FLOW LINE VARIATIONS IN ABYSSAL HILL MORPHOLOGY FOR THE PACIFIC-ANTARCTIC RIDGE AT 65-DEGREES-S

We present the results of a statistical study on the morphological characteristics of abyssal hills recently mapped along two adjacent segments of the Pacific-Antarctic Ridge at 65 degrees S. The studied area is a densely surveyed corridor (60 km wide by 600 km long) which is centered on the Pitman Fracture Zone (PFZ) and extends to 12 Ma crust on both sides of the ridge. Abyssal hill size parameters (RMS height H and characteristic width lambda) are estimated using Hydrosweep multibeam data. Variations in abyssal hill characteristics are compared with spreading rate history and crustal structure (as inferred from the mantle Bouguer gravity) in order to indirectly quantify the evolution of this ridge crest system. The magnetic data document an abrupt acceleration in spreading rate from similar to 36 to similar to 63 mm/yr (full rate) at Chron 3a (5.7-6.4 Ma). Our results indicate a statistically significant negative correlation between abyssal hill size parameters and full spreading rates. Abyssal hills formed during the slower spreading period (ages >8 Ma; full rates 36-44 mm/yr) are 31-86% taller and 21- >100% wider than hills created during the faster spreading interval (ages <4 Ma; full rates 52-63 mm/yr). The well-resolved positive correlation between H and lambda is interpreted as an indication of temporal changes in the flexural rigidity of the lithosphere near the vicinity of the ridge crest and, by implication, axial thermal structure. However, we cannot rule out that such positive trend is due to constructional volcanism. The lack of correlation between crustal thick-ness and abyssal hill size parameters is likely to be caused by the small magnitude of crustal thickness variations along flow lines (similar to 0.4 km in contrast to similar to 2 km reported in previous studies for the Mid-Atlantic Ridge). The most significant variations in crustal thickness are seen across the PFZ (thinning from north to south by 0.5-0.7 km), which coincide with a well-resolved increase in the averaged lambda estimate. The predictions of the detachment surface model in terms of morphological and structural inside/outside corner asymmetries are not supported by our observations. The main variations in H and lambda that cannot be explained in terms of either the spreading rate or crustal thickness effect include the following: (1) anomalously large abyssal hills north of the PFZ for 4-6 Ma age crust; (2) abyssal hill size estimates for crustal ages greater than 8 Ma show significant asymmetry for opposite ridge flanks north of the PFZ; and (3) toward the segment ends, H estimates are 27-68% larger, while lambda estimates either do not significantly change (to the north of the PFZ) or are up to 40% smaller (to the south of the PFZ). We suggest that the H and lambda changes seen toward the segment ends are related to either an increase in the amount of extension (without a corresponding increase in the strength of the lithosphere) or variations in the relative contribution of constructional volcanism to overall abyssal hill morphology.