Landfast Sea Ice
Examining Landfast Sea Ice on Alaska’s Northern Coast
by Andy Mahoney, Hajo Eicken and Lew Shapiro, UAF Geophysical Institute; and Allison Graves, Nuna Technologies
Landfast sea ice is the largely stationary ice attached to the coasts of ice-covered seas. It is important to arctic coastal communities, which use it as an extension of the land for travel and to hunt marine mammals associated with the landfast ice and the adjacent system of leads.
Its presence reduces the effects of coastal erosion, and nearshore oil and gas development rely on landfast sea ice as a stable platform. Along the northern coast of Alaska, landfast sea ice is a seasonal phenomenon.
Beginning in late autumn, the landfast ice undergoes a gradual increase in area toward a stable extent before it experiences a rapid break-up and retreat in the following spring.
Punctuating this more gradual response to the unfurling of the seasons are episodic events, such as break-offs and stable extensions, which can have profound and lasting effects on the landfast ice and its denizens. Synthetic aperture radar provides imagery independent of weather and daylight with sufficient frequency to make it ideal for studying such events.
More than 950 medium-resolution, calibrated GeoTIFFs of RADARSAT-1 ScanSAR data were used to generate collocated mosaics of the study area approximately every 10 days for the 8 ice seasons between 1996 and 2004. From the mosaics, we have delineated the seaward landfast ice edge (SLIE) 222 times. Th is has yeilded approximately 29 SLIEs per year between October and July as the landfast ice undergoes its annual cycle.
We apply a rigorous defiition of landfast ice in this study, consisting of two criteria determined from remote sensing data: 1) that the ice is contiguous with the coast, and 2) that the ice lacks identifiable motion in three consecutive mosaics, or approximately 20 days.
Initial analysis revealed the asymmetric annual development of landfast ice, as defined above, with great inter-annual variability. Short-period episodic events such as stable extensions and a number of ice break-outs were also captured. The consequences of break-out events are of deep concern to those who use the landfast ice for travel, hunting or development.
Further, stable extensions of ice are found persisting for over a month up to 250km from shore in water over 200m deep. Th ese extensions do not occur every year and the mechanisms responsible for their occurrence are not clear. However, the oceanographic and ecological implications of such a broad persistent barrier between the ocean and atmosphere warrant further investigation.
The data also revealed significant inter-annual variability in the maximum landfast ice extent according to our definition, but well inside this there is a zone in which the SLIE is more frequently located.
Within this zone the SLIEs converge at nodes where the variability throughout the year is less and the SLIE is more stable. These nodes lie along or close to the 20m isobath, which suggests they correspond to the locations of grounded ridges.
The figure to the left shows the landfast ice area associated with each SLIE stacked on top of each other to create maps indicating the number of occasions that landfast ice occupied any point in the study area for a given year. Th is can be viewed as the probability of finding landfast ice at any point between October and July. Each annual SLIE is shown in black, appearing as a probability contour. Stable nodes show up as regions of strong radients in probability where SLIEs converge.
A comparison of our results with data compiled at the Geophysical Institute in the 1970s indicates that the SLIE has occupied a more shoreward position in recent years.
Further investigation is required to determine whether this reflects a change in the landfast ice regime or is a result of a different method of delineating landfast ice requires further investigation.
Upon completion of this project, all the derived products such as SLIE positions and mosaics will be publicly available.