BP Gulf of Mexico Neogene Astronomically-tuned Time Scale ( BP GNATTS )

J.A. Bergen1, S. Truax III2, E. de Kaenel3, S. Blair4, E. Browning5,†, J. Lundquist6, T. Boesiger7, M. Bolivar5, and K. Clark5 Paleo at the Hill Country, Brenham, Texas 77833, USA Houston, Texas 77070, USA DeKaenel Paleo Research, Mont-sur-Rolle, Switzerland Blair Biostratigraphy, Cypress, Texas 77429, USA BP America, 200 Westlake Park Blvd., Houston, Texas 77079, USA Lundquist Stratigraphic, LLC, Houston, Texas 77065, USA Nebraska Oil and Gas Conservation Commission, Sidney, Nebraska 69162, USA


INTRODUCTION
Microfossils are an important, arguably integral tool in subsurface petroleum exploration.Conventional exploration has reached into new regions and basins, while further development takes place in stratigraphically-and structurallycomplicated fields with increased requirements for finer correlation in reservoir intervals (i.e., the need for increased resolution in the expanded sections that are typically targeted in offshore exploration).Application of published Cenozoic global biozonations (Blow, 1969;Martini, 1971;Okada and Bukry, 1980) in deep-water exploration was not ideal, especially with the combined effects produced by sediment dilution on microfossil recovery, different taxonomic concepts, and varied methodologies.This necessitated improvement beyond published global bio zonations and stimulated petroleum companies to support research that improved their biostratigraphic databases and frameworks.
The first industrial applications of microfossil biostratigraphy along the U.S. Gulf of Mexico Coast began with benthic foraminifera nearly a century ago (Loutit et al., 1988) and progressed from onshore to deep-water in re-sponse to exploration focus (Martin, 2013).Today, planktonic foraminifera and nannofossils are the primary groups used for time correlation in deep water wells and the construction of global Cenozoic timescales.Integrated Gulf of Mexico (GoM) industrial biostratigraphies and published global biozonations utilizing these two planktonic groups date back half a century, near the inception of a research coring program in the world's deep-sea basins (Deep Sea Drilling Project).During this time, many industrial staffs dedicated to the GoM developed their own internal Cenozoic biostratigraphic frameworks.The Deep Sea Drilling Project (DSDP) was later rebranded the Ocean Drilling Program (ODP) in 1983 and the Integrated Ocean Drilling Program (IODP) from 2003 to 2013.
Low oil prices from the mid-1980s through most of the 1990s impacted specialties such as biostratigraphy.Maintenance of taxonomic concepts and methodologies were challenged in this business environment.Foraminiferal and nannofossil taxonomic equivalency projects were formed through the Gulf Coast section of the Society for Sedimentary Geology (SEPM) amidst concerns about biostratigraphic data and terminologies (Picou et al., 1999).Around this time, Shell (Styzen, 1996) and Texaco (Lawless et al., 1997) published their GoM Cenozoic charts.The post-Eocene Shell Offshore Inc. biozonation established separate enumerations for surfaces represented by the tops of nannofossils and combined benthic-planktonic forami nif era events (Neogene resolution 307 k.y.) and was related to the time scale of Berggren et al. (1995aBerggren et al. ( , 1995b)).The Cenozoic Texaco zonation utilized fossil bases and more down-hole abundance increases in wells to enhance resolution (271 k.y. for the Neogene).
Cenozoic stratigraphic research was in the midst of a revolution during the 1990's.Research on, and formal ratification of, reference outcrop sections known as global boundary stratotype section and point (GSSP) resulted in precise global definitions of stratotype boundaries (see stratigraphy .org).The scaling of geologic time shifted from geomagnetic polarity time scales (GPTS) with the first applications of astronomical "tuning" (Lourens et al.,1996;Laskar et al., 2004) of sedimentary cycles in the construction of higher temporal resolution Late Neogene timescales (see Hilgen et al., 1997).Today, Neogene calcareous microfossil bio zonations (e.g., Backman et al., 2012) are founded almost entirely by astronomical ages.Although the accuracy and precision of ages for Neogene planktonic microfossil events also improved by an order of magnitude with these techniques, the resolution of global biozonations (~750-1200 k.y.) are still coarser than the lower limits of astronomical tuning achieved by the dominant 405 k.y.eccentricity cycle (see Hinnov, 2013).
Advances in timescale construction, combined with the upswing in oil prices around the beginning of the twenty-first century, provided new opportunities for industrial biostratigraphy in the GoM.BP America staff from the three heritage companies (BP, Amoco, and Arco Vastar) were charged with producing a single Neogene GoM chart from three independentlyderived GoM biostratigraphic frameworks.Staff exchanged taxonomic concepts and methodologies, which accounted for some differences between company schemes; remaining discrepancies were solved through in-house analyses of well samples.The resulting improved and unified internal framework, completed in 2007, aided exploration and development efforts within the stratigraphically and structurally complex GoM deep-water (DW).
It had long been realized that there were only a limited number of published geologic ages available for bioevents in GoM industrial schemes.Coupled with the late twentieth century "cyclostratigraphic revolution" and the availability of such reference sections, an internal research program was initiated in 2002 with the goal to derive astronomically-calibrated geologic ages for the entire BP Neogene biostratigraphic framework.Efforts culminated in the first "BP Gulf of Mexico Neogene Astronomically-tuned Time Scale" (BP GNATTS) in early 2007.Subsequent research through 2012 extended calibration into the lower Oligocene through sampling and study of the base Neogene GSSP in northern Italy and ODP Leg 154 cores, offshore NE Brazil (Fig. 1).In 2016, efforts were refocused on publication of BP GoM taxonomy and biostratigraphy, including new research on ODP Leg 154 materials to more thoroughly document biostratigraphic events for publication.

MATERIALS Localities
There are several thousand wells in the BP GoM biostratigraphic database.The vast majority are located onshore and on the continental shelf; several hundred wells are located in deep water.Deep-water exploratory drilling began in the 1970s and encountered progressively older section through time.By the 1990s, several major DW Miocene reservoirs had been discovered in the GoM.
Age calibration of the GoM Neogene biostratigraphic framework was based on the study of ODP core materials (Fig. 1) in the western Atlantic Ocean (ODP Leg 154) and western Mediterranean Sea (ODP Leg 161).Five sites (925-929) drilled as a depth transect on the Ceará Rise in the western tropical Atlantic Ocean during ODP Leg 154 recovered an important composite cyclostratigraphic reference section.The age model derived from ODP Leg 154 cores is based on the astronomicallytuned cycles (Shackleton and Crowhurst, 1997;Shackleton et al., 1999;Pälike and Shackleton 2000;Zachos et al., 2001) recalibrated to the orbital solution of Laskar et al. (2004).This composite section was sampled by de Kaenel from the lower Oligocene (30.679Ma) through lower Pleistocene (1.595 Ma) for nannofossils (1357 samples) and foraminifera (633 samples) between 2003-2011.We realize continued improvements of the ODP Leg 154 astronomical-tuning have been published for the upper Miocene to Pliocene since 2011 (Zeeden et al., 2013;Drury et al., 2017;Wilkens et al., 2017) and that problems remain with the orbital tuning of the Early Miocene (see Hilgen et al., 2012;Ogg et al., 2016), both of which could affect age estimates for the ODP Leg 154 samples by as much as 100 k.y.We have maintained the integrity of the astronomically-tuned age model used by de Kaenel, when he sampled the ODP Leg 154 cores at the IODP repository in Bremen, Germany.This makes direct correlation and comparison to Neogene nanno fossil (see Backman et al., 2012) and planktonic foraminifera (see Wade et al., 2011) bioevent ages derived from these cores possible.It also maintains syn chronic ity with nannofossil bioevent ages published recently in the Journal of the Nanno plankton Research (Bergen et al., 2017;Blair et al., 2017;Boesiger et al., 2017;Browning et al., 2017;de Kaenel et al., 2017).The ODP Leg 154 sites 925-929 were chosen because of their geographic proximity to the GoM and because individual samples could be equated to geologic time.They are referred to herein by their geologic age; individual sample information is given in the range charts (www .Pangaea .de).The remainder of the Pleistocene is calibrated to the orbital time scale based on the nanno fossil biostratigraphy (de Kaenel The BP GNATTS framework has been extended to the Mediterranean for the Oligocene-Miocene boundary interval.The first Mediterranean locality is the base Neogene GSSP of Lemme-Carrosio in northern Italy (Bergen et al., 2009;de Kaenel and Villa, 2010).Thirtynine samples were collected from this section in Fall 2008 with aid from the University of Parma.Biostratigraphic research on the GSSP was completed in 2009 and the nannofossils reexamined in late 2017 as part of this study.The second location in the eastern Mediterranean Sea are three exploration wells analyzed by the first two authors.

