GEOLOGIC SETTING

Escanaba Trough forms the southernmost segment of the Gorda Ridge (Fig. 1), extending 130 km northward from the Mendocino Fracture Zone at about 40°25€N latitude to a right-lateral offset near 41°35€N. The total spreading rate along the Escanaba Trough segment is ~2·3 cm/yr (Atwater & Mudie, 1973; Riddihough, 1980), which is characteristic of a slow-rate spreading center. Most of the rift valley is bordered by normal-faulted rift mountains up to 1500 m high. The valley floor is 3-5 km wide at the northern end and widens to 18 km near the junction with the Mendocino Fracture Zone. South of about 41°08€N latitude, the axis of Escanaba Trough is covered by turbidite and hemipelagic sediment (Moore & Sharman, 1970; Vallier et al., 1973; Karlin & Lyle, 1986; Morton & Fox, 1994). Terrigeneous sediments, originating at the North American continental margin, enter the valley at the southern end adjacent to the Mendocino Fracture Zone because topographic barriers in the form of rift mountains are absent in this region. Deep Sea Drilling Project (DSDP) Site 35, drilled slightly off axis in the central Escanaba Trough (Fig. 1), penetrated 390 m of Pleistocene turbidite sediment but did not reach volcanic basement (McManus, 1970).

Figure 1. Bathymetric map of Escanaba Trough showing sample locations and DSDP Site 35. NESCA and SESCA sites are indicated. Inset map shows location of oceanic ridge segments in the northern Pacific Ocean.

Within the trough, discrete volcanic edifices uplift, and locally disrupt, the sediment cover (Denlinger & Holmes, 1994; Morton & Fox, 1994). Mesa-like hills, ~70-100 m high and ~1 km in diameter, are thought to be formed by laccolithic intrusions at, or near, the sediment-basement interface, 400-600 m below the sea floor (Denlinger & Holmes, 1994). Deformation of sediment generates sets of imbricated thrust and normal fault scarps exposing layers of sediments at the flanks of the hills (Zierenberg et al., 1994). Both thrust and normal faults channel hot seawater that precipitates large deposits of massive sulfide near the base of the escarpments (Denlinger & Holmes, 1994; Zierenberg et al., 1994). Apparently, uplift pre-dated surface volcanic activity and basaltic feeders also followed fault traces to the sea floor, where eruptions produced lava flows on top of the sediments. These flows have been sampled in the sediment-covered southern Escanaba Trough.

Two of the larger volcanic sites were studied in detail by Morton & Fox, (1994). The southernmost of these two sites, referred to as SESCA, is located at about 40°45€N, and the other site, referred to as NESCA, is located at about 41°00€N (Fig. 1). Seismic data indicate that the volcanic edifices at NESCA and SESCA have both intrusive and extrusive components and the sediments above the intrusions have been uplifted as much as 100 m in some places (Fig. 2). At both sites, exposed basaltic rocks are predominantly pillow basalts, although flat, folded, sheet and lobate flows also occur (Ross & Zierenberg, 1994). Flows typically appear to originate near the base of the uplifted sediment hills. Fissures and faults cut the flows and some fault scarps are in turn covered by basalt flows, indicating contemporaneous faulting and extrusion (Ross & Zierenberg, 1994). The flows atNESCA appear to be of similar age (Ross & Zierenberg, 1994), <10-20 ka based on ²³⁸U-²³⁰Th studies(Goldstein et al., 1992). Volcanic rocks were dredgedfrom four additional, essentially sediment-free, sitesnorth of NESCA, which we refer to as northern sites.

Figure 2. Schematic cross-section showing uplifted sediment hill in Escanaba Trough as a result of magma intrusion. Faults provide pathways for basalt flows and for hydrothermal fluids, which form massive sulfide deposits at the base of the hill. Modified from Morton & Fox, (1994).

SAMPLING AND ANALYTICAL METHODS

Lava samples were recovered by dredging on R.V. S. P. Lee cruises L6-85NC, L2-86NC and L3-86NC, R.V. Kana Keoki cruise KK2-83NP, and by submersibles Alvin and Sea Cliff. Sediments were sampled using 3-m gravity cores on R.V. S. P. Lee cruises L5-86NC, L2-86NC, and L3-86NC. Forty-eight glassy rims of pillow and sheet flow fragments were selected for analysis by electron microprobe at the US Geological Survey in Menlo Park. The individual analyses and sampling details have been given by Davis et al., (1994). The 48 glasses were averaged into flow units if all elements were within analytical precision (Table 1). Selected glass samples were analyzed for chlorine using a sodalite standard (6·82% Cl). Analytical precision is 1-2% for major elements and 3-5% for minor elements. Precision for trace elements such as S and Cl is only about 10% and 30%, respectively, at these low concentrations. The interior of one basalt (Alv-2039-2) and sediments from Escanaba Trough were analyzed for major and some trace elements by X-ray fluorescence (XRF) in the Analytical Laboratory of the US Geological Survey in Denver.

