Present‐day kinematics of the Rivera Plate and implications for tectonics in southwestern Mexico

Abstract

We model the present‐day motion of the Rivera plate relative to the Pacific, North America, and Cocos plates, and combine the results with the NUVEL‐I model for Cocos‐North America motion to examine present‐day deformation in southwestern Mexico. The Pacific‐Rivera data, which include 25 three m.y.‐average spreading rates determined from original surface‐ship magnetic data and 22 azimuthal data along the Rivera transform, are systematically misfit by the NUVEL‐I Pacific‐North America and Pacifie‐Cocos Euler vectors, indicating that the Rivera plate is kinematically distinct from North America and Cocos. An F test shows that the improvement in fit to the Pacific‐Rivera data from a model with a distinct Rivera plate exceeds that expected solely from adding three model parameters. The Rivera‐North America Euler vector derived from closure of the Rivera‐Pacific‐North America plate circuit predicts slower and more trenchnormal convergence along the Acapulco trench than prior models. In addition, the model predicts convergence normal to the eastern Tamayo fracture zone at a rate 60% slower than the right‐lateral strike‐slip predicted by prior models. An observed systematic misfit of the Pacific‐Rivera Euler vector to azimuths from the western Rivera transform fault may have several causes. There may be a time‐averaging problem between the 3.0‐m.y.‐average rates and the shorter time‐average transform fault azimuths and earthquake slip vectors. Alternatively, Rivera transform trends may be biased by deformation within lithosphere adjacent to the transform valley, or future detailed mapping of the transform may reveal the misfit to be an artifact of the presently available bathymetry. Along the entire Pacific‐Rivera rise, spreading rates averaged over the past 0.7 m.y. are systematically faster than 3.0‐m.y.‐average rates, with the difference increasing southward along the rise. The post‐3 Ma change in the rate gradient has caused a southward migration of the Pacific‐Rivera Euler pole since 3 Ma, possibly in response to the ∼3 Ma completion of a spreading reorganization that doubled the length of the Rivera transform fault and halved the area of the Rivera plate. We hypothesize that deformation in southwestern Mexico, including opening along the Colima rift, is related to oblique subduction of the Cocos plate along the northern Middle America trench. Three Unes of evidence support this hypothesis. First, the sense of oblique subduction predicted by NUVEL‐I along the Middle America trench is consistent with a model in which part of western Mexico located southeast of the Colima rift moves to the southeast relative to North America. Second, field geologic data and Landsat imagery suggest that in the past few million years, the Chapala‐Oaxaco fault zone has accommodated several kilometers of sinistral motion and sinistral transtension has occurred along the Trans‐Mexican Volcanic Belt. Either or both of these fault zones could accommodate any southeastward coastal block motion that might result from oblique subduction. Third, the observation that earthquake slip vectors from the northern Middle America trench trend systematically counterclockwise from the predicted Cocos‐North America convergence direction is also consistent with southeastward motion. The discrepancy between the observed and predicted slip directions along the Middle America trench suggests sinistral slip is 0–10 mm yr−1. If oblique subduction is driving southeastward transport of a coastal sliver, the Colima rift is a passive, pull‐apart zone at the northwestern end of this sliver, rather than an incipient spreading ridge that will replace the Pacific‐Rivera rise.