The steady loss of sea ice around the Arctic over the past few decades is expected to have a strong impact on the climate. Recent studies have linked the loss of ice to a cooling over mid-latitudes in the northern hemisphere – a counterintuitive result, since the lost ice should itself have had a reflective, cooling quality. Scientists think that an initial warming caused by the ice loss might have weakened the sub-polar jet stream, which would have eased the flow of cold Arctic air southwards. In any case, the message seems clear: there is a complex interaction between sea ice and the climate, which scientists need to understand.
Models that explore the relationship between ice and climate have two ways of representing the extent of ice. One is to use a monthly average of ice values, calculated from several years' past data. In this so-called climatological sea-ice condition, the sea-ice value for, say, January will always be the same, regardless of the year. The other way, which is more data-intensive, uses actual sea-ice values for every month in every year; this is the so-called realistic sea-ice condition.
In the past, scientists have performed short-term analyses to determine how much better the realistic sea-ice condition is than the climatological sea-ice condition. Now, however, Yoo-Geun Ham of NASA's Goddard Space Flight Centre in the US, and colleagues from the University of Gothenburg in Sweden and the Korea Ocean Research and Development Institute, have performed a thorough, long-term analysis over 24 years.
Ham and colleagues used an atmospheric general circulation model to see how simulations of Eurasian climate from 1988 to 2011 based on realistic and climatological sea-ice conditions compared. They used a correlation method to determine the similarity in atmospheric simulations, and then a composite and regression method to determine the difference in simulated atmospheres when sea-ice values were known to be especially high or low.
The researchers found that the realistic sea-ice condition reproduced Eurasian winter temperature patterns better than the climatological sea-ice condition, and in particular captured the observed link between temperature anomalies across Eurasia and sea-ice concentrations in the Barents and Kara seas north of Russia. "The realistic sea-ice is one of the crucial factors to realistically simulate mid-latitude winter climate," the researchers said.
Currently, explained Ham, most simulations using atmospheric global climate models already use realistic sea-ice conditions. But this is not true of all simulations, such as those in ocean-atmosphere coupled global climate models, which predict their own sea-ice variability. "In that case, the simulation of the sea-ice is not satisfactory," he said. "This might cause some problems [in simulating] mid-latitude climate."
The researchers now plan to investigate how some of the record-breaking sea-ice retreats observed in recent years have influenced the climate.
The study is published in Environmental Research Letters.

Related links