Response of deep-sea biota to rapid global warming during the early Paleocene

The early Paleogene was characterized by temperatures higher than recent, especially at high latitudes. Superimposed on this warm background climate, several transient (<200 kyr) global warming episodes (hyperthermals) were demonstrated. The Paleocene-Eocene Thermal Maximum or PETM (56 Ma) was the most extreme of these warming events. The PETM is characterized by a warming of ~5°C and a negative carbon isotope excursion (CIE) of ~3‰ in marine sediments. For deep-sea benthic foraminifera, a mass extinction was recorded in which 30-50% of all species went extinct. More recently, lesser hyperthermals were discovered in the early Eocene and in the Paleocene. These hyperthermal events are the best analogs for studying global warming, as both the early Paleogene hyperthermals and the recent global warming are caused by increasing CO2 levels in the atmosphere.

The Latest Danian Event or LDE (~62.1 Ma) is marked by a ~0.7‰ CIE in benthic foraminiferal tests and a ~2°C deep-sea temperature rise. The LDE shares a few other characteristics with recent global warming, making this particular event more interesting. The background CO2 concentration prior to the LDE was at ~400 ppm, low compared to the early Eocene, but close to pre-industrial Holocene values (~280 ppm). In addition, the LDE was superimposed on a decreasing temperature trend, similar to the long-term (Myr) cooling trend from the Pliocene to the Holocene. As it was the Paleogene first hyperthermal event, faunal patterns at the LDE can be studied without any preconditioning from previous events, as during the early Eocene.

The main aim of this project is the evaluation of responses of benthic foraminifera to temperature rise and environmental changes during the LDE. Pronounced faunal change is expected, given the benthic foraminiferal extinction at the PETM and benthic foraminiferal patterns at early Eocene hyperthermals. Faunal patterns are compared between locations and with the PETM and early Eocene hyperthermals. Relative timing and magnitude of benthic foraminiferal (Nuttallides truempyi) isotopic excursions is evaluated to further compare the event to early Eocene hyperthermals. For this thesis, the LDE was studied in three deep-sea sequences distributed over paleodepth (~2000 m for Site U1407, 2000-2500 m for Site 1210 and 2500-3000 m for Site 1262), oceanic basins (Atlantic and Pacific) and latitudes of 35-40°S (Site 1262), 20-25°N (Site 1210) and 30-35°S (Site U1407).

The sediment at Newfoundland Ridge Site U1407 was partially silicified and benthic foraminiferal identification proved particularly difficult. For these sediments, ultrasound was tested as a preparation technique in addition to wet sieving (Chapter 2). Ultrasound treatment improved the ease of benthic foraminiferal identification, but also increased fragmentation. The ideal time for ultrasound treatment depends on the initial material and research goal and was determined as 2-5 minutes for the Newfoundland Ridge benthic foraminiferal study.

A detailed isotope record for Walvis Ridge Site 1262 (Chapter 3) confirmed that the LDE was a double-peaked event, with peaks LDE1 and LDE2. Similar to the early Eocene hyperthermals (e.g., ETM2 and H2), both peaks occur at a 100 kyr eccentricity maximum in one 400 kyr eccentricity maximum. A similar relationship between the δ18O and δ13C excursion as for the early Eocene events shows similar climate sensitivity and a similar source of reduced carbon for the LDE and early Eocene hyperthermals. Both the astronomical pacing and the relationships between the two isotope systems point to a similar astronomically controlled causal mechanism for the LDE and early Eocene hyperthermals.

Prior to the LDE, benthic foraminifera record oligotrophic uppermost lower abyssal assemblages at Walvis Ridge ODP Site 1262 (Chapter 3 & 4). The food flux shows a gradually rising trend during the late Danian. During the LDE, a reduction in the efficiency of the biological pump causes renewed oligotrophic conditions. However, only minor faunal change was recorded compared to the early Eocene hyperthermals. A more pronounced maximum in abundance of Paralabamina lunata was recorded right after the LDE recovery, indicating increased abundance of fresh phytodetritus at the sea-floor.

