Mapping Arctic sea ice in a polar projection

Arctic sea ice changes throughout the year, from winter maximums to summer minimums, but is generally shrinking. 2015 marked the lowest maximum and fourth lowest minimum since consistent satellite measurement began in 1978. Using open data from the National Snow and Ice Data Center and command-line tools, you can visualize historical sea ice in a polar projection.

This map shows ice extents in September, the typical summer minimum. The blue line gives total area, and the orange line tracks average global land-ocean temperature anomalies relative to a 1951-1980 baseline. Use the time slider in the lower left to compare changes from 1978-2014.

View full-screen map

Arctic sea ice coverage is both a symptom and a cause of broader climatological trends. 2015’s record low maximum corresponds with several years of unusual jet stream patterns and a general trend of rising temperatures in the Arctic and globally. Sea ice retreat also contributes to warming, as exposed seawater absorbs more solar radiation than higher albedo ice.

Sea ice is frozen seawater. Freezing begins at -1.8 degrees Celsius and progresses from slushy grease ice to fragile nilas sheets to first-year and then thicker multi-year ice. Sea ice is categorized as fast ice (attached to land) or drift ice (drifting with winds and currents). It’s different from ice shelves and icebergs, which form from freshwater glaciers.

Aerial photograph from a NASA P-3B Orion flying at 450 m altitude, showing darker first-year and whiter multi-year ice near Greenland. Photo: NASA.

Satellite-based passive microwave radiometers like the Special Sensor Microwave Imager/Sounders on Defense Meteorological Satellites have monitored sea ice continuously since 1978. These sensors detect the microwave signature of sea ice, day or night and through clouds, in 25 km2 grid cells. Sea ice extent is defined as all grid cells containing at least 15 percent ice.

To build the map we converted ice extent shapefiles in EPSG:3995 (a polar stereographic projection) to GeoJSON with gdal_translate. For visual context, we used Natural Earth shorelines, boundaries, and labels as well as Arctic shipping route polylines reprojected with ogr2ogr -t_srs EPSG:3995. Then we assigned a Web Mercator projection to the layers using ogr2ogr -a_srs EPSG:3857 to prevent automatic reprojection when uploaded to Mapbox. We styled the data in Mapbox Studio, and Mapbox GL keeps labels properly oriented when the map is rotated (hold ctrl while panning).

Check out @aj_ashton’s alternate projection code on Github, and sign up for the private beta of Mapbox Studio. Find me on Twitter @mtirwin or ping @chieflybrit if you have questions about the Arctic map or Mapbox tools.

Header photo shows researchers on the NASA-funded ICESCAPE mission measuring sea ice melt ponds. Photo: NASA.