Regional Arrays

Introduction

North America plate reference frame realizations serve several purposes in the analysis of GPS data from regional arrays in and around the North America continent.   First, North America plate realizations provide a tectonically motivated perspective from which to visualize velocity estimates, compare estimates against tectonic intuition, and otherwise qualitatively interpret estimates.    Second, North America frames provide a common ground for comparisons between solutions obtained by different analysis centers.  Third, North America reference frames provide an important kinematic boundary condition for studies of plate boundary (or intra-plate) deformation and an important link between plate boundary zone deformation processes and global plate tectonic motions.

Current practices

Although the specifics differ among regional array analysis centers, the North America plate frame is currently realized by most centers following the same basic steps:

(1)   Analysis of raw GPS data from the regional array together with data from stations within the stable interior regions of North America adopting a well-defined external reference frame such as ITRF2000.

(2)   Selection of a set of stations whose velocity estimates represent motion of the North America plate and estimation of transformation parameters so as to minimize the velocities at the stations deemed representative of the North America plate.

(3)   Make a decision about how to account for reference frame uncertainty in the variance-covariance matrix associated with the velocity estimates.  

An alternative to steps (1) and (2) above is to apply a “known” rotation from the regional velocity solution in a global reference frame to the North America frame (e.g., from ITRF2000 to North America using an NNR-NUVEL-1A-North America Euler vector).   Two examples of reference frame realizations, one for the PANGA network and one for the BARGEN network, were reviewed at the January 2004 SNARF Workshop (cf., snarf_2004_mmm.ppt; BennettSNARFJan2004.ppt).  Experience based on analyses of data from regional arrays exposes several limitations in the utility of present North America reference frame determinations as outlined in the next section. 

Limitations

There are a variety of factors that limit the precision (repeatability or stability) and accuracy of reference frame realizations using GPS.   These factors lead to differences between frame realizations determined (1) at different times by the same analysis center, and (2) by different analysis centers.  

Reference frame differences among successive solutions determined by a given analysis center can result from a variety of factors.   First and foremost, reference frames are based on a selection of available data, which evolves with time, both in terms of the amount of data at any given station and the number of available stations producing (useful) data.  Second, plate fixed reference frames are determined under the assumption that the secular motions of the stations used to define the plate frame represent plate motion only.   That is, any apparently secular motions of the sites due, for example, to strain associated with nearby plate boundary zone faults, intraplate faults, hydrological loads, Glacial Isostatic Adjustment, etc, can bias estimates for plate motion at a site.   Furthermore, periodic and episodic motions of sites, including apparent episodic displacements associated with equipment changes, etc, if not modeled or averaged carefully, can also lead to biased estimates for secular motion that depend critically on the time spans of the data.   Comparison of successive solutions, to assess the quality of the solutions for example, requires an assessment of the stability of the reference frame realization from solution to solution. 

Reference frame differences between analysis centers can arise from all of the data-set dependent factors affecting individual analysis centers listed above.   But in addition, comparisons between analysis centers must also take into consideration the specifics of the respective procedures used.   The transformation from global to North America reference frames (a) by estimation of transformation parameters unique to the data set, or (b) by application of a known set of transformation parameters in step 3 under “Current practices” above, provides an obvious example.   Quantitative comparisons between solutions determined by different analysis centers may require re-alignment of one or both frames if precision of the order of 1 mm/yr is required, even if identical subsets of data are analyzed by each center. 

Future needs

One of the primary goals of SNARF is to overcome or mitigate the limitations listed in the previous section.   Following is a summary the needs of regional network analysts as identified at the January 2004 SNARF Workshop: 

(1)   Decide upon a common set of procedures for determining a GPS-based SNARF, including specification of a set of high-quality GPS stations in the North America interior through which the SNARF frame can be realized, and a prescription for the proper determination of variance-covariance matrices associated with solutions in the SNARF frame.

(2)   Provide guidelines for authors on how to describe their reference frame determinations in order to facilitate comparisons among published solutions.

(3)   Provide “directions for use” for solutions provided in the SNARF frame.  The directions should be aimed at a non-geodesist audience, but should include a description of how to treat the variance-covariance matrices provided with SNARF solutions, and recommendations for how to use velocities provided in the SNARF frame to constrain geophysical models. 

Last modified: 2019-12-24  02:12:54  America/Denver