NWC REU 2016
May 23 - July 29



Photo of author

Arctic Weather and Abrupt Sea Ice Loss

Uriel Gutierrez, Steven Cavallo, and Nicholas Szapirol


What is already known:

  • The long term climate trend for Arctic sea ice is well established with the ice-albedo effect as a main driver.
  • Wind patterns from inter-seasonal oscillations (such as Arctic oscillation) have effects on sea ice motion and extent.
  • Abrupt sea ice loss events have been observed on time scales of days and coincide with surface cyclones.
  • A better understanding of year-to-year variability is important for improving future predictions of sea ice.

What this study adds:

  • Oscillations in change of sea ice extent (1979-2014) at synoptic time scales were shown to be statistically significant.
  • Synoptic time scale reductions in sea ice extent occur most frequently in July and December.
  • Composite of top 1% of abrupt loss in sea ice extent events revealed strong winds over loss area.
  • These conditions always occurred with a nearby surface cyclone; enhancement from anticyclones could sometimes also occur.


September Arctic sea ice extent is decreasing rapidly, especially over the past few decades. While the mechanisms contributing to this climate trend are relatively-well understood, the year-to-year variability is not. This study examines 2-d decreases in summer sea ice extent to quantify the year-to-year variability that is due to synoptic time-scale processes and isolate its possible source. It is hypothesized that the abrupt reductions in sea ice are a consequence of synoptic–scale cyclones, and in particular the anomalously strong surface winds over the periphery of the cyclones from a strong pressure gradient.


A spectral analysis of two-day changes in sea ice extent is performed to determine whether events at synoptic time-scales have significant contributions to sea ice loss with respect to red noise. Several significant periods are found at synoptic time-scales, at 5, 6, 8, 10, and 16 days. A Butterworth filter is then applied to high-pass periods shorter than 18 days to isolate the abrupt sea ice loss events corresponding to these high frequencies and compile a set of significant events. Defining the top 1% of the high-pass filtered two-day decrease in sea ice extent, there is found to be two annual maxima: July and December, and only summer cases (June-August) are retained for the present study. Composite sea level pressure of the 25 cases reveals the presence of a 998 hPa mean surface cyclone, which varies in strength from 999 to 978 hPa. While there is always a cyclone, there is often, but not always, a nearby anticyclone that can further enhance the pressure gradient over the sea ice loss region.

Full Paper [PDF]