Errors have been found in a recent study suggesting the oceans were soaking up more heat than previously estimated.
The initial report suggested that the seas have absorbed 60% more than previously thought.
But a re-examination by a mathematician showed that the margin of error was larger than in the published study.
The authors have acknowledged the problem and have submitted a correction to the journal.
What had the researchers initially claimed?
According to the last major assessment by the Intergovernmental Panel on Climate Change (IPCC), the world's oceans have taken up more than 90% of the excess heat trapped by greenhouse gases.
But the new study claimed that every year, for the past 25 years, we have put about 150 times the amount of energy used to generate electricity globally into the seas - 60% more than previous estimates.
The authors suggested that this had huge implications for the way that the world is warming.
If far more heat than previously believed was going into oceans, it also meant that far more heat than we thought has been generated by the warming gases we have emitted in the world to date.
Therefore more heat from the same amount of gas means the Earth is more sensitive to CO2.
What had these scientists done differently?
Since 2007, scientists have been able to rely on a system of almost 4,000 Argo floats that record temperature and salinity in the oceans around the world.
But prior to this, the methods used to measure the heat in the ocean had many flaws and uncertainties.
The team involved in the study used a new approach to measuring the amount of oxygen and carbon dioxide in the air. They believed that this allowed them to accurately measure ocean temperatures globally, dating back to 1991, when data from a global network of stations became available.
The key element in this methodology is the fact that, as waters get warmer, they release more carbon dioxide and oxygen into the air.
"When the ocean warms, the amount of these gases that the ocean is able to hold goes down," said lead author Dr Laure Resplandy from Princeton University.
"So what we measured was the amount lost by the oceans, and then we can calculate how much warming we need to explain that change in gases."
So how were the errors detected?
They were discovered by a mathematician, Nicholas Lewis, who has long been critical of climate models, arguing that they are over-sensitive to emissions of carbon dioxide.
He reviewed the paper when it was published and found what he termed "serious (but surely inadvertent) errors in the underlying calculations."
The key element deals with the amount of uncertainty in the measurements. As a result of the increased margin of error, the findings on the amount of heat being absorbed by the oceans can't be definitively supported.
How have the authors reacted?
They've put their hands up and acknowledged the error.
Ralph Keeling, who co-authored the study said in a statement that he takes responsibility for the "oversights" and would like to "thank Nicholas Lewis for bringing an apparent anomaly in the trend calculation to our attention."
"These problems do not invalidate the methodology or the new insights into ocean biogeochemistry on which it is based, but they do influence the mean rate of warming we infer, and more importantly, the uncertainties of that calculation."
The authors say they have now reworked their calculations and submitted a correction to the journal, Nature, which originally published the study.
What have the publishers of the report said?
They have issued an editor's note about the issue.
"We would like to alert readers that the authors have informed us of errors in the paper. An implication of the errors is that the uncertainties in ocean heat content are substantially underestimated. We are working with the authors to establish the quantitative impact of the errors on the published results, at which point in time we will provide a further update."
What does all this mean for ocean warming?
According to the scientists involved in the original study, the ocean is still probably warming more than the estimate used by the Intergovernmental Panel on Climate Change. But the errors mean that it has a larger range of probability - between 10% and 70%. This is in line with other studies.
"Our error margins are too big now to really weigh in on the precise amount of warming that's going on in the ocean," Ralph Keeling said. "We really muffed the error margins."
UPDATE: The original version of this story was published before the errors in the study came to light.