Evidence for volcanic eruption on the southern Juan de Fuca ridge between 1981 and 1987

William W. Chadwick, Jr.

OSU/CIMRS, Hatfield Marine Science Center, Newport, Oregon 97365, USA

Robert W. Embley and Christopher G. Fox

PMEL/NOAA, Hatfield Marine Science Center, Newport, Oregon 97365, USA

Nature, 350, 416-418 (1991)
Copyright ©1991 Macmillan Publishers Ltd.  Further electronic distribution not allowed.

(Text of paper)

The lavas that were erupted during the 1980s occur as a series of pillow mounds and ridges between 45°00.5N and 45°09.5 N along the northern Cleft segment (Fig. 1). The evidence for the recent eruption of these lavas was first discovered through a discrepancy between microbathymetry derived from a towed camera system and bathymetric charts based on Sea Beam surveys. Three camera tows (collected in 1989 from the National Oceanic Atmospheric Administration (NOAA) ship Discoverer) over the southernmost pillow mound (mound 1, Fig. 1) showed glassy lava flows forming a 35-m-high hill, ~1 km in diameter. Earlier Sea Beam bathymetry (based on data collected in May 1981 from the NOAA ship Surveyor) shows no such hill, but instead a gentle slope to the east (Fig. 2a). The Sea Beam and camera-tow bathymetry correspond closely in other areas.

Fig. 1. Sea Beam bathymetric map of overlap region between Cleft and Vance segments, southern Juan de Fuca ridge (50-m contours); CS is northern end of Cleft segment, VS is southern end of Vance segment. Solid black areas show locations of new pillow lava mounds. Numbers to right of each mound are for identification in text. Box around mound 1 shows outline of Fig. 2. On inset map, GR is Gorda ridge, BFZ is Blanco Fracture Zone, and JdFR is Juan de Fuca ridge.

Other surveys confirm the recent appearance of mound 1. Bathymetry from Sea Beam surveys conducted over the same area by the research vessel Atlantis II in September 1987 and by the Discoverer in August 1990 shows a 25-m-high mound in the exact location of' the glassy lavas delineated by the camera tows (Fig. 2b). Mound 1 was also imaged by a SeaMARC I sidescan survey in August 1987 from the Discoverer, and appears as an unfractured positive relief feature which covers preexisting fractures (Fig. 3). The date of eruption of mound 1 is further restricted by a single Sea Beam swath collected during a camera tow by the Surveyor in June 1983. Although the data only sample the western edge of mound 1, the coverage is adequate to show that the 1983 swath has the same pre-eruption depth contours as the 1981 Sea Beam survey, and does not show the contours of the new lava mound. These Sea Beam surveys definitely constrain the time of the eruption of mound I to between 1981-1987, and probably to between 1983-1987.

Fig. 2. Sea Beam bathymetric maps (5-m contours) of area including southern-most new lava mound (mound 1 in Fig. 1). a, Pre-eruption bathymetry from survey in 1981. b, Post-eruption bathymetry from survey in 1990. Mound 1 is a new feature near centre of map. c, Map of the significant depth differences between 1981 and 1990 Sea Beam surveys (see text for discussion), showing outline and shape of mound 1. This corresponds exactly with the location of a mound of fresh, glassy pillow lavas based on camera tows, Alvin dives and sidescan sonar data (Fig. 3)

Fig. 3. Detailed map (outline shown in Fig. 2c) of mound 1 based on SeaMARC I sidescan sonar data, and the locations of flow contacts as surveyed by camera tows and Alvin dive 2262 (thin dashed lines). Note close correspondence with outline of mound 1 in Fig. 2c, derived independently fom Sea Beam data.

Further comparisons of the 1981, 1987 and 1990 Sea Beam data sets to the north of mound 1 reveal that several other mounds appear in the 1987/1990 data that were not present in 1981 (Fig. 1). No significant changes are evident between the 1987 and 1990 Sea Beam surveys, but the 1987 survey only covered mounds 1, 7 and 8, and not the intervening area, whereas the coverage of the 1990 survey was complete. Therefore the time of eruption of mounds 4 and 5 is only constrained to between 1981 and 1989 (the year of the camera tow). Mounds 2, 3 and 6 are glassy pillow mounds mapped by camera tows in 1989, but are apparently too small to be resolved by Sea Beam, so their eruption date can only be limited to before 1989. Most of the volume of new lava was erupted between 1981 and 1987; whether the features were constructed simultaneously or in discrete events within this time period cannot be determined from these data. The entire string of new eruptive sites lies roughly in the direction N 15°E over a length of 16 km. This line of recent volcanic activity is clearly along the north extension of the Cleft segment [Embley et al., in press], but it is also within the zone of overlap of the Cleft and Vance segments, and the new eruptions extend into the axial valley of the Vance segment (Fig. 1).

Significant changes in the sea floor between Sea Beam surveys can be inferred qualitatively from obvious changes in the shape of bathymetric contours. Figure 2a and b shows bathymetric contour maps made from 1981 and 1990 Sea Beam data, using identical mathematical gridding and automated contouring routines. Mound 1 can be distinguished as a new landform in Fig. 2b which is not present in Fig. 2a. We also used a quantitative technique to compare digital bathymetric grids from various Sea Beam data sets so that we could objectively identify areas of significant bathymetric changes. Figure 2c was produced by the following steps: (1) the 1990 Sea Beam grid was subtracted from the 1981 grid; (2) the differences were weighted in inverse proportion to the slope of the sea floor at each grid cell; (3) a final grid was created consisting of only those original (unweighted) depth differences for which the slope-weighted differences exceeded a preselected threshold. The differences must be weighted because a given error in location (due primarily to navigation error) will create a larger error in depth differences over a steep slope than over a gentle slope. For Fig. 2c, a threshold was chosen such that the outline of the area of significant differences best corresponded to the edge of the glassy lavas as determined from camera tows. As this threshold value is slope-weighted, it does not correspond to a single depth difference; the smallest depth differences above the threshold range from 5 to 15 m. We believe the technique to be reliable because in every case where camera-tow coverage is available, all of the new mounds defined by the Sea Beam difference grids occur precisely where camera tows indicate fresh glassy lavas. Mound 8 is just beyond the limit of the 1989 camera tows and will be the target of future camera surveys.

