DISCUSSIONS AND CONCLUSIONS

During Leg 168, only Hole 1032A was logged. We used density and sonic velocity logs to generate synthetic seismograms and compare them with field seismic data, using a generalized nonlinear inversion scheme. We are able to correct the logs for poor hole conditions and to correct both continuous and discrete shipboard core measurements for elastic rebound. We find that both continuous and discrete laboratory measurements of bulk density after rebound correction agree well with the downhole density log. We also use the density logs to obtain pseudoporosity profiles. The intrinsic properties of solid grains and pore fluid are determined from shipboard core measurements of sediments. Using a theoretical model and these pseudoporosity logs, we derive pseudovelocity logs using the GNI inversion scheme to optimize a flexibility factor , which characterizes the softness of the formation. We find that the elastic rebound corrections made on laboratory measurements (IP porosity and DSV velocity) at shallow depths were underestimated and caution should be taken if these discrete measurements are to be used. This core-log-seismic integration at Hole 1032A provides confidence that shipboard continuous density records (GRAPE density) could be used to obtain high-resolution porosity-depth profiles at the other sites where no logs were available. The flexibility factor at each site is estimated through iterative comparison of synthetic and field seismograms using the generalized nonlinear inversion scheme. This enables us to obtain high-resolution velocity-depth profiles at all sites. Synthetic seismograms agree well with field seismic data not only in traveltime, but also in amplitude and waveform at all nine sites.

Through integrating core, log, and seismic measurements, we provide high-resolution velocity-depth and porosity-depth profiles for all the ODP Leg 168 sites. The high-resolution porosity-depth and velocity-depth profiles in the area can be interpolated to obtain a 2-D map of the depths to the sediment/basement contact. Further investigations of hydrological modeling and borehole observatory measurements could be aided using this information as input or constraints to construct accurate models of the sediment permeability and thermal conductivity structure in the area.

The flexibility factor we introduce in this simplified model provides a quantitative characterization of the softness or compaction state of the sediment column. For the Leg 168 transect, the values indicate that the sediment layers become more compacted as sediment thickness increases eastward and sediment ages become older. This qualitative observation between this flexibility factor and the sediment thickness and age may be expected; however, further investigation may unveil additional information from these profiles. The gamma value at Site 1032 is an exception to the general trend, because it is lower than values at Sites 1026 and 1027.

In this report, we did not use the seismic data to constrain the porosity-depth profiles from GRAPE density data. This could be done by a simultaneous optimization of both velocity and density profiles using more powerful, though more complicated, optimization schemes. Another important issue is the scale difference in bridging core, log, and seismic data. The core GRAPE density measurements and log data have a spatial resolution on the order of centimeters, whereas seismic resolution is on the order of 10 m, depending on the frequency. Although fine structures (<1 m thick) do not affect the traveltime of low-frequency seismic waves, a thin layer may strongly affect the amplitude and composite waveform of a seismic wavelet propagating through it. The mismatch in waveform and amplitude between the synthetic seismogram and field seismic data at Site 1027 is attributed to such thin layers, notably where high-resolution core information is missing. The noise in the field seismic data also affects the comparison of these results. Fortunately, the signal/noise ratio of the single-channel seismic data near the Leg 168 sites is quite acceptable for this comparison. It is expected, however, that an even better comparison could be obtained if seismic filtering had been applied to the data. This is not done, however, because filtering also distorts signal, especially at low and high frequencies that are vital to correlating the composite waveform and amplitude to core and log data.

We propose a velocity-porosity relationship for marine sediments along this 100-km-long transect of the Juan de Fuca Ridge. Compared with this relationship, Wyllie's time-average relation overestimates the velocity and Wood's equation underestimates the velocity at intermediate and low porosity. Wyllie's equation works well for intermediate porosity range and can be viewed as based on ray theory, although without a physical basis. Wood's equation is based on acoustic wave propagation and can be derived from the proposed model with some simplifications. Although the model parameters are site dependent, this relationship should be suitable for marine sediments in other similar environments. To a certain extent, this model can be viewed similarly to the Gassmann model, so that the results may be comparable with that obtained by Hamilton (1971) and Jarrard et al. (1989). However, their empirical fits for the frame bulk and shear moduli have only limited applicability. Here the frame moduli has a theoretical basis and is valid, in principle, for the full porosity range from 0% to 100%. In view of the assumptions used to derive these equations, the frame elastic moduli may not be adequate for all cases, but should have the same form as the general moduli discussed in previous sections. Further validation of this model and the methodology for core-log-seismic integration should be extended to sediments in other diverse environments and at other Deep Sea Drilling Project (DSDP) and ODP sites. In particular, the Leg 146 sediments in the Cascadia Basin, Leg 105 (Norwegian Sea), Leg 150 (New Jersey Margin), and other continental margin sediments may have adequate laboratory, log, and field seismic data to test the model without additional ship time.