Eos, Vol. 76, No. 1, January 3, 1995

EOS, TRANSACTIONS, AMERICAN GEOPHYSICAL UNION

VOLUME 76 NUMBER 1

JANUARY 3,1995 EOS PAGES 1-8

of rotation for this plate boundary lies to the Sonic Imaging Reveals New southeast (60.1°S, 178.3°W) [DeMetsetaL, 1990], the high-angle convergence along the Kermadec Trench changes to oblique conver­ Plate Boundary Structures gence of opposite polarity along the Puysegur Trench (Figure 1). Offshore New Zealand Between the trenches, the plate boundary crossing the New Zealand continental crust PAGES 1,4-5 is characterized by strike-slip and compres­ sive (i.e., transpressive) deformation along J.-Y. Collot, J. Delteil, R. H. Herzer, R. Wood, K. B. Lewis, and the Alpine Fault (Figure 1). On the GEODYNZ- Shipboard Party SUD cruise we studied the variations in deformation and sedimentation associated with changes from oceanic subduction to in- Recent bathymetry and sonar imagery were produced for each survey line. Swath tra-continental transpression. The following studies of offshore portions of the plate data were augmented by geopotential data observations were among the highlights: boundary transecting New Zealand allow the and six-channel seismic reflection data shot • widespread tectonic erosion along first confident mapping of detailed tectonic with two 75 cu in SODERA GI guns. the northern Hikurangi-Kermadec margin and sedimentary patterns of the region. The active margins of New Zealand re­ caused by the collision of ridges and vol­ Working in late 1993 aboard the R/V flect complex temporal and spatial canic seamounts; L'Atalante of the Institut Francais de Recher­ interactions between subduction and strike- • extensive offshore strike-slip faulting che pour TExploitation de la Mer (IFREMER), slip regimes along the Pacific-Australian along segments of the Hikurangi-Kermadec we recorded soundings of a wide swath of plate boundary. Because the PAC-AUS pole margin; seabed to elucidate major structural transi­ tions along the plate boundary. Results of the study, part of the GEODYNZ-SUD program developed jointly by institutions in France and New Zealand, will be complemented by New Zealand cruises to the Puysegur and Hikurangi margins. The total data set will be Fig. 1. Map of the processed and interpreted during the next areas surveyed during two years. the Geodynz-Sud A shipboard SIMRAD EM12Dual (EMI2D) cruise of the R/V multibeam system recorded 160,000 km2 of L'Atalante along New swath bathymetry and imagery over the Ker- Zealand's active mar­ madec-Hikurangi margin to the northeast gins with linearly and the Fiordland-Puysegur margin to the color-coded swath southwest of New Zealand (Figure 1). The bathymetry; red ar­ rows = PAC-AUS plate EM12D consists of separate multibeam echo- convergence direction sounders, each generating 81 stabilized [DeMetsetaL, 1990]; beams at frequencies of 12.6-13 kHz. This sys­ HP'=Hikurangi Pla­ tem allows simultaneous determination of teau; FM=Fiordland 162 measurements of phase giving depth and Margin; sawtooth 162 measurements of the energy back-scat­ lines = subduction tered. After on-board processing, swath maps fronts; heavy lines = of bathymetry and side scan sonar imagery major strike-slip faults; over a maximum 22 km-wide strip of seabed light blue = continen­ tal rocks, depth <1.5 J.-Y. Collot, Institut Francais de Recherche km; and dark blue = Scientifique pour le Developement en Coopera­ seafloor depth >1.5 tion, Villefranche s/mer, France; J. Delteil, In­ km. Original color im­ stitut de Geodynamique, Nice, France; R. age appears at the Herzer and R. Wood, Institute of Geological back of this volume. and Nuclear Sciences; and K.W. Lewis, Na­ tional Institute of Water and Atmospheric Re­ search, Wellington, New Zealand

