« First uptick IJIS | Main | The fat lady's singing »

Comments

Feed You can follow this conversation by subscribing to the comment feed for this post.

Bob Wallace

I think I can do a more detailed analysis of the state of the sea ice myself.

1) There's less of it.

2) What's left is spread out thin over the Arctic.

Unless some unforeseen force comes into play, look for massive melting next year.

I, for one, will not be surprised at a summer melt-out in the next 3-4 years.

William Crump

Bob:

What evidence do you have on the rate of decline of thickness of first year ice in the Central Arctic Basin that suggests the Arctic Basin will experience a "melt-out in the next 3-4 years"?

The volume data from PIOMAS reflects the massive loss of thick multi-year ice. It does not provide any information about the rate of decline in thickness of first year ice. First year ice in the Arctic Basin has replaced the multi-year ice, it has not disappeared. The October 4, 2010 NSIDC press release noted the loss of older ice, saying:

"The ice

Researchers often look at ice age as a way to estimate ice thickness. Older ice tends to be thicker than younger, one- or two-year-old ice. Last winter, the wind patterns associated with the negative phase of the Arctic Oscillation transported a great deal of multiyear ice from the coast of the Canadian Arctic into the Beaufort and Chukchi seas. Scientists speculated that much of this ice, some five years or older, would survive the summer melt period. Instead, it mostly melted away. At the end of the summer 2010, under 15% of the ice remaining the Arctic was more than two years old, compared to 50 to 60% during the 1980s. There is virtually none of the oldest (at least five years old) ice remaining in the Arctic (less than 60,000 square kilometers [23,000 square miles] compared to 2 million square kilometers [722,000 square miles] during the 1980s).

Whether younger multiyear ice (two or three years old) in the Arctic Ocean will continue to age and thicken depends on two things: first, how much of that ice stays in the Arctic instead of exiting into the North Atlantic through Fram Strait; and second, whether the ice survives its transit across the Beaufort and Chukchi Seas or instead melts away."


The MASIE graph shows no massive melt out occurring in the Central Arctic Basin in the last 5 years. As of September 14th, the extent for 2011 is approximately 2.95 million km2 which is nearly 200,000 km2 above 2007. 2011 is less than 75,000 km2 below 2008 and about 230,000 km2 below 2009 and 2010.

If the Central Arctic Basin consistently lost 250,000 km2 per year (a feat which has no historical precedence) it would still take 12 years for the Central Arctic Basin to be empty of ice in September.

The melt rate for this region over the last 30 days of approximately 120,000 km2 is unimpressive as the Arctic Basin was about 3.07 million km2 on August 16th; thus,there has only been a 4% drop in ice extent during this period.

ftp://sidads.colorado.edu/DATASETS/NOAA/G02186/plots/r11_Central_Arctic_ts.png

No "massive melting" exists in the Cryosphere Today area graphs for the Arctic Basin which show 2011 as being approximately 2.5 million km2, where it has pretty much been since the middle of August. While it is 1.0 million km2 below the 30 year 1979 to 2008 average, there is no indication that the remaining 2.5 million km2 of ice will collapse in 3-4 years.

http://arctic.atmos.uiuc.edu/cryosphere/IMAGES/recent365.anom.region.1.html

What "unforeseen force" is going to come into play that has not been in play during the five year period 2007 to 2011 that will "melt-out" the Arctic Basin?

My prediction based on the ice area data from Cryosphere today is that the Central Arctic Basin will grow to its maximum extent of approximately 4.2 million km2 by the beginning of December and it will stay at this level until the middle of April. It will be above 3.5 million km2 at the beginning of July and will be above 3.0 million km2 for most of July.

It is clear that regions outside of the Central Arctic Basin have reached an "ice free" state and will continue to stay that way, but there is no data base which indicates that the thickness of first year ice in the central Arctic Basin is declining at a rate sufficient to be "ice free" in 3-4 years.

Bob Wallace

William, each summer heat comes to the Arctic. As the atmosphere warms, more heat comes.

Heat melts ice. Each year we are starting with less and less ice.

Simple physics tells us that the Central Arctic's turn is coming. It has, as you acknowledge, thinned by a tremendous amount. The '20 feet thick' of last century is gone.
The protective rim around CA ice is decreasing over years. It's starting to melt away from the land up against which it is shoved.

Now, I have not predicted that we will see a summer melt out in 3-4 years. I simply said that I won't be surprised if it happens.