Data
Foraminifera taxonomy is based on heritage company concepts and supplemented with the standard industry concepts in Picou et al. (1999).The BP taxonomic concepts and biostratigraphy for 204 Oligocene-Recent nannofossil species were recently published as a series of five papers in the Journal of Nannoplankton Research (Bergen et al., 2017;Blair et al., 2017, Boesiger et al., 2017;Browning et al., 2017;de Kaenel et al., 2017), including 70 new species.Photographic plates of the BP GNATTS nannofossil markers species are included in the supplementary materials (Plates S1-S9) 1 .Also included in the Supplementary Materials are the following data tables: (1) the BP GNATTS calibrations for chart horizons and events (Table S1; see footnote 1); (2) BP taxonomic equivalencies (Table S1); and (3) Pleistocene ODP Leg 161 nannofossil events (Table S2; see footnote 1).All supplemental information (Plates S1-S9 and Tables S1-S2) is available in the GSA Data Repository, as well as a full-sized pdf of Figure 2 2 (see footnote 1).All stratigraphic range chart data from the Italian GSSP for nanno fossils and foraminifera (Tables S3-S4) and ODP Leg 154 (Tables S5-S23) are stored in the Pangaea web database (www .Pangaea .de).

METHODS
The preparation and examination of samples are two critical factors to results.Consistency is of primary importance in sample preparation because microfossil abundance estimates are fundamental.ALS Ellington and Associates, Inc. (Houston, Texas, USA) prepared all well samples and foraminifera ODP core samples, whereas ODP core samples taken for nannofossils were prepared by the research scientists.Historically, BP utilized microfossil abundance peaks in well correlation and the identification of event horizons within these peaks for use in basin-wide and field scale schemes.This methodology is outlined in Armentrout (1996).
Time is the most determinative factor in sample analyses because fossils are very rare at the ends of their stratigraphic ranges.In clastic settings such as the GoM, analysis times are skewed toward samples within abundance peaks, which maximizes results and makes it possible for an experienced analyst to average up to two samples per hour.For pelagic and hemipelagic settings, collecting accurate abundance estimates and range data for rare taxa from fossil-rich assemblages is more time consuming.
The primary ODP Leg 154 research was done by Eric de Kaenel and Jim Bergen for nannofossils (1202 samples), Sheila Barnette and Steve Truax for foraminifera (618 samples).The remaining authors re-examined selected research samples in the months prior to completion of an internal 2007 GoM Neogene chart and again for this publication.For nannofossils, 45-60 minutes were typically needed to document rare "marker" taxa, corresponding to ~1000 fields-of-view at 1000× magnification.For planktonic foraminifera, similar analysis times are recommended.Barnette examined 519 samples from the upper Oligocene (24.303Ma) to lower Pliocene (4.128 Ma).BP staff examined 253 samples for the presence/absence of marker species, while also extending stratigraphic coverage into the upper Oligocene (24.900Ma) and lower Pleistocene (1.720 Ma).
Abundance estimates (total assemblage and individual taxa) are routine for well and research analyses.For nannofossils, the cascading count system of Styzen (1997) is GoM industry standard (specimens/100 fields of view estimates) outside of BP.Within BP, a 0-100 based counting system has been employed for decades for the sole practical purpose in having to work with hardcopy histogram data (estimates based on specimens per field-of-view at 1000×).For foraminifera, relative abundance categories (10 categories between Present to Flood, with quantitative values estimated within categories) per BP standard were used in well analyses and the original research.The foraminifera research data was converted to presence/absence for this publication because it was not possible to normalize all count data from different researchers over a fifteen-year time span.
The final critical factor to achieving our results was the setting.These innovations were possible because the biostratigraphy team worked together daily in an integrated environment with geologists and geophysicists.These interactions gave biostratigraphers time to share concepts, identify business needs, fully integrate biostratigraphy with the description of the subsurface (geology), and provide essential one-onone training for the next generation.
Biostratigraphic events derived from sample analyses are often evaluated relative to their resolution, reliability, and synchronicity.Resolution is defined herein by the number of events per unit of geologic time.Sample resolution is the number of samples per unit of geologic time.For ODP Leg 154, where samples are related to geologic time through cyclostratigraphic methods, sample resolution equates to the precision of derived bioevents and estimates of error in their geologic age.Reliability refers to bioevents either within a biostratigraphic scheme or an individual well or section.For BP GNATTS, the reliability of biostratigraphic markers is evaluated relative to the number of times a bioevent has been reproduced in sequence between wells.For events used by all three heritage companies in the GoM, marker events in these schemes have been tested in hundreds to possibly thousands of wells.For postmerger BP, largely in the DW GoM, marker events in BP GNATTS have been tested in tens to hundreds of wells.The reliability of a microfossil top (highest occurrence) or base (lowest occurrence) in an individual section (well, core, or outcrop) is also related to its persistence in occurrence.Taxa that are very rare and sporadic in occurrence at the ends of their stratigraphic ranges are considered less reliable.Such taxa are certainly less desirable as marker events in biostratigraphic schemes, especially if they are sporadic in occurrence in fossil-rich sections (e.g., ODP Leg 154 cores) not affected by sediment dilution.Bioevents defined by significant abundance changes in taxa are determined more quickly in sample analyses, but must be evaluated relative to facies changes.This is certainly true in a large sedimentary basin such as the GoM, where lateral changes in microfossil abundances can occur between an original wellbore and its bypass or sidetrack hole.The synchronicity of microfossil events used in BP GNATTS has been tested in a number of ways.First, is their reproducibility in sequence within an enumerated Neogene framework having an average resolution of 144 k.y.Second, is their application within much higher resolution biostratigraphic frameworks at field scale or individual well bores, where correlations are further constrained by geologic log correlations.1 GSA Data Repository item 2018407, Tables S1  and S2 and Plates S1-S9, is available at http:// www .geosociety.org/datarepository /2018 or by request to editing@ geosociety .org.Tables S3-S23 are stored in the Pangaea web database (www .Pangaea.de).
2 Figure 2 is on a separate sheet accompanying this issue, and a PDF of the figure is also included in the Data Repository.Third, is that the large majority of events in BP GNATTS (~86%) have been reproduced in sequence from research on cores in the western tropical Atlantic Ocean (ODP Leg 154) and western Mediterranean Sea (ODP Leg 161).

RESULTS
The BP Gulf of Mexico Neogene Astronomically-tuned Time Scale (BP GNATTS) is presented in Figure 2 and a more detailed digital version is available in the Supplementary Materials (Table S1).The chart extends into the Paleogene, including ~2 m.y. of the Late Oligocene.The following summarize the BP GNATTS chart.

