A MULTIVARIATE ANALYSIS OF THE RECOVERY OF CALCAREOUS NANNOPLANKTON AND PLANKTONIC FORAMINIFERA FROM THE CRETACEOUS/PALEOGENE (K/P) MASS EXTINCTION

0
232
You can download this material now from our portal

A MULTIVARIATE ANALYSIS OF THE RECOVERY OF CALCAREOUS NANNOPLANKTON AND PLANKTONIC FORAMINIFERA FROM THE CRETACEOUS/PALEOGENE (K/P) MASS EXTINCTION

ABSTRACT

 

The Cretaceous-Paleogene (K/P) bolide impact decimated surface ocean ecosystems.

While the extinction has been well studied, there have been few investigations of the recovery of ecosystems from the K/P extinction.  The relationship between the post-extinction ecologies of nannoplankton and planktonic foraminifera can lend insight into how ecosystems were rebuilt after the K/P impact.  This study utilizes multivariate analysis on abundance counts from five sites with wide global distribution to compare the recoveries of these calcareous plankton groups.  The analysis shows there is a distinct geographical differentiation in the recovery of both groups, with assemblages in the northern hemisphere surface oceans diversifying more slowly than southern hemisphere surface oceans.  This is most likely a result of higher extinction intensity caused by the concentration of dust created by the bolide impact in the northern hemisphere.  The multivariate analysis results suggest a probable long-term influence on the post-extinction marine ecosystem was increased nutrient concentration in the surface ocean resulting from the collapse of the biological pump.  Nannoplankton recover before planktonic foraminifera in the open ocean, as expected, but after planktonic foraminifera on the shelf.  Dinoflagellates, which were common at many shelf locations across the K/P boundary may have provided another food source for planktonic foraminifera and hindered nannoplankton diversification by invading niche space emptied by the K/P extinction.  Ecological recovery of calcareous microplankton took place only after the biologic pump was restored ~300 Kyr after the K/P extinction, removing the enhanced nutrient concentrations from the surface ocean.

TABLE OF CONTENTS

 

 

List of Tables…………………………………………………………………………………….. v

List of Figures…………………………………………………………………………………… vi

Preface…………………………………………………………………………………………………………………..…… vii

Acknowledgements……………………………………………………………………………………………………….viii

 

MICROPLANKTON ECOSYSTEMS AFTER THE K/P BOUNDARY……………………….. 1

 

METHODS………………………………………………………………………………………. 5

Data………………………………………………………………………………………… 5

   Brazos River Nannoplankton Counts……………………………………………………….. 7

Multivariate Analysis………………………………………………………………………. 7

Diversity……………………………………………………………………………………..9

Age Model………………………………………………………………………………….. 9

Paleoecology..…………………………………………………………………………….. 11

 

RESULTS………………………………………………………………………………………. 12

NMS Results……………………………………………………………………………….12

Diversity and Extinction Intensitiy…………………………………………………………13

 

DISCUSSION…………………………………………………………………………………… 20

Environmental Stress after the K/P Impact……………………………………………….. 20

Trends in the Recovery of Post K/P Microplankton Ecosystems…………………………. 21

Ecological Recovery after the K/P Extinction……………….……………………………. 23

 

CONCLUSIONS……………………………………………………………………………………….26

 

References………………………………………………………………………………………. 28

 

Appendix A:  Comprehensive Taxonomy……………………………………………………… 40

 

Appendix B:  Paleoecology Table……………………………………………………………… 52

 

Appendix C:  Supplements to Thesis…………………………………………………………… 59

Detailed NMS Methods……………………………………………….……………………..60

Age Model ……..…………………………………………………………………………61

Range/Abundance charts of selected Danian taxa….………….……………………….. 67

MICROPLANKTON ECOSYSTEMS AFTER THE K/P MASS EXTINCTION

 

The mass extinction event at the Cretaceous-Paleogene (K/P) boundary (65.68 Ma) devastated oceanic ecosystems.    Planktonic protists that thrived in the Cretaceous were almost completely eradicated, with >90% of planktonic foraminifera and calcareous nannoplankton species going extinct across the K/P boundary (Kaiho, 1994; Bown et al., 2004).  Oceanic food webs and the biologic pump were disrupted and took millions of years to fully recover (e.g., Zachos and Arthur, 1986; D’Hondt et al., 1998; D’Hondt, 2005).  The taxonomic basis of the extinction event has been documented in great detail for both planktonic foraminifera (e.g.,

