Organic Carbon Fluxes and Ecological Recovery from the Cretaceous-Tertiary Mass Extinction

See allHide authors and affiliations

Science  09 Oct 1998:
Vol. 282, Issue 5387, pp. 276-279
DOI: 10.1126/science.282.5387.276

This article has a correction. Please see:


Differences between the carbon isotopic values of carbonates secreted by planktic and benthic organisms did not recover to stable preextinction levels for more than 3 million years after the Cretaceous-Tertiary mass extinction. These decreased differences may have resulted from a smaller proportion of marine biological production sinking to deep water in the postextinction ocean. Under this hypothesis, marine production may have recovered shortly after the mass extinction, but the structure of the open-ocean ecosystem did not fully recover for more than 3 million years.

A wide range of geochemical evidence indicates that the organic flux from the surface ocean to the deep sea decreased drastically at the time of the Cretaceous-Tertiary (K-T) mass extinction and did not recover for more than a million years. The evidence for this collapse of the deep-sea organic flux includes dramatic decreases in (i) carbon isotopic (δ13C) differences between carbonate skeletons secreted by planktic and benthic organisms (usually foraminifera) (1–6), (ii) δ13C differences between benthic foraminiferal skeletons (tests) from different ocean basins (2), (iii) δ13C differences between the tests of benthic foraminifera that lived on the seafloor and those that lived in the underlying sediment (3), and (iv) the accumulation of barium in deep-sea sediments (3).

Our study of planktic-to-benthic δ13C differences indicates that final recovery of the organic flux to the deep sea may have occurred more than 3 million years after the mass extinction (Figs. 1 and2). An early phase of recovery is marked by the return of planktic-to-benthic δ13C differences (and interbenthic δ13C differences) to low but relatively stable levels within the first several hundred thousand years after the extinction (1–4). This early stage of recovery has been identified at Pacific, Southern Ocean, and South Atlantic sites (1–4). After this early phase of recovery, differences between planktic and benthic δ13C values at South Atlantic Deep Sea Drilling Project (DSDP) site 528 remained below preextinction levels for more than 2 million additional years (Fig. 2). At site 528, the final recovery of planktic-to-benthic δ13C differences is exhibited by fine (<25 μm in diameter) CaCO3, near-surface planktic foraminifera and planktic foraminifera that lived deeper in the thermocline (7) (Fig. 2). The parallel nature of these planktic-to-benthic records provides strong evidence that surface-to-deep δ13C gradients did not fully recover for more than 3 million years after the mass extinction. Data from Central Pacific DSDP site 577 indicate that this long delay in final recovery was global (Fig. 3).

Figure 1

Cretaceous-Tertiary δ13C records from South Atlantic DSDP site 528. Magnetostrat., magnetostratigraphy; Ma, million years ago; mbsf, meters below sea floor. (A) δ13C records of benthic foraminifera (Gavelinella andNuttallides) (solid circles) and fine (<25 μm) carbonate (open circles) (23). (B) δ13C records of benthic foraminifera (solid circles), fine (<25 μm) carbonate (open circles), and planktic foraminifera. Open symbols represent near-surface planktic foraminifera: Morozovella angulata (right-pointing open triangles), Praemurica taurica (left-pointing open triangles), and Rugoglobigerina rotundata (open squares). Gray symbols represent deeper-dwelling planktic foraminifera: Eoglobigerina eobulloides (gray circles), Parasubbotina pseudobulloides (right-pointing gray triangles), and Pseudotextularia ultimatumida (gray squares). Hatched areas represent intervals with insufficient data to assign magnetic polarity. White “core recovery” intervals mark missing sections.

Figure 2

Cretaceous-Tertiary δ13C records from site 528 (23, 24). (A) δ13C differences between fine carbonate and benthic foraminifera. (B) δ13C differences between planktic and benthic foraminifera and between fine carbonate and benthic foraminifera. These differences are identified by symbols corresponding to those representing the planktic taxa of Fig. 1.

Figure 3

Cretaceous-Tertiary δ13C records from DSDP site 577 (25). Open circles represent δ13C differences between bulk carbonate and benthic foraminifera (Nuttalides). Other symbols represent differences between planktic and benthic foraminifera. The planktic foraminifera are as follows: Morozovella species (right-pointing open triangles), Praemurica uncinata(downward-pointing open triangles), Praemurica taurica(left-pointing open triangles), Rugoglobigerina rotundata(open squares), Parasubbotina pseudobulloides(right-pointing gray triangles), Subbotina triloculinoides(gray diamonds), and Pseudotextularia ultimatumida (gray squares). For further information, see Fig. 1 caption.

A low-productivity “Strangelove” ocean has often been invoked to explain a low organic flux to the deep sea during the earliest Tertiary (1–3, 8). A low-productivity ocean would naturally have occurred during any interval of darkness that resulted from the K-T impact of a large asteroid or comet. However, global darkness would have lasted only a few years (9). Because phytoplankton typically double on time scales of hours to days (10), it would have been difficult to maintain a low-productivity ocean once surface irradiance recovered.

