3. THE REALISATION OF NEPTUNE - THE REAL WORLD IS WORSE THAN WE THOUGHT

3.1. Stratigraphic and geographic coverage

The geographic distribution of the 165 holes included in Neptune is uneven. In some areas there is a very detailed coverage, like for example in some parts of the Antarctic Ocean (Fig. 2.1). On the other hand, no holes from the central north and southeastern Pacific Ocean are present in Neptune. (Holes from ODP Leg 145 now provide a transect across the north Pacific.) The mid- and high latitudes in the southern hemisphere and the tropical regions of the Atlantic Ocean are also not well represented. This is due in part to the uneven coverage of DSDP and ODP cruises and in part to the selection made for Neptune, which preferentially included holes with good biostratigraphy.

Each year, ODP organizes five to six drilling cruises which result in as many published Scientific Results. Although not all cruises retrieve micropaleontologically significant material, many of them provide a detailed biostratigraphy and data relevant to the scopes of Neptune. The present geographic coverage of holes in Neptune has been last updated in 1995 (Leg 135). Since then more than twenty-five volumes of Scientific Results have been published. From the beginning of the project, we were faced with the need to maintain a balance between keeping up with the new data produced by ODP and the need to analyze the data already in Neptune for biostratigraphic or micropaleontological studies. At present, I have decided to keep Neptune at its current, acceptable but not optimal, size in order to complete some of the studies that we had planned. If it will be decided to update Neptune in the future, it will be necessary to:

Another limitation of the database is given by the often incomplete often-incomplete recovery of sediments (Fig. 3.1). Before the advent of hydraulic piston coring, few continuously recovered sections were available. Core recovery has drastically improved in the more recent ODP holes but sediment loss at core breaks is still common even in continuously cored sections (Farrell and Janecek 1991).

In addition, there is an uneven distribution in the temporal coverage of the sections. Whilst Plio-Pleistocene sections are very well represented in Neptune (as they are in ODP holes overall), the detail of stratigraphic coverage decreases for older time periods, as naturally expected from the drilling procedure (Fig. 3.2, Spencer-Cervato 1998). This might be interpreted as a need to recover more Miocene and older sections, but this pattern actually reflects the number of studied sections and not simply the recovered sections. Therefore, I believe that the problem does not lie only in the ‘quantity’ of older sections drilled, but also in the ‘quality’ of their stratigraphy. The reliability of the stratigraphy provided for a section depends strongly on the availability of good calibrations, and these are currently available mainly for Neogene sediments. Figure 3.2 also shows that the number of well-studied sections does not decrease gradually and regularly with age, but shows peaks (around 2 Ma) and plateaus (e.g., between 20 and 32 Ma). This likely reflects the relative, unequal attention given to the Cenozoic stratigraphy through the history of DSDP and ODP.

3.2. Chronology

For the database, we have chosen to use a comprehensive biochronology based on deep-sea sections, therefore not considering land sections, which represent the type localities where stratigraphic series were first described. This may represent a limitation in the achieved biochronological calibration. The precision of the ages determined with the age models depends on various factors, some subjective and nonquantifiable, and some, like sample spacing, accuracy of biostratigraphic calibration, or core recovery, that can be quantified. A conservative estimate of the age model precision of 0.36 m.y. was determined for Neogene sediments (Spencer-Cervato et al. 1994). For Paleogene sediments it is about 0.66 m.y. (twice the average sample spacing).

Another important factor is the quality of the age model. The Neogene DSDP age/depth plots that we have published so far (Spencer-Cervato et al. 1993; Lazarus et al. 1995a) are a good example of the range of reliability of the line of correlation. The subjective ranking given in Table 2.2 varies from very poor or poor (wide scatter of events, straight line of correlation drawn across the middle of the cloud), to moderate (some scatter of a limited number of events, various possible lines of correlation), to good or excellent (40% of the holes: very good agreement of the event ages, abundant events to constrain the line of correlation, good agreement between magnetostratigraphy and biostratigraphy). Several factors can cause the scatter of events observed in most age/depth plots: reworking, downhole contamination, incorrect entry in the biostratigraphy file, typos in the range charts, diachrony of the calibrated event, , and sample spacing. Whilst most of these causes can be easily double-checked, diachrony is quite difficult to assess. The assumption of ‘globally synchronous events’ which is at the base of biochronology is validly established only for some selected, well documented events (e.g., Hays and Shackleton 1976; Thierstein et al. 1977; Backman and Shackleton 1983; Wei 1993; Spencer-Cervato et al. 1994). It is likely that more complete data collection and documentation would lead to the identification of more globally synchronous events. But in most cases, a calibration is valid only for the more or less restricted biogeographic province where it is done, and only a few events are truly globally synchronous, within the precision of the method adopted for calibration. The need for localized calibrations has long been known for siliceous plankton stratigraphy, but it is not widely accepted by biostratigraphers using calcareous plankton. To minimize this factor, we intentionally used multiple regional calibrations for diatoms and radiolarians. Even with this approach, the scatter is sometimes too large to provide a reliable line of correlation. For nannofossils and foraminifera only one general (low latitude) calibration is available (Berggren et al. 1985, 1995a, b). The advantage of this calibration is that it is based on several sites, while most of the regional calibrations are based only on one hole. An estimate of the diachrony/synchrony of Neogene events was done with a subset of the holes currently present in Neptune (Spencer-Cervato et al. 1994). This study indicated that calcareous nannofossils provide the most reliable biostratigraphic events, as they are mostly cosmopolitan and, if diachronous, the age margin is relatively small.

