The Canadian Ice Service (CIS) has archived weekly ice concentration data for the Eastern and Western Canadian Arctic. The temporal coverage spans from 1968 to 1998. Figure 4.1 shows the graphical representation of the eastern and western grid points where observations were made. The grid spacing is approximately 0.5° lat. x 1.0° lon. but nodes have been "strategically placed" within the Canadian Archipelago and other such complex regions.
Eastern and Western Arctic Grid points
Until 1980, ice concentration observations were made from June through October for each year. From 1980 onward, measurements were made all year round, but usually only one measurement was made per month during the winter (Nov. - May). Also, several weekly observations contained nodes with missing ice concentration values. For our purposes, all weekly observations with more than four missing data points were discarded. Data from 1997 onward was also ignored, as the east/west boundary was redefined in 1997.
Estimation of Ice Extent
Unlike the interpolated observed and model data discussed in Section 1, this gridded CIS data could not easily be interpolated onto an arctic model grid as a land mask with sufficiently detailed coastline was not readily available.By obtaining the total area covered by the grid points shown in Figure 1 from CIS, we were able to make the following estimate for the area of each grid cell in the Eastern and Western Arctic:
area= (cellarea)*cos(latitude) / cos(75°)
where cellarea is the estimated surface area of each grid cell, 1875 km2, and the cos(75°)term is used as a weight - cells above 75° latitude will have less area, and cells below 75° will have a larger area. The ice extent for each week is calculated by summing all of the grid cell areas, multiplied by their corresponding ice concentration value.
The following two figures display the estimated ice extent for each valid observation, from 1968 to 1996, inclusive. A weekly observation is deemed valid if it has less than five missing data points. Please note that as with other observed data sets, all ice concentration values of 0.15 or greater are 'assigned' a value of 1, and all other ice concentration values are set to zero.
Figure 1 - Western Arctic Figure 2 - Eastern
[click on images to enlarge]
Figures 1 and 2 showed an apparent increase in maximum ice extent from 1968 to 1980. It should be pointed out again that there were very few observations taken from November to May until 1980, so this apparent increase in ice extent is misleading. It can be noticed easily, however, that the estimated winter maximum ice extent from approximately 1980 onwards, shows no significant increase or decrease.
The estimated minimum ice extent for each year, from 1968 to 1996, inclusive, are plotted in the two figures below, with the same 'cutoff' concentration of 0.15.
Figure 3 - Western Arctic, Min Extent Figure 4 - Eastern
Arctic, Min Extent
[click on images to enlarge]
Linear regression analysis was performed on the data presented in Figures 3 and 4, but there was no statistically significant trend line or correlation coefficient for either data set. For this reason, the trend lines were not plotted.
As with Figures 1 and 2, the yearly minima for the Eastern and Western arctic show no significant trend in either direction. Although the arctic ice extent was only estimated, there is no indication from the CIS gridded weekly data that the arctic ice extent is increasing or decreasing.
Landfast ice thickness and snow depth data for various ice stations in the Canadian arctic, distributed publicly on the CIS web site, were also examined. To read the discussion of this study, either return to the Main Menu by clicking the link below, or click here to go directly to the landfast ice thickness page.
[back to main index]
Last Modified: February 2001
Maintained by Erica Muzzerall
Updated by Daniel Roberge