I discard any possible surprise based on the plot of annual volume by month. If the trend continues we see a melt out by 2015. (I expect the September curve to be re-plotted at the end of this month to hit 2014.)

http://neven1.typepad.com/.a/6a0133f03a1e37970b015435378c39970c-pi

Lacking any identification of some natural force expected to appear and stop the melt I've got to go with the curve. And I see amplifying factors which could speed the curve. Less albedo, more water absorption, more turbulence along the CA ice edges, more warm currents flowing under, ....

Something is making the curve curved and not a straight line. We don't need an unforeseen force to perform the CA melt, just a continuation of volume along the lines we've been observing.

You are predicting a refreeze each winter. I agree with that prediction. It will take a considerably larger amount of global warming to prevent winter freeze. But what is happening is that almost every winter we freeze less than we melted the year before.

Look at the April line.

Some area has to be the last to melt. The area best positioned to last the longest is the Central Arctic. That is hanging on longest is no surprise and I won't be surprised to see it melt soon.

William Crump

Bob:

I appreciate your thoughtful response, but the graph you are citing is not reliable and a quick analysis of the graph itself reveals its flaws.

While the lines show smooth running curves, the data is not so smooth as some years report upward spikes and some years report declines below the line. Another indication of trouble is that the graph is showing monthly averages and not the minimum for each month. This means the graph point for 2015 represents an Arctic that is ice free for the entire month of September, not just ice free for one day.

I suggest you contact FrankD, one of the originators of this graph, and see how much reliance he would put on the Arctic being ice free for the entire month of September by 2015.

While you can reply that the graph showing the entire month as ice free for September makes it more likely that one day in September will be ice free by 2015, my response is that if you do not believe the graph is accurate you can not base any conclusions upon the graph.

The graph shows both August and October as ice free in 2016. It has July and November as ice free by 2018. This is not going to happen. Do you trust the graph's indications that the Arctic will be devoid of ice for five months by 2018?

If the "predictions" for these months are not accurate why is the September line an accurate representation of how the arctic will behave?

It is possible the September volume line is starting to flatten, although it is too soon to be certain. Please correct the following figures if they are wrong, but I thought I saw a chart that Neven put out that reported the minimum volume for 2010 as 4.3 km3 and 2011 is 4.2 km3. I also saw in the post some numbers provided by O'Neil? that after the September 2007 crash, 2008 and 2009 volume minimums were higher than 2007. In 2010, the volume levels crashed; however, this was likely caused by the decline of multi-year ice in 2010 noted in the NSIDC October 4, 2010 report, as area and extent in 2010 were higher than 2007. While I am not predicting a higher volume for 2012 than 2011, it is within the realm of the possible and is more likely than a zero volume for the entire month of September in 2015.

You asked what is making the line curve and I believe the answer is the decline and export of thick multi-year ice. Since there is very little of this stuff left, the steep downward slope of the line can not be maintained. Much of this ice was lost due to export, not melting. It has been replaced by first year ice.

On of the oddities about the arctic basin is that water temperature increases as you go deeper. It never gets warm, but the layer below 100 meters can be a few degrees warmer than the water above 100 meters. I am uncertain of the exact physics involved, but it appears first year ice only requires a thin layer of cold water immediately beneath it to keep the heat from the warmer waters below from reaching it. By looking at the charts which show the percentage of ice made up of first year ice v. 2-3 year old ice and ice older than 3 years, it is clear that first year ice has a higher survival rate at the minimum than older ice. This may be counter-intuitive, but it is part of the reason I do not believe Arctic wide data for all types of ice can be used to predict an "ice free" Arctic.

There is no reason to believe that the shape of the curve will continue in the exponential fashion indicated on the graph. It is far more likely that by 2020 the curve will flatten out in what has been called an "S" shape, akin to a graph representing the economic concept of diminishing returns.

If you can produce a graph of the thickness of first year ice at the September minimum that shows an "ice free" September by 2015, then I will join your assessment, but for now, I would prefer that the line drawers use the September minimum figures for the Central Arctic Basin for both area and extent when predicting an "ice free" day than lines drawn using Arctic wide data.

The emphasis on a single year or when the Arctic becomes "ice free" for a day is misplaced. The most telling feature about the Arctic is that the last seven years have produced the lowest minimum figures in the satellite period. This is sufficient to indicate there is significant warming occurring in the Arctic. A single "ice free" day, regardless of when it occurs, is not going to make any difference.