Luterbacher and Premoli Silva, 1964; Smit and Hertogen, 1980; Smit, 1982; Keller, 1988; Smit,

1990; Molina 1994, 1995; Molina et al., 1996, 1998) and calcareous nannoplankton (e.g.,

Bramlette and Martini, 1964; Perch-Nielsen, 1969; Percival and Fischer, 1977; Romein, 1977; Perch-Nielsen et al., 1982; Pospichal, 1994).  Post extinction foraminifera were dominated by low-diversity assemblages made up of small taxa with simple morphologies (e.g., Luterbacher and Premoli Silva, 1964; Blow, 1979; Tourmarkine and Luterbacher, 1985).  Contemporaneous nannoplankton were dominated by resting cysts likely produced during high environmental stress—such as extremes in salinity or nutrients (e.g., Eshet et al., 1992; Tantawy, 2003).  Both planktonic foraminifera and nannoplankton took millions of years to recover from the extinction (Coxall et al., 2006; Fuqua et al., 2008), and diversity did not reach pre-extinction levels for as long as 15 million years after the event (Norris, 2001; Bown et al., 2004).  While the extinction has been well documented for both plankton groups, very few quantitative, ecological studies have been performed on the recovery.

The impact of a large bolide on the Yucatan peninsula in Mexico is thought to have caused the K/P mass extinction (Alvarez et al., 1980; Smit and Hertogen, 1980; Alvarez, 1987; Erikson and Dickson, 1987; Hildebrand et al., 1991; Smit, 1990; D’Hondt et al., 1994a; Toon et al., 1997).  Environmental effects of the impact led to the mass extinction of planktonic organisms, and, at the same time, delayed their recovery.  The bolide is thought to have struck at an oblique angle directed towards the north to northwest (Johnson and Hickey, 1990; Alvarez et al., 1995; Schultz and D’Hondt, 1996), ultimately concentrating ejecta in the northern hemisphere.  Models suggest that the K/P impact did not produce enough dust to cause global darkness, but ejected sufficient material into the atmosphere to hinder photosynthesis for several months or years (Toon et al., 1997; Pope, 2002).   The impact was also thought to hit a large area of evaporites which led to surface ocean acidification (Sigurdsson et al., 1991; Brett, 1992; D’Hondt et al., 1994a; Pope et al., 1997).  Global darkness, which lasted at most a few months, and ocean acidification, which would have been very brief due to the ocean’s buffering capacity (D’Hondt et al., 1994a), may have led to the mass extinction; however, these factors most likely would not have had long-term ecological impacts (Pope et al., 1997; Pope, 2002).   Another potentially important result of the impact may have been metal loading of the surface ocean.  The bolide impact may have added toxic levels of metals to the surface oceans (Erickson and Dickson, 1987).  Metal loading could have affected the surface ocean for an extended period of time and delayed ecosystem recovery (Jiang et al., in review).

The extinction event had a profound effect on the marine carbon cycle.  This is evident from the collapse of the δ13C gradient between the surface and deep ocean, which suggests a significantly weakened biologic pump (e.g. Zachos and Arthur, 1986; Stott and Kennett, 1989; Zachos et al., 1989, 1992; Hilting et al., 2008) and decreased organic matter delivery to the deep ocean (Hsü et al., 1982, Zachos et al., 1989).  There were two main recovery intervals in the carbon isotopic gradient after the K/P extinction:  a step ~300-500 Kyr after the boundary, when initial recovery of the gradient took place, and a second step ~3 Myr after the extinction, when the gradient reached pre-extinction values (D’Hondt et al., 1998; Adams et al., 2004).    The pattern of δ13C collapse and recovery in the surface ocean is seen in foraminifera and fine fraction globally after the K/P extinction.  The final recovery is thought to represent the evolution of larger plankton or grazers (D’Hondt et al., 1998; D’Hondt, 2005).

The return of diversity and dominance to pre-extinction levels is also a key aspect of the ecologic recovery of the marine plankton.  Jiang et al. (in review) showed there was a distinct geographic differentiation in the recovery of nannoplankton, and proposed that the surface ocean in the northern hemisphere recovered nearly 300 Kyr after recovery in the southern hemisphere.  Nannoplankton did not fully recover in diversity and ecology until the biologic pump was restored 3 Myr after the extinction (Fuqua et al., 2008).  In the earliest Paleocene, there were two periods of diversification in the planktonic foraminifera (Coxall et al., 2006).  Approximately 60% of total Paleogene diversity was reached 300 Kyr after the extinction, and most Paleocene foraminiferal morphologies evolved by that time.  This was followed by another rapid diversification 2-3 million years after the extinction which corresponded to a shift to an assemblage that was oligotrophic and similar in diversity to modern assemblages (D’Hondt et al.,

