PETM climate warming 56 million years ago strongly tied to igneous activity

Part 2 – Earth’s “Internal Bleeding”

This is the 2^nd part in a 3-part series on the PETM expanded from an article I originally wrote for Quanta Magazine and features quotes from interviews that appeared in that piece. Click here for Part 1.

The connection between volcanic activity and the release of organic carbon was first spotted in 2004 by Henrik Svensen of the University of Oslo, and colleagues. Examining seismic scans through the layers of sediments offshore Norway, they saw vents that led upwards from sheet-like bodies of solidified magma (“sills”) to craters that formed at about the time of the PETM. The vents resembled similar structures in the Siberian Traps and in South Africa (Karoo Volcanics), which have been linked to climate changes at the end-Permian and in the Jurassic respectively. Svenson et al reasoned that the Norwegian vents resulted from the hot magma in the sills baking organic-rich sediments to generate methane and CO₂ that erupted through the seabed to drive the PETM.

Crucially, that carbon would have been rich in carbon-12.

“The interesting thing about sills is that they are emplaced really, really fast, and even a thick sill can be fully emplaced within a timescale of a hundred years,” Svensen told me in 2016. “A timescale of a few thousand years includes emplacement, metamorphism, and also gas generation… we know that some of the sills … intruded into organic-rich sediments and so carbon gases must have been generated, and that’s a fact.”

Artist impression of methane and CO2 erupting from the Atlantic seabed during the PETM Modified with permission from original by Millett in Reynolds et al EPSL 2017

Artist’s impression of methane and CO2 erupting from the Atlantic seabed during the PETM Modified with permission from original by J.M. Millett in Reynolds et al, EPSL 2017

"you can’t drill them all"

But these structures lay deep beneath the seabed, under hundreds of meters of Atlantic Ocean, so getting samples to date and analyze was near impossible. In 2016 Frieling et al published data from a 1986 oil exploration borehole that had been drilled through 390 meters (1,280 feet) of ocean and 3,521 meters (11,552 feet) into the seabed. It is the only published borehole to date that samples one of those vents identified by Svensen and colleagues in 2004. Dating of samples showed that the vent formed in the middle of the PETM, not at the start, so it wasn’t quite a ‘smoking gun,’ but the carbon isotopes confirmed that the vent had emitted carbon rich in carbon-12.

“You drill one crater and its outside the PETM, then what can you conclude when this particularly crater postdates [the onset of the PETM]? But what about the 2,000 others? You can’t drill them all!” Svensen told me.

This lack of data meant scientists studying the PETM remained skeptical of the volcanic trigger. Each sill can only generate a limited amount of gas and although seismic scans reveal thousands of sills, the NAIP was active for several million years, and it wasn’t clear how carbon emissions were spread out or concentrated in that time. In contrast, sediments constrain the time it took to start the PETM to perhaps as little as 3,000 or 4,000 years.

"you can turn the tap on so fast - in five to ten thousand years"

“No one thought that it could deliver the amount of carbon on the timescales of the PETM,” said Tom Dunkley Jones, a coauthor on the new study by a team from the University of Birmingham, UK, which has found a new way to constrain the timing of the sills and their greenhouse gas emissions.

“That will be news to many paleoclimate people, that you can turn the tap on so fast - in five to ten thousand years,” said Stephen Jones, the study’s lead author.

Study	C isotope mix (‰)	total C input (Pg C)	onset duration
Zeebe 2009	−50	4,500	5,000 years
Bowen 2015	−55	3,284	∼3,000 years
Frieling 2016	−50 and −45	4,500	5,000 years
Gutjahr 2017	Avg −11	10,200	∼21,000 years
Kirtland Turner 2018	−35 and −6	9,660	3,000 years

Various estimates of the time it took to start the PETM. From Kirtland Turner 2018. Carbon isotope mix in δ13C notation, more negative is more carbon-12-rich (ie more organic-rich)

The Birmingham team used techniques and data from the oil industry to quantify all the pieces of the puzzle: the rate of magma generation by the mantle, the ‘internal bleeding’ of that magma into sediments as sills, and the resulting gas emissions:

“Relatively few large intrusions with this rapid recurrence time controlled by the plume flux can actually give you the carbon flux that you need, and that’s the real breakthrough!” said Dunkley Jones.

