The high-resolution seismic-reflection profiles were collected to delineate the sedimentary environments and facies beneath the lake and to define how those environments have responded to climate change. In addition, the selection and correlation of core sites for high-resolution paleoclimatic records were critically dependent on the seismic-reflection data, because of the complex facies distribution of sediments in Lake Baikal and the complicated pattern of recent faulting, erosion, and mass movement.
All seismic, navigation, and related equipment were provided by the USGS. All three expeditions were conducted aboard the R.V. Vereshchagin, which belongs to the Limnological Institute in Irkutsk. Because of the nature of the paleoclimatic records that were sought, the seismic-reflection surveys were focused on selected parts of the lake, primarily the Selenga Delta, Academic Ridge, and the North Basin (northern and central portions). The purpose of this paper is to describe the data collection and processing methods that we used and to provide examples of the data.
During the first (1990) expedition, the USGS used two primary systems for subbottom seismic-reflection profiling. One system was an ORE 3.5 kHz sub-bottom profiling system, which consisted of a towed four-transducer array that served as both the sound source and receiver, and an ORE Model 140, 10 kw transceiver. This data was graphically recorded on an EPC 4800 plotter and was recorded on a Hewlett Packard 8 track tape deck for playback capabilities and archiving. In most water depths, subbottom reflections were successfully acquired using this technique. However, over the deepest part of the lake (more than about 1400 m), the seismic data became somewhat distorted and weak due to several contributing factors, including signal attenuation in the water column, loss of energy by reflections from thermocline layers, and the limited beam-width output from the four-transducer array. During the 1991 and 1992 expeditions, the 3.5 kHz data acquisition system remained the same. However, during the 1992 expedition, two of the four transducers failed, so that adequate subbottom records could only be obtained in water depths shallower than about 500 m.
The second system used in 1990 was a Huntec shallow-towed, broad-band electro-mechanical "boomer system" with a surface-towed Benthos 100-element hydrophone streamer. The boomer system proved to be inadequate at maximum power output levels (one kilojoule (kj)), apparently due to signal attenuation in the water column and to high levels of ship noise in the frequency band of the boomer source.
After reviewing the results of the 1990 field expedition, we decided to use a 15-cubic-inch water gun and 200-element hydrophone acquisition system for subsequent trips to the lake. An Atlas Kopco electric compressor was used to provide 3000 pounds per square inch (psi) of air pressure at 20 standard cubic feet per minute (scfm) for the sound source. In discussions with Mr. Jim Hedger of Seismic Systems Inc., Houston Texas, it was decided that the best sound-source system configuration for Lake Baikal would be to operate the compressor to provide a pressure of 3000 psi (instead of the more commonly used 2000 psi) and to modify the ports of the water gun. The time necessary to increase the pressure in the water gun required that it be fired only every seven seconds, rather than every three or four seconds. The modification of the water gun ports consisted of using only the side ports of the gun, so that it would not expel water upward through the surface of the lake; the result is increased power and frequency content. This technique was originally developed for shallow-water, gas-charged sediments in the swamp areas of Louisiana and Texas. By towing the instrument at 15 inches below the water surface and firing it at 3000 psi, utilizing the special ports, it is possible to reach an upper frequency of 3000 Hz, with an increase in power output of 50 percent, from 1.0 to 1.5 kj.
However, because of the ship's excessive speed at low rpm (ca. 10-11 km/hr), the 15 inch sub-surface tow depth for the sound source could not be used. Instead, the sound source was towed at 45 inches sub-surface. This tow depth was chosen to minimize the bubble pulse of the gun and to reduce interference of the reflection from the water surface ("surface ghost"). A custom-made, 200 element, single-channel hydrophone was built using Benthos AQ-4 hydrophones spaced approximately 18 inches apart. This combination of water gun and 200 element hydrophone proved to be very successful. Due to the background noise from the ship and the water depths of the lake, the offset of the sound source from the head of the hydrophone array was set at a relatively high value, on the order of 180 meters. The water-gun data was graphically recorded on an EPC 4800 plotter and stored on a Hewlett Packard 8-track analog tape deck.
The analog field data from the water-gun system were of limited use, because the seven-second firing rate resulted in field data that was highly compressed in a horizontal direction, resulting in excessive vertical exaggeration. For this reason and for further digital processing, all of the water-gun data were digitized in the USGS, Woods Hole, laboratory using a Masscomp computer system. The data were digitized from the analog tapes using a sampling rate of 0.5 ms for the first 3 s of each shot, and the digitized data were stored in SEGY format. For routine plotting purposes, the data were subject to a digital band-pass filter, typically 30 to 500 Hz, and an exponential time-varying gain function. The data were plotted using a variable-area display with horizontal and vertical scales that resulted in a vertical exaggeration of about eight to one. This vertical exaggeration was a vast improvement over the original field records, which had a vertical exaggeration of 50-60 to one.
For short segments of some profiles, we have experimented with other types of processing, including use of a spherical divergence correction and weighted moving averages of shot traces. We have also used wiggle-trace displays of the data, and have made relative amplitude and phase plots. However, for most interpretive work, we have used the routine processing steps discussed above, which were used to produce the image files included in this report.
The 3.5 kHz profiles are of generally of excellent quality. They typically provide vertical resolution of less than 0.5 m and subbottom penetration of 30 to 50 m in mid-range (500-1000 m) water depths. Under favorable water-depth and sediment conditions (soft, fine-grained deposits), subbottom penetration was well in excess of 50 m. Difficulties in sediment penetration were encountered only in the deepest parts of the lake (>1200 m) and where the lake floor was underlain by sandy sediments. With only two of the four transducers functional in 1992, data quality in water depths less than about 500 m was only slightly diminished; this suggests that the primary limitation on the system was power to overcome signal attenuation in the water column.
The water-gun data were also of excellent quality, but they required processing to achieve their full utility. Typical sediment penetration in mid-range (500-1000 m) water depths was 200 to 400 m. Vertical resolution was less than one meter. Signal attenuation and sediment type affected subbottom penetration, but under favorable conditions, penetration of 500 to 600 m was achieved. Excellent records were obtained from the floor of the North Basin of the lake, which is underlain by sandy turbidites, indicating that coarse-grained sediments had less of an effect on the quality and penetration of the water-gun data than they did for the 3.5 kHz data.
Both types of seismic-reflection data have been useful for defining and mapping sedimentary environments in the lake (Colman et al., 1992, 1993), which include delta fronts, pro-delta areas, isolated areas of pelagic sedimentation, turbidite fans, and basin plains. These data have also been critical for locating the best sites for cores that provide a record of paleoclimate and paleolimnologic change (Lake Baikal Paleoclimate Project Members, 1992). Our seismic-reflection data were also used to choose the site for the ice-based drilling of two 100-m core holes by the Baikal Drilling Project in March of 1993.
Seen here in a satellite photo, Lake Baikal lies within an active continental rift, which is the target of ongoing studies of deep multichannel seismic-reflection studies (Hutchinson et al., 1992; Colman et al., 1993). Many of the structures seen on multichannel seismic profiles can be correlated with structures that appear in the high-resolution seismic profiles. The high-resolution profiles clearly show faults that cut the lake floor and the sedimentary effects of rapid rift subsidence. Thus, the high-resolution and multichannel data form complimentary sets for studying tectonic and kinematic questions related to active rifting.
Back 
Continue
[an error occurred while processing this directive]