RESULTS OF INITIAL STUDIES

The pilot studies described below were performed for two basic reasons: (1) to demonstrate that LAS is indeed a useful minerology indicator, and (2) to investigate sources of spectral distrubance, which could result in minerology calculation errors.

We prepared four end point mineral standards (opal, calcite, smectite, and illite) and 50:50 mixes of these end points. Our "opal" standard is a diatom ooze, and our "calcite" standard is a nannofossil ooze; both are from intervals that are moderately pure. These samples were measured using the instrumentation setup shown in Figure F1B.

Figure F2 shows LAS spectra for these mineral standards. The four end points have distinctively different spectra, each with diagnostic features. As expected based on LAS results for other minerals (Clark et al., 1990), the 50:50 mixes have generally intermediate spectral characteristics, but the spectral responses are nonlinear. In particular, LAS spectra are more sensitive to opal and smectite than to calcite and illite. Consequently, we anticipate that LAS resolution will be best for calcite and illite at high concentrations and best for opal and smectite at lower concentrations. Nonlinear responses necessitate use of several intermediate mixes for calibration of spectral responses.

Our second pilot study is a simple reconnaissance to determine concentrations of paleoclimatically significant minerals. To a first approximation, near-equatorial Pacific sediments can be thought of as consisting of calcite (nannofossils and/or foraminifers), opal (radiolarians and/or diatoms), and terrigenous components. We took 24 samples from Site 846, an eastern equatorial Pacific site in which the dominant components are calcite and opal. Samples were chosen at the same locations (±3 cm) as ones previously analyzed by Mix et al. (1995) for calcite, opal, and "other" (other = 100% - %calcite - %opal). Opal concentrations may be slightly underestimated because of incomplete opal dissolution (Farrell et al., 1995). These ground-truth measurements were used to quantify the responses of individual spectral features. We determined calcite based on the calcite absorption band at ~2350 nm. For opal, we used three spectral characteristics (depth of the 1900 nm water trough, drop between 1300 and 900 nm, and drop between 900 and 400 nm), recognizing that each could also be affected by a different component and would, therefore, be most accurate and useful in different environments.

This feasibility study demonstrates that LAS can achieve an accuracy of ±10% for calcite and opal determination at this site (Fig. F3). This accuracy is not directly portable to sites with quite different concentrations of calcite, opal, and other components, but we expect the same approach of local ground truth for LAS results to succeed in many other environments. The second part of this paper focuses on a similar, but more sophisticated, analysis technique (e.g., inversion) for determining mineralogy.

Clay minerals deposited in the Pacific Ocean are mainly derived from wind-blown continental dust (Rea, 1994). As dominant winds tap different source areas, different clay species are deposited in distinct areas of the ocean. A small pilot study was conducted to see if these different clays could be detected by the LAS technique. Clay concentrations were found by subtracting known concentrations of calcite and opal from 100% (Olivarez Lyle and Lyle, this volume).

Three high clay-content samples were chosen from the EW97909 cores (the site survey cores for ODP Leg 199). Sample A₁ and A₂ were taken near the top of core EW97909-21GC (25°N, 147°W) and contain 89% and 88% clay, respectively. Sample B was taken near the top of core EW97909-12PC (5°N, 140°W) and contains 80% clay (Fig. F4). XRD and LAS analyses were performed on all three samples. The XRD patterns were collected on a Rigaku diffractometer, using CuK radiation at 30 mA and 25 kV with a step size of 0.05°2. After the three clay samples were analyzed, they were exposed to ethylene glycol for 24 hr and then reanalyzed. This method helps in the identification of certain smectite peaks.

The XRD analyses showed that the two samples from EW97909-21GC (the northern core) are virtually the same. Both samples contain predominantly illite with lesser amounts of kaolinite, chlorite, and smectite. The illite peak at ~8.9°2 is relatively large compared to the small smectite peak at 7.1°2 for the nonglycolated sample and at ~5°2 for the glycolated sample. Chlorite is more abundant than kaolinite, based on peak intensities on either side of 25°2. In contrast, the sample from EW97909-12PC (the southern core) contains dominantly smectite with lesser amounts of illite and kaolinite but no chlorite. This sample shows a high-intensity smectite peak at ~7°2 for the nonglycolated sample and at ~5.2°2 for the glycolated sample. In contrast, the illite peaks have very small intensities.

The difference in the dominant clay mineral of the two cores is clearly seen from the XRD analyses. The northern clays are composed mostly of illite, whereas, the clays closer to the equator are composed mostly of smectite. This contrast is most likely attributed to the fact that different winds over the Pacific Ocean tap different on-land source areas. The northern illite-rich clays may have been carried by westerly winds originating in China, which dominate at latitudes higher than 20°N. The vast loess plains of Asia are the main source for wind blown illite-rich dust in the northern Pacific (Rea, 1994). The southern smectite-rich clays may have been carried by the northeast trade winds that dominate between 0° and 20°N latitude. The source of this dust is the Central and South American volcanic arcs (Fig. F4). These sediments are smectite rich because smectite is an alteration product of volcanic rocks. The small amount of chlorite found in the northern core, but not in the southern one, also fits this pattern. Chlorite might be expected in dust that originated from arid lands in Asia. However, chlorite would not be expected in sediments originating from volcanic regions.

LAS analyses of these samples (Fig. F5) indicate a definite spectral difference between the smectite-rich sample and the two illite-rich samples. This result confirms the results from our first feasibility study, which indicated that smectite and illite should be readily distinguishable with LAS. Because the spectral signatures of both smectite and illite are compositionally dependent (Clark et al., 1990), this study is more diagnostic of LAS usefulness for smectite vs. illite during Leg 199. Unfortunately, LAS could not recognize the small amounts of chlorite or kaolinite, but the dominant clay mineral was easily identified. This study shows that LAS can be a useful tool for rapid identification of clay minerals in cores, aiding reconstruction of Tertiary wind patterns.

The goal of this study was to determine whether the spectral signature of opal changes at the opal-A/opal-CT boundary. Rice et al. (1995) used Fourier transform infrared spectroscopy to detect absorption changes associated with the opal-A/CT transition, but neither opal-A nor biogenous cristobalite has been studied previously using LAS.

Seven samples were chosen from Deep Sea Drilling Project (DSDP) Leg 17, Site 166, which is located in the central Pacific Basin (3°N, 175°W). The samples range in depth from 7.7 to 198.0 meters below seafloor (mbsf), with the opal-A/opal-CT boundary present at ~190.0 mbsf. The upper five samples are late Miocene-middle Eocene radiolarian oozes, and the two deepest samples are middle Eocene cherts (Shipboard Scientific Party, 1973). To isolate the radiolarians from the clays, the samples were placed through a 75-µm sieve. Figure F6 shows that even though the crystallinity of the two zones is different, the spectra for opal-A radiolarian oozes are very similar to those of the opal-CT cherts. The total reflectance difference is attributable to the measuring technique and does not reflect true opal-A/opal-CT changes. Since the pieces of chert were smaller than the opening on the light probe, some of the light escaped instead of being reflected back into the spectrometer. In contrast, the powdered radiolarian oozes covered the entire opening, reflecting back all of the light (Fig. F1A). These results indicate that LAS-based opal determination from Leg 199 is not likely to be adversely affected by the opal-A/opal-CT transformation or other reworking.