Calcareous Nannoplankton Response to Surface-Water Acidification Around Oceanic Anoxic Event 1a

See allHide authors and affiliations

Science  23 Jul 2010:
Vol. 329, Issue 5990, pp. 428-432
DOI: 10.1126/science.1188886


Ocean acidification induced by atmospheric CO2 may be a major threat to marine ecosystems, particularly to calcareous nannoplankton. We show that, during the Aptian (~120 million years ago) Oceanic Anoxic Event 1a, which resulted from a massive addition of volcanic CO2, the morphological features of calcareous nannofossils traced the biological response to acidified surface waters. We observe the demise of heavily calcified nannoconids and reduced calcite paleofluxes at the beginning of a pre-anoxia calcification crisis. Ephemeral coccolith dwarfism and malformation represent species-specific adjustments to survive lower pH, whereas later, abundance peaks indicate intermittent alkalinity recovery. Deepwater acidification occurred with a delay of 25,000 to 30,000 years. After the dissolution climax, nannoplankton and carbonate recovery developed over ~160,000 years under persisting global dysoxia-anoxia.

The dissolution of an atmospheric CO2 surplus [that is, over 500 parts per million (ppm)] in the ocean lowers pH and reduces the CaCO3 saturation state, consequently accelerating carbonate dissolution in the deep sea (1). The effect of modern surface-water acidification on organisms with CaCO3-based skeletons or tests, such as calcareous nannoplankton, remains elusive (26). Throughout Earth’s history, there is evidence of large CO2 releases, greenhouse conditions, ocean acidification, and major changes in biota, particularly in marine calcifiers (7). In many cases, the geological record indicates that ocean biota can adapt to increased acidity; however, past examples of ocean acidification occurred over tens of thousands of years, giving time for life to adjust to CO2 concentrations as high as 2000 to 3000 ppm (7).

The early Aptian [121 to 118 million years ago (Ma)] represents a case history of excess CO2 derived from a major volcanic episode, namely the emplacement of the Ontong Java Plateau (OJP) (8, 9), which is marked by changes in the evolutionary rates, species richness, abundance, and calcite production of calcareous nannoplankton (1012). These changes occurred during Oceanic Anoxic Event 1a (OAE1a) (~120 Ma), which was a time of severe global warming (13, 14). Although global anoxia and enhanced organic matter burial are the most striking and intriguing paleoceanograhic phenomena during this event, OAE1a sediments reveal a sequence of CO2 pulses (15) and weathering changes (16). For example, the cutoff of carbonates during OAE1a is the result of volcanogenic CO2-related ocean acidification (7, 10, 17).

We analyzed calcareous nannofossil assemblages from two drill sites in the Tethys (Cismon core) and Pacific [Deep Sea Drilling Project (DSDP) site 463] Oceans (fig. S1) (18). At both sites, nannofossil changes integrated with geochemical and cyclochronological data (15, 19) identify and date the effects of acidification on calcareous nannoplankton. Shortly before magnetic chron M0 (Fig. 1), at 121.3 Ma (19), nannoconid abundance declined and nannofossil paleofluxes (tracing nannoplankton carbonate production and accumulation) decreased as response to a major injection of volcanogenic CO2. Later, a sharp nannoconid crisis at 120.25 Ma was part of a global calcification failure of planktonic and benthic calcifiers in pelagic and neritic settings under excess CO2 in the ocean-atmosphere system (17). During the 1-million-year-long interval between these two events, the geological record reveals subtle effects of ocean acidification traced only by nannofossils, and specifically by the heavily calcified nannoconids, with trivial effects on other coccoliths and apparently no evidence in the lithologic and geochemical records. Although the negative carbon isotopic event (CIE) at the beginning of global anoxia (~120 Ma) coincides with the drop in carbonate content, there was an increase in relative abundance of Biscutum constans, Zeugrhabdotus erectus, and Discorhabdus rotatorius, represented by dwarfed specimens (Fig. 1). Size variation was species-specific at both sites, because B. constans displays the most pronounced morphometric decrease (a volume/mass reduction of 50 to 60% for single coccoliths), whereas Z. erectus diminishes in size to a lesser extent (a volume/mass reduction of 30 to 40% for single coccoliths). D. rotatorius also exhibits smaller-than-normal sizes throughout the studied interval, reaching minimum dimensions in the CIE (a volume/mass reduction of 5 to 10% for single coccoliths in the study interval, but up to 70% with respect to the holotype). Watznaueria barnesiae shows minor changes in abundance and average size, but several malformed/deformed specimens occur in the CIE interval (fig. S2).

