Can bumps in the seafloor explain glacial-interglacial cycles?

Citation: Crowley et al, Glacial cycles drive variations in the production of oceanic crust (2015), Science 347: 1237, and subsequent commentaries.

Twice in my time as a graduate student has a scientific discovery truly blown my mind. One of these ideas was a theory for mass extinctions, which isn’t yet published (the author is still working on some of the details, and admirably wants to get it just right). Thankfully the other one recently was, so that’s what we’re going to talk about today.

Arguably the greatest mystery of earth’s climate is that of glacial-interglacial cycles. For the past ~800,000 years, the earth has switched between glacial periods (read: ice ages) and interglacial periods (read: not ice ages) with a full cycle taking about 100,000 years. We know this because scientists have devised clever ways to dig deep into polar ice and analyze the gas bubbles that get trapped when the ice freezes. We’ll have to be a bit fast and loose about this sort of thing today, because part of a mindblowing discovery is that it tends to bring a bunch of different things together in a neat way. The weird part is that before 800,000 years ago there were also glacial-interglacial cycles, but they took more like 40,000 years.

Screen Shot 2016-08-25 at 10.07.14 PM
A temperature proxy, signifying glacial-interglacial cycles of the past 800kyrs. Note the periodic behavior still visible through all the noise. Image credit: Ed Boyle.

Lots of theories have been proposed for why this happens, but they’re mostly speculative, and it basically reduces down to the fact that atmospheric carbon dioxide concentrations (which influence temperature) have to be oscillating every 40,000 or 100,000 years, and it’s hard to imagine why it would switch from one frequency to the other all of a sudden, not to mention it’s hard to imagine why it would ever oscillate like this anyways. If one could explain with a single mechanism how:

  1. i) an ice age could produce an increased atmospheric CO2 concentration,
  2. ii) a non-ice age could reduce atmospheric CO2 sufficiently for glaciers to form

iii) this could happen on either 40,000 or 100,000 year frequencies

one could have a pretty compelling explanation of this elusive and very important climatic mystery. Your inner sleuth might be able to detect that this is what we’re after today.

Ok, now let’s switch gears and talk about the earth’s orbit. Why? Because glaciers care. The earth rotates around the sun once a year, and the axis of rotation of the earth (the arrow pointing from the South Pole to the North Pole) is slightly askew (23.5°, to be exact) from the plane in which the earth rotates around the sun. However, this angle changes between 22.1 and 24.5° every 41,000 years, and this axis of rotation wobbles like a top, making a full rotation every 23,000 years. The pictures say it better than I can.

These two phenomena are called procession and obliquity, respectively, and they determine how much sunlight the high latitudes get; at certain times ice sheets get blasted with extra sun, causing lots of melt. Since these processes can influence formation and melting of glaciers, we should be keeping an eye out for their 41,000 year and 23,000 year signatures in whatever process we’re looking for.

Screen Shot 2016-08-25 at 10.07.24 PM
Obliquity. Image credit: wikicommons.


Screen Shot 2016-08-25 at 10.07.33 PM
Precession. Image credit: wikicommons



Ok, now let’s switch gears again and talk about volcanoes on the bottom of the seafloor. Why? Well, volcanoes are a great way to change atmospheric CO2. Most of the time scientists think about how the ocean might burp some extra carbon into the atmosphere as a way to change this CO2, but really most of the carbon on earth is under the mantle, and when volcanoes blow up, they can pump a lot of CO2 into the atmosphere. They’re under a lot of pressure, though, which damps their explosions.

Here’s where the authors of the present study come in. They saw a potential mechanism for glacial-interglacial changes fitting (i-iii) above in these volcanoes. They figured if there was lots of ice during a glacial period, that means sea level is lower (by 50-100 meters), which reduces pressure on the seafloor, which makes it easier for volcanoes to erupt, thus pumping CO2 into the atmosphere. Supporting this idea is the fact that the difference in pressure caused by ~50 meters of sea level is just enough to make a lot of seafloor volcanoes explode. This CO2 then warms the surface of the earth, melting the glaciers, increasing the sea level, adding pressure to the volcanoes, and then letting the CO2 concentration fall back down, until it gets cold enough again to start to form glaciers, and we have ourselves a glacial-interglacial cycle.

Screen Shot 2016-08-03 at 4.50.02 PM
Mechanism for glacial-interglacial cycles from sea level changes causing pressure changes in seafloor volcanoes. How cool is that! Adapted from paper.


How would one go about showing this happens, though? And what about the 40,000 vs. 100,000 year distinction? Well, since things below the seafloor happen very slowly, there’s a delay between this pressure change and the time it takes for the volcano to spew carbon into the atmosphere. Other scientists’ best guess for this delay time, considering lots of details about the earth’s curst and all, is 40,000-50,000 years.   So that would be the 40,000 year cycle. Then there’s another nifty thing about delayed oscillating systems like this one, which is that if they get an extra strong push, they can go into a slower but stronger oscillatory mode, which the scientists estimate would be at about 100,000 years. Obliquity made an unusually big kick about 800,000 years ago, which means this theory is coming together pretty nicely. It can explain (i-iii) above in a very simple fashion, i.e. with just a feedback between glaciers and seafloor volcanoes, as long as it can be demonstrated that seafloor volcanoes do really feel the effect of these sea level changes.

Screen Shot 2016-08-03 at 4.52.02 PM
Bathymetry data from the South Ocean – note the periodic and parallel ridges. Adapted from paper.


All that’s left, then, is to show that seafloor volcanoes do record the influence of the pressure differenced caused by glacier formation and melting. The characteristic signature of this influence would be seeing the strongly periodic behavior of the orbital cycles that influence glacial formation and melting, at 23,000, 41,000, and 100,000 years. The authors take some deep-sea seafloor maps from a Korean vessel’s 2011 and 2013 cruises, and they see periodic ridges along the seafloor next to a slow-spouting volcano, which slowly generates new rock each year and pushes the older rock away. By knowing the characteristic mantle-forming rate for this volcano (3 cm/yr), they can convert the distance of these ridges into time. In short, they convert the map of the seafloor bottom into a record of how fast the volcano was spitting out fresh magma back for hundreds of thousands of years. Guess what the frequencies are where they see big signatures: 23,000, 40,000 and 100,000! Other authors report similar behavior for volcanoes off the coasts of Chile and Mexico, too, so it’s not a unique phenomenon, and in fact the 100,000 year signature is not seen in ridges that are more than 800,000 years old, further supporting the theory! It should also be noted that the authors back this finding up with substantial computer simulation studies, showing that it’s not a mere coincidence that this is happening.

Screen Shot 2016-08-03 at 4.51.44 PM
Figure showing the strong periodic signals at 23,000, 41,000, and 100,000 year frequencies in the seafloor ridges. Adapted from paper.



Whoa, what just happened so fast? In short, the authors have shown that magma formation on the bottom of the seafloor is influenced by glacial formation and melting, by showing they have the same temporal patterns, themselves related to the earth’s orbital changes. This provides really strong support for a simple theory that can explain one of the greatest mysteries in climate, linking seafloor volcanoes with glacial-interglacial cycles. While the argument for the whole glacial-interglacial problem is not airtight just yet, it provides the context for the significance of this study. Even without this, it’s pretty cool that ridges on the seafloor are influenced by the wobble of earth’s rotation!

Leave a Reply

Your email address will not be published.