Because of its finite acoustic footprint and limited depth resolution, Sea Beam cannot detect all new lava flows. A few of the smallest patches of fresh lavas mapped by the towed camera system in 1989 do not produce a detectable change in bottom topography between the 1981 and 1990 Sea Beam surveys. Nevertheless, these lavas are identical in morphology and sediment cover to the other new lavas, and all the mounds occur along a linear trend. Therefore, we suspect all the glassy lava mounds were erupted during the 1981 to 1987 period.

A total volume for the new lava mounds of 0.05 km³ is calculated from the gridded Sea Beam differences. For comparison, this is similar to the volume produced during the 1977 eruption of Kilauea volcano, Hawaii (0.04 km³), which lasted 18 days [Moore et al., 1980]. During the 9-year rifting episode in northeast Iceland (1975-1984) [Björnsson, 1985; Sigurdsson, 1987], a total volume of 0.1-0.2 km³ of basalt was erupted in 9 out of 21 separate intrusion events; the volume of intruded dykes was much larger, ~1.0 km³.

Many apparently young, glassy lavas have been observed along the mid-ocean-ridge system, but it has not been possible to determine exact ages. Now we have a benchmark for the character of 'zero-age' lava. Bottom photographs taken in 1989 and observations ftom the submersible Alvin on dive 2262 in 1990 (Fig. 3) show that the new lavas are similar in appearance and sediment cover. Each new mound is primarily an accumulation of pillow lavas. The lava flows consist of high-standing bulbous and elongate pillows with corrugated surfaces, surrounded by cylindrical lobes with smooth glassy surfaces that resemble subaerial pahoehoe toes. The difference in microrelief of these two surface textures gives them a consistent difference in apparent sediment cover. The smooth pahoehoe toes have no sediment veneer, whereas small amounts of sediment are present along the extrusion marks on the bulbous pillows. Small sediment pockets between pillows are present in some areas. There is a sharp contrast in light reflectivity and sediment cover between the new lava flows and the surrounding older lavas.

In addition to the grey-brown pelagic sediment, there are two types of yellow deposits on the surface of the fresh lavas. One type consists of small globules of yellow material commonly located along small cooling cracks on individual pillows and ubiquitous throughout the mounds. This material seems to form from the direct interaction of sea water and hot lava. A review of the literature and data from other areas indicates that this type of deposit is not generally seen on older pillows and may be a good indication of very young submarine lava flows. The second type of deposit appears as a fine dusting of yellow-tinged material, and is commonly found surrounding small orifices near the edges of lava lobes, as if it had been deposited from fluids escaping from beneath. This type of deposit seems to be primarily associated with the crestline of the mounds and may be the product of low-temperature diffuse venting.

Camera tows photographed tube worm colonies living on older lava between the new mounds along a hydrothermally active fracture system [Embley et al., in press], and also found tube worms living on new lava on mound 1 (Fig. 3) and mound 2. This gives an absolute constraint on the maximum time necessary for the biological colonization of new vent sites (2-6 years on mound 1). The hydrothermal vents on the new lavas may be due to leakage of warm fluids through the permeable new lava mounds from the buried fracture system.

Our interpretation is that all the fresh lava mounds were erupted during one rifting episode, either in one event or in several events over a period of years, analogous to Icelandic volcanic systems [Björnsson et al., 1979]. The eruption clearly occurred along a prominent fissure system that is now hydrothermally active [Embley et al., in press], and it is reasonable to assume that the lava mounds were fed from, and are underlain by, a recent dyke intrusion. Monitoring of subaerial volcanoes suggests that the emplacement of this dyke produced horizontal extension across the northern Cleft segment of the order of metres and therefore represents an episode of active sea-floor spreading. No seismicity was detected associated with this eruption (R. Dziak, personal communication); however, all seismic data for the northeast Pacific is from land-based stations, and the detection threshold is m 4.0 (B. Presgrave, personal communication).

The Juan de Fuca megaplumes [Baker et al., 1989] were observed in 1986 and 1987, within the time interval for the eruption. This coincidence in time supports earlier models that megaplumes may be caused by sea-floor spreading [Baker et al., 1987; Baker et al., 1989; Embley et al., in press; Cann and Strens, 1989; Embley and Chadwick, 1990] and the conclusion [Baker and Lupton, 1990] that the 1986 megaplume was followed by degassing of fresh magma into the hydrothermal system. Horizontal extension caused by dyke intrusion is a reasonable mechanism for a sudden increase in the permeability of the crust in the vicinity of a pre-existing hydrothermal system leading to catastrophic release of hydrothermal fluids. Megaplumes may thus be indirect indicators of sea-floor spreading events.

Acknowledgments. D. Clague and J. Morton provided helpful reviews. This research was supported by the NOAA VENTS Program and the NOAA Undersea Research Program. Authorship of this paper is alphabetical.

Go back to abstract or forward to references

PMEL Outstanding Papers

PMEL Publications Search

PMEL Homepage