Automatic Segmentation of Geophysical Trackline Data Applied to
Pacific-Rivera Motion
Abstract
Magnetic trackline data are of central importance in estimating relative
plate motions, have allowed a detailed apparent polar wander path to be
determined for the Pacific Plate from skewness analysis, and provide
important constraints for many other geophysical investigations.
Interpreting trackline data typically requires processing, which often
involves dividing tracks into quasi-linear segments. Here we present a
method for automating this processing step by adapting the
Ramer-Douglas-Peucker (RDP) algorithm, originally developed for
polygonal approximation and cartographic generalization, for use on
trackline data on a sphere and demonstrate its ability to segment data
into quasi-linear (i.e., great circle-like) sections quickly based on
two intuitive parameters. The new procedure is largely automated and
requires minimal effort. As a test and proof of concept, we apply this
method to estimate an angular velocity for motion between the Pacific
and Rivera plates since 0.781 Ma using data from the NCEI geophysical
trackline database. Our modified RDP algorithm identified 400+ track
segments that intersected the Rivera Rise, of which more than 50
provided useful Pacific-Rivera spreading rates, which is roughly twice
the 26 rates used to estimate Pacific-Rivera motion in the MORVEL set of
geologically current plate relative angular . We compare the resulting
angular velocity to previous estimates of Pacific-Rivera angular
velocity and explore implications for Pacific-North America and
Pacific-Cocos relative motion and Rivera absolute motion.