VARIABILITY OF THE AXIAL MORPHOLOGY AND OF THE GRAVITY STRUCTURE ALONG THE CENTRAL SPREADING RIDGE (NORTH FIJI BASIN) - EVIDENCE FOR CONTRASTING THERMAL REGIMES

The Central Spreading Ridge (CSR) is located in the central part of th e North Fiji Basin, a complex back-are basin created 12 Ma ago between the Pacific and Indo-Australian plates. The 3.5 Ma old CSR is the bes t developed, for both structure and magmatism, of all the spreading ce nters identified in the basin, and may be one of the largest spreading systems of the west Pacific back-are basins. It is more than 800 km l ong and 50-60 km wide, and has been intensively explored during the Fr ench-Japanese STARMER project (1987-1991). The CSR is segmented into t hree first order segments named, from north to south, N160 degrees, N1 5 degrees and N-S according to their orientation. This segmentation pa ttern is similar to that found at mid-ocean ridges. The calculated spr eading rate is intermediate and ranges from 83 mm/yr at 20 degrees 30' S to 50 mm/yr at 17 degrees S. In addition, there is a change in the axial ridge morphology and gravity structure between the northern and southern sections of the CSR. The axial morphology changes from a deep rift valley (N160 degrees segment), to a dome split by an axial grabe n (N15 degrees segment) and to a rectangular Bat top high (N-S segment ). The Mantle Bouguer Anomalies obtained on the northern part of the C SR (N160 degrees/N15 degrees segments) show ''bull's eye'' structures associated with mantle upwelling at the 16 degrees 50' S triple juncti on and also in the middle of the segments. The Mantle Bouguer Anomalie s of the southern part of the ridge (N-S segment) are more homogeneous and consistent with the observed smooth topography associated with ax ial isostatic compensation. At these intermediate spreading rates the contrast in bathymetry and gravity structure between the segments may reflect differences in heal supply. We suggest that the N160 degrees a nd N15 degrees segments are ''cold'' with respect to the ''hot'' N-S s egment. We use a nonsteady-stare thermal model to test this hypothesis . In this model, the accretion is simulated as a nearly steady-state s eafloor spreading upon which are superimposed periodic thermal inputs. With the measured spreading rate of 50 mm/yr, a cooling cycle of 200, 000 yr develops a thermal state that permits to explain the axial morp hology and gravity structure observed on the N160 degrees segment. A s preading rate of 83 mm/yr and a cooling cycle of 120,000 yr would gene rate the optimal thermal structure to explain the characteristics of t he N-S segment. The boundaries between the ''hot'' N-S segment and its ''cold'' bounding segments are the 18 degrees 10' S and 20 degrees 30 ' S propagating rifts. A heat propagation event along the N-S segment at the expense of the adjacent colder failing segments, can explain th e sharp changes in the observed morphology and structure between the s egments.