Revisiting the sedimentary record of the rise of diatoms

Diatoms are a major primary producer in the modern oceans and play a critical role in the marine silica cycle. Their rise to dominance is recognized as one of the largest shifts in Cenozoic marine ecosystems, but the timing of this transition is debated. Here, we use a diagenetic model to examine the effect of sedimentation rate and temperature on the burial efficiency of biogenic silica over the past 66 million years (i.e., the Cenozoic). We find that the changing preservation potential of siliceous microfossils during that time would have overprinted the primary signal of diatom and radiolarian abundance. We generate a taphonomic null hypothesis of the diatom fossil record by assuming a constant flux of diatoms to the sea floor and having diagenetic conditions driven by observed shifts in temperature and sedimentation rate. This null hypothesis produces a late Cenozoic (∼5 Ma to 20 Ma) increase in the relative abundance of fossilized diatoms that is comparable to current empirical records. This suggests that the observed increase in diatom abundance in the sedimentary record may be driven by changing preservation potential. A late Cenozoic rise in diatoms has been causally tied to the rise of grasslands and baleen whales and to declining atmospheric CO₂ levels. Here we suggest that the similarity among these records primarily arises from a common driver—the cooling climate system—that drove enhanced diatom preservation as well as the rise of grasslands and whales, rather than a causal link among them.

Diatoms—phytoplankton that construct their shells out of silica—are critical to marine food webs and geochemical cycles. They account for ∼40% of marine primary productivity today (1), but are a relatively recent contributor to ocean ecosystems (2). Diatoms first appear in the fossil record in the Jurassic (3) and become ecologically dominant among phytoplankton during the Cenozoic (4, 5). Hypothesized explanations for their middle to late Cenozoic rise include the decline in atmospheric CO₂ concentrations over the last 40 My, sea level change, an increase in bioavailable silica reaching the ocean due to elevated continental weathering and/or the expansion of grasslands, and changes in nutrient focusing due to cooler temperatures, among others (5–9).The timing and cause of diatoms'' ascension is important beyond simply reconstructing the history of marine primary producers—it represents a major shift in Earth''s silica and carbon cycles. Diatoms are believed to have drawn down ocean silica concentrations to their lowest levels in Earth''s history (10), which, studies suggest, could have fundamentally changed climate regulation by altering marine authigenic clay formation (11, 12). A shift from calcifying to silicifying plankton also partially decouples inorganic and organic carbon and leads to a tighter coupling of organic carbon (along with nitrogen and phosphorous) to silica (3, 13, 14). In addition, the evolution of relatively large, well-protected phytoplankton lineages including diatoms, coccolithophores, and dinoflagellates, and their subsequent rise in ecological significance, is hypothesized to be the bottom-up impetus for massive ocean ecosystem restructuring in the Mesozoic and Cenozoic (15).Recent work (16–18) has called into question the classic timeline of diatoms'' increase in abundance and diversity (Fig. 1A). It has been assumed, based on fossil databases, that diatom diversity and abundance were generally very low at the beginning of the Cenozoic and increased toward the present, with a rapid rise beginning around the middle Miocene (23 Ma to 5 Ma) (8, 19, 20). Diatom abundance has, for the most part, been inferred from diatom diversity (21–23), although there is a similar increase in the relative abundance of diatoms in deep-sea sediments (5). Punctuating this long-term trend, siliceous microfossil (and radiolarian) abundance peaks in the Middle Eocene (5, 24) and is followed by a peak in diatom diversity in the latest Eocene to early Oligocene (10, 20, 21, 25, 26). Prior to the ecological rise of diatoms, radiolarians, a group of heterotrophic to mixotrophic protists, were the dominant pelagic silicifiers. As diatoms expanded, seawater silica concentrations are believed to have declined more than tenfold, leading to range contractions, reduced silicification, and reduced abundance in radiolarians and other silicifiers (11, 22, 23, 27–30). However, paired sponge and radiolarian silicon isotope work suggests roughly constant surface water silica concentrations between the latest Paleocene (60 Ma) and the Oligocene (33 Ma), at levels equivalent to modern surface ocean concentrations (<60 µM) (17) (Fig. 1). The Si isotope proxy builds from the observations that the extent of fractionation in sponges is strongly dependent on ambient dissolved Si concentration, while fractionation in radiolarians is mostly Si concentration independent (17, 31–36). In other words, these Si isotope findings suggest that any diatom-driven drawdown of silica must have occurred prior to the late Paleocene. Consistent with this alternative, Si isotope hypothesis, sponge reefs and hypersilicification in neritic sponges (indicative of high silica concentrations) disappeared in the Cretaceous to lower Paleocene (37). However, changes in Si isotope values in the Southern Ocean suggest yet another chronology, with diatom abundance increasing to near modern levels during the Eocene (10).Open in a separate windowFig. 1.Evidence for changes in the silica cycle over the Cenozoic. (A) Trends in diatom diversity according to two calculations dotted green line, Barron Diatom Catalogue; solid green line, Neptune Database (96); both as reported in ref. 20] roughly coincide with that of whales (blue line) (97). Similarly, the expansion of terrestrial grassland ecosystems (yellow bar) (98–100) and grazers (red dots, hypsodonty index from ref. 21) coincides with an increase in the relative abundance of diatoms (orange line is 1.5-My running median relative diatom to radiolarian abundances from ref. 5; orange envelope shows interquartile range; both begin at 48 Ma, before which data are too scarce). (B) Other normalized diatom diversity curves show a much earlier peak (and then drop) (21), and δ³⁰Si from radiolarians and sponges indicate consistent ocean Si] as far back as 61 Ma, suggesting diatoms did not increase their ecological impact since that time (17). Taxon silhouettes are from Phylopic.Here we consider whether secular change in the preservation potential of biogenic silica could reconcile this apparently conflicting evidence on diatoms and the evolution of the modern silica cycle. Two of the main factors determining whether biogenic silica makes it into the rock record are bottom-water temperature and sedimentation rate. Sedimentation rate determines how quickly silica is removed from the (diagenetically active) reactive zone beneath the sediment−water interface, and temperature determines the rates of dissolution in that zone (38). Both have changed through the Cenozoic (39–44), with declining deep-sea temperatures and increasing sedimentation rates correlated to declining atmospheric CO₂ levels, global cooling, and evolving ocean basins (45–49). Here we use a diagenetic model to investigate the effect of secular changes in deep-sea sedimentation rate and porewater temperature over the Cenozoic on opal burial efficiency (i.e., the proportion of opal rain reaching the seafloor that is permanently sequestered), and how this relates to the apparent rise of diatoms over the same interval. We generate a taphonomic null hypothesis for the fossil record of diatom abundance that explicitly assumes that there is no change in the flux of diatoms to the seafloor through the Cenozoic, and thereby quantifies and constrains the potential effect of changing diagenetic conditions on the interpretation of the siliceous microfossil record.