A No-Net-Rotation Model of Present-Day Surface Motions

Corné Kreemer and William E. Holt (SUNY Stony Brook)

A significant portion of the Earth's surface consists of zones of diffuse deformation. The interior regions of these diffuse zones of deformation move at distinctly different velocities from that of adjacent plates, and, because of their complexities, have been ignored in previous no-net-rotation (NNR) models (e.g., NNR-NUVEL1A). We have calculated a new NNR model from a continuous velocity field that incorporates both rigid plate motion and velocity gradients within plate boundary zones. The velocity field is obtained through a bi-cubic Bessel interpolation of almost 3000 geodetic velocities and strain rates inferred from Quaternary faults. The geodetic velocities are taken from about 50 different, mainly published, studies. For each study we have not adopted the original reference frame. Instead, we have solved for a rigid body rotation for each study that rotate the vectors of each study into a model reference frame in the process of satisfying a least-squares fit between model and observed velocities and model and observed strain rates. When compared with earlier NNR models we find significantly different angular velocities for many plates in our model. Differences between the NNR model presented here and earlier NNR models can be attributed to both the effect of including velocity gradients in diffuse plate boundary zones, as well as actual differences between geodetically derived, present-day, surface motions and geologic estimates. We find that for the Indian, Arabian, Nazca, Cocos, Philippine Sea, and the Caribbean plate the differences between our model and the NNR-NUVEL1A model are mainly due to differences between geodetic and geologic plate velocities. For the Eurasian plate the discrepancy that we find between our result and NNR-NUVEL1A model can not only be ascribed to the difference between geodetic and geologic velocities, but also to the significant effect of including plate boundary zones. The significantly different NNR rotation vectors that we find for the majority of plates suggests that caution is warranted when using the NNR-NUVEL1A model to change from an ITRF to a tectonic reference frame. Our new NNR results indicate that such practice may result in significant discrepancies in crustal velocities with respect to the chosen reference plate. Finally, similar to earlier NNR models, we find a significant difference between the NNR velocities and velocities with respect to hotspots.