Antarctic sea ice variability and trends, 1979–2010

The planet is warming. Arctic sea ice is collapsing, down dramatically over recent decades. This trend is consistent with every expectation from climate physics. And yet, at the other end of the Earth, sea ice has been growing. Not shrinking, but growing. This is not a measurement error. It is a real signal in the data, and it has held for more than thirty years. Understanding it requires looking past the single global number and reading the regional story hidden within it. That's exactly what Claire Parkinson and Don Cavalieri did. Their 2012 paper analyzes satellite passive microwave data from November nineteen seventy-eight through December two thousand ten, a thirty-two-year record, to track how Antarctic sea ice has changed across the entire Southern Ocean and within each of its major regional sectors. Passive microwave instruments are uniquely suited to this task: sea ice and open water emit differently at microwave wavelengths, so the satellites can distinguish them through clouds and polar darkness, day or night, year-round. Parkinson and Cavalieri drew on three successive instruments: the Scanning Multichannel Microwave Radiometer, known as SMMR, on Nimbus-7 from nineteen seventy-eight to nineteen eighty-seven, the Special Sensor Microwave Imager on several DMSP satellites through two thousand seven, and the upgraded Special Sensor Microwave Imager Sounder from two thousand eight onward to build a continuous, self-consistent record. From those raw brightness temperatures, the NASA Team algorithm calculates ice concentration: the fraction of each twenty-five by twenty-five kilometer grid cell covered by ice. Two summary metrics follow. Ice extent is the summed area of all grid cells where concentration reaches at least fifteen percent—any cell above that threshold gets counted in full. Ice area weights by concentration: you multiply each cell's area by its ice fraction and sum across cells. The distinction matters. When area and extent trends diverge, the gap tells you something about what's happening to the density of the pack itself, not just its spread. Over the full thirty-two-year record, the hemispheric verdict is clear: Antarctic sea ice extent increased at a rate of seventeen thousand one hundred plus or minus two thousand three hundred square kilometers per year. That trend is positive in every single month of the year, from a low of nine thousand one hundred plus or minus six thousand three hundred square kilometers per year in February to a peak of twenty-four thousand seven hundred plus or minus ten thousand square kilometers per year in May. The hemispheric signal never reverses sign. But the hemispheric number obscures a story of violent regional contrasts, and that story is the real finding. Parkinson and Cavalieri divide the Southern Ocean into five sectors: the Weddell Sea, the Indian Ocean, the western Pacific, the Ross Sea, and the combined Bellingshausen and Amundsen Seas. Picture the geography. The Ross Sea sits beneath the Pacific face of Antarctica, spanning roughly one hundred sixty degrees East to one hundred thirty degrees West. Immediately to its east, curving toward the Antarctic Peninsula, are the Bellingshausen and Amundsen Seas—two neighboring sectors. And they are moving in exactly opposite directions. The Ross Sea drove nearly all of the hemispheric increase, posting a trend of thirteen thousand seven hundred plus or minus one thousand five hundred square kilometers per year—statistically significant at the ninety-nine percent confidence level. The Bellingshausen and Amundsen Seas, sitting right next door, shed ice at eight thousand two hundred plus or minus one thousand two hundred square kilometers per year—also significant at ninety-nine percent, but in the opposite direction, behaving more like the Arctic than like the rest of Antarctica. The Indian Ocean contributed a modest positive trend, and the Weddell Sea was also positive. The western Pacific was more mixed. But the dominant drama is the Ross-Bellingshausen seesaw. The Ross Sea shows positive trends of at least seven thousand square kilometers per year in every single month. The Bellingshausen and Amundsen Seas are negative in every single month, with the largest seasonal loss—fourteen thousand three hundred plus or minus two thousand six hundred square kilometers per year—occurring in summer. The comparison between ice area and ice extent adds another layer of meaning to this picture. Parkinson and Cavalieri find that in every region, the ice area trend matches the sign of the ice extent trend—but the magnitudes differ, and those differences are informative. In the western Pacific, for example, ice area grew faster than ice extent on yearly, summer, and autumn timescales. That implies the pack there didn't just spread; it got denser. Concentration increased within cells that were already counted as ice-covered. Where extent grows faster than area, the opposite occurs: the pack is reaching farther but becoming more diffuse. Two numbers measure the same phenomenon, and the gap between them is the signal. So what's driving all of this? Parkinson and Cavalieri are frank: the causes of Antarctic sea ice changes are not yet fully understood. But the spatial pattern is a powerful clue. A pronounced seesaw between two adjacent sectors—one gaining, one losing—looks like a circulation fingerprint, not a uniform forcing. The paper points to several candidate mechanisms. One involves large-scale atmospheric variability: the El Niño Southern Oscillation, or ENSO, and the Southern Annular Mode, the dominant mode of atmospheric variability across the Southern Hemisphere, roughly analogous to the Arctic Oscillation in the north. Rind and colleagues had suggested that El Niño years tend to reduce ice in the Pacific sector while La Niña does the reverse. Stammerjohn and colleagues, analyzing from nineteen seventy-nine to two thousand four, found the Bellingshausen-Ross contrast becomes especially sharp when ENSO and the Southern Annular Mode interact—when an El Niño coincides with a negative index, or La Niña with a positive one. A second set of hypotheses ties the pattern to stratospheric ozone depletion. Thompson and Solomon linked warming of the Antarctic Peninsula and relative cooling elsewhere to a strengthening of circumpolar westerly winds, itself connected to the long-term trend in the polar vortex that ozone depletion produced. Turner and colleagues proposed a more specific mechanism: increased cyclonic flow over the Amundsen Sea, which would funnel warm air northward over the Bellingshausen Sea—driving ice loss—while simultaneously pushing cold air southward over the Ross Sea, building ice. The seesaw as a consequence of a single atmospheric circulation anomaly. It's a compelling picture. But Parkinson and Cavalieri don't endorse any single explanation. Sigmond and Fyfe's climate model simulations suggested the ozone hole would actually cause Antarctic sea ice to decrease—which means the observed increase must have other causes, or the models are missing something. Natural variability and forced change are genuinely hard to separate in a thirty-two-year record. Antarctica remains the hardest place on Earth to instrument, the satellite era is still young relative to the climate timescales in play, and the answer likely involves multiple interacting mechanisms. What the thirty-two-year record unambiguously provides is the baseline. Antarctic sea ice modulates exchanges between the ocean and atmosphere, reflects solar radiation back to space, and structures the marine ecosystem from krill upward. Deep water formation in the Southern Ocean—which helps drive global ocean circulation—depends in part on the density and salinity changes that sea ice growth and melt produce. Changes to the ice mean changes to all of that. The Parkinson and Cavalieri dataset—continuous, multi-satellite, gridded at twenty-five kilometers, spanning more than three decades—is the reference against which every subsequent shift in Antarctic sea ice will be measured. And subsequent shifts have been dramatic. After this paper's publication, Antarctic sea ice hit record high extents between two thousand twelve and two thousand fourteen, then collapsed to record lows after two thousand sixteen in a reversal that stunned researchers. Interpreting those extremes required exactly the kind of baseline this work established. The paper closes with a simple, pragmatic argument: the observational record must keep growing, and models and data analyses must mature together, if we are ever to cleanly separate what is natural from what is forced. The thirty-two-year record is not the end of the question. It is the foundation on which the answer has to be built. This lecture was created by ennepō. Go to https://ennepo.ai to Discover, Create and Follow the latest research in your field. Read when you can. Listen when you want to.