Geodetic measurements of convergence at the New Hebrides island arc indicate arc fragmentation caused by an impinging aseismic ridge F. W. Taylor Institute for Geophysics, University of Texas, 8701 North MoPac Expressway, Austin, Texas 78759-8397 M. G. Bevis Hawaii Institute of Geophysics and Planetology, University of Hawaii, Honolulu, Hawaii 96822 B. E. Schutz Center for Space Research, University of Texas, Austin, Texas 78713 D. Kuang J. Recy Institut Franc ¸ ais de Recherche Scientifique pour le De ´ veloppement en Coope ´ ration (ORSTOM), Laboratoire de Ge ´ odynamique sous-marine, B.P. 48, 06230 Villefranche sur mer, France S. Calmant Institut Franc ¸ ais de Recherche Scientifique pour le De ´ veloppement en Coope ´ ration (ORSTOM), B.P. A5, Noumea, New Caledonia D. Charley Institut Franc ¸ ais de Recherche Scientifique pour le De ´ veloppement en Coope ´ ration (ORSTOM), B.P. 76, Port Vila, Vanuatu M. Regnier B. Perin Universities NAVSTAR Consortium, P.O. Box 3000, Boulder, Colorado 80307 M. Jackson Department of Geological Sciences, University of Colorado, Boulder, Colorado 80309 C. Reichenfeld Institut Franc ¸ ais de Recherche Scientifique pour le De ´ veloppement en Coope ´ ration (ORSTOM), Paris, France ABSTRACT Global positioning system (GPS) measurements in 1990 and 1992 from two sites on the southern New Hebrides island arc give convergence rates with the Australian plate of 103  5 mm/yr and 118  10 mm/yr. In contrast, GPS measurements in the central New Hebrides indicate anomalously low convergence rates of 42 mm/yr. On geologic time scales, the mean central New Hebrides convergence rate has been 85–132 mm/yr. Elastic fault models with a locked interplate thrust zone indicate that maximum possible rates of horizontal elastic strain are insufficient to account for the anomalously slow convergence. Therefore, we propose that the central New Hebrides segment is moving eastward relative to adjacent arc segments at a rate of 36 – 83 mm/yr. This displacement is accommodated by crustal shortening at the eastern margin of the arc and strike-slip faults crosscutting the arc. Resistance to subduction of the aseismic D’Entrecasteaux Ridge system is the likely cause for horizontal forces sufficient to shove a large segment eastward and fragment the arc. This process demonstrates that subducting bathymetric features can impose funda- mental structural modifications on an arc that may represent the initial stages of arc polarity reversal. INTRODUCTION The tectonic effects of subducting bathy- metric features on arc tectonics are signifi- cant if their effects are cumulative, because subducting objects are numerous. Subduct- ing oceanic plateaus, island arcs, seamounts, ridges, or active sea-floor–spreading systems may affect arc tectonics by a variety of mech- anisms (e.g., Cloos, 1992, 1993; Geist et al., 1993). Tectonic erosion and underplating have permanent effects, but isostatically in- duced vertical motion may reverse following subduction of an object. Similarly, the ‘‘stick-slip’’ earthquake cycle of elastic- strain accumulation and release may not be a factor in permanent deformation. How- ever, an important issue addressed here con- cerns the magnitude and distribution of horizontal forces applied by an impinging feature and how far such forces extend from the zone of contact. The D’Entrecasteaux Ridge system, which intersects the New Hebrides arc (Figs. 1 and 2), is 100 km wide, consists of a par- allel northern ridge and a southern sea- mount chain, and has 3 km of relief. Sub- duction of this system probably causes the anomalous morphology of the central New Hebrides arc, including the absence of a trench, rapid uplift of both back-arc (sites PENT and MAEW; see Table 1 for island names) and fore-arc islands (SNTO and MLKL), and subsidence of the intervening basin (Collot et al., 1985; Taylor et al., 1987; Greene et al., 1988). Moreover, the Bou- gainville Guyot is currently indenting the edge of the arc by about 10 km (Fig. 2), causing a fore-arc tectonic bulge hundreds of metres high (Fisher et al., 1991). We present here global positioning sys- tem (GPS) measurements of convergence with the Australian plate of six New Heb- rides sites, and we interpret the influence of D’Entrecasteaux Ridge impingement on convergence rates within the New Hebrides arc. If there was no disruption of the arc by subduction of the D’Entrecasteaux Ridge, then convergence between the central New Hebrides and the Australian plate should equal Pacific-Australian convergence rates or exceed them due to back-arc extension in the North Fiji Basin. CONVERGENCE RATES AT THE NEW HEBRIDES ARC GPS convergence rates for the southern New Hebrides with the Australian plate are 118  10 mm/yr (TANN) and 103  5 mm/yr (EFAT) and azimuths are N113° 1°W and N112° 1°W, respectively (Ta- ble 1 and Fig. 1). In contrast, convergence rates at the central New Hebrides arc at sites SNTO, MLKL, and MAEW are only 39  4, 41  4, and 41  2 mm/yr, and azimuths range from N115° 5°W to N123° 2°W. PENT has a slightly different convergence vector (47  4 mm/yr; azimuth = N100° 2°W). The mean convergence rate for the four sites is 42 mm/yr. Analysis of local mo- tion among these four sites finds slight con- vergence between MAEW and PENT, which indicates only minor tectonic defor- mation within the polygon formed by the four sites. However, the data from PENT are fewer than for the other sites, so this is a tentative interpretation. The global plate motion model NUVEL- 1A predicts Pacific-Australian convergence rates of 79 – 86 mm/yr for the New Hebrides arc from TANN northward to SNTO (DeMets et al., 1994; Table 1). GPS conver- gence rates at southern New Hebrides sites TANN and EFAT (118  10 mm/yr and 103  5 mm/yr) clearly exceed these NUVEL-1A rates. The EFAT and TANN GPS convergence azimuths (N112°–113°W) differ significantly from NUVEL-1A Aus- tralian-Pacific azimuths of N99°–100°W, but do not differ significantly from the average earthquake slip vector azimuth of N104° 11°W (e.g., Isacks et al., 1981), given the un- certainty in the slip vector. For the central Geology; November 1995; v. 23; no. 11; p. 1011–1014; 3 figures; 1 table. 1011