This page may be freely copied. Eos, Vol. 76, No. 1, January 3, 1995

Mets, 1990]. On the Pacific plate, the 10-15 partitioning across the margin. The compres­ km thick Hikurangi Plateau, which has areas sive component is taken up mainly within of abundant seamounts and sediment-filled off-scraped Plio-Pleistocene trench sediment troughs, is subducted westward at the south­ of the accretionary wedge. Behind this wedge ern Kermadec Trench and Hikurangi Trough the Cenozoic slope sediment takes up much of [Davy, 1992]. On the Australian plate, south the strike-slip component of the strain. of the Kermadec island arc, the eastern edge At the southern end of the accretionary of the New Zealand continental crust is a wedge, where the margin is almost parallel to compressed wedge of mid-Cenozoic slope the convergence direction, strike-slip faults sediments, with local outgrowth of an accre- in the upper margin converge southwestward tionary wedge of off-scraped trench in a horse tail tectonic pattern. Farther south, sediments [Lewis and Pettinga, 1993]. in the zone of continental collision, the mar­ The GEODYNZ-SUD cruise focused on the gin is deformed mainly by compression. On above-mentioned transition zones of the Ker­ the adjacent land a wide zone of transpres- madec-Hikurangi margin (Figure 1) to clarify sion links south with the Alpine Fault that the relationships between subduction and extends to Fiordland. strike-slip motion and to assess the impact of seamount and scarp subduction on the evolu­ The Flordland-Puysegur Margin tion of the margin. Observations in the northern part of the Southwest of New Zealand, the Austra­ survey area suggest that the 1000-m-high lian plate is subducted eastward along the scarp forming the northern edge of the Pacific-Australian plate boundary. The Hikurangi Plateau has swept southward GEODYNZ-SUD study focused on the struc­ along the trench. North of the scarp-trench tural transitions between the Fiordland junction, the sweep has left a collapsing in­ continental margin (trending N30°^45°E) at ner trench wall that is offset 15 km westward the southern end of the Alpine Fault, the con­ from the trench to the south. South of the tinent-to-ocean transition zone of the junction, the Hikurangi Plateau has uplifted Puysegur Bank (trending N10°-15°E), and the Kermadec forearc and trench axis by 1.5 the N20°-25° E trending Puysegur Ridge (the km to form a series of narrow en echelon, northern part of the Macquarie Ridge Com­ Fig. 2. Generalized multibeam bathymetric sediment-starved basins (Figure 2). On the plex) and Puysegur Trench (Figures 1 and 3). map (contour interval 250 m) of the northern northern Hikurangi Plateau, there are several Along these segments, the relative plate mo­ Hikurangi-Kermadec margin showing the up­ previously unrecognized, highly reflective, tion is very oblique (20^0°) to the lifted, en echelon southern Kermadec trench, presumably volcanic seamounts and ridges deformation front, and the convergence rate the indented northern Hikurangi margin and trending N140°-168°E and ranging up to 1000 is uniform (3.7-3.2 cm/yr) [De Mets etai, the location of strike-slip deformation; saw­ m high. Some of the seamounts lie in front of 1990]. On the lower plate, the Resolution tooth line = subduction front; large arrow = a newly recognized indentation of the mar­ Ridge (Figure 3) is thought to separate the PAC-AUS plate convergence direction; shaded gin that is 270 km long (37°45'to 40°S) and up Cretaceous-Paleocene Tasman Sea oceanic zone = ridge damming the sediment of the to 25 km wide. This indented margin has an crust (to the northwest) from a southeastern Hikurangi Trough; andNI=North Island. abnormally steep inner slope and abundant Eocene-Oligocene wedge of Indian Ocean slump and collapse features, suggesting ex­ oceanic crust [Weisseletal, 1977]. On the tensive tectonic erosion of the margin by upper plate, the dextral transpressive Alpine • a ridge forming a dam that prevents seamount collisions. This process causes fault [Wellman, 1953] is thought to lie along movement of turbidity from the Hikurangi slope failures that feed the Hikurangi Trough the continental shelf of Fiordland (Figure 1). Trough to the Kermadec Trench with avalanche deposits. These deposits, Beneath Fiordland, an 80°SE dipping Benioff • a major strike-slip fault zone extend­ ponding in the Hikurangi Trough, fail to zone [Smith and Davey, 1984] defines the ing along the summit of the Puysegur Ridge reach the Kermadec Trench because they morphology of the subducting Australian (the Puysegur Fault); are impeded by a ridge at 37°45'S. plate. South of Fiordland, although subduc­ • a relay zone of splay faults linking the We recognized strike-slip faulting, pre­ tion is marked by the Puysegur Trench and a Alpine Fault and the Puysegur Fault; and viously known only onshore, at least 70 km single, Quaternary andesitic volcano, there is • seafloor spreading fabric with three east of the shore. This faulting appears to be no clear Benioff zone. Focal mechanisms distinct orientations on oceanic crust west of widespread along the northern and southern show both strike-slip and thrust type deforma­ the Puysegur Trench. Hikurangi margins that accommodate part of tion [Anderson etai, 1993]. the oblique subduction of the Hikurangi Pla­ New data collected in the area adjacent teau (Figure 2). In the north, the evidence for to Fiordland show that submarine canyons, The Kermadec-Hikurangi Margin strike-slip faults includes scarps and mor­ which incise the upper margin west of the East of New Zealand, changes in structure phologic lineaments that define northward presumed Alpine Fault trace, appear to be lat­ and sedimentation along the PAC-AUS plate diverging horse tail tectonic patterns near erally offset from the mouths of fiords. This boundary reflect transitions from intra-ocean 37°Sand near40°S. offset indicates recent dextral strike-slip dis­ subduction at the Kermadec Trench, to sub­ In the south, dextral strike-slip is evident placement along the upper margin. Several continental subduction at the Hikurangi as offset ridges with small "pull apart" basins linear fault scarps, (presumably strike-slip), Trough, to continental collision of the behind the accretionary wedge. In contrast, splay southwestward in horse-tail fashion Chatham Rise. These changes also indicate at the southern end of the accretionary across the continental slope (Figure 3). relative plate motion from a 60° angle across wedge, the new data show extremely linear, These scarps splay away from the presumed the Kermadec Trench to a more oblique back-tilted thrust ridges, striking 45° to the main trace of the Alpine fault and appear to angle (20°) along the southern Hikurangi plate convergence direction. Thus a geo­ reach the subduction deformation front. This Trough. Here the convergence rate south­ graphic separation between strike-slip pattern suggests that the Fiordland margin is ward also decreases from 6 to 3.9 cm/yr [De faulting and thrusting suggests some strain obliquely segmented into continental strike-