Also the uppermost abyssal (~2000 m) Newfoundland Ridge IODP Site U1407 in the North Atlantic (Chapter 4) is characterized by oligotrophy in the late Danian. Silicification in the onset of the LDE indicates high biogenic silica production, followed by benthic foraminiferal assemblages dominated by P. lunata. This indicates high abundance of fresh phytodetritus. High Nodosaria spp. abundances in LDE1 indicate a raised year-round food abundance. After the LDE, oligotrophy is restored.

At Shatsky Rise Site 1210, the disappearance of Praemurica spp. and Morozovella praeangulata is recorded at the base of LDE1 (Chapter 5). Enhanced stratification is reconstructed starting ~170 kyr prior to the LDE. High abundances of Parasubbottina pseudobulloides/variospira indicate increased abundance of nutrients during the LDE, interpreted as evidence for the presence of a deep chlorophyll maximum, where nutrients reach the euphotic zone below the thin stratified layer.

At the sea-floor, Site 1210 (Chapter 6) shows stable oligotrophic conditions well before the LDE. Increased efficiency of the biological pump caused increased food flux at the upper abyssal site ~200 kyr prior to the LDE1 onset. Tappanina selmensis increased in abundance, marking episodic peaks in food abundance at the first warming, ~40 kyr prior to the LDE1 onset. Oligotrophic conditions are restored in the LDE core. High abundances of Bolivinoides huneri possibly record increased abundance of refractory organic matter. Renewed episodic food fluxes after the LDE are recorded by a second dominance period of T. selmensis. About 430 kyr after the LDE1 onset, oligotrophic conditions are gradually restored.

On a taxonomic level, the same species of benthic foraminifera are present at all abyssal sites, but they strongly differ at the bathyal Caravaca section. However, also at the abyssal sites, relative abundances vary. Abyssal benthic foraminiferal distribution patterns at the three studied sites are consistent with a latitudinal control (Chapter 7), with similar assemblages at similar latitudes on both hemispheres. Paleodepth differences turn out to be less important in structuring these assemblages. An increased food flux directly after the LDE relative to the pre-LDE interval is observed at Walvis Ridge and Shatsky Rise, in agreement with the increase in abundance of endobenthic morphotypes recorded in the Southern Ocean. No global first or last occurrences were recorded across the LDE at the studied sites.

The faunal response to the LDE was at all sites similar to the faunal response to the PETM and early Eocene hyperthermals. However, the faunal response was minor compared to the PETM and, at least at Walvis Ridge, compared to early Eocene hyperthermals. The faunal response was at all sites lesser than the extent of the faunal response to ETM2 at Walvis Ridge. We propose that the lower background temperatures and absolute temperature related thresholds are responsible for this minor faunal response. Other possible causes are the decreasing temperature trend prior to the LDE and the PETM extinction. Because of the decreasing temperature trend, benthic foraminifera were well-adapted to the higher deep-sea temperatures that prevailed during most of the Danian. The PETM extinction caused reduced diversity in the early Eocene benthic foraminiferal assemblages. The limited redundancy in these assemblages may have caused less resilient benthic foraminiferal communities.

Between LDE1 and LDE2, there is only limited isotopic and faunal recovery. This contrasts with the full recovery between the similarly spaced Eocene ETM2 and H2. Three options were proposed to account for this slower recovery. Rates of physical processes, such as carbonate dissolution and silicate weathering are dependent on temperature and CO2 concentration, causing a slower recovery in the lower LDE temperatures. Another option is the limited presence of opportunistic species during the LDE, causing a reduced abundance of organic matter, relative to early Eocene hyperthermals. Finally, also thermogenic methane release at the North Atlantic Igneous Province during the LDE could cause an additional CO2 input, and a slower recovery.

The benthic foraminiferal response to the temperature rise and environmental change during the LDE is geographically different. However, the event seems to be characterized by oligotrophy at middle-lower abyssal locations, while bathyal locations, and locations closer to the continent show increased productivity. This reflects the expansion of the trophic resource continuum.