Dates of sea ice movement and sea ice distance in McMurdo Sound, Antarctica from MODIS and SSMI imagery between 1978-2015 (McMurdo Predator Prey project)

Abstract

Fast/sea ice movement was quantified from visible-wavelength images from the Moderate-resolution Imaging Spectroradiometer (MODIS) aboard the Aqua and Terra satellites (250 m resolution; processing occurred for 2002/03-2014/15 seasons and Terra satellite date from 2000-2002 were not used) and sea ice concentration derived from the Scanning Multichannel Microwave Radiometer- and Special Sensor Microwave Imager-family passive microwave sensors (SSM/I; 25 km resolution; 1978/79-2014/15). MODIS data were acquired in one of two ways, from either processing of Level 1 swath data into “true color” images using SeaDAS software v. 6.4 (2002-2012), or from the Corrected Reflectance (True Color) layers of the NASA Worldview website (http://worldview.earthdata.nasa.gov/; 2012-2015). Fast ice areas were generated manually from clear-sky images by drawing polygons in GIS software; pack ice was excluded from analysis. The fast ice in MODIS images was sometimes obscured by clouds, so for days with missing imagery we interpolated linearly between valid data. From the MODIS imagery, we also measured the direct linear distance between McMurdo Station and the nearest open water. For SSM/I, daily or bi-daily fractional sea ice cover was extracted from data available at the National Snow and Ice Data Center (NSIDC). SSM/I ice concentration was retrieved from the NSDIC web site and ftp site (http://nsidc.org/data/seaice/). To minimize the biases inherent to the different data processing algorithms and in order to reduce the daily variability introduced by the movement of pack ice, we took the maximum of either the Bootstrap or NASATEAM processed values (Comiso, 2000; Cavalieri and others, 2015), and then used a 5-day median filter to smooth changes in sea ice concentration. To further compensate for short-term oscillations we masked ice concentrations greater than 80% when extracting the dates of changes in sea ice cover. For detecting the timing of sea ice changes, sea ice concentrations below 15% were excluded from our analysis, following the methods of Comiso and Steffen (2001).> To simplify discussion in the following, we use the inclusive term “fast/sea ice” to refer to fast ice as determined by MODIS and sea ice as determined by SSM/I. Fast/sea ice area was plotted over time, and the following sequential pattern of fast/sea ice events is identified: (1) initial fast/sea ice retreat from winter maximum; (2) final rapid fast/sea ice retreat to minimum extent; (3) fast/sea ice cover minimum in the entire McMurdo Sound; and (4) fast/sea ice advance. From the MODIS data, we additionally determined (5) fast ice cover minimum on the west side of the Sound; and (6) fast ice cover minimum on the east side of the Sound. For a complete list of measurements, refer to the supplemental document 'Field_names.pdf', and a full dataset description is included in the supplemental file 'Dataset_description.pdf'. The most current version of this dataset is available at: http://www.bco-dmo.org/dataset/674819