This page may be freely copied. Eos, Vol. 76, No. 1, January 3, 1995

n—i—|—i—i—p 165'

TASMAN SEA

Fig. 3. Generalized multibeam bathymet- ric map (contour in­ terval 250 m) of the Fiordland-Puysegur margin. Faults splay southwestward away from the pre­ sumed offshore trace of the Alpine Fault, and south of the Puysegur bank, the strike-slip, fan- shaped pattern of faults merge south­ ward with the Puysegur Fault; thick solid lines = faults; sawtooth line = subduction front; short dashed lines = N60°E trending oce­ Fig. 4. (top) Side scan sonar imagery (high anic fabric; large ar­ reflectivity is shown as black), and (bottom) row = PAC-AUS plate structural interpretation of the Puysegur Fault, convergence L 'Atalante Fracture Zone and oceanic fabrics direction. (trending N120° andN85°E) west of the Puysegur Trench; the Puysegur Ridge is shaded; crosses = bathymetric highs; large sawtooth line = thrusting front; small sawtooth line - reverse faults; lines with ticks = scarps or normal faults; large arrow=plate convergence [De Mets et al., 1989]; double arrow=strike-slip motion slip slivers that are underthrust by the Austra­ Evidence for eastward subduction, local along the Puysegur Fault. lian plate. tectonic accretion and tectonic erosion also South of Puysegur Bank, in the Puysegur stems from the newly acquired data. At the change and the disappearance of the Ridge, the tectonic pattern suggests that front of the continental strike-slip slivers, east trench—may be an indication of the process strike-slip is distributed throughout a fan- of Fiordland, small accretionary lobes ex­ of subduction initiation in this region sug­ shaped, fault-sliver, valley-and-ridge terrain tend into the trench, a 10-20 km wide and gested by Ruff et al. [1989]. on the overriding plate. These faults con­ 4000-m-deep sediment-filled basin. The outer The clear spreading fabric on the oceanic verge south of latitude 48° S to form a sharply trench wall is deformed by Recent trench-par­ Australian plate (including fracture zones localized, linear zone of deformation parallel allel normal faults caused by plate flexure. such as L'Atalante Fracture Zone) containing to the ridge about 55 km east of the trench The Resolution Ridge, a series of en echelon three different spreading directions was an (Figure 3). The narrow shear zone continues blocks, crosses the trench opposite the south­ unexpected discovery. The fabric trends southward into the summit region of the very ern end of Fiordland. South of the ridge the N120°E near 49°30S, N85°E near48°30S (Fig­ linear, highly reflective Puysegur Ridge as an inner trench shows possible damage from ure 4) and N60°E near 4 7°S (Figure 3, lower axial valley, trending N25°E, 20-33 km east of previous subduction of a segment of the left). The N60°E trend is parallel to the late the trench (Figure 4). This perched valley is a ridge. From latitude 47°S, the Puysegur Eocene-early Oligocene magnetic anomalies series of elongated troughs that relay and lo­ Trench extends southward as an en echelon, farther southwest that were recognized by cally overlap along the strike of the ridge, narrow, flat-bottomed trough deepening Weissel etal. [ 1977], but the trends we ob­ isolating uplifted and downthrown blocks. from 5500 to 6200 m near 48°S and shallow­ served are new discoveries. We have not We believe that most, if not all, of the ing farther south to 5200 m where the trench identified any triple junction. These trends strike-slip motion on this part of the plate disappears. The inner trench wall from 47° to appear to represent either discrete or gradual boundary takes place along the axial valley. 48°S shows evidence for indentation and changes of spreading direction, at least in This valley represents the trace of a major slope failures suggesting frontal tectonic ero­ this region, during the Cenozoic. fault that we named the Puysegur Fault. The sion where the Puysegur Ridge is the widest. overall tectonic pattern at the transition be­ Farther south of 48°30\ where the ridge nar­ rows and shoals to 125 m, the inner trench tween the Fiordland margin and the Scientific Party Puysegur Ridge suggests that a relay zone of wall becomes steeper and regular. The south­ splay strike-slip faults links the continental Al­ ward changes—narrowing and shallowing of J.-C. Audru, B. Mercier de Lepinay, M. pine Fault to the oceanic Puysegur Fault. the ridge, its along-strike morphologic Popoff, E. Ruellan, and M. Sosson, IG, Nice,

