Current motion and deformation of the Nazca Plate: new constraints from GPS measurements

We use new GPS data to determine an updated Euler pole describing the present-day motion of the oceanic Nazca Plate. Our solution includes continuous GPS (cGPS) measurements at Malpelo Island offshore Colombia, two sites in the Galapagos archipelago, Easter Island and Salas y Gomez Island in the western part of the plate and Robinson Crusoe Island offshore Chile. A careful analysis of geodetic time-series reveals that (1) previous estimates using former cGPS site EISL are biased by several millimetres per year eastward due to station malfunctioning (2) north velocity component of cGPS site GLPS at Santa Cruz Island in the Galapagos is impacted by volcanic deformation at the 1-2 mm yr(-1) level, probably caused by the recurrent volcanic activity of the Sierra Negra volcano. In addition, we find shortening at similar to 1 mm yr(-1) between Easter Island (cGPS ISPA) and Salas y Gomez Island (cGPS ILSG), consistent with the elastic deformation induced by rapid opening at the East Pacific rise. cGPS site at Robinson Crusoe Island shows similar to 4-5 mm yr(-1) abnormally fast East velocity induced by the visco-elastic relaxation following the Maule M-w 8.8 2010 earthquake. Using this information, we determine a new Euler pole (longitude: -90.93 degrees E, latitude 56.19 degrees N, 0.588 deg Myr(-1)) describing the present-day Nazca-South America Plate motion, using five sites (Malpelo Island, two sites in the Galapagos archipelago, Easter Island and Salas y Gomez Island). The proposed Euler pole provides a weighted root mean square (wrms) of residual velocities of 0.6 mm yr(-1), slightly higher than usually observed for other major tectonic plates and accounting for the uncertainty of potential volcanic-tectonic deformation. Our model predicts a maximum convergence rate at 65.5 +/- 0.8 mm yr(-1) at latitude similar to 30 degrees S along the Chile trench, decreasing to 50.8 +/- 0.7 mm yr(-1) in northern Colombia and 64.5 +/- 0.9 mm yr(-1) in southern Chile (1 sigma confidence level). Comparison with the geological models NUVEL1A and MORVEL indicates constant decrease since 3.16 Ma of opening rate along the Nazca-Antarctic Plate boundary spreading centres at similar to 1 cm yr(-1) per Myr. Combined with the ITRF2014 pole for the Pacific and Antarctic plates, our derived Euler pole predicts closure at the similar to 1 mm yr(-1) level for Pacific-Antarctic-Nazca Plate circuit. However, combining our results with MORVEL estimates for the Cocos Plate, the non-closure of the Pacific-Cocos-Nazca Plate circuit is 9.7 +/- 1.6 mm yr(-1), 30 per cent lower than the 14 +/- 5 mm yr(-1) reported in MORVEL model, but still significant. A small (similar to 1.5 mm yr(-1)) velocity residual at Malpelo Island neither supports the hypothesis of an independent Malpelo microplate offshore Colombia nor large scale internal deformation induced by thermal contraction. Our solution rather suggests that non-closure of the Pacific-Cocos-Nazca Plate circuit arises from the determination of the Cocos Plate motion in MORVEL, an hypothesis further supported by the large discrepancy between MORVEL's prediction and the observed GPS velocity observed at Cocos Islands (cGPS ISCO).