Projects Help with Projects Past Projects AfricaArray BARGEN COCONet Dataworks for GNSS EarthScope EarthScope - PBO GeoEarthScope GEON GSAC GSRM - GEM GSRM - ILP INTERFACE L1 Project NLAS PBO GPS/GNSS Campaigns PBO Nucleus SNARF SSARA SuomiNet TEQC TLALOCNet

SNARF Working Group - First Release (SNARF 1.0)

A Stable North American Reference Frame (SNARF): First Release

The SNARF Working Group: G. Blewitt, D. Argus, R. Bennett, Y. Bock, E. Calais, M. Craymer, J. Davis, T. Dixon, J. Freymueller, T. Herring, D. Johnson, K. Larson, M. Miller, G. Sella, R. Snay, and M. Tamisiea


Starting in 2003, NSF funded a series of small workshops by UNAVCO's Stable North American Reference Frame (SNARF) Working Group to define the reference frame to be used for EarthScope. Such a frame would be important to describe relative motions of Plate Boundary Observatory sites spanning the North America - Pacific plate boundary. The goal was to facilitate geophysical interpretation and inter-comparison of geodetic solutions through standardization and documentation.

Fundamentally a reference frame is required because GPS alone does not provide unambiguous coordinates: GPS data are relatively insensitive to global rotations of the entire system. Fixing the rotation according to a well-documented scientific rationale and procedure can facilitate geophysical interpretation. Early on, the SNARF working group identified that current frames such as NUVEL-1A have significant deficiencies, particularly as the East African Rift was not taken into account. Moreover, glacial isostatic adjustment (GIA) is known to produce greater intraplate deformations than plate tectonics across a large portion of the North America, and so GIA would need to be considered in the reference frame. Moreover, research by the SNARF group indicated that GIA models are very sensitive to model parameters, and it is important to define a frame that does not come into systematic conflict with GPS data from well-established sites. For example, models of lateral variations in Earth structure can change predicted horizontal motions by a few millimeters per year in some locations.

Here we report on the anticipated release of the first version of SNARF, including scientific rationale, procedures, with discussion on how to use the products. Driving the design of SNARF at the highest level are the "big questions" that EarthScope is being designed to answer. We identified four big questions:

(1) Where does the plate boundary begin, and why? What is the extent of the stable plate interior, and how tectonically stable is the plate interior. Is the Colorado Plateau still rotating, and how active is the Rio Grande Rift? Is Alaska rigidly attached to North America (as current empirical evidence is weak).

(2) What is the vertical velocity field across North America? What is GIA versus tectonic? What is the role of body forces and mantle dynamics? A deceptively simple question whose answer is completely reference frame dependent: Is the Basin and Range going up or down?

(3) What signals are natural versus anthropogenic? What signals are due to ground fluid withdrawal, aquifer deformation, hydrological and atmospheric pressure loading? How do we disentangle these signals from GIA and tectonics? How do we define an unbiased reference frame in light of these effects?

(4) How can we design geodetic products that are stable over decadal time-scales and beyond? Will we be able to detect a > 5 year transient? Can we detect the "ghosts" of historic earthquakes? Is tectonic activity steady state, or does it switch on and off spatially and temporally? Can we confidently compare and understand differences between geodetic rates to geologic rates?

Guided by these questions and the ensuing research, the SNARF Working Group has identified and tackled several major issues, including (1) the production of a GPS velocity field that is accurate (representative) and relatively dense to select a base model for GIA, (2) the selection of "frame sites" based on geological and engineering criteria, (3) the selection of a subset of "datum sites" that represent the stable plate interior and will be used to define a no-net rotation condition, and (4) the definition of products to be distributed for general use. The vertical datum of SNARF will be consistent with ITRF2000, in that the center of mass of the whole Earth system is taken to be the origin, and the horizontal datum will differ by a rotation rate that brings the rotation of stable North America to rest. In this first release of SNARF 1.0, the product consists of (1) a rotation rate vector that transforms ITRF2000 velocity components into SNARF velocities, (2) an initial reference frame, defined as a list of selected sites, epoch coordinates, and site velocities, in the geocentric Cartesian system (X, Y, Z). SNARF will be adopted by PBO Data Analysis Centers which are scheduled to be in production-mode by October 2005. The intention is to incrementally improve SNARF through further research over the next 2-3 years, before handing off the production and maintenance of SNARF to the national geodetic agencies (the National Geodetic Survey and Natural Resources Canada). It is anticipated that SNARF will take on legal status as part of the bilateral North American Reference Frame (NAREF). At that point, NAREF will become the preferred name of the reference frame. Feedback by users of PBO products will be an essential ingredient toward improving the SNARF products.

Workshops for Establishing a Stable North American Reference Frame (SNARF) to Enable Geophysical and Geodetic Studies with EarthScope: Annual Report 2004-2005 PowerPoint
Stable North American Reference Frame (SNARF): Version 1 PowerPoint


[Back to SNARF Home]

     Comments or questions about this page? Send e-mail to Geoff Blewitt (


Last modified: 2019-12-24  02:12:52  America/Denver