Stratigraphic Terminology
In 2009, the International Union of Geological Sciences (IUGS) ratified the Quaternary System (Period) as a formal stratigraphic unit (Gibbard and Head, 2010), truncating the top of the Neogene System and abandoning the informal term Tertiary.We have chosen a twofold division of the Cenozoic, referring the Miocene through Holocene Series (Epochs) to the Neogene, following decades of conventional use in marine micropaleontology and that advocated by Hilgen et al. (2012) for the Neogene Period.To us, the terms Quaternary and Tertiary should be paired and are an alternate way to subdivide the Cenozoic.For further clarification, we also follow Hilgen et al. (2012) by placing the Pliocene-Pleistocene boundary at the top, as opposed to the base, of the Gelasian Stage.We have maintained an age of 23.03 Ma for the Oligo cene-Miocene boundary, following both Hilgen et al. (2012) and Ogg et al. (2016).Stadial terminology is not typically used in the GoM Basin, where Series/Epochs (e.g., Pliocene) are the basic stratigraphic units.

Horizons and Event Terminology
BP biostratigraphic Chart Horizons are both chronostratigraphic terms and mappable surfaces in the GoM.In application, a Chart Horizon (Fig. 2) is a surface that includes section down to the top of the next Chart Horizon (a topdown or drilling sense).A biozone is bounded by two surfaces and its application functions in a similar manner.A BP biostratigraphic Chart Horizon is set apart from a biozone in that it is multidisciplinary (i.e., both nannofossil and foraminifera events) and often utilizes multiple event criteria and intra-horizon (or intrazonal) events.
The terminology of BP GoM Chart Horizons has a long history prior to the BP mergers and has evolved somewhat haphazardly.Letter designations refer to the Pleistocene (PS), Pliocene (P), Miocene (M), and Oligocene (O) Epochs.Horizons within each epoch are then numbered bottoms-up, akin to global biozonations.Epoch enumerations in the original BP framework ranged from 1 to 100, but this is no longer true because epoch boundary criteria have changed over the decades.Some Chart Horizons are subdivided into upper (U) and lower (L) and/or subdivided by lettering (from A to D), both in a topdown sense.
Horizons not yet proven across the entire basin are referred to as "Locals."For the Neogene, there are 160 Chart Horizons and 14 Locals.Eight Oligocene Chart Horizons (O85-O75) are also included on BP GNATTS.
Historically, GoM microfossil events have been limited to fossil tops (highest occurrences) and abundance increases (in a down-hole or drilling sense) because ditch-cuttings routinely "cave" down the well bore.Down-hole caving is less problematic in modern DW GoM wells relative to those drilled before the latter part of the twentieth century.We speculate that this could be due to the use of synthetic muds that do not break down the ditch-cutting samples and by improved drilling parameters and pressure predictions.Remobilization and redeposition of sediments, the latter referred to as reworking by microfossil specialists, are more problematic in a terrigenoclastic basin with salt tectonism such as the GoM.In the GoM, the reworking of microfossils is most often observed as sporadic occurrences of rare specimens involving only a few species.Such processes are easier to recognize in wells when applying a much higher resolution biostratigraphic scheme such as BP GNATTS, which then provides a foundation for still higher resolution biostratigraphic schemes developed for expanded reservoir sections in the GoM.When all this is considered along with the consistency provided in sample preparations, it is now possible to more fully utilize the entire abundance profile of individual species (Fig. 3) in correlation.Biostratigraphic resolution can be further improved by utilizing non-standard marker taxa and describing new species.
The documentation of fossil appearances (first occurrences) and extinctions (last occurrences) in a stratigraphic section is of primary importance, where they are expressed as lowest occurrences (bases) or highest occurrences (tops), respectively.Two-thirds of the microfossil events used in BP GNATTS are lowest or highest occurrences, with the latter being far more predominant.Ninety percent of the bioevents utilized among the six Neogene calcareous microfossil biozonations discussed herein (see "Calibration, Global Biozonations" section) are based on microfossil appearance or extinction events, including all zonal bioevents for planktonic foraminifera.For nannofossils, age estimates in the Backman et al. (2012) biozonation were chiefly derived from astronomically-tuned cyclostratigraphies using the semiquantitative methods of Backman and Shackleton (1983).These counting methods provide precise and reliable bioevents that are easily determined; however, they emphasize speed, which is counter-intuitive to documenting the rare and sporadic occurrences that typify the stratigraphic extremities of individual species.We believe these semiquantitative methods actually have produced both initial and final abundance changes as proxies for nearly all of the Neogene nannofossil appearance and extinction events in Backman et al. (2012).Thus, we introduce two biostratigraphic events for these abundance changes: (1) the lowest increase occurrence for the initial abundance increase; and (2) the highest increase occurrence for the final abundance decrease.This fundamental difference of opinion about the stratigraphic expres- sion of nannofossil appearances and extinctions explains why many of our geologic ages are outside those ages presented in Backman et al. (2012), many of which were derived from the same reference cores (ODP Leg 154).
The following standard abbreviations (Figs. 3  and 4) are used for bioevent terminology: LO (lowest occurrence) and HO (highest occurrence); abundance modifiers are: R (regular or persistent) for LRO and HRO, I (increase) for LIO and HIO, F (few) for LFO and HFO, C (common) for LCO and HCO, A (abundant or acme) for HAO and LAO, and Ab (absence) for LOAb and HOAb.Other paired terms utilized herein include EXIT and RE (re-entry), disappearance (DA) and reappearance (RA), and increase (INC) and decrease (DEC).For planktonic foraminifera, coiling directions are abbreviated as S (sinistral) or D (dextral).Event terminology is in a "top-down" or "down-hole" sense because of its use in drilling (Figs. 4 and  5).These terms should not be confused with those used in a "bottoms-up" or depositional sense for deep-sea research cores (DSDP, ODP, IODP) and outcrops (see Appendix).
Abundance categories for calcareous nannofossils (Fig. 3) are based on estimates of the number of specimens per fields-of-view (FOV) at 1000× magnification, where from lowest to highest are: P (present) being 1 specimen in >100 FOV; R (rare) then being 1 specimen in ≤100 FOV; F (few) then being 1 specimen in ≤10 FOV; C (common) then being ≥1 specimen per FOV; and A (abundant) then being ≥10 specimens per FOV.

Stratigraphic Boundaries
We have used the ages for stage boundaries in accordance with the most recent geologic time scale of Ogg et al. (2016), who followed the IUGS and positioned the base of the Pleistocene Series-and Quaternary System-at the base of the Gelasian Stage.However, we have chosen to maintain a 3-fold subdivision of the Pliocene and place the Pliocene-Pleistocene boundary at the top of the Gelasian.This conforms to placement within previous GoM charts (Shell and Texaco) and global planktonic foraminifera and nannofossil biozonations.
Industrial historic placement of Neogene Epoch boundaries in the GoM Basin are lithologic, corresponding to major mappable seismic surfaces.Placement of microfossil biostratigraphy relative to these surfaces has varied among companies, partially caused by different global usages of boundary microfossil criteria.This has sometimes caused confusion in communication between companies and partners; for ex-ample, when stating "a well has penetrated the Oligocene."Various GoM microfossil events that have been used to pick the top Oligocene vary over 3.55 m.y.(Bergen et al., 2009), representing the interval spanned by BP Horizons M4 to O79 (Fig. 2).Within BP, this boundary has moved from a purely lithologic definition in the twentieth century (M4/old O90), to a "global" definition in 2001 (O80), to where it is now and calibrated internally to both the base Neogene GSSP in northern Italy and the ODP Leg 154 chronometer between Horizons LM3C and O85 (old M2).

