Ice Ages and Climate Change

Read any long Facebook argument about climate change and someone inevitably writes “of course the climate is changing, it’s natural, it’s always changing”. If the argument drags on a bit further, you’ll eventually hear, “climate change is a good thing, we’d be headed for an ice age otherwise”.

This post explains the theory of ice ages, and just how wrong comments like that are.

Just to review the basics, the Earth goes around the sun in an ellipse, and the Earth is tilted relative to it’s orbit. The seasons come mostly from the tilt, the orbit is close enough to circular that it doesn’t matter much.

But all those details change over time. The other planets are pulling on the Earth, and some of the other planets weigh a lot more than the Earth, so they can tug our orbit into slightly different shapes.

I programmed a simulation of the solar system, then tried running the simulation a million years into the future, to log how the orbits would change.

These are the results I came up with, for the orbit’s eccentricity, x axis is years, y is eccentricity (0 is a circle, higher means it’s a skinnier ellipse):

This is inclination, the plane of the orbit moves up and down by a few degrees:

Ice ages are influenced by fluctuations in the Earth’s orbit. There are a few natural cycles which drive the Earth’s climate. This graphic from wikipedia summarizes the major ones:

Note that this roughly matches my data — my eccentricity graph is about the same as the right side of the green graph here, though I scaled mine differently. So, that’s cool, I can sit at home and program a computer to predict where the Earth will be in a million years.

The math and science are pretty straightforward, a lot of modern science is more complicated, and I can’t reproduce all of it. The theory of ice ages goes roughly like this:

We have some guesses as to what the Earth’s temperature was like throughout history. We’ve drilled deep holes into glaciers in Antarctica. The snow has layers from each prior year, like rings on a tree. Each layer has a different composition (of gasses, of isotopes) and we can make some guesses about the atmosphere and the climate, in the year that layer was formed. Here’s the graph from one experiment, going back 400,000 years:

There’s 3 superimposed cycles of fluctuations in the data.

The shortest one comes from the precession of the Earth’s poles. The Earth spins, but it spins like a wobbling top, so the north pole doesn’t always point to the same star. While the Earth is spinning quickly (once a day…) the axis is moving slowly, once every 26,000 years. That should make winters a little bit warmer or colder, because the peak of winter will happen at a different point in the Earth’s elliptical orbit. The effect should be opposite, in each hemisphere. But the Earth is asymmetrical, right now there’s more land in the northern hemisphere, so I guess the net effect doesn’t have to be balanced.

There’s a second cycle, of about 41,000 years. The tilt of the Earth increases and decreases periodically by a degree or two. I don’t understand this one as well — I’m not sure if the tilt is just changing relative to the plane of the orbit or if the tilt of the earth is also changing in some fixed frame of reference (this would be called nutation, for a spinning top). Either way, winters are a little bit warmer or colder.

The longest cycle is around 100,000 years. Most people think that’s because of eccentricity changes in the orbit, but there are other theories — the lecturer (Richard Muller) that originally got me interested in this claimed that it was from a 100,000 year cycle in the inclination of the orbit, with the Earth regularly passing through a fixed plane of dust in the solar system.

I think the ice data only goes back about a million years. The ice sheets themselves aren’t much older. The Earth is not always in an ice age, though we are in one now. Specifically, we’re at a warm intermission within an ice age. Looking back at the temperature graph, notice that warm interglacials are shorter than the long cold periods in between. What we think of as normal is a brief respite from a colder Earth, at least that’s been the pattern for the last million years.

Over an even longer term, there are bigger climate trends that aren’t cyclical. This is an attempt someone did to graph climate history over a much longer period. (I think this comes from layers of sea sediments, and a variety of other geological clues, I’m not that familiar with all the science)

The longer term changes happen for various reasons. Continents move over time. If the continents allow for easy transfer of water between the equator and poles, the Earth tends to be warmer. Right now, we are in a cold configuration — Antarctica sits over the south pole, and the arctic ocean is mostly surrounded by land, so it doesn’t get much water flow. The shift colder over the last 60 million years probably has something to do with continents moving into that position. Things may have gotten colder a few million years ago, with the formation of central america, which prevents the atlantic and pacific from mixing. The uplift of the Himalayas, over the last 40 million years, may have also slowly cooled the planet. (supposedly that increases total rainfall, which slowly washes carbon dioxide out of the atmosphere) Over really long time scales, the sun will get brighter (about 10% per billion years), and the Earth should heat up.

From that long term graph, you’ll notice that there have been eras much hotter and much cooler than the present day. At one extreme, the Earth was maybe 25 degrees warmer, 50 million years ago. Palm trees grew north of the arctic circle. That’s a little after the era of the dinosaurs, but their time was pretty warm as well. At the other extreme, we’ve had a few other ice ages, some worse than the present one. It’s possible the entire Earth was frozen, in the coldest one. That was about 600 million years ago, after life had started, but before it had gotten really complex (before the cambrian explosion).

Right now, there are short term changes that are happening faster than any of the natural cycles:

That is, of course, a global warming graph. People are warming the planet rapidly, but the total change has only been a couple degrees so far. We don’t entirely know what that will do, and we don’t agree on what to do about it, and we’ve elected a president that doesn’t even believe it’s happening. But, you know, that’s hopefully just a 4 year cycle that will correct itself.

One thing that I think is missing from the usual conversations about global warming is some sense of the time scales that changes will happen over. Some people talk about the danger of sea level rise. If Greenland and Antarctica melt entirely, the ocean could rise about 200 feet. Most forecasts say that we’re only expecting a few feet of sea level rise by 2100, though.

