Sunday :: Feb 11, 2007

Waiting for the Ice Age (part one)

by Christina Hulbe

When and how the "next ice age" might start is a topic that comes up with some frequency in the global warming and climate change conversation (and in comments here) so I thought it would be useful to provide an introduction to the processes at work. These are the basics we need to understand in order to put contemporary events into context and to build and use climate system models. There are two broad issues: global cooling driven by long-term cycles (10,000's of years) in Earth's orbit around our sun and northern hemisphere cooling driven by a change in North Atlantic Ocean circulation. I'll take these one at a time, celestial mechanics in part one today and ocean circulation in a few days in part two.

astronomical forcing of Earth's climate

Earth's orbit around the Sun varies over 10's of thousands of years due to the gravitational attractions of all the objects in our solar system. Thanks to the careful observations of Tycho Brahe and the careful mathematics of Johanes Kepler (and the political pressures of the Thirty Years War), astronomers have understood planetary motion (and masses) since the early 17th Century. The mathematical theory of gravitationally-driven variations in planetary orbits was developed by Urbain Le Verrier, a French mathematician who specialized in celestial mechanics. The importance of such variations to long-term climate change was first proposed by another French mathematician, Joseph Adhemar, and then expanded by James Croll, a 19th Century Scottish hotel manager, insurance agent, museum janitor (so he could read the books) and scientist.

Croll's hypothesis predicted multiple past ice ages, evidence of which was at the time (the latter half of the 1800's) being discovered in the geologic record. Indeed, sea floor sediment records show us that the climate of the last 2.6 million years or so has been characterized by profound cycles of glaciation and deglaciation in the northern hemisphere. Recently, the period of each complete cycle has been about 100,000 years, with gradual global cooling toward the glacial maximum followed by relatively rapid warming (over thousands of years). The most recent interglacial (like our own time: relatively high sea level & atmospheric CO2) was about 128,000 years ago. The last glacial maximum (relatively low sea level & atmospheric CO2) was about 18,000 years ago and the big North American ice sheet that reached its maximum extent at that time was mostly gone 10,000 years later. (On geologic time scales, that's a rapid retreat.)

Croll didn't have all the details right but the fundamental importance of orbital variations was recognized. The correct calculations were eventually carried out by Milutin Milankovic, a Serbian mathematician. He worked out much of his astronomical theory of the ice ages while a prisoner of war in WWI. His Cannon of Insolation and the Ice Age Problem was published in 1941. It took until the 1970's for the theory to be validated, in a study of oxygen isotope variations in deep sea sediment cores (Hays, Imbrie, and Scackleton, 1976, Variations in the Earth’s orbit: Pacemaker of the ice agesScience, v 194; a paper of singular importance in paleoclimatology).

Three aspects of Earth's orbit are of concern in driving long-term climate change: the tilt of our spin axis with respect to the plane in which we move about our sun; the eccentricity of the orbit (how elliptical it is); and the precession of the equinoxes about the orbit (is a particular hemisphere near or far from the sun in a given season?). Together, these variations determine the intensity of incoming solar radiation ("insolation") at any location on the surface of the planet at any time and set the pace for long-time-scale climate cycles. Feedbacks within the Earth system modulate (and amplify) the system's response to this external orbital forcing.

The most recent northern hemisphere insolation peak was about 9,000 years ago and the corresponding "climate optimum" is generally taken to be from about 9 to 5 thousand years ago. So does this mean that we are "due for an ice age?" On long enough time scales the answer is certainly yes, but a more refined answer requires additional information.

We must look back through four glacial cycles to find the most suitable point of comparison with our present non-anthropogenic climate setting. (This is because the longest period in the orbital forcing is 413,000 years.) The interglacial most orbitally similar to our own began around 420,000 years ago and is called "marine isotope sage 11" (MIS11) by paleoclimatologists. Climate proxy records that extend back that far indicate that MIS11 lasted between 20 and 30 thousand years, far more time than we've yet spent in the current interglacial.

A very important point to make here is that climate, and climate cycles over time, are driven not just by celestial mechanics. The large amplitude glacial-to-interglacial cycles in temperature, sea level, atmospheric greenhouse gas content and so on are a modulation (and amplification) of the orbital forcing by Earth's climate system. A short list of the components of that system includes ice sheets, sea ice, atmospheric circulation, atmospheric chemistry, the terrestrial and marine carbon cycles, ocean circulation, and ocean chemistry. It would (will?) take many posts to work through these. Perhaps you all could make suggestions about what's most interesting. In any case, promised to write about Greenland and I'll do that first.

The bottom line is this: orbital forcing in and of itself is very gradual and the paleoclimate record suggests that the un-perturbed system might have been in for another 10,000 years or so of relatively warm conditions. Finally, we need to recognize that the excess CO2 we've added to the atmosphere far exceeds the range of variability for the un-perturbed (orbital forcing + feedbacks) system (pdf of a figure showing six cycles, for comparison: the present CO2 is 383 ppmv, the pre-industrial level was 278 ppmv, and the typical glacial-to-interglacial change is between 80 and 100 ppmv).

Christina Hulbe :: 12:12 PM :: Comments (5) :: Digg It!