1994b; Norris, 1996).  Both of these events were likely related to the reestablishment of the biological pump (Coxall et al., 2006; D’Hondt, 2005), and the final recovery of planktonic foraminifera occurred at roughly the same time as that of the nannoplankton (Fuqua et al., 2008).     There is some indication that the recovery of planktonic foraminifera was geographically dependant, similar to nannoplankton.   Planktonic foraminifera quickly diversified in high latitudes while species richness remained low in low latitudes for several million years (Keller et al., 1993; Barrera and Keller, 1994).  However, there has been no comparison on how the recovery of planktonic foraminifera and nannoplankton relate.  Theoretically, because nannoplankton are a food source for foraminifera (e.g. Hemleben et al., 1989), foraminiferal diversity should be restored after the return of nannoplankton productivity opens up niche space for the planktonic foraminfera (Solé et al., 2002).

This work expands significantly on the previous work of Jiang et al, who showed there were distinct geographic patterns in the recovery of nannoplankton after the extinction.  Whereas Jiang et al. (in review) focused on the nannofossils, the base of the food web, we compare the recovery of nannoplankton with planktonic foraminifera, thereby expanding the focus to multiple trophic levels.  By looking at multiple trophic levels, we can better understand how food webs were rebuilt after the K/P extinction.  Therefore, while our work is similar to Jiang et al. (in review), it is a more intricate look into how trophic relationships were rebuilt as the marine ecosystem recovered.  Our goal is to utilize multivariate analysis of assemblage data from a transect of latitudes and oceanic environments to better understand how the recovery of nannoplankton and planktonic foraminifera were coupled in the 300 Kyr interval following the K/P boundary.  We constrain the relative timing of the resurgence of nannoplankton and planktonic foraminiferal diversity, and we use the suite of sites to determine how the base of the marine food web was rebuilt in the wake of the K/P mass extinction.

METHODS

Data

We have complied assemblage counts for nannofossils and planktonic foraminifera from five sites with a wide geographical distribution (Fig. 1; Table 1).  These include:  Shatsky Rise

(Central Pacific), Walvis Ridge (South Atlantic), Kerguelen Plateau (Indian Ocean), and El Kef,

Tunisia (Tethys Sea).  To augment these data, we counted nannofossils from the Cottonmouth

Creek K/P section of the Brazos River, Texas (for a stratigraphic overview of this section, see

Yancey, 1996 and Schulte et al., 2006).  Planktonic foraminiferal counts from the Cottonmouth Creek section were obtained from C. Liu through T. Yancey (unpublished data).  Nannofossils and foraminifera in these sections were not sampled at the same levels, and because of this, two separate multivariate analyses were run for each fossil group.  These data differ from those used in Jiang et al. (in review) by adding planktonic foraminifera and by adding new counts from the Brazos River.

A comprehensive taxonomy was established to combine data from multiple authors with differing taxonomic concepts (Appendix A).  In most cases, species were combined into genera to facilitate comparison; only those species prevalent at all sites were kept separate.

Early Paleocene nannoplankton samples are dominated by calcispheres (predominately calcareous dinoflagellates) which disaggregate readily and are present almost always as fragments.  There is wide discrepancy on how to count calcisphere fragments.  Jiang and Gartner (1986) counted every calcisphere fragment as one specimen, whereas other authors have counted every three fragments greater than eight microns as one individual (e.g. Bernoala and Monechi,

2007).  Because in most of our samples, the average size of the calcisphere fragments is around

4-5 microns, one individual was counted out of every three fragments greater than or equal to 5 microns.  All calcisphere counts from the published data were also corrected to this counting method.  Braarudospheres also are found mostly as fragments.  For this study, three braarudosphere pentalith segments or one pentalith made from three or more segments were considered one individual braarudosphere.

Fig. 1:  Paleogeographic reconstruction of the Cretaceous/Paleogene boundary (65.5 Ma) showing the locations used in the study.  Map constructed using software available at http://odsn.de/odsn/services/paleomap/paleomap.html based on paleogeographic reconstructions from Hay et al., 1999.

