3.5-kHz SUBSEAFLOOR REFLECTIONS

The 41-m borehole at Site 1223 recovered brown clay and turbidites in the upper 13 m and more consolidated vitric tuffs and claystones below. Based on the physical property measurements, the cored interval represents ~36-ms reflection traveltime, nearly the entire interval during which reflections were visible on the ship's 3.5-kHz echo sounder (Fig. F82).

Reflections from beneath the seafloor are commonly observed in sediments using echo sounders with transducers in the 1- to 5-kHz frequency range. Reflections from continuous interfaces as deep as 100 ms have been observed on traverses of many kilometers in some sediments. A sizable descriptive literature on seafloor sediment deposition has been developed using these reflection profiles. The assumption in these analyses is that the reflections occur at sharp boundaries between thick (in terms of wavelengths) layers. This is not likely the case at Site 1223.

Some of the constraints in correlating the soundings with the cored material are as follows:

1. A large area of the seafloor is insonified because of the wide beam width of the transducer and the 4230-m water depth.
2. The 43-cm wavelength of the sound does not readily discriminate between more closely spaced interfaces.
3. The chirp processor of the sounder has anticipatory and trailing side lobes, limiting resolution to ~3 ms.
4. On the 1-s window presented on the EPC graphic recording, the entire echo sequence observed is only about an inch wide.

Extensive reflecting interfaces were observed as we approached and while we were on the station at Site 1223. We continued to run the echo sounder during drilling operations, and we observed a temporal variability of the reflected returns. The variability of the reflected returns with the time of observation prompted our making a recording for 10 hr.

This variability is somewhat surprising considering the water depth (4235 m), the seafloor's modest slope, and the modest amount of ship movement. A 7-hr-long recording and its navigation files were examined. The time interval was 1258 to 2007 hr on 21 December. These are UTC times. Local time equals UTC time - 10 hr. The ship's position is maintained using a five-element receiving array below the hull and a Datasonics seafloor beacon emitting a stream of 16-kHz pulses. These and the Global Positioning System (GPS) data, taken at 1-min intervals, indicate the ship kept station within 20 m (Fig. F83). The GPS and acoustic positioning systems gave comparably precise measurements for the ship's position.

The Raytheon receiver processor applied a "slow" time-varying gain to the returning signal. The matched filter chirp processor gives rise to the anticipatory precursor to the seafloor reflection. The ship's heading was between 106° and 109° for the duration of these recordings. The periods of the ship's roll, pitch, yaw, reach, sway, and heave are all shorter than the 1- to 4-min changes in the reflection sequences observed.

Even with the variability observed, some observations common to much of the recording can be matched to the Hole 1223A summary shown in Fig. F2:

1. The seafloor reflection is followed by a relatively "acoustically transparent" interval of ~16 ms. Assuming a velocity of 1.7 km/s, this would be ~12 m thick. This appears to correspond to cored Units 1 to 4, which are volcaniclastic turbidites, dark-brown clay, and black sands. The acoustic impedance contrasts in these units are smaller than the seafloor and other interfaces.
2. Underlying this, there are a variety of reflections in the interval from 16 to 28 ms (~15 m thick). Several of these reflections are strong, but their appearance on the recording comes and goes with observing time. These probably correspond to the more indurated Units 5 through 7. The strength of these reflections suggests a greater impedance contrast, and the variability suggests that the interfaces may be of limited spatial extent.
3. Beneath this interval, there is a 6-ms (~6 m thick) interval in which there are only a few weak reflections.
4. The deepest reflections observed at 36 to 40 ms (32 m and deeper) are weak, and their appearance is discontinuous on the recording. These weaker reflections would correspond to the tuffs and claystones in the lower two cores.

It is possible that Units 8, 9, and 10 may come from the bottom of Core 200-1223A-3X and sit on top of Units 11 to 14. In this case, there may not have been any core recovery for an interval as thick as 13 m between 16 and 29 mbsf.

The wavelength in water of 3.5-kHz sound is 43 cm. The reflections seen are presumably from acoustic impedance contrasts comparable or greater in size.

A 212-ms-long traveltime window of the graphic recording, which includes the seafloor and subseafloor reflections, was imaged. Plots were made of the ship's position, and the incremental movement between fixes was calculated. The response time of the ship's dynamic positioning system is comparable to the 1-min fix rate; therefore, applying a smoothing function to the 1-min fixes would probably be inappropriate.

The graphic recording was examined, and the times at which noticeable changes in the reflection sequence occurred were noted and tabulated. The location of these changes were then compared with the positions (Fig. F84) of the ship during the whole period. There were no apparent groupings of the positions at which changes in the recording occurred. The incremental movement of the ship (Fig. F85) (average = 2 m) at these locations was comparable to or less than the incremental movement between successive fixes for the whole survey period.

For a water depth of 4235 m, the 40-ms time period of the seafloor and subseafloor returns could represent scattered returns from a circular area of the seafloor as large as 1400 m in diameter. Table T20 shows the size of the "footprint" from which returns could be received as a function of the length of the echo sequence. From this, the subseafloor volume that could return sound can be estimated using the velocity of sound in the substrate.

Preliminary conclusions are as follows:

1. The variability in the seafloor return does not appear to be an artifact of the ship's transducer, heading changes, the chirp correlator processor, or the graphic recording.
2. The seafloor turbidites have an acoustic impedance greater than seawater and less than the underlying tuff and act like a coated lens to increase the amount of sound reaching the deeper interfaces.
3. The seafloor in this area may be acting as a nonspecular reflector as evidenced by the intermittent nature of the reflection that was found in the indurated tuff. Viewing the seafloor and its underlying interfaces as a complex acoustic projector of the dimensions shown above, it would have to have a directivity of 0.2° or greater for the shipboard receiver to observe significant differences with a horizontal movement of a few meters. The diameter of the area that could return sound over the 40-ms interval observed is eight times greater than the minimum required to achieve that degree of directivity.
4. The subseafloor reflectors might be an interference from a sequence of reflectors spaced less than a wavelength apart (the dominant wavelength of the sound is 43 cm).
5. Small changes in the 3.5-kHz source frequency content or in the separation of a sequence of layers could make the apparent reflector disappear or move. A mechanism for producing a frequency difference in the source over the periods observed is not apparent. Although the flow of water past the 3.5-kHz transducer pod from thrusters three and four varies, the pod is a massive piece of steel and the surface waters of the ocean here are well mixed to >100 m depth by the recent storm waves.
6. A possible model for these observations might be numerous marbles or prisms disposed on an irregular acoustically transparent substrate. The buried extrusive volcanics found at this site could have such a complex shape. In addition to the dynamics of the tuff deposition, faulting and fracturing give rise to interfaces with varying shapes.