William Crump

Bob:

When you put the September minimum ice maps side by side you can see that some areas that are devoid of ice in one year will have ice in the following years. The Arctic is "messy" like this, it does not decline at a steady pace.

Many regions have reached an effectively "ice free" state at the minimum. These regions can not contribute to any future decline in volume as they have reached their ultimate minimum of zero.

Including data for regions which have already reached a zero volume point in the graph predicting Arctic ice demise will necessarily provide a false reading. If possible, draw a volume graph for first year ice in the Central Arctic Basin and see how much later it takes before a zero volume point is reached.

Wayne Kernochan

@William Crump -

I am sorry to butt in like this, but I find it extremely frustrating that you continue to have difficulty in understanding both what the volume metric means and the fact that much of the ice in the Basin is not the same from year to year -- despite the fact that I and others have pointed this out to you.

You are quite correct that the volume metric will flatten out as it approaches zero. However, this will be because of the variability in average thickness, and because "negative thickness" will show up as a decline in area. Effectively, the point at which volume reaches zero is the point at which half the area of the ice in, say, 1980 at that point is gone. However, the slow decline of the area metric shows that the variability of ice thickness is relatively small. Therefore, it is reasonable to project that in two years beyond that point, less than 5% of the ice area (two standard deviations)will remain -- and that is indeed average, through September.

The fact that multi-year ice is mostly gone is irrelevant, because the degree of multi-year ice has been shown to not affect the overall trend in volume -- again, because it moves on, and does not sit there and accumulate in the Basin.

These considerations are why Maslowski framed it (iirc) as "less than 5% of the ice in the Arctic remaining in 2016". I guesstimate that would mean less than 15% of the area in the Arctic Basin has ice in 2016.

Finally, this is the most likely prediction -- and therefore a speedup is as likely as a slowdown. My conclusion is that present trends suggest less than 5% of the Arctic Basin having ice on average in September somewhere between 2016 and 2019.

For a further discussion of the math and the model, see my blog post (http://waynekernochanblog.blogspot.com/).

Wayne Kernochan

One other point I just noticed - but I guess those who haven't noticed it would appreciate the clarification: The minimum volume in 2011 in the graph is as of August 31st. Assuming a continuing acceleration in melt between then and the usual time of minimum volume, the actual minimum should be between 3.8 and 3.9, a decline of 10-12% year to year. I don't find this at all comforting, especially since the last 5 years have shown two large lurches downward in minimum volume (2007 and 2010). A comparable lurch from this point on would leave less than 2mkm3 in the Arctic, and one in the next four years plus three more years of .5 mkm3 declines would leave us in the "zero volume" situation I discussed in my previous comment.

crandles

William does seem to keep asking questions like:

"What evidence do you have on the rate of decline of thickness of first year ice in the Central Arctic Basin that suggests the Arctic Basin will experience a "melt-out in the next 3-4 years"?"

and from SIE18
http://neven1.typepad.com/blog/2011/08/sie-2011-update-18-ten-yard-line/comments/page/5/#comments

"Those charting the demise of the Arctic ice at the September minimum by 2016 using a data set for the Arctic as a whole need to explain why the charts for area and extent for the Arctic Basin, which constitutes more than 80% of the ice that remains at the September minimum, do not show a rate of decline sufficient for this to happen."

Lots of answers were given. At some point we will have to agree to disagree.

I seems to come down to William believing
"The point I am making is that drawing trend lines using Arctic wide data will give a false trend line as the central Arctic Basin is not declining as fast as the rest of the Arctic."

While most people believe that excluding the other areas gives a false trend and explanations have been given.

Asking the same questions time after time does not seem likely to make progress. So I tried asking why William thinks the volume is declining but will it will not get down to a meltable volume?

No reply yet, but perhaps this isn't the right question. Is there anywhere else for this discussion to go? Or is it just time to agree to disagree?

Chris Biscan

Prelim maps today show the warmer winds compacting the pack. These winds for the next 4 days will be twice as strong. Probably see the first drop in a while today with extent.

Bfraser

I'm trying to post some data, but it keeps failing. I'm going to try writing it in paragraph form. The volume minima for the sea ice in the Central Basin (per Topaz) in the last five years has been 5.016, 5.042, 4.366, 3.795, and 2.757. Quite a decline!

bill

Neven

Wayne, cool to see you have your own blog!

Bob Wallace

William, you posted...

