free, A point of interest here is that within 2 hours after this ice 
was observed at 1000 LST by helicopter the areal distribution changed 
very rapidly to that indicated in figure 9. This change occurred with 
the approach of high tide and 10-knot surface winds from the east and 
southeast. Average tide range during this period (9 to 10 October) 
was about 6 feet. 
On 10 October, westward movement of the ice was noticed at ap- 
proximately 1000 LST. Easterly winds at this time had increased to 
19 knots with gusts to 25 knots. At 1100 LST, aerial reconnaissance 
made possible a detailed synoptic analysis of the ice as shown in 
figure 10. Numerous cracks, leads, and polynyas had developed. Con=- 
siderable amounts of slush ice had formed in the newly developed water 
areaSe Ice thickness on this date measured 5.0 inches. By 1300 LST, 
the wind speed was 32 knots with gusts to O mots. By 1600 LST, the 
ice picture was radically different. The ice had moved out of the local 
harber area, except along the east shore, as shown in figure 11, Com- 
paratively strong easterly winds continued throughout the next day. 
During the period from 1 to 12 October, the USS ATKA (AGB-3) was 
anchored in North Star Bigt. Weather observations were taken regularly 
aboard the ship during the time and can be compared directly with the 
observations taken at Thule Air Force Base. These observations are given 
in table IIe On 10 October, when the ice changed radically as shown in 
figures 10 and 11, winds recorded on the ATKA were considerably stronger 
than those recarded at the Air Force station. In checking the winds for 
1030, 1330, and 1630 IST, it was found that the wind speeds at the land 
station were only 5) percent of those at the ship. Similar conditions also 
existed on nearly every day. This wind speed differential plays an im- 
portant role in forecasting ice distribution during breakup as well as 
during the period of freezeup. Since nearly all of the wind information 
used by the Hydrographic Office in making ice forecasts in harbors is 
derived or inferred from observations at neighboring land stations, it 
is evident that local harbor studies are necessary to determine the 
relative applicability of each land-station record to the forecasting 
of sea ice conditions in the surrounding areas. 
Grease ice developed on the newly formed water area during the 
morning of 12 October, at which time there was about five-tenths grease 
ice coverage. The distribution of the ice between the harbor and Wol- 
stenholme # on this date is shown in figure 12. Ice that formed on this 
newly exposed water area will be known as "new" ice hereafter, whereas 
the ice that formed the first part of the month will be known as "old 
ice. On this date (October 12), the old ice was 6.0 inches thick. Much 
rafting had occurred in this ice southeast of Saunder § to the main- 
land coast. By 21 October, the new ice attained a thickness of 8.5 
inches and the old ice 11.2 inches. No polynyas, cracks, or leads were 
present from Kap Athol to the dock area. On 23 October (fig. 13), the 
old ice was 12.0 inches thick with 1.5 inches of snow cover. No snow 
cover was evident on the new ice, even though some very light snow 