Fig. 1

Average coccolith size (±1σ) of W. barnesiae, B. constans, Z. erectus, and D. rotatorius in the Cismon core (A) and DSDP site 463 (B). Nannofossil abundances are plotted against litho-, bio-, magneto-, and chemostratigraphy available for both the Cismon core (15, 19, 29) and DSDP site 463 (20, 30). Photographs of W. barnesiae document coccolith variations in the pre-CIE, CIE, and post-CIE intervals.

In the Cismon core, approximately 25 to 30 thousand years (ky) after the nannoconid crisis and just before the first OAE1a black shale, nannofossil assemblages in pseudonodular limestone show a relative increase in abundance and diversity, a minimum of nannoconids, and increased relative abundance of B. constans, Z. erectus, and D. rotatorius (Fig. 2, interval 2). The δ13C curve is stable through this interval, and therefore an individual volcanogenic CO2 pulse is unlikely. We suggest that this pseudonodular limestone is evidence of intense dissolution and reflects a shallowing of the calcite lysocline up to ~1200 m of paleowater depth in the Tethys Ocean, indicating a delayed effect of acidification on deep waters and position of the calcite compensation depth (CCD).

Fig. 2

Close-up of the onset of the OAE1a in the Cismon core. Lithology is an original photograph of the split core (100% core recovery). Nannofossil total abundance and diversity, nannofossil paleofluxes, single taxon relative abundance, and morphometric data are plotted against CaCO3, C, and O stable isotopes.

The beginning of global anoxia was likely caused by a paroxysmal volcanic event, evidenced by a decrease of the δ13C curve and carbonate content. Nannofossils did not experience extinction, and insignificant changes in nannofloral abundance and diversity occur simultaneously with the beginning of dwarfism in B. constans, Z. erectus, and D. rotatorius (Fig. 2, interval 3). The smallest coccoliths are not systematically in samples with their highest relative abundance before and after CIE (Fig. 1); therefore, we presume a specific cause for the secretion of dwarf coccoliths. A negative δ18O shift that marks interval 3 (lasting <10 ky) suggests a major warming event and is consistent with a coeval episode reconstructed at DSDP site 463 (20). Nannofossil paleofluxes show a progressive decrease at this time, reaching values close to zero in the CIE core (Figs. 2 and 3). Nannofossil paleofluxes do not fully mirror total abundance; heavily calcified taxa, in particular nannoconids, dominate W. barnesiae in terms of mass. Likewise, B. constans, Z. erectus, and D. rotatorius increase in abundance, but their coccoliths are too small to affect or compensate for losses in nannofossil calcite paleofluxes.

Fig. 3

Details of the CIE interval in the Cismon core. Nannofossil abundance and paleofluxes, relative and absolute abundances of the Assipetra/Rucinolithus group, CaCO3 [weight percent (wt %)], and δ13C are plotted versus age (Ma) following the age model of (19). The initial position of the calcite lysocline and CCD are based on the reconstruction by (21) for the Cretaceous Tethys Ocean. The history of the calcite lysocline and CCD reflects the carbonate chemistry evolution in the Tethys Ocean relative to the Cismon paleo-water depth (29).

Shortly after the beginning of OAE1a, in an interval of non-anoxic sediments, a temporary increase in nannoconid abundance and nannofossil paleofluxes, corroborated by a CaCO3 increase (Fig. 2, interval 4), indicates a ~15-ky-long stasis in the CO2 input. Indeed, this interval corresponds to less-negative δ18O values, suggesting a relative cooling interrupting the initial warming episode. Apparent discrepancy relative to a previous reconstruction (15), which implies a warming pulse derived from biomarkers [interval II of (15)], is possibly caused by different sampling resolution. Although there is one temperature anomaly at a slightly higher level within the CIE at DSDP site 463 (20), the δ18O data show a first warm peak followed by less-negative values at the same stratigraphic levels depicted in the Cismon core.