I used to read those two numbers and think: what we’re doing is mild, the amount of warming is still small relative to what would really cause damage. That interpretation was wrong, though — the main reason that sea level rise is slow is not because the world isn’t warm enough, it’s because the ice in Greenland and Antarctica is two miles thick, and it takes a really long time to melt something that big. We’re on track for larger sea level rises, it’s just something that takes hundreds/thousand of years to happen fully.

This paper tries to model the future melting of continent sized ice sheets:
http://mahb.stanford.edu/library-item/consequences-of-twenty-first-century-policy-for-multi-millennial-climate-and-sea-level-change/

Basically, for a given level of carbon put into the atmosphere, the Earth is committed to a certain amount of sea level rise. It takes a few thousand years for that rise to fully occur, though.

If we all quit burning fossil fuels tomorrow, we’d already be committed to maybe 10 feet of sea level rise. We won’t stop tomorrow, of course, in the best case we will probably continue burning for decades, maybe increase the amount for a while, and then slowly cut back as some new technology comes around. Suppose we take 100 years to stop, and our use averages about the current level for that time. If we do that, we’ll be on a path for 60 feet of sea level rise, over the course of the next thousand years or so. That’s a strange thing to think about, that our actions today could flood cities that far into the future.

Here’s a few maps showing what 60 feet of rise looks like:

That’s one of the things we’re discussing, when we talk about global warming. Do we want more economic growth and cheap fuel now? Or do we still want to have Florida a thousand years from now?

It’s hard to even reduce this to an economic problem. How much does it cost to move a billion people’s homes and cities, over a thousand years? Will people come up with a cheaper fix than that (some opposite form of geo-engineering)? How much do you discount the future costs, relative to what it costs today to prevent it?

We tend to talk about environmental problems with cliches like “we must act now to save our children”. When it comes to sea level rise, the reality is something more like “our generation and our children must make sacrifices to protect people for many generations into the future”.

So, how do ice ages and global warming interact? When will the climate naturally start cooling again? And where is it going, if we are warming the planet now? Over thousands of years, which trend will be stronger?

We’ve been in a warm period for the last 12,000 years or so. I don’t know if there’s a scientific consensus on how long it will last. Most interglacial periods don’t last much longer than ours, if you just compare us to averages, it’s due to end in a few thousand years. Some other papers try to calculate based on orbital changes, and say that we’re at an abnormally good point in the cycle, and things might naturally stay warm for tens of thousands of years more.

Either way, this should naturally end, and the Earth will get 10–15 degrees (Fahrenheit) colder. Just as glaciers don’t melt quickly, the Earth doesn’t freeze quickly. Instead, every winter there’s some snow, the snow doesn’t all melt the next summer, the next winter it piles up a little higher, and after thousands of years, you have a mile thick sheet of ice covering Canada, Russia, and the northern United States.

If people are still around, and civilization is doing well, we probably won’t want that to happen. So, we’d… maybe pump a whole lot of greenhouse gasses into the atmosphere, to fix it? (or maybe we could bulldoze Panama, if that ocean circulation theory is true)

That’s another funny thing to think about. Global warming could be a technique for good, someday. But the timing is off, we’re doing it thousands, maybe tens of thousands of years too early.

This paper tries to model the Earth’s future temperature, based on various amounts of carbon emission in the present day:
http://onlinelibrary.wiley.com/doi/10.1029/2004GC000891/full

This image gives a quick summary of what they modelled — the amount we’ve emitted so far is not enough to skip the next ice age. If we keep burning fossil fuels for another 100 years or so, we’ll maybe keep the Earth warm for the next 100 thousand years. Past a point, C02 will weather out of the atmosphere, and we’ll be back to the usual cycles.

So, that’s one silver lining to what we’re doing now. Of course, if we were doing this intentionally, we’d wait a few millenia, and only warm the planet in a reactive way.

I’m not writing this to take a stance denying or favoring global warming. I’m just making two points. Some of the consequences of global warming are much slower than you’d intuitively expect. Also, over really long time scales, we may want to intentionally alter the climate.

For the most part, I worry more about global warming’s consequences on nature than its effects on people. As you might have noticed from the graphs above, life can exist on our planet in a wide variety of temperatures, at least from 15 degrees colder to 25 degrees warmer. Plants and animals move north or south as the temperature changes. New animals evolve to take advantage of new opportunities. That’s assuming gradual change, though, with lots of time for life to adapt.

With huge, sudden changes, you get this:

The real risk I worry about is that the changes we’re causing might happen too fast for nature to react. The Earth has seen 5 major extinctions in the past, we’re possibly in the middle of another. It takes millions of years for biodiversity to recover from one, and life is never the same, when it does recover. If we wreck the ecosystem we have, we will never get it back.

Global warming is not the only driver of extinction, I’m not sure it’s even the biggest one. I think the biggest one may be habitat loss, as forests are cut down for agriculture, and more land turns into roads and cities. There’s other factors — big animals in Africa and Asia are dying out because of hunting/poaching. Marine life is dying out from overfishing. We accidentally or intentionally transport invasive species from place to place, and local species die out. (this is a big problem on islands, but can be an issue from continent to continent)

Preventing extinction is the big environmental issue for our time. Slowing down global warming is one big part of that, but there’s a whole lot of other work to be done. That also means stopping deforestation, poaching, overfishing, and other kinds of pollution. It means getting people to care about leaving some land untouched, or preserving wildlife, even if there’s no economic benefit to it.

(Incidentally, that dinosaur image came from the onion, which might be able to put things in perspective better than I can?)