 

Site Location ODP/DSDP

Site

Latitude Longitude Source of data
Shatsky Rise Central Pacific ODP 1210 32º13’ N 158º16’ E Bown, 2005
DSDP 577 32º27’ N 157º43’ E D’Hondt and Keller, 1991
Walvis Ridge South Atlantic ODP 1262 27º11’ S 1º34’ E Bernoala and Monechi, 2007
DSDP 528 28º32’ S 2º19’ E D’Hondt and Keller, 1991
Kerguelen Plateau Indian Ocean ODP 738 62º43’ S 82º47’ E Huber, 1991;

Pospichal, 1993

El Kef,

Tunisia

Tethys Sea N/A 36º11’ N 8º43’ E Pospichal, 1994

Arenillas et al., 2000

 

Table 1:  Location, modern day latitude and longitude and source of data for Shatsky Rise, Walvis Ridge, Kerguelen Plateau and El Kef.

 

 

Brazos River Nannoplankton Counts

The K/P section of Cottonmouth Creek (near Brazos River, Falls County, Texas) was sampled in detail.  The Danian section at Cottonmouth Creek was sampled starting at the top of the K/P sedimentary complex thought to be an impact tsunami deposit (Bourgeois et al., 1988; Yancey, 1996).  The lowermost Paleocene of Cottonmouth Creek is typified by calcareous mudstones with intermittent calcareous and iron concretion horizons (a detailed stratigraphy is given in Yancey, 1996 and Schulte et al., 2006).  Extensive paleontological, geochemical and sedimentological studies have been done on the Cottonmouth Creek section and the nearby “Brazos 2” core (Hansen et al., 1987; Hansen et al., 1993a,b; Keller, 1989a,b).  Biostratigraphy indicates this section is fairly complete and corresponds to NP1 and lowermost NP2 nannofossil biozones of Martini (1971).  Added to the samples from Cottnmouth Creek are nine samples from nearby Frost Bluff (Milam County, Texas) (see Gardner, 1933 and D’Agostino and Yancey, 1996 for a detailed overview of the stratigraphy of this section).  The lower seven meters of the Frost Bluff section contains abundant nannofossils.  This section correlates roughly to the top of the Cottonmouth Creek section with some overlap.  The sampled section at Frost Bluff lies entirely in the NP2 biozone.  Samples were disaggregated in buffered water and counting slides were produced for each sample using glass beads as a counting standard (after Okada, 1992).

 

NMS Analysis

Nonmetric Multidemsional Scaling (NMS) in an ordination method that describes a large, multivariate dataset with a smaller set of variables (or dimensions) using the ranked distance between samples.  The use of a ranked distance lessens the biases in the analysis by eliminating

7

many assumptions about the structure of the data (Clarke, 1993; McCune and Grace, 2002).  Many other techniques, such as principal components analysis or detrended correspondence analysis, assume a linear relationship among the datapoints or that distances between samples are related by a set formula.  Because NMS avoids these assumptions, it was chosen as the least biased method to extract the structure of the data.

Because most of the data from the early Danian are dominated by a few abundant taxa, the NMS results can be biased.  Also, the rarer taxa are useful in making determinations of ecosystem properties.  To eliminate the dominance of the abundant taxa in the analysis, the datasets were first standardized by converting number counts into percent abundance for each sample.  These percentages were then log transformed.  These steps lessen the bias of the abundant post-extinction taxa while still maintaining the underlying structure in the data

(McCune and Grace, 2002).  The NMS analysis was run using the Vegan package in R software (www.r-project.org).  Two separate NMS analyses were run on the nannoplankton and planktonic foraminifera datasets, and variance represented in the NMS results was calculated (Appendix C).  Results were then rotated around the origin after the fact to show stronger relationships between the samples and the NMS dimensions.  Nannoplankton and plankonic foraminifera taxa were plotted in NMS space in R using the function “wascores,” which is built in the metaMDS algorithm.  To understand how ecology factors into the recovery, the paleoecology of each taxa was then color coded.  A more detailed outline of the NMS methods is given in Appendix C.

 

 

 

Diversity

To visualize the changes in diversity across the recovery interval, several “Whittaker plots” were constructed by plotting the natural log of each species’ relative percent abundance by the species rank (Whittaker, 1965).  These plots show how each sample’s species richness differs as well as the evenness of the assemblage.  Seven samples were chosen from each site for both nannoplankton and planktonic foraminifera to show how diversity changed throughout the recovery interval. To eliminate possible bias of reworked fossils, all taxa considered “Cretaceous Taxa” (i.e. Percival and Fischer, 1977; Bown, 2005; Bernoala and Monechi, 2007) were eliminated from Paleocene samples.