Horizons and Events
The geologic ages for BP GNATTS were derived through orbital scaling of ODP Leg 161 (Horizons PS107-LPS60) and ODP Leg 154 (Horizons PS50-O75) based on our internal research on these cores.The base Holocene age is assigned to PS108.Direct ages are those based on the same type of event in both the research and the GoM.For example, the highest occurrence (HO) of the nannofossil Discoaster brouweri is used in both BP GNATTS and the research on the ODP Leg 154 cores.Ages are considered indirect when a GoM event (e.g., HO or HCO) is sequenced to different event type (e.g., HRO or HIO, respectively) in the research.Associated ages are those assigned ages only through correlations established in the GoM framework; all benthic foraminifera events fall into this category.Bioevent ages derived from sampling of ODP Leg 154 and Leg 161 are maintained at three decimal precision, when expressed in mega-annums (Ma).Errors for these ages are the differences in assigned ages for the next sample analyzed upwards or downwards in the composite section (Supplementary Materials; Tables S5-S16, S18-S23).
Most of the 174 Chart Horizons (including Locals) are defined by multiple events and various combinations of events from each of the three microfossil groups.The age for a Chart Horizon is the youngest geologic age determined for either a planktonic foraminiferal or calcareous nannofossil event associated with that Chart Horizon (Fig. 2; Table S1).Direct age calibrations to ODP Leg 154 have been made for 86% of the Chart Horizons.An extreme example of an indirect calibration would be Hori zon M82, where the HO of the nannofossil Disco aster bellus in the GoM has been tied to its HCO in the ODP Leg 154 research.
There are 468 Neogene events on BP GNATTS, which yielded an effective resolution of 101 k.y. for nannofossil events, 397 k.y. for planktonic foraminifera events, and 344 k.y. for benthic foraminifera.About 60% of these events are fossil tops.Chart Horizon resolution is highest in the Pleistocene (67 k.y.) and lowest in the Early Miocene (261 k.y.)-also typical of global calcareous microfossil biozonations and other GoM frameworks.

Global Biozonations
BP's routine application of fossil bases in GoM wells enabled full calibration to global biozonations, which are weighted toward fossil appearances.Three nannofossil biozonations are included on BP GNATTS.Emendations of the NN zonation of Martini (1971) have resulted in additions of Neogene subzones (Rio et al., 1990b;Raffi and Flores, 1995;Blair et al., 2017) and a new zone (NP26) at the top of the Paleogene (de Kaenel et al., 2017).These emendations yield 794 k.y.resolution for the Neogene.The CN zonation of Okada and Bukry (1980) has comparable Neogene resolution (853 k.y.) and shares many events with Martini (1971).The CNM-CNPL zonation of Backman et al. (2012) We have derived ages for all zonal and subzonal events in each of the three nannofossil biozonations shown on BP GNATTS from our internal research.This information is summarized in Table 1, along with all remaining nanno fossil events presented in Backman et al. (2012).We share 47 dated events from ODP Leg 154 sites (including the Oligocene) with Backman et al. (2012), but have significant age differences (>100 k.y.) for 22 events.As previously discussed, our methodology that underpins BP GNATTS-investing significant sample examination time to establish nannofossil appearances and extinctions-is responsible for our longer stratigraphic ranges.More restricted taxonomic concepts can be invoked to explain why we may have established shorter stratigraphic ranges.For example, Curry et al. (1995) dated the LO of the nannofossil Helicosphaera ampli aperta at 20.43 Ma in Hole 926B (ODP Leg 154), whereas our determination in Hole 926B for this event was 19.115 Ma (Boesiger et al., 2017).However, we determined the LO for a very similar species, Helicosphaera scissura, at 20.350 Ma and comparable to the age determined by Curry et al. (1995).Four nannofossil bioevents have discrepancies greater than one million years between Backman et al. (2012) and our study.We have suggested alternative event types or taxonomic fixes for all such discrepancies (Table 1).We believe only the HO of Helicosphaera recta (base Zone NN1) and the HO Triquetrorhab dulus carinatus (base Zone NN3) are problematic in global nannofos-sil biozonations, the latter also within the GoM Basin.When operating ~100 k.y.resolution, it is necessary to more precisely define taxa.For example, the total range of the nannofossil Discoaster hamatus defines an upper Miocene zone used in all three Neogene schemes (NN9, CN7, CNM13).Our stratigraphic range for larger specimens (>15µm) is 1.512 m.y.(Browning et al., 2017), 0.67 m.y.longer than reported by Backman et al. (2012).For even smaller specimens of Discoaster hamatus, the total stratigraphic range is estimated around 2.280 m.y. in the ODP Leg 154 cores.
Three planktonic foraminifera biozonations are included herein (Fig. 2).The emended N-P Blow (1969) zonation (Fig. 2) has 1152 k.y.resolution for the Neogene.The M-Pl-Pt zonation of Berggren et al. (1995aBerggren et al. ( , 1995b)), which has 886 k.y.resolution, listed 135 Neogene events calibrated to the GPTS of Cande and Kent (1995).Wade et al. (2011) drew from many sources, the ultimate being British Petroleum, and summarized the history of tropical and subtropical Neogene planktonic foraminifera zonations.They listed 108 Neogene events and their sources alongside the ages of Berggren et al. (1995aBerggren et al. ( , 1995b)).The Wade et al. (2011) ages are used herein, in conjunction with those derived from in-house research on ODP Leg 154 cores (Table 2).The upper Oligocene zones (P22, O6, O7) on BP GNATTS are extensions of the three aforementioned Cenozoic frameworks into the Paleogene.Significant differences in age determinations for planktonic foraminifera bioevents between Wade et al. (2011) and our internal research, as with the nannofossils, are best explained by methodology and taxonomy.

Curves
Three columns outside biostratigraphy were included for reference and calibration.The Earth's major orbital eccentricity periodicities (~100 k.y. and 405 k.y.) derived from the Laskar et al. (2004) solution are shown near the left side of BP GNATTS.Geomagnetic polarity appears to the left of the orbital scale (from Ogg et al., 2016;Hilgen et al., 2012).The eustatic sea level sequences and transgressiveregressive cycles at the right side of the chart are calibrated to BP GNATTS through Time-Scale Creator GTS 2016 by Purdue University, West Lafayette, Indiana, USA, (engineering .purdue.edu/stratigraphy /tscreator) founded on the SEPM Chart #2 by Hardenbol et al. (1998).

DISCUSSION
BP GNATTS is a step-change in stratigraphic resolution (144 k.y.) relative to pub lished Neogene calcareous microfossil schemes.It doubles stratigraphic resolution relative to published industrial GoM Neogene biostratigraphic charts (Styzen, 1996;Lawless et al., 1997) and is distinguished from all other industrial schemes by being "fully-tuned" to orbital periodicities.Neogene resolution for nannofossil events marking BP GNATTS Chart Horizons is 142 k.y., whereas the effective resolution for all nannofossil events on BP GNATTS is 101 k.y.These are five-and fourfold increases in resolution relative to the zonal and event level resolution of the Backman et al. (2012) global scheme.For planktonic foraminifera, Neogene resolution on BP GNATTS Chart Horizons is 435 k.y., doubling that of the highest resolution global biozonation of Berggren et al. (1995aBerggren et al. ( , 1995b)).
BP GNATTS has utility beyond a biostratigraphic framework for the GoM Basin, where it has impacted the entire value chain in the GoM from exploration to production and is the foundation for yet higher resolution, reservoir biostratigraphic frameworks.BP GNATTS has been tested in selected portions of the geologic column outside the GoM Basin, in both research and industrial settings.Extending this improved resolution geographically has new found application for the interpretation and correlation of cyclostratigraphic, magnetostratigraphic, and eustatic records.

