2. Data Report: Logging While Drilling Data Analysis of Leg 171A, a Multivariate Statistical Approach¹

C. Bücker,² J. Shimeld,³ S. Hunze,² and W. Brückmann⁴

ABSTRACT

In the northern Barbados accretionary wedge, several Deep Sea Drilling Project (DSDP) and Ocean Drilling Program (ODP) legs (DSDP Leg 78 and ODP Legs 110, 156, and 171A) targeted the décollement and the seaward extension of the décollement, the proto-décollement. During Leg 171A, the logging while drilling (LWD) technique was used to determine the physical properties variations along a profile across the deformation front. Because of the unstable borehole conditions in accretionary wedges, LWD is the most effective method for the measurements of physical properties in these poorly consolidated sediments. LWD data are acquired just above the drill bit a few minutes after the formation has been drilled, yielding measurements as close to in situ conditions as possible.

The large amount of LWD data and the demand for a quick, objective, and reliable evaluation calls for the application of multivariate statistical methods. The multivariate factor analysis is a method of reducing the amount of logging data while giving them a new integrated meaning with no loss of important information, resulting in factor logs that are helpful tools for further interpretation. The cluster analysis of the two or three most significant factors proved to be a useful and objective method to identify and confirm significant logging units. The main objective of the application of multivariate statistical methods in this study is twofold. First, Leg 171A was a stand-alone logging leg, where no cores were retrieved. The factor analysis was used as an objective tool for a classification of the drilled sequences based on their physical and chemical properties. The new factor logs mirror the basic processes behind the measured geophysical properties and make them easier to interpret. Second, in the succeeding cluster analysis, similar geophysical properties are grouped into one cluster, reflecting one logging unit. These objectively defined logging units can be compared to statistical electrofacies, which are helpful in differentiating lithologic characterizations. In particular for LWD measurements, the multivariate statistical methods of factor and cluster analysis are helpful tools for a fast, reliable, and objective definition of logging units, which should be considered for future legs.