A new study extends global average surface temperature records back to 2 million years ago, further back in time than any previous climate research. The research suggests that global warming from human emissions of greenhouse gases will cause far more warming, up to 9 degrees Celsius, or 16.2 degrees Fahrenheit, during the next several millennia than previously expected.

However, prominent climate researchers refuted that more disturbing conclusion in interviews with Mashable.

The study, published Monday in the scientific journal Nature, uses nearly five dozen ocean sediment cores to develop a record of Earth's global average surface temperature dating back to 2 million years.

The study's author, Carolyn Snyder, a scholar at Stanford University and an official at the U.S. Environmental Protection Agency, used these sediment cores to create more than 20,000 climate reconstructions from specific points in the ocean.

Like a forensic investigator examining a crime scene, Snyder used a variety of methods -- in this case statistics and modeling -- to extend the temperature reconstructions to land areas as well.

"Previously, the longest continuous global temperature reconstruction was only 22,000 years long," Snyder told Mashable. "Before that in time, only a few isolated individual windows in time had global temperature reconstructions."

While outside scientists praised the temperature reconstruction as a valuable addition to climate science, several researchers contacted for this article were critical of how Snyder used the temperature reconstructions to infer the climate's sensitivity to changing ice sheets, carbon dioxide levels, and other factors.

The study estimates what is known as the "Earth system sensitivity," which encompasses a variety of feedbacks within the climate system, from the response of the atmosphere and oceans to fluctuations in greenhouse gases to the ways that ice sheet expansion or melting can alter global temperatures.

However, this metric is a correlation between events, and doesn't pinpoint whether one event caused another. Still, the study estimates an Earth system sensitivity of 9 degrees Celsius, or 16.2 degrees Fahrenheit, per a doubling of atmospheric carbon dioxide levels over millennium timescales.

In more simple terms, this means that over the long, long-term, our planet will see its global average surface temperature increase by up to 9 degrees Celsius if greenhouse gases in the atmosphere were to double, which they are currently on course to do.

The study found that if all greenhouse gas emissions were to cease today, the climate would still warm by about 5 degrees Celsius, or 9 degrees Fahrenheit, during the next several centuries.

However, the Earth system sensitivity metric is not the same as the similarly named, but altogether different, scientific metric known as climate sensitivity. That metric is defined as how much the globe would warm if greenhouse gas concentrations in the atmosphere were to double.

Climate sensitivity considers the influence of greenhouse gases, such as carbon dioxide, alone, while Earth system sensitivity involves a variety of feedbacks between the land, oceans and atmosphere, some of which are not well understood.

With climate sensitivity, carbon dioxide and other greenhouse gases are in the driver’s seat, whereas with Earth system sensitivity, there are many drivers, with cars going in different directions and sometimes colliding head on.

Estimates of climate sensitivity tend to be much lower than 9 degrees Celsius, closer to about 3 degrees Celsius.

The problem, Snyder as well as several outside scientists told Mashable, is that it's not clear exactly what was driving temperature changes during some time periods in the past.

"[Earth system sensitivity] is a useful metric that summarizes a combination of interactive feedbacks in the climate system (including temperature, greenhouse gases, ice sheets, vegetation, and dust)," Snyder said in an email. 

"But it is a correlation observed in the past, not a test of causation," she said.

Michael Mann, a climate researcher at Penn State University who has published influential studies on the planet's climate history, said he views the new study as "somewhat of an outlier." Mann was not involved in the new research.

"The estimate of earth system sensitivity (9C for CO2 doubling) is so much higher than the prevailing estimates (5-6C) that one has to consider it somewhat of an outlier, and treat it with an appropriate level of skepticism," he told Mashablein an email.

One major problem with the study, Mann said, is that the sensitivity estimate is dominated by glacial and interglacial cycles during the past 800,000 years, and it's tough to untangle the roles played by carbon dioxide in such variations.

This is because carbon dioxide both causes and responds to temperature changes that are driven by other factors, such as variations in Earth's orbit around the sun.

"It is unclear that an estimate of the relationship between global temperature and carbon dioxide under those circumstances is an appropriate measure of the response of temperature when carbon dioxide alone is the major driving force, as it true today," Mann said.

"So I regard the study as provocative and interesting, but the quantitative findings must be viewed rather skeptically until the analysis has been thoroughly vetted by the scientific community."

Gavin Schmidt, a climate scientist who directs NASA's Goddard Institute for Space Studies in New York, was more blunt in his views on the new publication.

"The temperature reconstruction is great, but the claims about sensitivity are just wrong," Schmidt, who was not involved in the new research, said in an email. "This is not an argument about methods or what to present in public or whether you like models or observations, it is just wrong."