Table 1. Location and composition of Escanaba basalts

Rare earth elements (REEs) on a representative set of basalt and sediment samples were analyzed by inductively coupled plasma mass spectrometry (ICP-MS) in the same place. Methods, precision and accuracy have been described by Lichte et al., (1987). Six samples were analyzed for ⁸⁷Sr/⁸⁶Sr, ¹⁴³Nd/¹⁴⁴Nd and Pb isotopic composition at Cornell University. Precision and accuracy of the methods have been given by White et al., (1990, and references therein). Oxygen isotope analyses were performed on eight selected samples at the US Geological Survey in Menlo Park. Oxygen was extracted by reaction with ClF₃ at 550°C in nickel bombs using methods similar to those described by Clayton & Mayeda, (1963). All oxygen isotope measurements were performed in duplicate and values are reported as permil deviation relative to SMOW (Standard Mean Ocean Water).

ESCANABA BASALT

Petrography

All samples appear fresh with mm- to cm-thick glass rinds; palagonite is thin or absent. One sample recovered by Alvin (Alv2039-2) has baked clay selvages on the glassy rinds. This glass is somewhat darker in color and less transparent than the other samples but shows no other visual evidence of devitrification. Except for plagioclase-phyric (>20%) samples from the northern sites (83-16, 3-86-10), samples are nearly aphyric or sparsely porphyritic with euhedral plagioclase and olivine microphenocrysts. Modal mineralogies of representative samples are included in Table 1. Minute (<200 µm), euhedral to subhedral spinel crystals occur in most samples but are especially abundant and in equilibrium with their host glass in NESCA samples. Olivine and plagioclase are texturally and compositionally similar to those from the northern Gorda Ridge (Davis & Clague, 1987, 1990) but spinel in NESCA samples is higher in TiO₂ and ferric iron (Fig. 3; Davis et al., 1994) than spinel from the northern Gorda Ridge or typical N-MORB spinel from other slow-spreading centers (Sigurdsson & Schilling, 1976). Plagioclase in the high-K₂O sample (Alv2039-2) has measurable K₂O of ~0·06% (Davis et al., 1994), compared with other plagioclase crystals from Gorda Ridge basalts, which have K₂O at or below detection limits (Davis & Clague, 1987, 1990).

Figure 3. Representative compositions of chromian spinel in Escanaba basalts show that spinel in NESCA basalts ([squf]) has higher calculated ferric iron than typical for spinel in basalt from the northern sites in Escanaba Trough ([utrif]) and in other Gorda Ridge basalts ([circle]). Data from Davis & Clague, (1987, 1990).

Major element chemistry

The basaltic glasses from Escanaba Trough range from relatively primitive {mg-number [100Mg/(Mg + Fe²⁺)] ~67} to moderately fractionated (mg-number ~56, Table 1). The compositional range is similar to, but slightly less than that in basalt from the northern Gorda Ridge (mg-number 69-55; Davis & Clague, 1987). All glasses, except one, are low-K₂O N-MORB (Fig. 4a). Sample Alv2039-2 has a K₂O content of 0·43, is much higher than is typical for N-MORB and is the first such enriched composition reported from the Gorda Ridge. Besides a slightly elevated P₂O₅ content and lower, but variable, Na₂O abundance, this sample appears ordinary with respect to the other elements analyzed. As this sample has baked clay selvages on the glassy rind, a piece of the interior of the pillow was analyzed by XRF to determine if the high K₂O content is due to incipient secondary alteration. However, the interior of the pillow has only slightly lower K₂O content of 0·36% (Table 1) and the higher K₂O contents of the plagioclase confirm that the higher K₂O content is magmatic in origin. Although the other glasses have K₂O contents within the range typical of normal MORBs, at least two groups with distinct K₂O contents are apparent (Fig. 4a). All samples at the NESCA site have higher K₂O contents than i