Simulation vs Reality: Researchers Modify Climate Model to Explain Expanding Antarctic Sea Ice

Scientists at Scripps Institution of Oceanography at UC San Diego and the California Institute of Technology are one step closer to understanding an enigma in climate change science: why do climate models predict Antarctic sea ice retreating although observations show it expanding?

Between 1979 and 2015, sea ice in the Southern Ocean steadily expanded – a head-scratching phenomenon climate scientists are trying to understand as models suggest it should be retreating in a warming world. Scripps oceanographer Ian Eisenman and California Institute of Technology postdoctoral researcher Shantong Sun wanted to understand why the models are behaving so differently from satellite observations. Their findings were published February 16 in the journal Nature Communications.

“It's vital to understand the processes that lead to these discrepancies between models and observations,” said Sun, who completed this research while he was a PhD student at Scripps.  “A number of explanations have been proposed previously, but the enigma really remains unresolved. So, in this study, we investigated the hypothesis that this discrepancy occurs due to simulation biases in how fast the sea ice is moving.” 

In order to understand the discrepancies, the researchers used a global climate model that relies on the observed rise in atmospheric carbon dioxide – and factors including other greenhouse gases – over recent decades and ran the simulation, with the output showing the commonly-cited retreat in Antarctic sea ice as seen in other current climate models. These models include simulated sea ice motion, which is based on simulated surface winds, simulated ocean currents, and the tendency of sea ice floes to resist being compressed or otherwise deformed. They then ran the model again, but this time instead of using simulated sea ice movement they used ice motion based on satellite observations, with the result showing sea ice expansion, similar to what scientists have been seeing in observations during the past few decades. 

“I was excited by the results of these simulations, which rather cleanly lay the blame for this enigma on problems with the simulated sea ice drift,” said Eisenman. “We’re still working though to understand exactly what it is about the observed ice drift that caused the ice cover to expand. The relationship between changes in sea ice drift and changes in sea ice area extent is not well understood.”

Climate models simulate processes in the oceans and atmosphere with specified atmospheric carbon dioxide levels and other factors, to characterize how Earth systems may interact and change in the future. Relying on complex mathematical equations based on physical principles, they allow climate scientists to test theories and project how a warmer world may operate under various conditions. Due to limitations of model resolutions in both space and time, processes that occur on smaller scales cannot be explicitly simulated and their effects are usually approximated in climate models. These approximations could lead to biases in climate model simulations.

The results of the study show that the poor validation results from climate models lie in the simulated sea ice motion, putting a finer point on one of the issues with these models. Whereas some have speculated that the problem could be related to meltwater flux from Antarctic glacial discharge or ocean heat uptake, Sun and Eisenman argue that more accurate simulation of sea ice movement in the climate models could bridge the gap between models and observations. 

This study was funded by the National Science Foundation.