This page may be freely copied. Eos, Vol. 76, No. 1, January 3, 1995

France; P. Barnes, NIWA, Wellington, New Foreign Affairs for funding and supporting De Mets, C, R. G. Gordon, D. F. Argus and Zealand; G. Lamarche, J.-F. Lebrun, B. Pon- this collaborative work, IFREMER for provid­ S. Stein, Current plate motions, Geophys. J. Int., 101, 425, 1990. toise, and B. Toussaint, ORSTOM, ing R/V L'Atalante ship time and equipment, Lewis, K. B. and J. R. Pettinga, The emerg­ Villefranche s/mer, France; S. Calmant and and GENAVIR officers, technicians and crew. ing, imbricate frontal wedge of the B.Pelletier, ORSTOM, Noumea, New Caledo­ Hikurangi margin, in Basins of the South­ nia; F. Chanier and J. Ferriere, University of References west Pacific: Sedimentary Basins of the World, vol. 3, edited by P. F. Ballance, Lille, France; E. Chaumillon, Laboratoire de Anderson, R, T. Webb and J. Jackson, Focal 1993. Geodynamique, Villefranche s/mer, France; mechanism of large earthquakes in the Ruff, L. J., J. W. Given, C. O. Sanders, and M. Coffin UTIG, Austin, Tex.; B. Davy and C. South Island of New Zealand; Implications C. M. Sperber, Large earthquakes in the Uruski, IGNS, Wellington, New Zealand; D. for the accommodation of Pacific-Australia Macquarie Ridge Complex: Transitional tec­ Christoffel, Victoria University, Wellington, plate motion, Geophys. J. Int., 115, tonics and subduction initiation, Pure and 1032, 1993. Applied Geophysics, 128, 72, 1989. New Zealand; S. Lallemand, University of Davey, F. J., M. Hampton, J. Childs, M. A Smith, E. C. C. and F. J. Davey, Joint hypo- Montpellier, France; A. Mauffret, CNRS, Paris, Fisher, K. B. Lewis and J. R. Pettinga, Struc­ center determination of intermediate depth France; A. Orpin and R. Sutherland, Otago ture of a growing accretionary prism, earthquakes in Fiordland region, New Zea­ University, New Zealand Hikurangi margin, New Zealand, Geology land, Tectonophysics, 104, 127, 1984. 14, 663, 1986. Weissel, J. F., D. E. Hayes and E. M. Her- Davy, B. M., The influence of subducting ron, Plate tectonics synthesis: The displace­ plate buoyancy on subduction of the ment between Australia, New Zealand, and Antarctica since the late Cretaceous, Mar. Acknowledgments Hikurangi-Chatham Plateau beneath North Is­ land, in Proceedings of the Halbouty Con­ Geology, 25, 231, 1977. We thank INSU, ORSTOM, Foundation for ference, edited by J. S. Watkins, F. Wellman, H. W., Data for the study of Re­ Zhiqiang and K. J. McMillen, 75 pp., Ameri­ cent and late Pleistocene faulting in the Research, Sciences & Technology of New Zea­ can Association of Petroleum Geologists, South Island of New Zealand, N. Z., J. land, IGNS, NIWA and the Ministry of French Tulsa, Okla., 1992. Sci. Technol, B34, 270, 1953.