Within the GoM
Per BP methodology, microfossil abundance peaks are the fundamental unit used in correlations.In theory, biostratigraphic resolution is then only limited by the number of abundance cycles that can be uniquely defined by their microfossil content and stratigraphic position.The influence of facies on microfossil recovery and preservation is usually a primary practical limitation on biostratigraphic resolution, affecting the abundances of species to entire microfossil groups, as well as the number of taxa that may be utilized in correlations.This is less problematic in DW GoM Neogene, although significant changes in facies affecting microfossil recovery and diversity can exist between an original hole and its sidetrack or bypass holes.One of the most important influences on stratigraphy in DW GoM is salt tectonism.This structural component affects the placement, orientation, and continuity of stratigraphic section.The effects of salt can be dramatic, involving thousands of feet of repeated, folded, or inverted section.The mixing of fossil assemblages and remobilization of sediments through salt movement must always be considered in biostratigraphic analyses and interpretations in the GoM.Salt tectonism and redeposition can cause discontinuities in the sequence of biostratigraphic events.The effects of such geologic processes are better detected by higher resolution schemes utilized in well-to-well correlation (see Denne, 2009).Detection of "out-ofsequence" events such as reworking, down-hole caving of drill cuttings, faulting, and repeated stratigraphic section is much more likely with improved resolution.Higher resolution biostratigraphy increases the number of correlations between sections, which in turn, improves the precision and accuracy of these correlations.Errors in correlation are not always obvious when lower resolution biostratigraphic schemes are applied to expanded GoM sections.In such settings, biostratigraphic events can appear to be in sequence, but in reality, there may be errors of hundreds to more than a thousand feet because of a combination of high sedimentation rates, reduced microfossil recoveries, and the low number of events.
The BP GNATTS chart has been built upon a succession of BP heritage GoM Neogene biostratigraphic charts.These heritage charts have never been static frameworks.Efforts in twenty-first century post-merger BP focused on improving both the reliability and resolution of GoM Chart Horizons, which are integrated into subsurface mapping and wells.BP GNATTS is strengthened by the next level of biostratigraphic data, whose entire detail is beyond the scope of this publication.However, three examples are presented below.
Both the resolution and reliability of Chart Horizons are enhanced by the addition of biostratigraphic events through well analyses.The evolution in the definition and recognition of two upper Miocene Chart Horizons (M70 and M68) is presented in Table 3.In pre-merger BP, three events marked the M70 Horizon and a single event marked the M68 Horizon; none were directly calibrated to geologic time.One of these four events could not be proven in post-merger BP through analyses of DW GoM wells, nor by research on ODP Leg 154 cores (the HO of the nannofossil Helicosphaera walbersdorfensis is one of only two Neogene nannofossil events in heritage company schemes that could not be confirmed post-merger).A new microfossil abundance cycle (i.e., Local), observed in several wells across DW GoM, has been included on BP GNATTS.Six calcareous microfossil events, all calibrated to geologic age through ODP Leg 154 research, now marked two Chart Horizons and one Local on BP GNATTS.Four additional microfossil events for Chart Horizon M70, three of which have been tied to geologic age in our research, have been observed in a limited number of DW GoM wells and were not included on BP GNATTS (Table 3).
The use of multiple types of biostratigraphic events, as in Chart Horizon M70, provides additional benefits.Using multiple microfossil groups for confirmation is standard practice in the GoM.Abundance changes of individual taxa or taxa groups have practical utility in their speed of recognition, but must be evaluated relative to lateral changes in facies in the GoM that can affect microfossil abundances.The sequencing of fossil bases (lowest occurrences) into the biostratigraphic framework has proven critical to interpretations of redeposition in DW GoM wells.
A second way to illustrate the use of different types of events and biostratigraphic resolution beyond BP GNATTS is through the lens of an individual taxon (Table 4).Standard industrial biostratigraphic schemes are founded on microfossil tops (HOs), supplemented by down-hole abundance increases in taxa, due to concerns about the caving of materials in well bores.We have observed that down-hole caving is not generally problematic in DW GoM wells drilled since the turn of the century.At first pass, this brings into play the use of fossil bases (LOs) and down-hole abundance deceases (DECs); this effectively doubles biostratigraphic resolution and makes full calibration to published academic biozonations possible.The use of multiple events for a single taxon has been taken to the extreme for the nannofossil Reticulofenestra pseudoumbilicus (Table 4).Here, twenty events recognized among DW GoM wells have been calibrated either directly through the ODP Leg 154 research or by "asso ciation" with marker events on BP GNATTS through well analyses.Ten of the more proven events mark Chart Horizons on BP GNATTS.Four events are related to the exit and re-entry (RE) of R. pseudoumbilicus from the GoM Basin during the early Middle Miocene and early Late Miocene.The timing for three of these four events involving the two disappearances of R. pseudoumbilicus is different between the GoM and the ODP Leg 154 sites near the equator (see Table 4).The disappearance of R. pseudoumbilicus in the early Late Miocene has long been studied and considered by some to represent a significant reorganization in the Neogene carbonate producing community (Rio et al., 1990b;Young, 1990;Takayama, 1993;Raffi and Flores, 1995;Backman and Raffi, 1997;Kameo and Bralower, 2000;Krammer et al., 2006).For the planktonic foraminifera Catapsydrax dissimilis, two of the five events recognized in GoM wells have been incorporated into BP GNATTS (Table 4).
A third example of biostratigraphic resolution beyond BP GNATTS is related to geographic proximity.Extension of BP GNATTS to field scale and individual well correlations

Extending outside of the GoM
All the research ages used to ground BP GNATTS are from sections outside the GoM Basin.Pleistocene ages for Chart Horizons above PS50 were derived directly from nannofossil research on ODP Leg 161 sites in the western Mediterranean Sea (de Kaenel et al., 1999; this study).While maintaining consistency in both methods and taxonomy, as de Kaenel was the primary researcher on both the ODP Leg 161 and Leg 154 cores, these data also contained the requisite resolution and detail to constrain all but two of these 24 Pleistocene Chart Horizons (Fig. 2; Tables S1-S2).Late Oligocene to earliest Pleistocene ages for nannofossil and planktonic foraminifera bioevents were based on our own research on ODP Leg 154 sites in the western tropical Atlantic Ocean (Tables S5-S23).
BP GNATTS was fully applied to the base Neogene GSSP in northern Italy, where all seven Chart Horizons (LM3B-O79) were recognized and dated by ODP Leg 154 research ( Table 5).A total of 21 events were established in sequence between the GSSP and ODP Leg 154, including 13 correlations involving 15 GoM events (135 k.y.resolution).From this triangulation, the calcareous microfossil event closest to the Neogene/Paleogene boundary is the HO of the nannofossil Sphenolithus capricornutus, dated at 22.998 Ma and 0.5-0.9m above boundary in the GSSP (Aubry and Villa, 1996;Bergen et al., 2017).The base of planktonic foraminifera Paragloborotalia kugleri helps bracket the boundary interval in both ODP Leg 154 and the Italian GSSP (Table 5).
BP GNATTS was applied to a broader Oligocene/Miocene boundary interval through analyses of ditch-cutting samples from three exploration wells in the eastern Mediterranean Sea.All 30 BP GNATTS Chart Horizons from the upper lower Miocene (M35) down into the upper Oligocene (O75) were established in these offshore wells by the first two authors, including the sequencing of 57 bioevents on BP GNATTS at 131 k.y.resolution.

