Arctic Ice Growth, 2008 – How Much?

A Guest Post By Steve Goddard

In my most recent article in The Register, and also posted here on WUWT, I incorrectly speculated that NSIDC graphs appeared to show less growth in Arctic ice extent than had actually occurred.  My calculations were based on counting ice pixels from Cryosphere Today maps.  Since then, I have had further discussions with Dr. Walt Meier at NSIDC and William Chapman at Cryosphere Today, to try to understand the source of the problem.   Dr. Meier has confirmed that counting pixels provides a “good rough estimate” and that NSIDC teaches pixel counting to CU students as a way to estimate ice extent.  William Chapman has confirmed that the projection used in CT maps is very close to what it appears and to what I had assumed it to be.  It is an astronaut’s view from about 5,000 miles above the north pole.

What I have learned

In 2008, CT and NSIDC maps show excellent agreement – as can be seen in this video which overlays an August 14 NSIDC map on top of the August 14 CT map.  The borders of ice extent are nearly identical in the two maps. (The videos show overlays of the two maps.)

The discrepancy occurred in August, 2007, when agreement between NSIDC and CT was not so good.  The equivalent video from August 15, 2007 shows that the CT map was missing a significant amount of low concentration ice on the Canada/Alaska side.  I have since confirmed from AMSR maps and NASA satellite photos that the NSIDC map is probably more accurate than the CT map.

The reason that CT provides their side-by-side image viewer is apparently to encourage visitors to make a visual comparison of two dates, which is exactly what myself and others here did when we observed the discrepancy vs. NSIDC graphs.  The human brain is quite good at making estimates of relative areas from images, and pixel counting is nothing more than a way to quantify what has already been observed.  Since writing The Register piece I have made adjustments to the CT pixel counts for map distortion, and as I expected that makes the discrepancy slightly larger.

Because CT maps showed less ice in 2007, the increase in 2008 ice extent appears to be much greater. There is little doubt now that the NSIDC reported ice growth is absolutely correct.  But wasn’t the ice supposed to shrink this year due to an excess of “thin first-year ice?”  In May, NSIDC’s mean forecast (based on previous year’s melt) was that Arctic ice extent would be 13% lower than last year.  (NSIDC has more recently posted on their web site some reasons why they believe the May estimates didn’t work out.)

Click for a larger image

The next graph shows the NSIDC May forecast superimposed on the AMSR graph of what has actually happened this year.  Ice extent has increased rather than decreased, and is not tremendously lower than most other years this decade.   Note that the AMSR and NSIDC graphs bottom out at 2Mkm2, not zero, which creates the visual impression that ice extent is lower than it actually is.

Click for a larger image

Looking at the NSIDC map below, several things are apparent.

Firstly, the widely publicised news stories predicting a possible “ice free Arctic,” an “ice free North Pole,” and a record low extent aren’t likely to happen. (See this WUWT post with a quote from Dr. Meier) The North Pole is nearly as far away from ice free water as in any other year.  In order to get to ice free water from the pole, one would have to travel over 500 miles of ice to near Svalbard.  Lewis Pugh’s kayak trip, as reported by the BBC August 30 – “Swimmer aims to kayak to N Pole” is going to be an impossible one.  That critical error didn’t stop the BBC from highlighting it on the front page of their web site this weekend.

Secondly, while Arctic ice is well below the 30 year mean, it is above expectations and nowhere near gone.  NSIDC graphs show Arctic ice extent at greater than 70% of normal – hardly a six sigma event.  As of August 1, NSIDC was even considering a possible return to normal extent this summer.  If not for a few weeks of stormy Siberian weather, the map (and story) would be quite different today.

There has been lots of press publicity for predictions of an ice free Arctic by the year 2013 or within the next ten years.  Looking at the CT ice area graph below, that would clearly require some major non-linear changes to the 30 year trend.

Click for a larger image

So how do we know if the trend is linear, exponential or sinuous? As seen below, many long-term GISS Arctic temperature records (except for stations close to the Bering Strait) show the last 30 years as being the warming leg of a possibly cyclical pattern – with current temperatures no warmer than 70 years ago.  Most stations close to the Bering Strait show a sharp upwards shift of several degrees (corresponding to the PDO shift) in 1976, and relatively flat temperatures since.  University of Alaska data shows that on average, state temperatures have been nearly flat since 1977.  In fact, parts of Alaska are coming off one of their coldest summers on record – possibly corresponding to another shift of the PDO to pre-1976 conditions.

Is it possible that the 30 year satellite record coincidentally represents only one leg of a waveform?  Greenland temperature records would hint at that.  If you examine only one leg of a waveform, you will absolutely come to the wrong conclusion about the long term behaviour – just as some did during the 1970s ice age panic.

Click for a larger image – original source image from NASA GISS

We know from official US Weather Bureau records that it was possible to sail as far as 81N latitude in ice-free water, during a similar warm period in 1922.  That is about as close to the pole as you can sail now.  Temperatures around Spitzbergen, Norway warmed a remarkable 12C during a few years prior to 1922.  From the November, 1922 Weather Bureau report – “He says that he first noted wanner conditions in 1915, that since that time it has steadily gotten warmer, and that to-day the Arctic of that region is not recognizable as the same region of 1865 to 1917.“

Is there cause for concern?  Perhaps.  But unfortunately much of the press coverage has been little more than science fiction so far.  How do we separate the science from the fiction?  Dr. Meier has graciously agreed to answer that question (and others) in my next article.

One thing we can state with a degree of certainty, is that there likely will be more multi-year ice in 2009 than there was in 2008.  This is  because the 2008 melt season is ending with more ice area than 2007.  Barring asteroid impact, after the winter freeze there will be (by definition) more multi-year ice than what we started with this year.  Any ice which survives the summer will be classified as multi-year ice in 2009.  If next year is cool like this year, is it unreasonable to hypothesise that ice extent will again increase?  Or are we on a non-linear trend which will lead to ice-free summers and a collapsing Greenland ice sheet?  Hopefully Dr. Meier can help sort this out for us.