Very few sections are actually continuous, and long stratigraphic gaps are common (Spencer-Cervato 1998) (Fig. 3.3). Two-thirds of the selected holes contain at least one hiatus, and on average they each contain three hiati of various lengths (Fig. 2.1). The presence of these hiati results in artificially older or younger ages for the samples adjacent to the gap. This does not allow one to automatically (‘blindly’) search the database for e.g.,such information as species ages, but requires that every output is be checked and compared with the age models.

The final and probably most necessary improvement of the chronology of Neptune is given by the life-timelifetime of the biochronology selected for the age model calibration. We initially based the age models on Berggren et al. (1985). An updated magnetostratigraphy was published later (Cande and Kent 1992) but it did not provide the combination of biochronology and magnetostratigraphy available from Berggren et al.'s (1985) work. We thus decided to continue using Berggren et al. (1985) throughout the first phase of the project (DSDP Neogene sediments). However, ten years after the first biochronology compilation, a new updated biochronology was published (Berggren et al. 1995b) and the chronology of Neptune became suddenly outdated. The iterations to update Neptune’s chronology were greatly helped by additional programming of Neptune by Dave Lazarus and an auxiliary computer program (not part of Neptune and written ad hoc by Bernhard Brabec) which created a correlation function between the old and the new master biochronology. This function was applied to all age model files and new revised age models were created. . Then, all biostratigraphy files were updated using a ‘find - replace’ routine with lookup tables (i.e.: if code in column 3 is equal to xYwz, replace age in column 4 with corresponding value in lookup table). While we could directly use the new calibrations for calcareous plankton as lookup tables, it was necessary to recalibrate to the new time scale all regional templates used for siliceous plankton events. Finally, before the new age models could be imported into Neptune, all the age/depth plots were redone by myself and eventually adjusted to fit the new event ages.

3.3. Taxonomy

Among the other reasons mentioned above, if a bio-event recorded in a specific hole plots far outside the area where the line of correlation can be drawn, it could be due to its taxonomic identification. Many authors have put together the hundreds of range charts that were used for Neptune and not all agree in the detailed taxonomic identification of all the 8800+ taxa included in Neptune. Indeed, taxonomic identification is subjective. The time pressure under which biostratigraphers are during a leg is also an important limiting factor in the number of species described in a range chart, which is often limited to biostratigraphic markers. The extent to which this taxonomic problem has affected the data in Neptune can be judged by experts in particular cases but cannot be easily quantified.

Starting from the biostratigraphy filesbiostratigraphic records assembled for the chronology, we assumed that the taxon associated with one event and described in the range chart was the one we were looking for. Further, we needed to consider the occurrence of synonyms. It sometimes happens that the name used by one author for a taxon corresponds either to a different taxon according to another author, or that a different name is used by a second author for this specific taxon (synonymy). For example, the foraminifer species Globorotalia truncatulinoides has been also called Truncorotalia truncatulinoides. To account for this, we have used the literature, personal experience and extensive consultation with taxonomic experts to identify valid taxon names. Three thousand of the 8810 names listed in Neptune (Fig. 3.4) are considered valid (i.e., are legal names in the framework of the ICZN and ICBN, and are known to be real to at least one of the experts). Synonyms to these valid names were then identified (with the corresponding valid name). They constitute 31% of the total number of names. In several cases we could not unequivocally identify a specific name and marked it as ‘unknown’ (15% of all names). Only 43 names (0.5%) were considered invalid. This information is available in the ‘Species Names’ table of Neptune. The synonymizsation is subjective (the initials of the person who identified each species name is also given in the ‘Species Names’ table) and the names list does not at all pretend to be a thorough or complete taxonomic revision of marine plankton. It merely represents a working table that gives us a first approximation of plankton taxonomy. A ‘real’ taxonomic database would need complete taxonomic descriptions (with history) for each taxon and a series of images to illustrate them. Cathy Nigrini, Jean-Pierre Caulet, and Dave Lazarus are currently working on a detailed taxonomic database for radiolarians, but it is well beyond the scopes of Neptune to even attempt anything like this for all groups.The taxonomic list also needs continuous update: every time a new hole is added to Neptune, the biostratigraphic range charts carry with them new names, sometimes several ones. These need to be added to the ‘Species Names’ list and identified as valid or not.

3.4. Biostratigraphy

There is an uneven distribution in the number of reports by plankton group in Neptune. Over 60% of the 225 articles from which data for Neptune have been extracted (Table 2.4) are on calcareous plankton, almost equally distributed between nannofossils and foraminifera. Radiolarians follow with about 21% and diatoms trail with only 16%. At the same time, biostratigraphic work on siliceous plankton is underrepresented in Paleogene sections, and most often limited to the Oligocene and younger sections (Fig. 3.5). This unevenness represents a bias for evolution studies where we would like to compare calcareous and siliceous plankton occurrences. Whether this distribution represents the average abundance of fossil plankton in deep-sea deposits or is instead the reflection of staffing decisions by DSDP and ODP is yet to be determined.

One of the limitations of Neptune as a comprehensive micropaleontological database is given by our decision to include only four plankton groups. The DSDP and ODP Reports include many articles on benthic foraminifera, silicoflagellates, dinoflagellates as well as palynology. At the moment, there are no plans to include their occurrence data in the database, which in itself would not be a huge task.

Go to Next Section