Climate Change: Could It Be Random?

Several large international projects have succeeded in drilling ice-cores from the top of the Greenland inland ice through the more than 3 km thick ice sheet. The ice is a frozen archive of the climate of the past, which has been dated back all the way to the previous interglacial Eem-period more than 120.000 years ago.

The ice archive shows that the climate has experienced very severe changes during the glacial period. During the glacial period there were 26 abrupt temperature increases of about 7-10 degrees. These glacial warm periods are named Dansgaard-Oeschger events after the two scientists first observing them.

The global warming we experience presently will cause a temperature increase of perhaps 2-5 degrees in the next century if greenhouse gas emissions continue, researchers claim. This will lead to increased sea levels and more severe weather with terrible consequences. The temperature rise during the glacial period were much larger and happened much faster.

Temperature increased by 10 degrees in less than 50 years with changes to the ocean currents and the whole ecosystem. These changes have caused sea level rises up to perhaps as much as 8 meters and large changes to the vegetation.

Climate shifts looks periodic

The 26 climate shifts are apparently periodic. They seem to occur with a period of 1470 years. Every now and then a period is skipped and the shifts occur 3-4000 years apart. Professor Peter Ditlevsen at the Niels Bohr Institute, Copenhagen University wanted to investigate the periodicity of the climate shifts. He asked: "Could it be that the shifts are chaotic and random, they just look periodic by pure coincidence. How probable is that?"

Using mathematical models of the climate shifts he calculated the probability of the periodicity. He focused on the time intervals between the climate shifts. How regular are they really? As a baton, periodically beating, how far from the beating are the climate shifts? If the distances are perfectly periodic 100% is obtained. It turned out that the climate shifts hit the beats of the baton by 70%.

Computer simulation

He then had the computer spreading the shifts over the ice age randomly. He did this 1000 times with different random time intervals. In this he got between 40% and 90% right hits. The major part of the calculations was between 55% and 75%.

Then he calculated the opposite assumption, that the climate shifts has a period. Again he made 1000 calculations, but this time the numbers came out between 80% and 100%. The major part came out above 90%. But 90% is not the regularity for the real climate changes, they occur with 70%.

The conclusion drawn by Peter Ditlevsen is that the probability of hitting 70% is less if the climate shifts are periodic than if they are random. This is very important for understanding the cause of the climate changes and especially for predicting climate shifts. If they are random and chaotic they are fundamentally unpredictable.