Colliding ice floes explain Arctic sea ice's mysterious movement patterns

Arctic sea ice does not move as a single sheet. Instead, it breaks into separate pieces called floes—some just meters across, others stretching several kilometers—that drift independently across the ocean, pushed by wind and bumped constantly by their neighbors. For decades, scientists have watched these floes move in ways that simple wind-based models could not explain. The ice traveled faster than predicted in some cases, slower in others. It spread across the ocean more gradually than the equations suggested it should. Researchers proposed various culprits: unusual wind patterns, ocean eddies, cracks in the ice. But a new study published in Physical Review Letters points to something far simpler: the floes are colliding with each other, and those collisions are the key to understanding how the ice actually behaves.

Bryan Shaddy, now at the University of Southern California but formerly an undergraduate at UC Riverside, led the research alongside materials scientist Alex Greaney and mechanical engineer Bhargav Rallabandi, both at UC Riverside. Their insight was to treat the problem like a physics problem that extends far beyond the Arctic. Imagine grains of sand moving through a silo, pushed by an external force. Now imagine those grains are floating on water, being shoved by turbulent winds, and you have the basic picture. The researchers built a computer model that included wind, ocean drag, and the repeated collisions between neighboring floes. When they fed the model real wind and ice conditions from the Fram Strait—the passage between Greenland and Svalbard where enormous volumes of Arctic ice drift toward the Atlantic—something remarkable happened. Using just one additional parameter that had minimal effect on the results, the simulation reproduced three observations that had long puzzled scientists: the actual rate at which sea ice spreads, the range of speeds at which individual floes move, and the way ice motion changes over timescales from hours to days.

The reason collisions matter so much comes down to density. When Arctic sea ice becomes tightly packed, individual floes collide with their neighbors far more frequently than the wind itself can accelerate them. Each collision transfers energy from one piece of ice to another and slows the overall spread. This repeated jostling explains why the ice field does not simply keep accelerating in response to wind—the collisions act as a brake, removing energy that would otherwise drive the floes farther apart. Rallabandi described it plainly: if you gather enough ice floes in one place and add wind, they bump into each other and transfer energy to their neighbors. That single process, he said, is all you need to explain the observations that had previously seemed mysterious.

The implications for climate science are significant. As the Arctic continues to warm, the amount of ice covering the ocean will shrink, and the size of individual floes may change. Both of these factors influence how often collisions occur, which in turn affects how rapidly the ice field can spread. The model does not yet predict exactly where Arctic sea ice will end up as the climate shifts, but it opens a path toward answering questions that matter: whether changing ice conditions might allow floes to spread apart more easily and drift into warmer waters, where they would melt faster. For climate modelers, the work offers something equally valuable. Global climate models cannot track every individual floe across the Arctic—there are simply too many. Instead, scientists need ways to represent the combined behavior of countless small pieces. A physics-based model of how floes interact could capture processes that are otherwise too small to simulate individually, making the models more accurate without requiring impossible computational power.

What struck Rallabandi most was how a relatively simple physical process could explain the complicated motion observed in the real world. The principle is not limited to sea ice. Any system involving many objects that collide while being pushed by an unpredictable force could behave the same way. The researchers point to avalanches, landslides, materials science, and even particle-filled inks used in three-dimensional printing as potential applications. The model needs only a noisy source of force and objects that experience collisions. As Arctic conditions continue to shift, this framework may help scientists understand not just how the ice moves, but where it goes and what happens when it gets there.