D&A Lab Working Team

D&A Working Team meetings

Supervising Scientist: Peter Blum

JOIDES Resolution Depth and Age Lab

Working Team Meetings

The ultimate goal of most geological field programs is to collect data in form of time series. For drill holes and cores, the initial independent variable of the sampled and measured time series is some type of depth scale. Data acquisition from cores or downhole measurements is associated with particular depth measurements including intrinsic errors ("depth 1"). Intervals measured or recovered from multiple holes can be correlated and spliced using composite depth scales ("depth 2") to obtain more continuous time series for a site. "Depth 1" or "depth 2" type scales can be further improved by applying corrections such as elimination of artificial gas voids or linear compression of a scale. Using age information obtained from cores and the most appropriate depth scale, one or more age models can be constructed for a hole or site. The age model then allows calculation of age as an independent variable for core and downhole data.

The following procedural steps explain the existing depth conventions and possible future scenarios for depth correlation and data integration (see also overview diagram). The generic depth categories "depth 1" and "depth 2" refer to intrinsic and correlated depth, respectively. Each category includes three possible depth types, which may appear confusing to the uninitiated. However, in practice any given data set will at any time only be associated with one type of each category. For instance, if a composite depth has been constructed for a site the present web-based data reports already return two depth columns: "depth 1" = "core sample mbsf"; and "depth 2" = "core sample mcd".

Proposed applications development tasks ("Janus" Review, December 1999) are highlighted (red and underlined).

• The intrinsic depth scale for each run is based on the wire length between downhole tool and shipboard winch.
• Multiple wireline logs are routinely correlated by BRG personnel using natural gamma ray measurements logged with every run.
• In general, the resulting "logging depth 1" ([L]mbsf) deviates only slightly from the depths measured with individual runs and supercedes the latter (BRG database).

• The intrinsic depth scale is based on the length of drill pipe lowered below the drill floor. The measurement is made by eye.
• The resulting "drilling depth 1" ([D]mbsf) has a precision of ~0.2 m for each stand and an accuracy of a few meters for any segment in a hole. The quality of these measurements may improve in the future as rig instrumentation becomes fully implemented.
• It may become possible or necessary to correlate multiple drill pipe entries into holes in the future as measurement-while-coring and logging-while-drilling data become more available. Correlation may impact accurate deployment of long-term monitoring devices, packers, etc..

• Driller's core top measurement ("drilling depth 1") is the primary reference for each core. Placement of a core at this depth scale carries the driller's measurement error (0.2~4 m).
• Sample depth below core top is added to the core top depth and carries the much smaller curation scale error (0.1 ~ 5 cm).
• The resulting "core sample depth 1" ([C]mbsf) is therefore a hybrid and can be improved by placing the core top reference in a more accurate framework (see below).
• An important convention is to place the top of the recovered core at the top of drilled interval even though we don't know (particularly in rotary coring) where the recovered intervals come from. Only careful correlation of core and downhole data may place the recovered intervals at a more accurate depth (see below).
• Cores (particularly APC) expand upon recovery and the recovered interval often exceeds the cored interval (AKA >100% recovery). This stratigraphic problem can be resolved by either constructing an expanded composite depth scale (multiple holes required; see below) or compressing the recovered depth scale to match the drilled interval (if only one hole exists).

• If a hole was cored and wireline logged, AND if the same or similar type of data were collected by downhole logging and from the cores, AND if significant sections have sufficiently good downhole and core data, THEN core and downhole data from those hole sections can be correlated

• In general, correlation to/from sea floor cannot be made (no wireline logs).

• This depth 2 scale will presumably be the same as the downhole logging depth for the hole.
• The new application "Sagan" can presumably be used for this type of shipboard correlation (Leg 189?).
• ODP/TAMU may have to implement a database correlation table in the future, similar to either the mcd (offset, splice) tables or the age-depth model. More definition and a few successful shipboard applications are needed before any effort should be expended.

• If cores from multiple holes are recovered, core logging data can be correlated to construct a (more) complete stratigraphic section for the site. The assumption is that no significant lateral changes in geology occur from one hole to the other (~20-50 m distance), which for sediments is true for all but very rare cases.
• A shipboard procedure to construct the "core depth 2" (coring meters composite depth, [C]mcd), has been applied effectively and successfully since Leg 138.
• One of the results of a shipboard mcd scale is the ability to construct a sampling splice, which is one representative section for the site spliced together from the best recovered intervals from 2-4 holes.
• Core composite sections are expanded relative to the cored interval because cores expand physically upon recovery. For stratigraphic purposes this is usually irrelevant, at least until a more detailed age control has been established postcruise.
• This procedure shifts entire cores relative to each other, it does not correct for differential stretching and squeezing within cores as a result of the coring process. This is initially irrelevant because the shipboard sampling splice will represent a unique, most complete section for the site.
• The program Splicer has been used since Leg 154 for mcd and sampling splice construction.
• It is important that construction of a composite can happen rapidly on the ship so coring depths can be directed in real-time to optimize recovery overlap between holes
• Input of core logging data from the database into Splicer has been cumbersome. "Better Integrate Splicer" was proposed as a development project for the "Janus review" but can be eliminated now that BRG appears to have made significant progress in alleviating the problems (Leg 189 deployment?).
• The Splicer program outputs two files: *.affine (cumulative offset table) and *.splice (splice tie point table). Affine tables must be uploaded to the database on the ship so mcd can be returned for all data reports. The splice table is presently uploaded but cannot be used to query data (see below). Both affine and splice tables can be uploaded to the database by the shipboard scientists using the old "Janus" interface. However, it seems there is a problem with bringing these data back home.
• Output of splice data from the database was not possible until last month, when the "Fix/Create Splice Reports" project was completed. Note that splices can be displayed by tie points or by intervals.
• An effective and efficient conversion program is needed to compute mbsf and mcd depth types from sample ID's and vice versa. The "Create X-Converter" project was proposed with detailed functional specifications. The X-Converter should eventually include all depth types that can be generated on the ship as well as age. At present, however, it would be very useful to create this capability for "core mbsf" - "core mcd"- "sample ID" conversions.
• Users should be able to retrieve a spliced data set from the database. The "Splice as Query Parameter" project has been proposed and discussed but is not in the works at present.

