1. Examples of Lead Features
Characteristic lead pattern progressions in differ ent regions of the study area
2. Examples of Seaward Landfast Ice Edge
SLIEs with and without a flaw lead, a mid-winter b reakout, and an anomalous stable extension
Chukchi Lead Progression | February 13 -20, 1994
For the 6 years of imagery that have been examined, we identified 11 sequences which began when a single lead (or zone of leads) appeared along the Chukchi Sea coast and extended northeastward past Point Barrow and reached into the Beaufort Sea as far as about 74N. New leads may then open parallel or sub-pa rallel to the initial lead. The pack ice west of the lead (system) moves to the west or southwest, and there may be some movement of the ice east of the lead in the same sense. However, the displacements in that area are smaller.
In 6 of the 11 examples, the next step in the sequence was the format ion of a series of curving leads that were tangent to the fast ice edge east of Point Barrow, and concave to the west. Then, new leads opened nearly normal to t he coast near Barter Island or further east and extended northward to 74N or fur ther. Subsequently, more leads with the same or similar orientation formed furth er east of Barter Island, as well as within the near shore pack ice, as the pack drifted west. The time required for this process was variable, and at most a fe w days.
The same sequence was followed in 2 other examples, but in those, the leads normal to the coast at Barter Island and eastward did not form. A possibl e explanation is suggested by the observation that that during those sequences, a wide area of open water or thin ice separated the pack ice from the coast betw een Mackenzie Bay and Cape Bathurst so the pack ice possibly was not in firm con tact with the fast ice as was apparently the case in the other examples. Thus, t he mechanical boundary conditions were different and may not have supported the development of the high angle leads in these two examples.
In 2 of the remaining 3 sequences, the process did not progress past forming a long, arcuate lead that replaced the initial, straighter lead system a nd then was stable and did not change. In the last sequence, new leads did form east of the initial lead system, but the pattern was different from that describ ed above as noted in alternative (2) above. In the last example, new leads did f orm to the east of the initial lead but the pattern was different from those des cribed above.
The six AVHRR images shown above illustrate the complete process from the initial lead formation along the Chukchi Sea coast to the west drift of the pack ice in the near shore Beaufort Sea after opening the leads east of Barter Island. The first two images predate the start of the process, but are included to show that for several days before the lead formed along the Chukchi Sea coast , the pack ice in both the Chukchi and Beaufort Seas was moving eastward and com pressing against the coast east of Mackenzie Bay. In the days following the last image shown here the pack ice continued to drift to the west but there was sign ificant cloud cover, so the images are not included here. Note that the time bet ween formation of the various leads that make up the sequences was variable and generally longer than was the case in the example shown here.
Point Barrow - Banks Island Lead Progression | Janua ry 14-23, 1998 | animation
Numerous definitions of landfast ice exist in the literature (e.g. Weaver 1951, Barry et al. 1979, Stringer et al. 1978). These are described in a separate document (Definition of Landfast Sea Ice) that also det ails our definition, which can be summarized by the following criteria:
- the ice is contiguous to the coast
- the backscatter signature remains constant over a period of 20 d ays
According to these criteria, the SLIE cannot be identified from a s ingle image since there needs to be some way of determining that the ice is stat ionary. The technique we employ to delineate the SLIE uses 3 consecutive Radarsa t mosaics, approximately 10 days apart (see Method for Delineating Seaward Landfast Ice Edge ). However, there are occasions when the SLIE corresponds to an ice edge that can be identified in a single Radarsat mosaic. Figure 1 shows the SLIE coincidin g with a coastal flaw lead along almost its entire length during April 2003. Thi s appears to be quite common at this time of year as the pack ice in the Beaufor t Sea begins to loosen up and withdraw slightly from the coast.
Figure 1: The SLIE determined from mosaics between 3 and 27 April 2 003 overlain on the mosaic produced from scenes between 12 and 17 April 2003. Se e text for description.
