U.S. Dept. of Commerce / NOAA / OAR / PMEL / Publications
A global review of the distribution of hydrothermal plumes [Baker et al., 1995] found a significant linear correlation between ph and us extending from spreading rates of 20 to 150 mm yr1. Baker et al. [1995] emphasized that because of the highly variable distribution of venting among adjacent segments, this relation depended on averaging hydrothermal activity over many segments, typically over distances of several hundred kilometers. As is apparent from Figure 4e, there is little correlation between ph and us if only individual second-order or smaller segments of different us are considered.
Averaging the Axs and Enet data regionally for the three study areas considered here also yields linear correlations with ph (Figure 9a). The uniformity of the three trends in Figure 9a indicates that over broad spatial scales, and by implication over broad temporal scales as well, both low-frequency (us) and high-frequency (Axs and Enet) indexes of the magmatic budget correlate linearly with the relative incidence of hydrothermal activity in these three areas. Regional means of MgO wt% and AMC percent coverage do not correlate significantly with ph.
Figure 9. (a) Scatterplot showing that at the regional scale, plume incidence ph is linearly correlated with mean values of spreading rate us, cross-sectional area Axs, and net elevation Enet. (b) Plot of Axs versus us. (c) Plot of Enet versus us. The linear relation in Figure 9a exists because for the three areas studied (solid symbols) both Axs and Enet are linearly correlated with us. This relation does not hold for other similar-sized regions of the superfast spreading southern EPR.
How general is the agreement between the low- and high-frequency indexes in Figure 9a? If us, Axs and Enet are correlated all along the EPR, then a prediction of the large-scale distribution of venting along the entire EPR may be straightforwardly made using any of these indexes. To test this possibility, Figures 9b and 9c show mean values of Axs and Enet against us for the JDFR and discrete units of the EPR. The northern EPR, 5°18°N, is quantized into 3° units to match the 9°12°N study area, while the southern EPR, 4°33°S, uses approximately 5° units to match the 13.5°18.5°S study area (the Easter microplate is not considered). For full spreading rates of 60 to 120 mm yr1 the correlation between us and Axs is nearly perfect (r2 = 0.9995). Moreover, the least squares fit for intermediate to fast rates extends precisely through the value for the 13.5°18.5°S superfast area. Elsewhere on the southern EPR, however, the inflation values are well below that predicted from this fit. At the extreme, the 18.5°23°S section, with a spreading rate virtually identical to the 13.5°18.5°S section, has a mean Axs only one-third as great, less than every section except the JDFR. The situation for Enet may be similar, except that the fit is poorer for the northern EPR between us of 80120 mm yr1 (Figure 9c). Alternatively, Figure 9c can be interpreted to demonstrate an effectively constant Enet except at extreme high and low values of us.
Figure 9 demonstrates that at the regional scale we presently have insufficient data to deduce whether a low-frequency (us) or high-frequency (Axs, Enet) index of the magmatic budget is the more general predictor of hydrothermal activity. If us, then virtually all large sections of the southern EPR should have a ph > 0.5. If Axs or Enet, the 13.5°18.5° section will have by far the largest ph, and the superfast section immediately south will have a ph lower than any other section of comparable size on the EPR.
Geologic predictors of hydrothermal activity on slow spreading ridge segments will be much different than those for faster ridges. The indexes Axs and Enet will not have the same interpretation for slow spreading ridges because axial topographic relief changes from positive to negative at spreading rates below ~5580 mm yr1, depending on the axial depth [Malinverno, 1993]. AMC percent coverage is a similarly ineffectual index because a reflector has been reported at only a single isolated location on the Mid-Atlantic Ridge (MAR) [Calvert, 1995]. Finally, not only is there no broad correspondence between hydrothermal observations and detailed petrological sampling along the MAR, but existing data suggests that because lavas on slower spreading ridges have more chemical diversity than those on faster ridges, a general relationship between MgO wt% and axial depth is not likely [e.g., Batiza, 1996].
While no thorough review of geologic indexes for slow spreading ridges will
be attempted here, there are two obvious possibilities for hydrothermally diagnostic
indexes: mantle Bouguer anomalies (MBA) and tectonization of the seafloor. Recent
geophysical surveys along the MAR find that the midpoints of tectonic segments
usually correspond to MBA gravity lows or "bull's eyes," interpreted
as arising from focused accretion of new ocean crust [e.g.,
Lin and Phipps Morgan, 1992; Detrick
et al., 1995]. Along the northern MAR, vent fields at 23°12
N
(Snake Pit) [Ocean
Drilling Program Leg 106 Scientific Party, 1986], 29°10N
(Broken Spur) [Murton
et al., 1994], 36°27N (AMAR) [German
et al., 1996], 37°17N (Lucky Strike)
[Fouquet
et al., 1994], and 37°50N (Menez Gwen)
[Fouquet
et al., 1994] all occur at segment bathymetric and MBA minima. Though
a gravity survey has not been published around the TAG field at 26°10N
[Rona
et al., 1986], it too is found near a segment bathymetric minimum [Karson
and Rona, 1990]. Crustal accretion at segment centers apparently results
from enhanced magma upwelling, and these areas may be focal points of hydrothermal
activity.
A contrasting model was offered by German et al. [1996], who found that five of seven hydrothermal sites identified by plume surveying on the MAR between ~36° and 37°N occurred in highly tectonized areas at or near the ends of second-order ridge discontinuities. They suggest that the dominant control on vent field location is not the crustal thermal field but the presence of cross-cutting fault fabrics at segment termini. These fault zones allow access to the deeper heat sources characteristic of slow spreading ridges [e.g., Purdy et al., 1992]. While faulting is clearly a fundamental control on vent field location on both slow [e.g., Karson and Rona, 1990] and fast [e.g., Wright et al., 1995] spreading ridges, the relative importance of thermal and permeability factors remains uncertain.
Return to previous section or go to next section