In recent summers the late summer volume loss in PIOMAS has been less than in the past, why is this?
Volume anomalies are, as usual, calculated against the baseline average for 1980 to 1999. The plot below shows how since 2010 there has been an aggressive spring volume loss, but this has been followed by a relaxation, as anomalies rise. This is because, after late June, volume losses in late summer have been less than the average.
Calculating the daily change of volume (rate of change), the current date's volume minus the previous date's volume, shows the spring volume loss that is so notable in the above graph after 2010. The purple plot of rate of change is more negative than in previous decades. However after late June, the rate of change for the period after 2010 rises above the averages for the previous decades, demonstrating a lesser rate of loss.
Both Chris Randles (a fellow amateur) and Dr Jinlun Zhang (the main scientist behind PIOMAS) have suggested that this late summer reduction in losses is probably due to there being less ice to melt. This has seemed reasonable but I've never looked into it.
I've re-framed the above graph by expressing the daily rate of change as a percentage of the total volume on each data.
The above graph shows that when expressed in terms of the overall daily volume, the late summer shows more loss than the average for previous decades. In other words, the late summer decline in volume loss seems to have disappeared.
However the original observation was in terms of anomalies from the 1980 to 1999 mean. So I've recalculated the above graphs as anomalies in terms of the 1980 to 1990 average of rate of change.
The anomaly of daily rate of change shows that the post 2010 average displays the behaviour implied by the top graph of this post: During spring volume loss rates are much greater (more negative) than the baseline 1980 to 1999 average. However there is a sudden shift to positive anomalies as the rates of change turn positive, this shows that in late summer the anomaly of rate of loss is positive, relative to the baseline, losses in late summer are reduced, which results in the increase in anomalies seen in the first graphic of this post.
However, as is to be expected from the preceding, if I express the anomalies of the rate of change as a percentage of daily volume, rates of change remain remarkably low relative to the baseline period (1980 to 1999) and to previous decades. Note the higher early freeze season rates of ice gain, due to lower volume fostering proportionately greater ice growth - the ice still grows to fill the Arctic Ocean.
So it seems to me that the hunch of both Chris Randles and Dr Zhang is correct: the late summer anomaly increase in PIOMAS does seem to be due to there being less ice to melt. Late summer melt may reach a period where the 'law of diminishing returns' applies to ice melt late in the season as the ice retreats to the northern coast of Greenland and the Canadian Arctic Archipelago. The spring volume loss ends at the time of maximum volume loss (maximum insolation), as shown in this graphic:
This suggests to me that the spring volume loss is due to anomalous pre-conditioning, after which the ice is already 'conditioned' for melt and further gains in loss do not happen, rather losses in absolute terms are lower than average because there is less ice to melt. It is only when one factors in the total volume available from which to melt ice that losses are seen to be a greater percentage of the available volume than in the past. With respect to the above graphic, note how the post peak volume losses show a downwards trend.
It is worth bearing in mind that despite the poor weather for ice melt in 2013, the same pattern as in other post 2010 years has persisted. This suggests that ice dynamics are the prime mover during the summer following the 2010 volume loss.