New Applications in Correlation
Initial geographic extension of the BP GNATTS framework into the western tropical Atlantic Ocean and Mediterranean Sea is encouraging for its potential application in other geographic locations.The ability to establish correlations at ~130 k.y.resolution has new found applications-provided taxonomic and methodologic rigor are maintained.Impact on the calibration and correlation of interpreted eustatic records, bridging industry and academia, is obvious.It may now be possible that calcareous microfossil biostratigraphy be utilized to evaluate and correlate Neogene cyclostratigraphic and magnetostratigraphic records, the latter previously used to assess the synchronicity of biostratigraphic events.Eustasy (change in global sea level) was initially inferred from observations of distinct, relatively synchronous intervals of deposition and erosion throughout the geologic record.One of the foundations for deciphering records of eustasy involves determining the timing and magnitude of sea level events, and often the basis for these age models is biostratigraphy.To investigate sea level on time scales of 10 5 -10 4 k.y., a standardized high-resolution biostratigraphic zonation scheme is crucial.Without an adequate age model, it is difficult to determine sedimentation rates, compare sea level proxies to global events, and to correlate sequence boundaries and flooding surfaces within a specific study area and to the global sea level curve.One of the future applications for the BP GNATTS, with resolution comparable to that of fourth and fifth order parasequences, could be its utilization in sea level studies to help better calibrate global records.Additionally, microfossils are also a tool in deciphering facies and the magnitude of sea level changes in outcrop, core, and the well bore.The construction of far more accurate age-depth profiles based on biostratigraphy equates to better estimates of sedimentation rates, hence identification of condensed sections and hiatal surfaces fundamental to sequence stratigraphic interpretations.
Advances in sequence and seismic stratigraphy during the last three decades of the twentieth century provided a template for better understanding of global sea level change and the creation of a global sea curve for the Phanerozoic (e.g., Vail et al., 1977;Vail and Hardenbol, 1979;Haq et al., 1987;Hardenbol et al., 1998).Subsequently, both curves have undergone intense scrutiny with studies focused on improving methodology, as well as testing and finding inconsistencies in both the timing and magnitude of sea level change (e.g.;Miller et al., 1998Miller et al., , 2005;;Zachos et al., 2001Zachos et al., , 2008;;John et al., 2004John et al., , 2011;;Kominz et al., 2008;Tcherepanov et al., 2008;Fielding et al., 2011;Passchier et al., 2011), effects of changes in regional tectonics (Browning et al., 2006;Moucha et al., 2008;Raymo et al., 2011), glacial isotopic adjustments associated with continental margins (Raymo et al., 2011), and far field sea level amplification (Mitrovica et al., 2009).Poor age control in many studies has generally led to the assumptions of incomplete records of Miocene eustasy (Isern et al., 2002;Eberli et al., 2010) and has made it difficult to compare records among various sites (Miller et al., 2005;Kominz et al., 2008).Conversely, certain high-resolution biostratigraphies (John et al., 2011), when applied to records previously considered to record stratigraphic discontinuities (e.g.; Marion Plateau, Australia ODP Leg 194 sites 1192Leg 194 sites -1195;;Isern et al., 2002;Eberli et al., 2010) have not recognized similar hiatal surfaces within the section, and then calibrated them to other sites and global sea level curves.
Late twentieth century Cenozoic geologic time scales (e.g., Berggren et al., 1995aBerggren et al., , 1995b) ) were scaled on magnetostratigraphy, which has been used to test the geographic synchronicity of biostratigraphic events (e.g., Berggren et al., 1985;Backman and Shackleton, 1983).The ages and durations of most Neogene polarity chrons are now adjusted to cyclostratigraphic calibrations (Ogg et al., 2016) and were presented in the 2012 Neogene Time Scale (ATNTS2012) by Hilgen et al. (2012).In general, global microfossil biozones can be subdivided by at least two polarity chrons (Ogg et al., 2008).Conversely, the stratigraphic resolution of Chart Horizons on BP GNATTS (144 k.y.) nearly double the resolution of geomagnetic polarity chrons (248 k.y.) for the Middle Miocene to Recent, although their resolution is roughly equivalent for the Early Miocene (Table 6).There is an overall decrease in biostratigraphic resolution on BP GNATTS with geologic age.This suggests the potential for increasing biostratigraphic resolution in the GoM lower Miocene.
BP GNATTS is scaled through sampling and analyses of cyclostratigraphic reference sections.Miocene to Pliocene section from ODP Leg 154 was sampled at 18 k.y.resolution for nannofossils and 32 k.y. for foraminifera.The sampling intervals from ODP Leg 154 equate to errors in geologic age determinations that are comparable to the orbital periodicities of precession (19 k.y. and 23 k.y.) and obliquity (41 k.y.).Stratigraphic resolution of the nannofossil and planktonic foraminifera events on BP GNATTS are comparable to the major periodicities of eccentricity (~100 k.y. and 405 k.y.).The stratigraphic resolution and precision of BP GNATTS can be improved by sampling and analyses of: (1) ODP Leg 154 cores to reduce error in age determinations; (2) other cyclostratigraphic reference sections in the circum-North Atlantic Basin; and (3) modern GoM wells in targeted portions of the geologic column, such as the lower Miocene.Two factors are key to improved accuracies of derived geologic ages herein: (1) the astronomical "retuning" of the cyclostratigraphic record preserved in the ODP Leg 154 cores; and (2) full astronomical tuning in continuous overlapping sections with magnetic polarity for the lower Miocene to uppermost Oligocene (Ogg et al., 2016).In its present form, BP GNATTS has sufficient biostratigraphic resolution to evaluate the completeness and correlation of cyclostratigraphic records at the level of eccentricity.

CONCLUSIONS
The twenty-first century BP post-merger biostratigraphy research program has resulted in the first published astronomically-tuned industrial biostratigraphic chart (BP GNATTS).This BP Neogene GoM chart represents dec ades of contributions from numerous biostratigraphers and a step-change in biostratigraphic resolution and calibration.Consistent methodology and uniform taxonomic concepts have minimized two factors that can plague biostratigraphic resolution and correlation, as exemplified by discrepancies in age estimates of bioevents from the astronomical chronometer of ODP Leg 154 sites in the BP research relative to most recent published ages (e.g., Wade et al., 2011;Backman et al., 2012).Geologic setting, coupled with the high number of wells, has eliminated poor preservation and recovery as factors.The use of multiple events for Chart Horizons has enhanced their reliability, but also the ability to determine when an individual microfossil event-or "pick"-is out of sequence in a well.
For the GoM Basin, a more reliable and higher resolution framework translates into enhanced detection of stratigraphic discontinuities, such as redeposition, slumps, unconformities, and faults.Geographic proximity of wells within the basin translates to still higher resolution and ability to correlate individual sands in expanded sections.The resolution of the integrated BP GNATTS at the level of Chart Horizons (with Locals) is five to eight times greater than that of stand-alone global calcareous microfossil biozonations.Although only portions of BP GNATTS have been tested in the Mediterranean Sea, we hope that this new resolution can be extended both stratigraphically and geographically to enhance correlation of other time scale parameters, as well as eustatic records.For example, BP GNATTS nannofossil and planktonic microfossil resolution are comparable to the major periodicities in eccentricity of ~100 k.y. and 405 k.y., respectively.
The use of the term "high resolution" is ubiqui tous in biostratigraphy and suggests technical limit and a mature state of the science.The demonstration of a significant jump in resolution in the most researched part of the geologic column for these two planktonic micro fossil groups contradicts the assessment of the discipline's maturity.Our experience has shown that biostratigraphic frameworks are not static; they improve with each repetition in both the research and well analyses.Technical transfer and the ability to reproduce results are best done with daily contact over a period measured in weeks, months, or years.This is a challenge for biostratigraphy, which has become a scarce specialty in both industry and academics.Yet methodologies, such as those presented, can be applied outside of the GoM and certainly outside of industry, as the need for better calibration of geologic and climatic events has become of paramount interest in the academic community (Pälike and Hilgen, 2008).Results presented here lend conviction to the promise that microfossil biostratigraphy is far from the end of its constructive growth, rather it is a discipline with great current utility and with a realistic expectation for developing new and exciting applications.