We propose that once sunlight returned, biological productivity also returned, but planktic-to-benthic δ13C differences remained low because a smaller fraction of marine production sank to deep waters in the postextinction ocean. Such a reduction in the organic flux to deep waters could have been a natural consequence of the ecosystem reorganization that resulted from the mass extinction, because a general absence of large pelagic grazers or a decrease in the mean size of phytoplankton would have greatly reduced the packaging of biomass into the large particles that sink to the deep ocean. Application of this hypothesis to our data implies that final recovery of the open-ocean ecosystem structure occurred more than 3 million years after the mass extinction (11). Such recovery may have required the evolution of new species at multiple trophic levels to replace those lost during the mass extinction.

Total organic production in the modern open ocean is between 30 × 1015 and 50 × 1015 g of carbon per year (12, 13). Most of this carbon is oxidized in relatively shallow waters; only about 10% sinks to a water depth of 200 m, and only 1% actually reaches deep-sea sediments (12). Most of the organic carbon in surface sediments is oxidized by biological activity. Only about 0.3% of open-ocean production is eventually buried in deep-sea sediments (12).

For the Strangelove model to explain postextinction maintenance of the mean surface-to-deep δ13C gradient at 50% of its preextinction level, earliest Tertiary marine biomass production must have been 50% lower than latest Cretaceous production—and the ratio of total production to downward organic flux must have been constant. In contrast, our model allows postextinction production to approximate preextinction production and explains the low postextinction δ13C gradient by just slightly increasing (from 90 to 95%) the fraction of total production that was degraded in the upper 200 m of the ocean.

The proportion of organic production that sinks from the surface ocean is primarily controlled by the ratio of phytoplanktic respiration to photosynthesis, the size of the phytoplankton that define the base of the trophic structure (small plankton sink too slowly and are degraded too quickly to settle to deep waters), the ability of phytoplankton to aggregate into larger particles, and the size and trophic efficiency of the animals that repackage biomass into larger aggregates that can sink more rapidly (14).

An ocean characterized by high rates of picoplankton production (phytoplankton with a diameter less than 2 μm) or low abundances of relatively large grazers (such as macrozooplankton and fish) would be characterized by low rates of biomass sinking to deep waters and high rates of biomass recycling in the upper water column (15). In such an ocean, an increased fraction of total production would be shunted through the microbial food web. By allowing essential nutrients to remain in easily remineralized forms (such as tiny microbially grazed plankton) in the euphotic zone, such changes may also increase rates of nutrient recycing and open-ocean biomass production.

The K-T mass extinction radically altered the open-ocean ecosystem. Most species of planktic foraminifera and calcareous nannoplankton went extinct at that time (16, 17). The mean accumulation of calcareous nannofossils in deep-sea sediments decreased by up to 85% (17). This decrease in nannofossil accumulation resulted from an equivalent drop in the production of calcite-secreting phytoplankton (17). The collapse of higher trophic levels is suggested by the fossil record of larger organisms. Ammonites (shelled nektonic cephalopods) suffered complete extinction (18), as did mosasaurs and sauropterygians (plesiosaurs and pliosaurs) (19). Marine osteichthyans (bony fish) and selachians (sharks and rays) underwent tremendous extinction (19).

Despite radical K-T changes in the open-ocean ecosystem, at least two lines of evidence suggest that biological production was relatively high during the 3-million-year interval of reduced δ13C differences. First, despite the mass extinction of planktic foraminifera and a drastic decrease in the test size of planktic foraminifera at the K-T boundary, the mean flux of foraminiferal tests to deep-sea sediments remained relatively stable across the K-T boundary and throughout the earliest Tertiary (17). Tiny (<100 μm in diameter) foraminiferal tests are abundant in sediments immediately above the extinction horizon. High concentrations of radiolarian tests and diatom frustules are seen in postextinction marine sediments of New Zealand (20). These findings suggest that within a few thousand years of the mass extinction, marine biological production returned to high enough levels to support abundant small zooplankton.

Second, planktic and benthic δ13C values stabilized within a few hundred thousand years after the K-T extinction (1–4) (Fig. 1). Their failure to continue shifting to more negative values suggests that the burial of organic matter recovered to as much as 90% of its preextinction rate (21). Final recovery of planktic δ13C values approximately coincided with the final recovery of planktic-to-benthic δ13C differences (Figs. 1 and 2). This coincidence suggests that the final recovery of organic burial may have been closely coupled to the final recovery of the organic flux to deep waters.

The sequential recovery of carbon burial and planktic-to-benthic δ13C differences is consistent with the idea that organic flux to shallow sediments recovered long before the recovery of the organic flux to the deep ocean. This staged recovery of the vertical organic flux is in turn consistent with progressive recovery of open-ocean trophic structures. For example, if abundant grazers became progressively larger over time, the sinking of organic carbon to shallow sediments and water depths of a couple hundred meters would have recovered before the sinking of organic carbon to greater water depths.

Because the organic flux to the deep sea is a major sink for atmospheric CO2 and biologically limiting nutrients from the surface ocean, the long delay in recovery of planktic-to-benthic δ13C differences suggests that the global biogeochemical cycles of carbon and other biologically active elements were also altered for up to 3 million years by the pattern of K-T mass extinction and biological recovery. The long-lived δ13C anomaly associated with the Permo-Triassic mass extinction (22) may have similarly resulted from the cycling of carbon by the altered ecosystem of a postextinction ocean.


View Abstract

Stay Connected to Science

Navigate This Article