• If downhole logging data exist from multiple holes, they can potentially be correlated to construct a composite depth scale for the site using incremental depth shifts.
• The resulting "logging depth 2" (logging meters composite depth, [L]mcd) has not been completed shipboard in the past because sufficient time is usually not available.
• If different intervals are logged successfully in different holes, constructing a splice for the site may be very useful.
• This procedure could be accomplished using a generic correlation program (e.g., Analyseries). The program Sagan (in development at BRG) supposedly can help doing the job in a more customized way.

• ODP/TAMU may have to implement a database correlation table in the future, similar to either the mcd (offset, splice) tables or the age-depth model. More definition and a few successful shipboard applications are needed before any effort should be expended.

• The final and scientifically most rewarding data integration step is to correlate the core depth 2 ([C]mcd) to the downhole depth 2 ([D]mcd), at least for the intervals where this is possible and useful.
• The resulting "integrated depth 2" (integrated meters composite depth", imcd) scale provides a common stratigraphic framework for all core and downhole logging data from all holes at a site.
• At present this kind of data integration is accomplished by individual scientist's postcruise research.
• The application "Sagan" can supposedly accomplish this procedure, however, it is doubtful that this will be possible shipboard because of time constraints.
• Construction of a core-downhole splice is also feasible, splicing core data where no downhole logs exist and vice versa..

• ODP/TAMU may have to implement a database correlation table in the future, similar to either the mcd (offset, splice) tables or the age-depth model. More definition and a few successful shipboard applications are needed before any effort should be expended.

• Age data are derived from core material (with very rare exceptions) and are primarily referenced to a sample ID.
• The depth assigned to an age datum depends on the level of correlation possible on the leg (and/or postcruise): Depth 1 (mbsf) if no correlation was possible, depth 2 (mcd) if at least cores from multiple holes could be correlated, or depth 2 (imcd) if core and downhole composites could be correlated.
• The depth scale applied to age data will also be the one assigned to the age-depth control points and the age model.

• Age-depth control points are derived from the list of datums determined by biostratigraphers, paleomagnetists, and other age data specialists. They are stored in a separate database table to be used for age modeling.
• Stratigraphers decide which of their datum calls are reliable ones and thus become control points. Through a database table and appropriate user interface specified by the "Age-Depth Control Points" project, they specify the depth and age errors associated with each control datum.
• A December 1999 meeting refined the specifications for the control points table input, formats, output, and user interface. The project is presently progressing at a slow pace and the X-LWT should review it frequently.

• Age models are simple tables mapping age to depth. The depth type will be the same as the one used for the age-depth control points.
• Age models are created by scientists using a data analysis tool of their choice, such as a spreadsheet, and a method of choice.
• The simplest models consist of a few straight-line segments fit between control points with relatively little error. More general and statistical models use least squares curve fitting through control points with consideration of (depth) errors and specific corrections of the core catcher sampling bias.
• The age-depth series should be sampled at increments that realistically represent the original data resolution (10 m if only core catchers were examined).
• Age models may be by hole or by site. Several age models may exist for a hole or a site. Age models are identified by Site/Hole, author name, date created, and depth type used.
• The proposed "Age Model Upload/Report" project specifies the need for a simple database upload utility for the completed age models, and the database report to view/download age models.
• Once the age models are in the database, age can be calculated from depth. This is outlined in the proposed "Age Independent Variable" project. Only the depth type used for creating the age model can be used to calculate age. If Depth 1 (mbsf) was used, the age model is strictly only applicable to the particular hole data used.

• Once the independent depth/age variables for different data sets are normalized, actual data integration can begin.
• Effective shipboard data integration requires that the wealth of data immediately accessible from the database can easily and dynamically be displayed with any of three x-variables: Depth 1, Depth 2, or age.
• ODP should provide a tool that overcomes the tremendously time-consuming and ineffective task of downloading, transferring, and formatting data to display them in various inadequate plotting and/or drafting programs.
• The proposed "Implement Integrated Display" project would implement a powerful, user-friendly tool, such as Winlog.