APPENDIX
Three topics are presented in the Appendix: (1) revised estimates in geologic ages for selected calcareous nannofossil events; (2) two zonal emendations for the Miocene portion of the Backman et al. (2012) calcareous nannofossil biozonation; and (3) more discussion on the directional sense of bioevents.

Revised Geologic Age Estimates for Select Calcareous Nannofossil Bioevents
The BP taxonomic concepts and biostratigraphy for over 200 Oligocene-Recent nannofossil species were recently published as a series of five papers in the Journal of Nannoplankton Research (Bergen et al., 2017;Blair et al., 2017, Boesiger et al., 2017;Browning et al., 2017;de Kaenel et al., 2017).Since their publication in November 2017, subsequent research on Ocean Drilling Program (ODP) Leg 154 cores led to revisions of the geologic age estimates for a limited number of bioevents presented in four of these five manuscripts (Table 7).Data from the base Neogene global boundary stratotype section and point (GSSP) near Lemme-Carrosio in northern Italy was also revised to reinforce correlations to ODP Leg 154 (Table 5).Our range chart data for the GSSP is introduced here for the first time (Table S3).For the Oligocene/Miocene boundary interval, the base of medium-sized Discoaster druggii and the corresponding base of Zone NN1 has been adjusted down one sample in ODP Hole 926B from that presented by de Kaenel et al. (2017) and is now dated at 23.038 Ma (versus 23.030 Ma).For the genus Sphenolithus, the top of Sphenolithus delphix has been adjusted up one sample in Hole 926B and the base of Sphenolithus paratintinnabulum down two samples in the same hole section.The remaining 19 revisions were the result of focused "marker hunts" and abundance counts (500 fields-of-view at 1000×), in order to complete BP quality control on the ODP Leg 154 research for the upper Miocene to basal Pleistocene between BP Chart Horizons M103 to PS42.

Emendations to CNM Zonation of Backman et al. (2012)
The results from our research showed it necessary to emend two Miocene zonal events in the Backman et al. (2012) nannofossil biozonation; both involve significant differences in age determinations for the lowest occurrences (LO and LO absence) of two species in this biozonation.Instead, the placement of these two events in published studies appear to represent abundance changes of the two species.Such discrepancies can occur when taxa are extremely rare and sporadic at the ends of their stratigraphic ranges, in combination insufficient analyses times.Specimen counts of the relevant taxa were done on smear slide preparations for 500 fields-of-view at 1000× magnification to document these abundance shifts.A total of 1000 fields-of-view were examined to document these rare occurrences.Emendations to the Cenozoic nanno fossil biozonation of Martini (1971) were done by Blair et al. (2017) andde Kaenel et al. (2017).

CNM1-Sphenolithus conicus Partial Range
Zone-emended Original definition: Interval from top of Sphenolithus delphix to the base of Sphenolithus disbelemnos.Emended definition: Interval from top of Sphenolithus delphix to the lowest increase occurrence (LIO) of Sphenolithus disbelemnos.Authors: Backman, Raffi, Rio, Fornaciari, and Pälike (2012).Original Reference Section: ODP Hole 1218A, central tropical Pacific Ocean.Additional Reference Sections: ODP Hole 926B, Lemme-Carrosio (northern Italy).Remarks: This zone was first defined in the Mediterranean Sea (Italy) zonation by Fornaciari and Rio (1996) as the Sphenolithus delphix-Sphenolithus disbelemnos Interval Subzone (MNN1c).The top of this zone was constrained by the base (LO) of Sphenolithus disbelemnos by Fornaciari and Rio (1996) and later extended into the Pacific Ocean by Backman et al. (2012) in their low to middle latitude Neogene biozonation.In contrast, we have observed significant overlap in the stratigraphic ranges of Sphenolithus dis-belemnos and Sphenolithus delphix in Gulf of Mexico (GoM) deep-water wells, as well as our research reference sites on the Ceará Rise offshore Brazil (ODP Leg 154) and the base Miocene GSSP section in northern Italy (Table 5).Both of these bioevents actually fall within the terminal Oligocene in Gulf of Mexico wells and these two reference sections (Zone NP26).Backman et al. (2012) gave on age of 23.06 Ma for the extinction of Sphenolithus delphix based on ODP Site 1218B, which is within sampling error (0.021 Ma) for our age of 23.072 Ma at Site 926B (Table 1).However, Backman et al. (2012) also determined an age of 22.41 Ma for the appearance (base or lowest occurrence [LO]) of Sphenolithus disbelemnos in ODP Hole 1218A.We have dated the base of Sphenolithus disbelemnos at 23.274 Ma in ODP Hole 926B, nearly one million years older (Table 1).The LIO of Sphenolithus disbelemnos in ODP Hole 926B (Fig. 6) is instead more congruent with the age determined for its LO by Backman et al. (2012), differing by only 90 k.y.(Table 1).The Oligocene-Miocene boundary section of Lemme-Carrosio in northern Italy provides a solution to this discrepancy.Fornaciari and Rio (1996) placed the LO of Sphenolithus disbelemnos at 57 m in this section (subjacent sample 55 m), twenty-two meters above the base of the Miocene.We placed its LIO in this section at 22.1 m (subjacent sample 20.2 m) above the base of the Miocene, but observed the actual base of Sphenolithus disbelemnos 9.4 m below the base of the Miocene (Table 5).
Although the upper Oligocene base for Sphenolithus disbelemnos can be explained by its very rare occurrences in the lower part of its stratigraphic range, taxonomy must also be considered.Our concept of Sphenolithus disbelemnos is more restricted than other authors, likely including Sphenolithus paratin-  tinnabulum Bergen andde Kaenel, 2017, in Bergen et al. (2017).The lowest increase occurrences of both species are the same in ODP Hole 926B.Spheno lithus paratintinnabulum has both an older appearance ( Table 5) and younger extinction than Sphenolithus disbelemnos (Bergen et al., 2017).