To determine how extinction intensity influenced the recovery, the percent of species that went extinct over the K/P was calculated for all sites.  This was done by first selecting the youngest five samples from the Cretaceous and the oldest five samples of the Paleocene (or 0.5 m above and below the K/P boundary depending on sample spacing).  The percent of species that went extinct was calculated by dividing the number of species that did not cross the K/P boundary in these samples by the number of total Cretaceous taxa.  The error of this calculation was determined using the method outlined in Raup (1991).

 

Age Model

An age model is needed to correlate the different sites and interpret the NMS analysis.

This age model was constructed using established orbital chronologies of Shatsky Rise (ODP Site 1210) and Walvis Ridge (ODP Site 1262) (Westerhold et al., 2008).  These two sites were correlated to one another using this orbital chronostratigraphy and a composite section was developed.  The holes used for planktonic foraminifera samples at Shatsky Rise (DSDP Site 577) and Walvis Ridge (DSDP Site 528) were then correlated to this composite section based on the biostratigraphy of those sections.  Because of their close spatial relation, biostratigraphic datums between ODP Site 1210/DSDP Site 577 and ODP Site 1262/DSDP Site 577 should be practically concurrent, and therefore allow for a precise correlation between the DSDP sections to the composite standard.  Well defined orbital chronologies do not exist for Kerguelen Plateau, El Kef or Brazos River, and so correlating these sites to the composite standard is a problem.  The only available correlation method between these sites is biostratigraphy; however, correlating these sites with the composite standard using biostratigraphy relies on the assumption that biostratigraphic events are not time transgressive over great distances.  To lessen some of the inaccuracy of this correlation, the precision of the age model at Brazos River, El Kef and Kerguelen Plateau was kept low.  All interpretation comparing the timing of events at these sites must be done with the caveat that the correlations may not be entirely accurate.  To correlate these sites with the composite standard, key nannofossil and planktonic foraminiferal datums (Martini, 1971; Berggren et al., 1995) were tied to the composite standard, and their ages were determined (Table 2) and correlated between Brazos River, El Kef, and Kerguelen Plateau. By assuming that sedimentation rates were constant between these tie points (which also adds to the inaccuracy of the correlation), age models were established for these three sites.  An extended and more detailed method of the construction of this age model is in Appendix C

 

 

 

 

 

Event CS Age (Ma)

(after Westerhold et al., 2008)

K-P Boundary 65.680
FO Pv. eugubina (P0/Pα) 65.66
LO Pv. eugubina (Pα/P1a) 65.63
FO S. triloculinoides(P1a/P1b) 65.51
FO P. inconstans (P1b/P1c) 65.34
FO C. intermedius (NP1/NP2) 65.32

 

Table 2:  Composite section standard ages (after Westerhold et al., 2008) and datums (Martini, 1971; Berggren et al., 1995) used to correlate all sites.

 

Paleoecology

Another important part of understanding how the NMS results relate to the ecological recovery is to understand the paleoecology of the organisms used in this study.  Some taxa, such as Braarudosphaera bigelowii, are extant and their ecologies are at least partially known (Kelly et al., 1993).  For the extinct taxa, other methods are needed to estimate paleoecology.  One way of determining paleoecology is based on a taxon’s paleogeographic distribution.  This has been done to estimate the paleoecology of several Cretaceous-aged nannofossils.  Taxa that are constrained to areas thought to represent high rates of upwelling, where waters are often eutrophic, are considered as eutrophic taxa (e.g. Roth and Bowdler, 1981; Thierstein, 1981; Watkins, 1989).  Taxa that are commonly found in the open ocean gyres, which today  are characterized by low-nutrient levels (Hallock, 1987), are thought to be oligotrophic (i.e. Thierstein, 1981).  In some cases, extant species with similar morphologies were used to determine the paleoecology of extinct taxa, as was the case for Guembelitria cretacea (Keller et al., 2002).  Ecology can also be determined through overall skeleton size.  Because smaller cells are more efficient at taking up nutrients, they are most likely to be prevalent in eutrophic environments (Hallock, 1987; Reynolds, 2006).  All of these factors were combined and used to determine the overall paleoecology of the taxa used in the NMS analysis.  A more complete overview of this process, along with a table showing the paleoecologic affinity of each taxon is given in Appendix B.

A MULTIVARIATE ANALYSIS OF THE RECOVERY OF CALCAREOUS NANNOPLANKTON AND PLANKTONIC FORAMINIFERA FROM THE CRETACEOUS/PALEOGENE (K/P) MASS EXTINCTION

Leave a Reply