Nannoplankton Extinction and Origination Across the Paleocene-Eocene Thermal Maximum

See allHide authors and affiliations

Science  15 Dec 2006:
Vol. 314, Issue 5806, pp. 1770-1773
DOI: 10.1126/science.1133902


The Paleocene-Eocene Thermal Maximum (PETM, ∼55 million years ago) was an interval of global warming and ocean acidification attributed to rapid release and oxidation of buried carbon. We show that the onset of the PETM coincided with a prominent increase in the origination and extinction of calcareous phytoplankton. Yet major perturbation of the surface-water saturation state across the PETM was not detrimental to the survival of most calcareous nannoplankton taxa and did not impart a calcification or ecological bias to the pattern of evolutionary turnover. Instead, the rate of environmental change appears to have driven turnover, preferentially affecting rare taxa living close to their viable limits.

Modern biodiversity studies predict huge losses of flora and fauna (1) associated with anthropogenic carbon emissions and projected climate change (up to ∼35% of species by 2050) (2). Paleontological records from analog events in geologic history provide a way to test the ecological basis of these predictions. The Paleocene-Eocene Thermal Maximum (PETM) is attributed to a rapid increase in atmospheric CO2 levels, perhaps caused by the exhumation and oxidation of methane from marine sediments (35). Within less than 10 thousand years (ky) (6, 7), ocean temperatures rose by 5° to 8°C, marine and terrestrial carbon isotope values (δ13C) decreased by 3 to 8 per mil, and the calcite compensation depth (CCD) (8) in the deep sea shoaled by up to 2 km (912). Subsequently, δ13C values returned to near-background levels within 110 to 210 ky (6, 13). The PETM was accompanied by dramatic reorganization in marine and terrestrial ecosystems (9, 1419), with the most extreme response being the catastrophic extinction of 35 to 50% of benthic foraminiferal species (17).

To understand biotic change across the PETM more fully, we estimate the rate of evolutionary change within the calcareous nannoplankton; i.e., the number of species that appeared and disappeared through time. Calcareous nannoplankton are ideal for testing the organismal response to the PETM because their surface-water habitat renders them highly sensitive to environmental change. Moreover, their fossil record is exceptionally complete during this time interval, both taxonomically and stratigraphically (18, 19). Our records are at a resolution of up to 10 ky, allowing us to resolve patterns on the time scale of environmental disturbance.

Our data come from open-ocean sites in the paleoequatorial Pacific and the Southern Ocean [Ocean Drilling Program (ODP) sites 1209 and 690] and two sections on the New Jersey paleoshelf (U.S. Geological Survey drill hole at Wilson Lake and ODP leg 174AX drill hole at Bass River) (fig. S1). By combining nannofloral data sets with cyclostratigraphic age models (6), we calculated species extinction and origination rates per unit depth and per unit time (20). Two methods were used to calculate evolutionary rates per unit time. First, we applied the widely used proportional rates method, dividing the number of originations or extinctions by time, normalized for diversity [(20), following (21)]. Second, per-capita rates were calculated using the natural log of the ratio of taxa that range through a time bin to either those that only cross the bottom boundary of the bin (for extinction) or those that cross the top boundary (for origination) (20, 22, 23). This latter method has the advantage of removing spurious variation in rates created by unequal duration of time bins (22). Both methods return rates per unit time (24) and are thus not biased by variation in sediment accumulation (25).

At all sites, the pattern of change recorded across the PETM in evolutionary rates and species richness (Figs. 1 and 2 and table S1) is similar, pointing to global rather than local species turnover (26). In the ∼70 ky before the PETM, origination was low and there were virtually no extinctions, resulting in a gradual increase in species richness. Origination and extinction rates increased during the first 70 ky of the PETM, defined by the interval from the onset to the peak of the carbon isotope excursion (CIE, dark shaded zone in Figs. 1 and 2). During this interval, the inferred rate of CO2 absorption by the oceans was greatest, and up to 18 species (from a maximum of 84) appeared or disappeared. The synchronous increase in both per-unit-depth and per-unit-time rates demonstrates that the abrupt change in turnover cannot be attributed to changes in sedimentation rate (e.g., section condensation resulting from dissolution at the base of the PETM). At three locations [site 1209, Wilson Lake (WL), and Bass River (BR)], average proportional (per unit time) rates of origination and extinction were 1.6 and 1.7% per 10 ky, respectively (table S1), compared to an average of 0.5 and 0.1% per 10 ky in the pre-event background interval (Fig. 1). At site 690, Maud Rise, Southern Ocean, a more finely resolved record of the CIE is available (24), and the evolutionary rate data reveal that proportional origination and extinction rates were 11 and 5% per 10 ky, respectively, during the first 10 ky after the event onset (Fig. 2). During the remainder of the PETM, despite continued environmental perturbation (9, 10), origination and extinction rates rapidly returned to pre-event levels, with a maximum of 10 species appearing or disappearing over the next ∼150 ky. Proportional rates of origination and extinction in the recovery interval remained low (averaging 0.2 and 0.5% per 10 ky, respectively, table S1) with a gradual drop in species richness (Fig. 1).

Fig. 1.