CNM14-Reticulofenestra pseudoumbilicus Partial
Range Zone-emended Original definition: Interval from the top of Discoaster hamatus to the base of the interval of absence of Reticulofenestra pseudoumbilicus.
Emended definition: Interval from the top of Discoaster hamatus to the base of the interval with decreased numbers of Reticulofenestra pseudoumbilicus.Authors: Backman, Raffi, Rio, Fornaciari, and Pälike (2012).Reference Section: ODP Site 926 (western tropical Atlantic Ocean) Remarks: Backman et al. (2012) defined the top of this zone as "the base of the interval of absence of Reticulofenestra pseudoumbilicus."This definition is highly problematic for several reasons.First is the precise definition of this event.Backman et al. (2012) also referred to "the interval of almost total absence of R. pseudoumbilicus in upper Miocene sediments (the so-called "R.pseudoumbilicus paracme")" in their remarks on assemblages in the description of this zone.None of these three terms are synonymous: interval of absence, interval of almost total absence, or paracme.Backman and Raffi (1997) dated the paracme of R. pseudoumbilicus in ODP Hole 926B from 7.09 Ma to 8.80 Ma, using a size of >7 µm for the species.These ages are confirmed by our counts of specimens of R. pseudoumbilicus from Hole 926B (Table 1), but are constrained by abundance increases that in neither instance represent common occurrences or acmes (Figs.7A, 7B).A second consideration is methodsspecifically the time spent searching for rare specimens-which directly affect whether the "paracme" interval appears entirely devoid or almost entirely devoid of specimens.Third is the size used to define the species, as the age of Reticulofenestra events clearly vary with size.A general rule is that smaller specimens have longer stratigraphic ranges; this is true both for Paleogene and Neogene reticulofenestrids.We were unable to determine an interval either absent or almost entirely absent of specimens of Reticulofenestra pseudo umbilicus (>7 µm) in upper Miocene sediments of ODP Leg 154 cores (Table S7 (1973) was defined from the highest occurrence of Discoaster hamatus up to the lowest occurrence of Discoaster neorectus.Bukry (1975) later added the lowest occurrence of Discoaster loeblichii as a second event to mark the top of this Subzone.The Discoaster bellus Subzone was later codified (CN8a) by Okada and Bukry (1980).Backman et al. (2012) illustrated the lowest occurrences (bases) of both Discoaster loeblichii and Discoaster neorectus (their figure 2) in the uppermost portion of Zone CNM14.The base of Discoaster loeblichii is a very reliable event in GoM deep-water wells, in addition to the ODP Leg 154 research materials.This event is dated at 8.738 Ma in Hole 926B (Table 1) and is suggested as a secondary marker for the top of Zone CNM14.

Directional Sense of Bioevents
Some terminology is in a "top-down" or "downhole" sense when referring to wells and drilling and must be distinguished when using in a "bottomsup" sense for outcrops and deep-sea research cores (Fig. 4).Two associated paired terms and abbreviations are: DEC (decrease) and INC (increase), DA (disappearance) and RA (reappearance).Increase (INC) and decrease (DEC) are used herein only in a top-down sense.They are avoided in a bottoms-up sense and the terms highest increase occurrence (HIO) and lowest increase occurrence (LIO) are preferred for outcrop and cores.The use of various terms to describe the exit and re-entry of an individual species from a stratigraphic section are necessary so that ages can be related to published studies.Otherwise, the ages can be offset a single sample (see Fig. 4).For example, Wade et al. (2011) used the terms disappearance (DA) and reappearance (RA) when referring to certain planktonic foraminifera events, whereas nanno fossil studies (Backman et al., 2012;Agnini et al., 2014) prefer the use of the term "absence."Events are also offset one sample when referring to changes in planktonic foraminifera coiling directions (Fig. 5) in either a "top-down" or "bottoms-up" sense; the terminology is also reversed dextral to sinistral (D/S) versus sinistral to dextral (S/D).

ACKNOWLEDGMENTS
We thank many people in BP Gulf of Mexico (GoM) Exploration and Production for their support and encouragement over the years-and Graham Vinson, Liz Jolley, Tim Hill, and John Farrelly for permission to publish.We are grateful to all the GoM biostratigraphers from each of the three heritage companies whose efforts spanning decades contributed to this work.Appreciation is extended to our Egyptian colleagues in BP Cairo for helping to calibrate the Early Miocene to Late Oligocene and Integrated Ocean Drilling Program staff at the Bremen Core Repository in Bremen, Germany.We extend special gratitude to Sheila Barnette and Jim Newell for their contributions to the GoM and the original foraminiferal Ocean Drilling Program Leg 154 BP research.Thank you to the authors' families, as efforts to get this work into publication along with the proceeding five taxonomic nannofossil papers took untold time and dedication to the science in effort to pass this stratigraphy to the next generations.Special thanks to the two reviewers and Geological Society of America editors for their many helpful suggestions to improve the quality of this manuscript.Lastly, we would like to thank the GoM Exploration and Production teams who have fostered an integrated approach to subsurface description with biostratigraphy, geophysics, and geology.G. caribbeanica A2/G.caribbeanica (>3) (inc) GO assemblage/Gephyrocapsa omega (<4) (inc) Umbilicosphaera rotula GS assemblage/Gephyrocapsa spp.(<3) (inc) lower GS assemblage/G.caribbeanica (inc) Pseudoemiliania lacunosa "1"(delicate) lower Pseudoemiliania lacunosa "1" Pseudoemiliania lacunosa "1.5" lower Pseudoemiliania lacunosa "1.5" (inc) Gephyrocapsa omega (>4) Pseudoemiliania lacunosa "1.5" (inc) P. lacunosa "2"/Crenalithus japonicus P. lacunosa "2 large"/Crenalithus asanoi P. lacunosa "3"/Crenalithus doronicoides (HRO) P. lacunosa "3"/Crenalithus doronicoides (inc) Gephyrocapsa caribbeanica "B" (>6.
Figure 6.Total abundance counts of Sphenolithus disbelemnos from Ocean Drilling Program (ODP) Hole 926B offshore NE Brazil.LIO-lowest increase occurrence; LFO-lowest few occurrence.

Locality map of mate- rials used in this publication.
al., 1999)from ODP Leg 161 sites 974-977 and site 979 in the western Mediterranean Sea.

TABLE 1 .
Backman et al. (2012)OSSIL BIOEVENT AGE ESTIMATES (continued) Events shaded in light grey are secondary zonal markers.Primary order of events by age as determined byBackman et al. (2012), some events inBackman et al. (2012)are from Figures (no ages provided).Underlined ages indicate discrepancy in the order of event ages; alternate BP events and taxa are suggested for these discrepancies.*Agnini et al. (2016) Oligocene age.Abs.-absence; Dec.-decrease.

TABLE 2 .
GLOBAL NEOGENE PLANKTONIC FORAMINIFERAL BIOEVENT AGE ESTIMATES Primary order of events by age as determined by Wade et al. (2011).Underlined ages indicate discrepancy in the order of event ages.Coiling change (S/D or D/S), disappearance (DA), and reappearance (RA) events are indicated from "bottom to top" .Site/Hole designations shown as just Ocean Drilling Program LEG 154 indicates the use of more than one Site/Hole.The following abbreviations are used for bioevent terminology: S-sinistral; D-dextral, LO-lowest occurrence; HO-highest occurrence; HRO-highest regular occurrence; ch.-chambers; LCO-lowest common occurrence.

TABLE 4 .
INCREASED BIOSTRATIGRAPHIC RESOLUTION: TWO EXAMPLES OF MULTIPLE EVENTS FOR A SINGLE TAXON BP horizon markers are in boldface.Direct/indirect calibrations are relative to Ocean Drilling Program Leg 154; associated calibrations are relative to BP Gulf of Mexico (GoM) events.Increased Resolution: Multiple Events per Taxon.HO-highest occurrence; HRO-highest regular occurrence; HIO-highest increase occurrence; LCO-lowest common occurrence; LFO-lowest few occurrence; LRO-lowest regular occurrence; RE-re-entry; HAO-highest abundant occurrence; LO-lowest occurrence.
Steininger et al. (1997)ype section and point (GSSP) depths:Aubry and Villa (1996);Steininger et al. (1997).† Highest common occurrence (HCO) is a maximum value in GSSP because of low abundance.§ Highest abundant occurrence (HAO) is highest increase occurrence (HIO) in GSSP because of low abundance.Underlined events are out of sequence in GSSP and ODP Leg 154.

TABLE 7 .
REVISED GEOLOGIC AGES FOR SELECTED NANNOFOSSIL EVENTS FROM NEW RESEARCH ON ODP LEG 154* ODP-Ocean Drilling Program; HO-highest occurrence; HRO-highest regular occurrence; INC-increase (top-down); HIO-highest increase occurrence; LOlowest occurrence; HAO-highest abundant occurrence.