In the last post we looked at the extent, area, and volume of Arctic sea ice. We also mentioned that we can derive other quantities from these, namely the average thickness as the volume divided by area, and what I called the “spread” which is the extent divided by the area. I’ve also been looking at the difference (rather than ratio) of sea ice extent and area, which I’ve dubbed the “split” (for lack of a better term). I’ll take up spread and split in another post, at the moment let’s see how thickness has changed over time.
Note that thickness can decrease even when total ice (i.e. its volume) increases, if the ice is spread over a larger area so the average thickness is smaller. Likewise it can increase even when total ice decreases, especially if the thin ice melts away much more than thicker ice.
Therefore, although extent, area, and volume all tend to be greatest around March and smallest in September, the annual cycle of average thickness will peak and trough at different times of year. Here, first of all, is the monthly time series of thickness computed as sea ice volume divided by area:
If we “fold” this data with a period of one year (i.e., plot it as a function of month rather than of time) we can see the annual cycle (click this, or any, graph for a larger, clearer view):
For individual months, I’ve plotted thickness values in time order so this is actually 12 “time series plots” (one for each month) squeezed into a single graph. I’ve also plotted (in red) the average value for each month in order to highlight the average seasonal cycle.
Averaged over the 34 years of data coverage, the thickness cycle has peaked in July and reached its minimum in December. Note also that average thickness has decreased over time in all 12 months of the year, the decrease has been much greater during summer and fall months. It has also been most severe quite recently, since 2010. All this has changed the shape and the timing of the annual cycle of thickness. We can see this even more plainly if we compare the average annual cycle for times prior to 2010, to that after 2010:
Averaged over the last three years, the annual thickness cycle has peaked in May and bottomed out in November, shifting its timing by about 2 months. It has also changed its shape. Prior to 2010, the cycle showed a slower rise to maximum and faster drop to minimum, but recently the rise and fall have been at about the same rate so the cycle is more symmetrical about its peak (and is more “sinusoidal”).
It’s also worth noting that the annual maximum thickness is now about equal to the pre-2010 average annual minimum. The decline in sea ice thickness has combined with the decline in sea ice area to bring about the amazing decline of Arctic sea ice volume.
The overall decline is clear from a time series plot of anomaly values:
This also reveals that 2010 was the year that the “bottom dropped out” of sea ice thickness. Lest anyone suggest that thickness has stabilized since then, here is the time series plot of annual average thickness (not anomaly):
Even since the precipitous 2010 drop, sea ice thickness has continued its decrease, and is now more than 40% below its 1979 value.
Windowed Fourier analysis shows the changes in the annual cycle, which identifies 2010 as not only the year that the “bottom dropped out” of average value, but also the year that the annual cycle underwent dramatic change. The cycle’s amplitude decreased:
So too did the relative phases of the 3rd and 4th Fourier components (shown here in red and green with the relative phase of the 2nd component in black):
The change in the timing of the 3rd and 4th Fourier components relative to the 1st (the fundamental) is related to the pronounced shape change of the annual cycle. The timing of the annual cycle also showed extreme change in 2010:
All in all, the thickness shows a consistent downward trend throughout the satellite era combined with a sudden and extreme change in its character in 2010, when it decreased sharply and the size and shape of its annual cycle altered.
We usually think of 2007 and 2010 as the “banner years” in Arctic sea ice change. For extent, area, and volume, they are, especially since the annual minima in those years were such extreme record-breakers. Also, it was in 2007 that their annual cycles changed, increasing in amplitude. But when looking at thickness data, it’s 2010 which stands out above the crowd. I’m not sure why, but it occurs to me as a possibility that it was then that much of the older, multi-year ice disappeared, leaving behind nothing but much younger, thinner ice. I’m not sure where to get data for the amount of multi-year ice (perhaps someone knows?).
This much is clear: even after its banner year 2010, thickness has continued to decline. So have extent, area, and volume. The one thing we can say for sure about the most recent changes in Arctic sea ice is that it is most assuredly not “stabilizing.”
In part 3 I’ll look at the sea ice “spread” and “split.”