Clues from Red October’s graveyard

Stephen Jones had earlier worked on underwater ridges and valleys near Iceland, where Cold War submarines stalked in the movie “The Hunt for Red October.” Iceland today sits on top of a mantle plume – a wide conduit of hot rock that rises all the way from the core-mantle boundary. Pulses of hot mantle periodically ride up the Iceland Plume to create characteristic “v-shaped ridges” of lava on the Atlantic seafloor.

“The one important thing about the Iceland Plume creating the modern v-shaped ridges is it’s a smoking gun! It’s actually doing it at the moment,” said Jones.

Although they demonstrate the pulsing behavior of the Iceland Plume, the v-shaped ridges do not go back to the time of the PETM, so Jones et al used knowledge gained from oil industry studies of an important reservoir for oil and gas in the region, known as the “Forties Sandstone,” that formed at the time.

North Atlantic at the time of PETM showing plume head

The North Atlantic at the time of the PETM when the Iceland Plume lifted the seabed into a land bridge between Scotland and Greenland. Ellipses show the expanding pulse of the Iceland Plume. Paleogeography from Maclennan & Jones EPSL 2006, Iceland Plume from Jones et al 2019.

"the uplift we know really well from seismic data across this area"

The Forties Sandstone formed when a wide area of seabed between Scotland and Greenland was lifted into the air by a colossal pulse of hot mantle delivered by the Iceland Plume. It resulted in a temporary land bridge that enabled the dispersal of mammals and plants between Europe and America, and it developed rivers that transported the eroding seabed offshore, where it settled as the Forties Sandstone.

“The uplift we know really well from seismic data across this area,” said Dunkley Jones. “We can see marine sediments being uplifted and exposed above the sea level. You can actually see downcutting of river systems and features that tell you about the uplift, and that’s been known about for ten to fifteen years.”

The Great Atlantic Bake Off

As the pulse of hot mantle spread beneath the crust it melted, bleeding magma into sediments through interconnected sills. Differences in the timing of the seabed uplift across the region told the Birmingham team how fast that ‘internal bleeding’ spread, but to figure out the baking power of the sills they first had to find and measure them. That burden fell to coauthor Murray Hoggett and a PhD student named Karina Fernandez, who studied tens of thousands of kilometers of seismic scans to record 11,000 to 18,000 sills in the region.

Seismic image of sills from Jones et al supplementary fig 2 (c)

Caption: Seismic scan showing a vertical slice through sediments beneath the eastern Atlantic Ocean, offshore Ireland and Scotland. Red lines are sills, yellow line is the seabed at the time the sills were intruded. From Jones et al 2019 supplementary fig 2 (c) CC BY 4.0

“Until we had that database of geometries and dimensions, we couldn’t even tell you how fast or how regular they needed to be to get to the right carbon release,” said Sarah Greene, a co-author.

They then used a standard oil industry calculation of the gas that can be cooked from oil source rock, called a “kinetic maturation model,” to calculate the rate individual sills would bake off gas from sediments, combined with a statistical technique called “Monte Carlo Simulation” to calculate the rate that the sills would emit gas, collectively.

“The kinetic kerogen maturation model is what the oil industry is based on,” Jones told me. “It’s well-constrained because the oil industry models are constrained at the very slow timescales of burial and they’re [also] constrained by lab data. Sills are halfway between, in terms of the time.”

"You need a bunch to be active at the same time to sum up to the kinds of total release that we see"

“We don’t ever know which sill was the first sill that intruded, which was the second, which was the third, so we’re essentially kind of sampling out of a bag randomly and adding those up,” explained Greene. “Each sill is small and generates a small amount of carbon. You need a bunch to be active at the same time to sum up to the kinds of total release that we see.”

They show that the spread of magma from the Iceland Plume was fast enough to intrude a fresh sill into oil-rich sediments at a rate of 1 fresh sill every 2 to 6 years, each one baking gas for centuries from an expanding halo around it, which collected and “fracked” to the surface, erupting through vents on land and on the seabed.

The North Atlantic would have bubbled like a boiling pot.

To be continued in part 3