square slope of ocean ripples with wind ve­ Pairing of Radar Instruments on locity. The radar returns are, of course, from ripples over the illuminated area, but only those with length longer than the radar wave­ Satellites Could Provide Optimal length can contribute to specular returns. From Cox and Munk's results, Wu [1990a] de­ Mapping of Sea Surface Winds duced the mean-square slopes contributed by wave components with their length longer PAGE 3 than 2.5 cm. The X band radar used for the Geosat altimeter had a wavelength of about 2.16 cm, differing only slightly from the wave­ JinWu length discussed above. Wu [1994] derived an analytical altimeter wind algorithm using these spectrally resolved results of normally Scientists pairing two radar instruments performs most effectively at the low wind distributed ripple slopes. Wind speeds calcu­ with different sensitivities to ripples on the speeds in these areas [ Witter and Chelton, lated using the algorithm compared favor­ ocean surface may be able to measure the en­ 1991; Wu, 1994]. This sensor was designed ably with those tablulated with the algorithm tire range of sea surface wind speeds—vital primarily for measuring sea-surface eleva­ proposed by Witter and Chelton (1991) for parameters that affect climate and the ocean tions [Chelton etai, 1989]. Seasat altimeter data, except at very light environment. The altimeter may be used for measrring winds (Uio < 0.3 m s"1). Ocean ripples detected via remote radar low speed winds (U < 5 m s"1) and the scat- sensing are effective tracers of sea surface terometer, for winds of higher velocity. The wind velocity [Woiceshyn etai, 1986; Wu, scatterometer can detect the wind direction Detected Trends at Low Winds 1990b]. The scatterometer, which detects the effectively at all wind speeds from the pat­ scattering of radar waves by ripples on the tern of its returns at various azimuth angles Scatterometer Woiceshyn et al. [1986] ex­ amined the scatterometry wind algorithm on sea surface, is deployed on Earth-orbiting sat­ [Wu, 1990a]. the basis of in situ wind measurements from ellites to map wind velocities over the the ocean platform during the Joint Air-Sea world's oceans. The scatterometer/wind Remote Sensing of Sea-Surface Winds Interaction Experiment (JASIN) in the North anemometry relationship has been derived Atlantic Ocean, and from buoys operated by from data collected mainly under intermedi­ Scatterometer The Seasat A scatterometer the National Data Buoy Ofice (NDBO) off the ate wind velocities. system (SASS) employed for measuring sea surface wind speeds was a 14.6-GHz (2.1-cm Gulf coasts and in the North Atlantic and Pa­ Radar returns for the scatterometer are er­ cific oceans. This comparison, organized ratic, however, at low wind speeds wavelength) active microwave radar. Radar sea returns at medium incidences follow the into intervals of increasing wind speed, in­ [Woiceshyn etai, 1986], which prevail over first-order Bragg scattering, wherein the cluded 500 pairs of SASS/JASIN measure­ the oceans [Harrison, 1989]. Winds are espe­ ocean wave number is related to the radar ments and 100 pairs of SASS/NDBO cially light over tropical regions [Sadler etai, wave number and the angle of microwave in­ measurements. 1987], where most of the heat transfer from cidence. The SASS wind velocities, expressed in oceans to the atmosphere takes place. In Altimeter The backscattering of radiation terms of the radar cross section, are com­ these regions, precise wind measurements at small incidences from the sea surface is pri­ pared to the in situ wind velocities at the are essential for evaluating long-term climate marily specular reflection—e.g., reflection of standard anemometer height in Figure 1 changes. The altimeter, which operates on microwaves from sea surface ripples that are [ Woiceshyn et al., 1986]. At high winds, scat­ the basis of specular reflections from ripples, small compared with the wavelength of the terometer returns increase linearly with the incident radiation. In their classical study of logarithm of wind velocity. However, this re­ Jin Wu, Air-Sea Interaction Laboratory, Gradu­ 1 ate College of Marine Studies, University of sun glitters on the sea surface, Cox and Munk lationship breaks down at about Uio< 6 m s" 1 Delaware, Lewes, DE 19958 [1954] calculated the variation of mean- for JASIN results, and at about Uio< 4 m s" for

This page may be freely copied. Vol. 76, No, 1, January 3, 1995

165E HOE HSE

358 Fig, I, Map of the areas surveyedduring the Geodynz-Sud cruise of the RIV L'Atalante along New Australian Platle Zealand's active mar­ gins with linearly color­ coded swath 408 bathymetry; red ar­ HP rows = PAC-AUS plate convergence direction [De Metset al., 1990]; HP=Hikurangi Pla­ teau; FM = Fiordland Chatham Rise Margin; sawtooth lines= subduction fronts; heavy lines= major strike-slip faults; light blue =continen­ Pacific Plate ta/rocks, depth <1,5 km; and dark blue = seafloor depth >1,5 km,

Page 1