...the graph you are citing is not reliable and a quick analysis of the graph itself reveals its flaws.

While the lines show smooth running curves, the data is not so smooth as some years report upward spikes and some years report declines below the line. Another indication of trouble is that the graph is showing monthly averages and not the minimum for each month.

Of course the data points lay above and below the curves. The curved line is a mathematical line of best fit. The actual data is noisy. The curve is a calculation which shows us that the observed decline is not happening in a straight line, but accelerating.

The curve is based on monthly averages, but I don't see how that has any bearing on accuracy. Monthly averages are dropping and unless something intervenes volume will drop to zero.

The exact date when we hit "essentially zero ice" will most certainly be up to the vagaries of the weather. Essentially zero is when the cake is gone, there will still be crumbs.

Now will the September curve flatten out as it reaches zero? My guess would be no. Why would it? The less ice at the end of August means that the collected heat of the summer has less ice to melt. And the lesser ice of the summer months means that more incoming heat will have been stored up in the open water.

(There is always the possibility of some previously unseen phenomenon triggered by lower ice levels that will 'save the ice', but I've asked a few times and no one has suggested anything likely. At least nothing that has started to make an appearance to date.)

October, I would expect, will flatten some. By October we are back into the freezing season and it will likely be some years before the Arctic can store enough heat to make up for the lack on incoming solar heat.

The winter months, I would expect them to flatten. Volume will somewhat stabilize at the thickness of first year ice and slowly drop as the freeze zone shrinks. But a significantly shrinking freeze zone is likely many years away.

But those are only my guesses. I don't know enough to predict....

Chris Biscan

Jaxa has a -25,000km2.

more like a 50,000km2 loss or more with the running average.

Wayne Kernochan

@Neven - Thanks, but that's a bit like Einstein congratulating me for getting a B in high school physics :)

William Crump

Thanks for all the responses and sorry I did not get back sooner. I will try to address as many as I can.

Wayne:

Thanks for your efforts to educate me. You stated that:

"Effectively, the point at which volume reaches zero is the point at which half the area of the ice in, say, 1980 at that point is gone."

I am not sure why you are using ice area to determine when a zero volume point is reached. To me zero volume means zero area/extent. Can you explain? Also what is "negative thickness"?

I know that the ice is not the same from year to year. It looks like 50% of the ice at this year's September minimum consists of ice that did not exist last year - isn't that what first year ice is?

I have made repeated references to the fact that the ice does not just not just sit in place like your water glass example; and I have provided many references to ice transport into the Central Arctic Basin from other regions.

The quality of the ice has changed dramatically as the charts showing the declines in the age categories (age generally is matched by thickness) of the ice indicate. I believe it is the decline in the thickest ice that has driven the PIOMAS volume decline rate and not the decline in thickness of first year ice, for which there is no data other than anecdotal references of limited measuring points.

You stated:

"The fact that multi-year ice is mostly gone is irrelevant, because the degree of multi-year ice has been shown to not affect the overall trend in volume"

Can you provide a link to a paper that supports this - the recent paper cited by Neven appears to contradict this claim.

I looked at your excellent September 12, 2011 article, but I did not see a lot of references to science papers. You mention "tipping point" as if it is an established fact, it is not, and several science articles dispute the existence of a tipping point. In any event the phrase is vague and can be defined in various ways.

Maslowski is not the sole source of predictions of when the Arctic will become ice free and his predictions concern volume. Is there a specific number forecast for area and extent by Maslowski as to what he considers "ice free" for these two measure? The math says you can have 2,000,000 km2 of half meter ice and be at 1,000km3, but I doubt if you showed a picture of 2,000,000 km2 of ice that the average person would call that "ice free", as that is bigger than Alaska.

From the literature, the balance of papers indicate a much later date than 2019 for an "ice free" day in the Arctic, so my view of a later date is more mainstream. Time will tell, but what is the specific number threshold and which measuring agency are you using to determine when the area/extent figure will be "ice free"?

Regardless of which agency and which measure you use, I conclude that the minimum area/extent will not fall below 500,000 km2 before 2020. I only need a volume of 250 km3 for this to happen with .5 meter thick ice.

As for using the August 31 volume amount to approximate the September average volume, I do not think that it will be that far off. I think you have misread the chart as it is monthly averages not the minimum for the month which the graph is showing. I stand by the estimate that the September average volume will be close to the 4,300 km3 average for 2010. Perhaps the minimum will be as low as you suggest, but you are mixing averages with minimums.

