IAG-IASPEI 2017

Presentation information

Oral

IAG Symposia » G01. Reference frames

[G01-3] Reference frame methodology and implementation

Wed. Aug 2, 2017 1:30 PM - 3:00 PM Room 502 (Kobe International Conference Center 5F, Room 502)

Chairs: Johannes Böhm (Technische Universität Wien) , Zuheir Altamimi (Institut National de l'Information Géographique et Forestière)

1:30 PM - 1:45 PM

[G01-3-01] Robust realization of a no-net rotation reference frame on a deforming tectonic plate

Geoffrey Blewitt, Corne Kreemer, William C. Hammond (Nevada Geodetic Laboratory, Nevada Bureau of Mines & Geology, University of Nevada, Reno, USA)

We present a robust method of fixing a reference frame to a moving and deforming plate, using the example of North America. Scientific applications motivate that we fix the frame with a scale consistent with the SI system, an origin that coincides with the Earth system's center of mass, and with axes attached to the rotating interior of the plate. To realize this no-net rotation condition, a typical approach is to select a subset of “core stations" that have small residual velocities. Stations in the far field of plate boundaries and glacial isostatic adjustment can be selected to satisfy these criteria. However, such regions of the Earth's crust may appear to be rigid (e.g., ~0.3 mm/yr residuals for core stations in our NA12 frame), but the entire region can move relative to plate motion. Such common-mode motion biases plate rotation estimation, with implications on scientific interpretation. Our scheme presented here (1) starts with robust estimation of station velocities using our MIDAS algorithm, (2) then uses median spatial filtering to remove outlier stations that do not reflect the regional trend, and (3) performs simultaneous estimation of strain-rate tensor and rotation components for any chosen point within the plate using our new algorithm, MELD (“median estimate of local deformation"). MELD uses velocities from all possible triangles of stations enclosing each point within 2 or 3 nearest neighbors of the Delaunay Triangulation. From the parameter estimates of all local triangles, the median value for each strain rate and rotation component is then taken. From this analysis, we then take an area-weighted sum of rotations over all points to estimate plate rotation, which is then subtracted to define a no-net rotation, plate-fixed reference frame. This procedure estimates the rotation pole robustly, and the residual velocities prove to be significantly different than in frames realized using subsets of stations that appear to be internally rigid.