Another CO2 pulse [interval III of (15)] corresponds to further reduction of nannoconids and nannofossil paleofluxes and a decrease in carbonate content, δ13C, and δ18O values (interval 5, ~5 ky). Methane hydrate destabilization may be the cause for the most negative δ13C values measured in carbonates [interval IV of (15)]; here, nannofossil assemblages remain rare and characterized by dwarf coccoliths, whereas nannoconids show a final exclusion (interval 6, ~5 ky). In this interval, although rare, the heavily calcified Assipetra and Rucinolithus nannoliths maintain nannofossil paleofluxes above zero (Fig. 3).

In the following ~45-ky-long interval 7 (Figs. 2 and 3), nannofossil assemblages are stable, whereas δ18O values gradually increase. The next change in nannofossils suggests strong acidification and dissolution, because dwarf coccoliths exist in the middle part of CIE where nannofossil abundance, diversity, and paleofluxes are least, and CaCO3 is close to zero. The dissolution-acidification climax (intervals 8 to 10) protracted for ~70 ky, although minor fluctuations occurred. Many samples are barren of nannofossils or exclusively contain rare dissolution-resistant W. barnesiae, Assipetra, and Rucinolithus (mainly in interval 9).

At the beginning of interval 8, the rapid drop in nannofossil abundance, diversity, and paleofluxes, which are substantiated by the short-lived decrease of δ18O values, suggests another CO2 pulse rather than a progressive deterioration of environmental conditions. A minimal increase in CaCO3 and nannofossil abundance and fluxes, associated with dwarf coccoliths of B. constans, Z. erectus, and D. rotatorius, might represent an acidification pause, perhaps under cooler conditions as suggested by less-negative δ18O values (interval 9). Indeed, the coeval increase in relative and absolute abundance of Assipetra and Rucinolithus (Fig. 3) may suggest a temporary ~35-ky-long slight alkalinity recovery. A further episode of strong dissolution (interval 10, ~20 ky), almost identical to interval 8, is presumably initiated by another CO2 pulse, as inferred from δ18O and small δ13C decreases.

The initial recovery (Fig. 3, interval 11) corresponds to a minor increase in nannofossil abundance, paleofluxes, and CaCO3 during a ~75-ky-long time interval, followed by a more pronounced return to pre-CIE conditions, becoming again suitable for nannoconids (Fig. 3, interval 12). Dwarfism is absent after the full recovery of nannofossil abundance and paleofluxes, mirroring a drop in B. constans, Z. erectus, and D. rotatorius relative abundances (Fig. 1).

Changes in nannofossil assemblages allow the discrimination of surface-water acidification from bottom-water acidity, which postdates the early effects recorded by calcareous nannoplankton. Evidence of bottom-water acidification at 120.22 Ma suggests severe dissolution at the sediment/water interface, presumably induced by the shallowing of the calcite lysocline up to ~1200-m water depth. In the first phase of the CIE, nannofossil and CaCO3 changes enabled a reconstruction of further progressive shoaling of the calcite lysocline and CCD (Fig. 3). The dissolution climax, matching maximum ocean acidification, was ~75 ky after the beginning of the δ13C negative anomaly and persisted for ~70 ky. We calculate that this dissolution maximum resulted from shoaling of the Early Cretaceous CCD (21) up to ~1200 m at a rate of 17 m/ky. Recovery occurred at similar rate as the CCD and calcite lysocline deepened to >1200 m. Although the inferred shoaling and deepening of the CCD and calcite lysocline appear to be symmetric, the early acidification pulse recorded by the nannoconid crisis suggests more rapid shoaling at a rate possibly up to 70 m/ky.

