Another crack in the wall. From PhysOrg:
Sunlight has more powerful influence on ocean circulation and climate than North American ice sheets
The distribution of sunlight, rather than the size of North American ice sheets, is the key variable in changes in the North Atlantic deep-water formation during the last four glacial cycles, according to the article. The new study goes back 425,000 years, according to Lorraine Lisiecki, first author and assistant professor in the Department of Earth Science at the University of California, Santa Barbara.
Lisiecki and her co-authors studied 24 separate locations in the Atlantic by analyzing information from ocean sediment cores. By observing the properties of the shells of tiny marine organisms, called foraminifera, found in these cores, they were able to deduce information about the North Atlantic deep water formation. Scientists can discern historical ocean temperature and circulation patterns through the analysis of the chemical composition of these marine animals.
Previously, scientists relied on a study called "Specmap," performed in 1992, to find out how different parts of the climate system interacted with one another during glacial cycles. Specmap analyzed ocean circulation at only one place in the Atlantic.
"What I found was that the one site that the Specmap study used actually didn't match most of the other sites in the Atlantic," said Lisiecki. "They just happened to have a strange site that didn't behave like most of the other sites. The other sites show that the circulation is not responding to the ice volume, but that it is responding to changes in the distribution of sunlight."
For people to base their models on one site when so much other data are available makes me immediately think: "Cherry-picking your Data" which is a cardinal sin in Science. Nice to see this finally hit the light of day and in Nature no less... The paper is here: Atlantic overturning responses to Late Pleistocene climate forcings but you need to subscribe or pay to see the full document. Best to check a local library...
