none|
|Southeast Asia||Earth Observatory of Singapore||none|
|Northeast Asia||Styron et al., in prep.||none|
|North Africa||Poggi et al., in prep||none?|
|Central America and Caribbean||Styron et al., in revision||CCAF|
|Global (various regions)||Christophersen et al., 2015||Faulted Earth|
|Global (plate boundaries)||Bird, 2003||none|