The reconstructed CO2 pulses (Fig. 3) suggest a stepwise accumulation of CO2 in the ocean inducing progressive acidification. The possibility exists for a major acceleration of weathering in the lowermost part of OAE1a (16). Indeed, a weathering-induced drawdown of CO2 may contribute to abundance peak of Assipetra and Rucinolithus nannoliths (interval 7), possibly representing alkalinity recovery and cooling.

The occurrence of dwarf coccoliths may be the response to a pressure of CO2 (pCO2) above threshold levels for most species, but which actually accelerates photosynthesis and growth in some modern r-strategists (organisms that rapidly adapt to and reproduce in unstable conditions) (22). In this scenario, nannoconids were less CO2 tolerant, whereas B. constans, Z. erectus, and D. rotatorius were more tolerant, and W. barnesiae was even more adaptable. B. constans, Z. erectus, and D. rotatorius are mesotrophic Cretaceous indices, and their synchronous abundance increase in the Tethys and Pacific Oceans (Fig. 1) suggests a global fertilization event. However, dwarfism is not systematically associated with increased abundance of these mesotrophic taxa, and during OAE1a, the highest primary productivity was recorded above the CIE. Shallow-water calcifiers also experienced a calcification crisis, which was most pronounced in the Atlantic and Tethys Oceans; it coincided with the “nannoconid crisis” and OAE1a (17, 23), indicating that surface waters were loaded with excess CO2. The slow and smooth CaCO3 recovery after the CIE implies a stasis of OJP emplacement and gradual buffering of ocean acidification or a decrease in volcanogenic CO2 emissions and consistently higher CO2 drawdown through Corg burial and/or weathering.

A similar, albeit more pronounced, δ13C anomaly occurs ~55.8 Ma at the Paleocene Eocene Thermal Maximum (PETM) and also suggests ocean acidification (24). However, the initial decrease in δ13C occurred over a few thousand years for the PETM and over ~30 ky for OAE1a. This discrepancy could reflect different causal mechanisms, namely methane hydrate dissociation for PETM (25) and major volcanogenic CO2 injections for OAE1a, or condensation and hiatuses for most PETM sections. Indeed, severe dissolution in intervals preceding PETM occurs at several deep and intermediate sites (7, 24, 26). The difference in CIE duration (~100 to 150 ky and ~200 ky for the PETM and OAE1a, respectively) might reflect subsequent pulses accumulating excess CO2 in the ocean-atmosphere system. The recovery phase is also longer for OAE1a (~160 ky versus ~30 to 80 ky), possibly because the more protracted CIE perturbation required a longer time interval for deep-ocean buffering.

The response of calcareous nannoplankton to ocean acidification in the PETM is unclear, because transient, perhaps malformed, taxa may be of solely local importance (27) or indicative of unsaturated surface waters at global scale (26). Reconstructed rates of originations and extinctions at the onset of PETM indicate modest effects of ocean acidification on evolutionary trends (28). It has also been suggested that a major nannoplankton turnover occurred during OAE1a (11). However, our data demonstrate that rising pCO2 and surface-ocean acidification during OAE1a triggered false extinctions (a so-called Lazarus effect) among calcareous nannoplankton. Conversely, a major origination episode starts approximately 1 My before global anoxia and persists through OAE1a and associated acidification (10). Increasing pCO2 triggered coccolith malformation and solicited production of r-strategist taxa, which secreted dwarf coccoliths as a strategy to overcome acidification.

Supporting Online Material

Materials and Methods

Figs. S1 and S2

Tables S1 to S3


References and Notes

  1. Materials and methods are available as supporting material on Science Online.
  2. E.E. and C.B. were funded through MIUR (Italian Ministry of University and Research), Research Programs of National Interest (PRIN) grant 2007-2007W9B2WE 001. H.J.W. and C.E.K. were funded by the Swiss Science Foundation (project 200021-113687) and by ETH Zurich. E.E. and C.B. investigated calcareous nannofossil assemblages, H.J.W. and C.E.K. analyzed C and O isotopes. All authors contributed to the synthesis and paleoceanographic model. Samples from DSDP site 463 were supplied by the Integrated Ocean Drilling Program.
View Abstract

Navigate This Article