In the middle of winter however, several weeks commonly go by witho ut any sign of open water in the Beaufort Sea, when the SLIE does not correspond to an edge in any of the three mosaics from which it was determined. Figure 2 s hows a typical mosaic from early March. No open water was visible for several we eks either side of this time and the SLIE seems to coincide with regions of high backscatter, which most likely correspond to ridges and fields of deformed ice. Further work will be required to establish the fraction of time when there is o pen water immediately offshore of the SLIE.
Figure 2: The SLIE determined from mosaics between 23 February and 19 March 2003 overlain on the mosaic produced from scenes between 7 and 9 March 2003. See text for description.
The location of the SLIE varies with the seasons and is apparently governed in part by the nearshore bathymetry, since long-term stability of the ice depends upon the position of grounded ridges. However, even in mid-winter th e landfast ice can become unstable and break away from the coast. Figure 3 shows a break-out of the landfast ice east of Point Barrow captured in a mosaic of sc enes between February 6 and 9 2001. In order to understand what caused this unse asonal event, it will be necessary to examine the prior development of the landf ast ice as well as available meteorological and sea ice motion data. It is appar ent, however, that this event led to instability of subsequent landfast ice in t his area, which was observed to break out again between March 20 and 22. The sam e area broke out for a third and final time at the end of May, which is more usu al.
Figure 3: A mosaic of Radarsat scenes from between February 6 and 9 2001 capturing the break-out of a 150 km section of landfast ice east of Point Barrow. In this mosaic, the section remains intact and measures 70 km at it wide st point.
In contrast to a break-out, there are other occasions where the landfast ice, according to our definition, ext ends far beyond its normal position. Figure 4 shows a highly unusual edge extending beyond the study area boundaries. This example raises two important issues regarding the delineation of landfast ice from remote sensing data. The first issue surrounds our definition, which does not take account of any real geophysical processes. Although the edge identified in Figure 4 persisted for the specified period of time, it is hard to conceive that the ice was held fast by grounded ridges. Instead it appears that the pack ice came to rest over a large area for a long period of time. Stringer (1980) observed such stable extensions 30 or 40 km beyond the 20 m isobath. His suggestion is that they are caused by "an absence of sufficient winds, currrents and internal forces within the ice sheet to keep individual pans within the pack ice from freezing together'. The occurrence, extent and location of these stable extensions will be a topic for f urther study, but in accordance with our definition the SLIE has been defined ac cording to where a lack of ice motion was observed.
The second issue is discussed in the Section 1 of Method for Delineating Seaward Landfast Ice Edge and arises from the fact that each mosaic represents a number of snapshots of the study area over a p eriod of a few days. Therefore the mosaicking edges between the parent scenes, w hich can be clearly seen in all the figures in this section, represent temporal boundaries within the data. Thus, ice movement between two mosaics may not be ca ptured by all the parent scenes and on such occasions, it may not be possible to delineate the SLIE continuously across mosaicking edges.
Besides this technical aspect of mosaicking boundaries, there is a lso the practical aspect of trying to distinguish ice motion by flickering conse cutive mosaics. Geolocation errors are typically distinguished from ice motion b y differential motion, since geolocation error will cause an entire scene to mov e by the same amount, whereas ice motion will exclude the land and landfast ice. However, with each parent scene potentially having a different geolocation erro r it is possible to see differential movement of ice along a mosaicking boundary , making it confusing to the eye and harder to distinguish real ice motion from error. In these instances, it comes down to the skill of the user to identify th e location of the landfast ice, which obviously introduces an additional degree of subjectivity into the result.
Figure 4: A mosaic of Radarsat scenes acquired between March 16 and 18 2000 displaying SLIE extending beyond the boundaries of the study area. In t he center of the mosaic the SLIE coincides with a lead separating moving sea ice from ice that had remained stationary since February 23. To the east of this, t he SLIE coincides with a line along which differential movement of ice was obser ved with little or no open water.
Jump to top span>