Also, if you believe the ice behaves in a monotonic fashion you are not looking at the data.

I appreciate your bold conclusion:

"My conclusion is that present trends suggest less than 5% of the Arctic Basin having ice on average in September somewhere between 2016 and 2019."

But when I look at the September area and extent chart trends for the Central Arctic Basin, I see no trend that supports this conclusion and therefore, I believe the dynamics of the process can not be captured by using lines drawn on Arctic wide data.

As of September 17th, the Central Artic Basin per MASIE is nearly 3,000,000 km2. That is within 100,000 km2 of 2008 and 200,000 km2 of 2009 and 2010. It is 300,000 km2 above 2007. I see no trend line that would get it to 150,000 km2 (5% of 3,000,000 km2) by 2020.

ftp://sidads.colorado.edu/DATASETS/NOAA/G02186/plots/r11_Central_Arctic_ts.png

After dallying around 2.5 million km2, the Arctic Basin per Cryosphere today has started to rise and is nearing 2.75 million km2. It will not be 125,000 km2 (5% of 2.5 million km2) by 2020. I would agree, that the 5% levels for area and extent are close enough to be called "ice free".

Crandles:

I think you are right, it is time to just disagree as there is no data set showing first year ice thickness beyond the Icesat data which showed a volume increase for first year winter ice from 2004 through 2008 - (see blue line in the chart):

http://www.nasa.gov/images/content/365871main_earth3-20090707-full.jpg

While this data agrees with my position, it was over too short a period of time and may be reflecting the expansion of first year ice into areas that previously had multi-year ice rather than a measure of thickness. Time will provide the answer as to which assessment/guess was more correct.

Bob Wallace:

It is not clear why you think October will flatten but not September.

As for a reason why the line for all months will flatten, I have several, but these are primary:

1. The amount of thick multi-year ice remaining, the loss of which could contribute to future volume declines, is substantially diminished, so the rate of decline can not be sustained.

2. Recent papers posted by Neven indicate that as the ice thins a negative feedback can occur as the rate of heat from the system that is lost to space will increase.

3. Regions which have already reached a zero volume level and have contributed to the volume decline rate in the Arctic wide data in the past have nothing left to contribute to future volume declines. They can not have negative volume.

4. The purported volume decline is not showing up in the rate of decline of area/extent for the Central Arctic Basin.

5. There are significant differences in the forces and their magnitude (for example, temperature and albedo are much reduced above 80 degrees latitude and water depth in the basin is much deeper) acting upon the central Arctic Basin compared to other regions of the Arctic. Drawing trend lines which include data for other regions will show an exaggerated rate of decline.

Thanks to all for listening and engaging.

crandles

Re "I think you are right, it is time to just disagree as there is no data set showing first year ice thickness beyond the Icesat data which showed a volume increase for first year winter ice from 2004 through 2008 - (see blue line in the chart):

http://www.nasa.gov/images/content/365871main_earth3-20090707-full.jpg

While this data agrees with my position, it was over too short a period of time and may be reflecting the expansion of first year ice into areas that previously had multi-year ice rather than a measure of thickness. Time will provide the answer as to which assessment/guess was more correct."

That graph is of volume and there is only one significant increase from 2006 to 2007. The multiyear ice is decreasing far faster so the first year ice is taking over more area. Thus that graph is entirely consistent with a declining thickness of first year ice. But it isn't clear how much, if at all.

I do not expect first year ice to decline in thickness much because of the ice thickness negative feedback. So even if slow to start to freeze over, it catches up towards March.

Do you accept that the volume melted between the maximum volume point and the minimum volume point will increase as the volume of multiyear ice in the total being melted goes down?

Derek

Is there any way to determine the area of the first year ice shown in this series.

http://www.nasa.gov/images/content/365871main_earth3-20090707-full.jpg

We have the volume so we only need the area to give an average thickness. It might show a trend.

My impression is that the area of first year ice is increasing faster than its volume.

Wayne Kernochan

@William Crump: While I appreciate the politeness and thoroughness of your remarks, I find a large amount of misunderstanding of what I said in them. To cite one particular case, I am in fact arguing that the data do not support any notion of a "tipping point", which necessarily implies that one is passing from one "equilibrium" state to another, and that the "tipping point" is the point beyond which it is suddenly much easier to stay in the second state than to return to the first state.

I am therefore going to follow crandles' advice and give up.

The comments to this entry are closed.