GEOPHYSICAL RESEARCH LETTERS, VOL. 19, NO. 1, PAGES 13-16, JANUARY 3, 1992 TI-IE SPREADING RATE DEPENDENCE OF THREE-DIMENSIONAL MID-OCEAN RIDGE GRAVITY STRUCTURE Jiarl Lin WoodsHole Oceanographic Institution Jason Phipps Morgan Scripps Institution of Oceanography Abstract. We analyze over 1300 km of highresolution along-axis gravity profiles at ridges withhalf-spreading rates ranging from 1.2 to 5.5 cm/yr. The results show consistently higher along-axis gradients of mantle Bouguer anomaly at the slow-spreading Mid-Atlantic Ridge(MAR) (0.3-1.2mgal/km) than at theintermediate- to fast-spreading Cocos-Nazca Ridge and East Pacific Rise (EPR) (0.1-0.2mgal/km). The regional peak-to-trough amplitude of mantle Bouguer anomaly is also greater alongthe MAR (30-60 mgal) thanthe Cocos-Nazca Ridge and theEPR (10-20mgal). With increasing spreading rate, the regional peak-to-trough amplitude of axial seafloor depth decreases from 1000-1700 m to 200-700 m. 3-D numerical experiments suggest that mantle contributions to the gravity can besignificant only near large-offset transforms. At the more commonlyobserved non-transform offsets,gravity anomalies will reflect crustal thickness variations. The along-axis gravity datathus indicate thatthe amplitude of along-axis crustal thickness variation decreases with increasing spreading rate. We propose that thisspreading rate dependent crustal accretion style mayoriginate in themantle: finite-amplitude mantle upwelling is intrinsically plume-like (3- D) beneath a slow-spreading ridge but moresheet-like (2-D) beneath a fast-spreading ridge. Sucha transition in mantle upwelling may occurif the relativeimportance of passive upwelling overbuoyant upwelling increases with increasing spreading rate. Small amplitude 3-D upwellings mayoccur at a fast-spreading ridge, buttheir effects on crustal thickness variations will be significantly reduced by along-axis melt flows along a persistent low-viscosity crustal magma chamber. In contrast, thelargecrustal thickness variations dueto 3-D mantle upwellings will be maintained at a slow-spreading ridge because less along-axis melt flows can occur in the colder and morerigid crust there. Data Analysis and Results In the past fewyears, a number of ge. ophysical surveys have mapped the gravity field over portions of mid-ocean ridges indense two-dimensional grids. To reveal more subtle density variations beneath the seafloor, these studies reduced gravity data to mantle Bouguer anomalies (MBA) by removing thegravity effects of seafloor topography, aswellasthe effects of the crust-mantle interface assuming a 6-km thick uniform crust (Figure la, Linet al., 1990; lb, Phipps Morgan and Kleinrock, 1991;1 c,Madsen et al., 1990). Tosearch for systematic along-axis variations inthe gravity and bathymetry, weextracted more than 1300 kmof along- axis MBA and seafloor depth profiles from data of Figure 1, and additional data from the MAR at 31-34øS [Kuo and Forsyth, 1988], 25ø-26ø30'S [Blackman and Forsyth, 1991], and 22ø30'-23ø30'N [Morris and Derrick, 1991 ]. The profiles in Figure 2 show a number of first-order features that are related to ridge segments. The positive MBA is typically greater, and the seafloor isdeeper, near first- and second-order ridge-axis discontinuities such as transforms, overlapping Copyright 1992 bythe American Geophysical Union. Paper number 91GL03041 0094-8534/92/91GL-03041503.00 spreading centers (OSC), and non-transform offsets. The seafloor depth profiles atall ridges "mirror" the MBA profiles. We define thehalf-segment amplitude asthedifference in MBA fromthesegment mid-point to a distal end. Alongthe MAR, half-segment MBA amplitudes correlate positively with half-segment lengths (Figure 3a). The half-segment anomalies at theMAR (5-50mgal) are consistently greater than those of the Cocos-Nazca Ridgeandthe EPR (5-20 mgal), although only a few data points are available for intermediate- and fast- spreading rates. To better quantify spreading-rate effects, we normalize thehalf-segment MBA amplitude by half-segment length. Results in Figure 3bshow a range of 0.3-1.2 mgal/km for the MAR spreading segments. These values are consistently greater than those of theCocos-Nazca Ridge and the EPR (0.1-0.2 mgal/km). Among MAR segments, greater along-axis changes occurnear large-offset transforms - gradients near the Cox, Rio Grande, Kane, and Atlantis fracture zones (FZ) all exceed 0.7 mga!/km. Regional amplitude is another measure of along-axis variability. We define regional amplitude as the peak-to- trough variation (40) of eachregional profile (Figure 4). Results in Figure 3c show that the regional MBA amplitude is consistently greater at the MAR (30-60 mgal) than at the Cocos-Nazca Ridge and the EPR (10-20mgal). Similarly, the regional amplitude of axial seafloor depth is greater at the MAR (1000-!700 m) than at the Cocos-Nazca Ridgeandthe EPR (200-700m) (Figure 3d). Thusbothalong-axis MBA gradients and regional amplitudes suggest strong along-axis variations in the density structure of slow-spreading ridges; these variations decrease with increasing spreading rate. GravityModels Wehave estimated specific mantle and crustal contributions to theMAR gravity.Based on a passive upwelling model [Phipps Morgan andForsyrh, 1988], we show that mantle effects due to thermal expansion predict only1/3of the MBA atthe Atlantis transform and verylittle anomaly at small offsets (Figure 5). The combined effects ofmantle thermal expansion andmelt extraction canexplain about1/2 of the Atlantis transform anomalies (Figure 5); butthepredicted anomaly is still too smallto explain that observed at non-transform offsets. Recent 3-D numerical studies of buoyant upwelling also predict little mantle anomaly at small ridge offsets [Parmentier and Phipps Morgan, 1990;Sparks and Parmentier, 1990]. These studiesshow that significant buoyant upwelling and melting can develop in a low-viscosity mantle without producing large direct gravity signals. Thus the large anomalies atMAR non-transform offsets will reflect crustal density structure. Theresidual anomaly at the MAR,if due solely to crustal thickness variation, isequivalent tonearly 50% reduction in the crustal thickness from segment center to distalends[Kuo and Forsyth, 1988;Linet al., 1990]. In contrast, the EPR anomalies (Figure lc,d) imply little along-axis crustal thickness variation. Wepropose that the contrasting crustal accretion styles at fast- and slow-spreading ridges may originate in the mantle: mantle hpwelling may have a 3-Dplume structure beneath a slow-spreading ridge but a 2-D sheet structure beneath afast- spreading ridge (Figure 6). The upwelling magnitude of a buoyant plume can be characterized by the terminal velocity of 13