Per-unit-depth (% per sampling interval) and per-unit-time (proportional and per-capita) evolutionary rates across the CIE for (A) BR, (B) Wilson Lake, and (C) ODP site 1209. Extinction rates (Re), blue; origination rates (Ro), red; proportional, solid line (% per 10 ky); per capita, dashed line (lineage per 10 ky, using the same scale as the proportional rates but × 10–2). The sampling intervals are described in (20). Diversity is expressed as total species present (species richness) per sample with the 70-species level marked on each plot (yellow). The shaded intervals correspond to time elapsed from the onset of the CIE, with dark gray being the onset-to-peak interval and light gray being the recovery interval. Mbsf, meters below sea floor; mbs, meters below surface. At both BR and WL, hiatuses in the uppermost part of the recovery interval artificially elevate rates per unit depth above the disconformities. The sources of the carbon isotope records in Figs. 1, 2, and 4 are detailed in (20).

Fig. 2.

Per-unit-depth (% per sampling interval) and per-unit-time (proportional, solid; and per-capita, dashed) evolutionary rates across the onset-to-peak interval of the CIE at ODP site 690. Per-capita rates use the same scale as the proportional rates but × 10–2. Abbreviations and symbols are the same as in Fig. 1.

In Fig. 3, we compare the rates that we have measured for the PETM with long-term records of nannofossil diversity binned into 3-million-year (My) intervals (21). In the long-term record, the bin containing the PETM [56 to 53 million years ago (Ma)] shows the highest rates of turnover for the Cenozoic, with both origination and extinction rates at 54%. Comparison with the results of our study suggests that much of this turnover was probably focused in the onset of the short-lived PETM event. This finding implies that the importance of geological events marked by rapid evolutionary turnover, such as the PETM, is likely to be systematically underestimated within low-resolution records of biodiversity change.

Fig. 3.

High-resolution PETM evolutionary rate estimates compared to low-resolution (3 My) Paleogene nannofossil biodiversity. (A) Paleogene diversity data replotted from (22) at the midpoint of each 3-My time bin. Total nannofossil diversity (pink line) and proportional Re, blue, and Ro, red, rates per 3-My interval are shown. Evolutionary rates for the interval from 62 to 68 Ma reflect species losses and biotic recovery associated with the Cretaceous-Paleogene boundary (K/Pb). (B) Proportional evolutionary rates from 62 to 26 Ma recalculated as per-capita rates (lineage per ky). (C) PETM high-resolution (HR) per-capita rates (lineage per ky, circled symbols) from the event onset at site 690 (table S3) and HR per-capita rates for pre- and post-event intervals labeled as background (table S1). Note the break in the per-capita rate scale. This scale applies to both (B) and (C). It is necessary to use per-capita rates rather than proportional rates for comparison of time-interval bins of different durations (20, 22).

Despite the strong ecological responses of nannofossil assemblages to environmental changes at the PETM (18, 19), the pattern of evolutionary reorganization that we see is surprising in two ways. First, the rate of turnover is relatively modest, considering the magnitude of environmental change that has been inferred for the PETM and in light of findings from laboratory culture experiments and ocean acidification models (27, 28). Second, the turnover lacks an obvious ecological or calcification bias. Both oligotrophic warm-water–favoring taxa (e.g., several species of Discoaster and Fasciculithus) and inferred mesotrophic cool-water–favoring taxa (e.g., several species of Neochiastozygus and Prinsius bisulcus) appear and disappear (table S4). Furthermore, there is no obvious evidence of an evolutionary decrease in lith calcification or an overcompensation in robustness, despite increasing geologic evidence for massive carbonate undersaturation in the oceans (11, 12). Heavily calcified as well as fragile nannofossils both appear and disappear within the interval from the onset to the peak of PETM conditions (Fig. 4, table S4), providing no obvious evidence for geologically sustained inhibition of surface-water calcification. Apparently, surface-water saturation state was not perturbed across the event to a point that was detrimental to the survivorship of most calcareous nannoplankton taxa. On the other hand, taxa that became extinct were rare (<1% abundance in pre-event assemblages) compared to the dominant taxa that survived. Presumably, the rare taxa lived closer to their ecological limits than the more common taxa and were therefore more susceptible to a contraction of population numbers below viable limits (29), with originations rapidly filling the empty ecological niches. The lack of ecological bias in our data and the clustering of bioevents at the initiation of the PETM suggest that it was the rate of environmental change that drove evolutionary turnover during this event, rather than change in any given individual environmental factor.

Fig. 4.

Examples from BR of species that originated and became extinct during the PETM interval. Abundances are % abundances of total nannofossils. The species illustrated are examples of robust and delicate taxa (table S4). In this context, robust refers to nannofossils that are either relatively large or are constructed of large, blocky calcite units. Fragile refers to nannofossils composed of small and/or thin calcite units. The arrows indicate appearance (red) and disappearance (blue) levels. The light-microscope images are all at the same scale, with the bar representing 2 μm. Neochiasto., Neochiastozygus; Rhom., Rhomboaster.

In comparison to many plants and animals, calcareous nannoplankton should be relatively resilient to environmental perturbation because of their planktonic habit, near-cosmopolitan distribution, large population sizes, and short life cycle. Yet our data demonstrate that even these organisms display a prominent acceleration in both origination and extinction rates across the PETM (Figs. 1, 2, 3). However, the measured present-day rates of CO2 increase, and those predicted for the coming century are even greater than for the PETM (30), with projections of increased surface-water acidity of 0.7 pH units over the next 300 years (31), possibly even resulting in aragonite undersaturation of the high latitudes by 2050 (28). Although there was no catastrophic loss of calcifying plankton at the PETM, the evolutionary rates that we have documented are likely to be modest compared to those that will accompany projected surface-water acidification over coming centuries.

Supporting Online Material

Materials and Methods

Figs. S1 and S2

Tables S1 to S4


References and Notes

Stay Connected to Science

Navigate This Article