Sunday, 4 August 2013

July Status: PIOMAS Volume

PIOMAS Volume is now available from the main series, I use the term 'main series' to differentiate from the gridded data, the main series data are available from here, bottom of that page. 2013 shows a gain in volume over previous years

At 7.104k km^3 on the 31 July 2013 PIOMAS volume is currently above 2010 for the same day (7.026k km^3), although by such a small margin it can be considered about the same.

Aggressive spring melt has been a feature of the post 2010 years. Anomalies from the long term mean (1980 to 1999) show that this year's spring melt has been muted, it has been followed by a lower melt rate than average resulting in an increase in anomalies, the same pattern as the other post 2010 years, and to a lesser degree other recent years before 2010.

Taking the period from day 110 (20 April) to day 180 (19 June) as the typical period of the spring melt volume losses over this period are presented in the following graph.

This shows that despite appearances of sea ice area or extent, 2013 still had a strong spring melt and aside from 2010 to 2012 was the strongest spring melt in the series. 2013 remains in the set of recent years.

The volume loss calculated using monthly average volumes shows that July losses were average, July not having a strong trend and being the month of greatest volume loss.

The same approach but using ten day increments, link, reveals that the loss over the ten day period leading up to 29 June volume losses were at record lows. Given this data I have changed my mind about the issue of dispersion and re-freeze in the recent hiatus in CT Area loss, I now think it is largely due to re-freeze and cessation/slowing of melt over much of the pack. Note that in the anomaly plot of PIOMAS volume, shown as the second graphic in this post, the uptick at the end of the series starts from 25 July 2013, the date that atmospheric conditions shifted to strongly anti-cyclonic and temperatures dropped.

That the current PIOMAS volume is at the same levels as 2010 does not however mean we're in a recovery of the ice,similarly that CT Area is now above all of the post 2007 years does not mean ice state is back to conditions before 2007.

The above graphic is from PIOMAS gridded data for March, the figures along the bottom row are thicknesses, and the graphic shows the volumes of ice at those thicknesses, e.g. for 2m that's 2 to 2.5m, and in 2000 the volume of grid cells reporting that thickness of ice was 1.978 k km^3. The exception being 3.5m which is 3.5m and thicker.

This data shows that in 2010 there was a volume loss event that lead to a large drop in grid boxes reporting ice over 3.5m thick. Ice of this thickness would take many years to regrow back to the sort of volumes reported in the early 2000s. The drop in 2006 would seem to hold hope of a rapid re-growth, however that was associated with an increase of ice thicknesses in the 2.5m to 3m range, in other words there was a dispersion that reduced the average thickness of grid boxes. Similarly in 2010 there was a drop of 3.5m thick ice caused by a dispersion leading to a peaking of volume in the 2.25 to 3m ranges. This was caused by reduced thickening over winter and a mass export into the Siberian sector, link, what followed in 2010 was that much of this ice melted out, so did not contribute to a recovery in 2011, and lead to a series of low volumes thereafter.

With volume levels currently at 2010 levels it may seem that we will see a winter recovery to volume levels similar to winter 2010. However the ice at present will regrow thermodynamically to the order of 2m thick, this limit is due to the ice insulating atmosphere from ocean and reducing heat flux more efficiently the thicker it grows over the available time in a freeze season. There will be some ridging leading to thicker ice, but the ridging needed to create volumes of ice over 3.5m before 2010 takes years. So to draw the conclusion that this winter will see volume and thickness levels typical before 2010 would not be sound.

There has been some questioning about these PIOMAS results and whether the model is now producing inaccurate results. I see no reason to doubt the picture painted for 2013. The muted spring melt has lead to a summer that is above the PIOMAS volume of the other post 2010 years, and a cool summer has compounded this, in my opinion there is no reason to resort to claiming the model is inaccurate. It is worth re-stating the uncertainty found by Schweiger et al 2011, +/-1.35k km^3 although in the abstract this is stated to refer to October, in section 6.3 this is applied to the September minimum. If we apply this uncertainty to the 31 July 2013 volume we get a range of 5.75 to 8.45k km^3 which makes it impossible to say whether volume is above or below the range of 31 July volumes for 2010 to 2012, add in the uncertainties to those years and the situation becomes even more uncertain. I will continue to use and assess PIOMAS data as an index, knowing it is not 'reality', but accepting it as a good proxy for sea ice volume. On that basis I accept the current PIOMAS estimate as showing that 2013 has not seen a further decline and do not see problems of internal consistency, or consistency with other indices this season.

The multi-page Excel spreadsheet has now been updated with the PIOMAS data, it is available here.


iceman said...

Hi Chris,

The recent post at ASIF on "Arctic sea ice albedo diminishing" led me to review your May post on "The CT Area June Crash" (referencing Rosel & Kaleschke on albedo w.r.t. melt ponding). Following your inference about melt pond fraction from CT area, particularly in the CAB, it looks like 2013 had considerably less ponding than 2012. Comparing to the PIOMAS volume anomaly, I see some correspondence - in both timing and depth - between CT-CAB and PIOMAS for 2012 vs. 2013. Do you find this to be the case in general? Any indication that melt pond fraction has as much influence on the seasonal low (not just the June anomaly) for volume as for area or extent?

Chris Reynolds said...

Hi Iceman,

I'd love to have a regional breakdown of CT Area to investigate what you say better.

I don't find any indication melt ponding has an effect on PIOMAS volume. The PIOMAS spring volume loss that has been a feature of post 2010 years (a year of volume loss rivalling 2007) starts much too early to be due to melt ponding. Indeed the June Cliff, which I take as indicating rapid spread of melt ponds, happens in the second week of June, while post 2010 the PIOMAS anomalies level out around the solstice 21 June, giving little time for melt pond impact on PIOMAS volume. After the solstice PIOMAS anomalies rise (less than average volume loss) then level (about average volume loss).

I'm not sure from the point of view of the PIOMAS model how much it would be affected by melt ponds in the real world - would they be modelled? But I have suspected a role for melt ponds in forming the annual minimum.

2007 and 2010 to 2013 all show a June cliff.
Of these years 2007, 2011 and 2012 were new records for CT Area, and some other datasets (considering extent indices 2011 can best be described as a tie with 2007). Now 2013 has a cliff, but does not appear to be likely to challenge even the previous records. This apparent problem can be explained by considering why there seem to be more melt ponds post 2010 - this appears to be because flat first year ice is more conducive to melt ponding.

As a back of envelope calculation - If I take June losses to be indicative of the size of the June Cliff, and compare with the losses from June to the season minimum.

Year JtoMin June
2012 7.71 3.327
2010 6.83 3.159
2007 7.23 3.039
2011 7.15 3.037
2009 7.05 2.590
2008 6.87 2.469

JtoMin is loss from 1June to min, the table is ordered by June loss.

It may be significant that aside from 2010 JtoMin ends up in descending order.

For 2013 the June loss (JtoMin) was 3.063M km^2, which would suggest a June to minimum loss of the same order as 2007 or 2011. The loss so far is 6.19M km^2, post 2007 average loss to min from now is 0.95 (stdev 0.09), which would put 2013 at 7.14 - right in the region the region the above table would suggest. But still less of a melt than I'd been anticipating given ice state over winter.