By Steven Goddard
[see important addendum added to end of article ~ ctm]
[Note: The title and conclusion are wrong due to bias in the start/end point of the graph, the mistake was noted by Steven immediately after publication, and listed below as an addendum. I had never seen the article until after the correction was applied due to time difference in AU. My apologies to readers. I’ll leave it up (note altered title) as an example of what not to do when graphing trends, to illustrate that trends are very often slaves to endpoints. – Anthony]
JAXA Arctic Ice measurement just had its 8th birthday. They have been measuring Arctic ice extent since late June, 2002.
http://www.ijis.iarc.uaf.edu/en/home/seaice_extent.htm
We normally see year over year ice graphs displayed in the format above, with each year overlaid on top of previous years. The graph below just shows the standard representation of a time series, with the linest() trend.
As you can see, Arctic ice extent has been increasing by nearly 50,000 km² per year. Over the eight year record, that is an increase in average ice extent of about the size of California. More proof that the Arctic is melting down – as we are constantly reminded. Spreadsheet is here.
How do we explain this? There has been more ice during winter, paralleling the record winter snow in the Northern Hemisphere. Meanwhile in the Southern Hemisphere, ice extent is at a record high for the date.
Size matters, but I’m guessing that Nobel Prize winner Al Gore didn’t share this information with his masseuse.
Addendum:
I realized after publication that this analysis is biased by the time of year which the eighth anniversary occurred. While the linest() calculation uses eight complete cycles, it would produce different slopes depending on the date of the anniversary. For instance, had the anniversary occurred in March, the trend line would be less steep and perhaps negative.
This is always a problem with graphing any cyclical trend, but the short length of the record (8 years) makes it more problematic than what would be seen in a 30 year record.
Yes, starting and ending dates for a cyclical activity heavily influence linear trends.
I could be a rich man if I could get a dollar for every time I see CAGW proponents making a linear line out of convenient points of a cyclical trend.
Do I detect a trend?
Thanks to WUWT for showing what you do when (as everybody does) you make a mistake. You correct it ASAP and with due prominence. A poster above asked whether the correction preceded or followed criticism by others. Not really the point unless you have evidence of deliberate delay in retraction – I didn’t notice any. The hockey stick team by contrast tried to shrug off cogent criticism for ages. That is the worry, especially for something that was affecting policy at the global level.
It just goes to show that a person’s mind is clearer at 2 PM than at 5 AM ;^)
Steven Goddard: You should have converted to anomalies first:
http://i49.tinypic.com/25jeemw.jpg
stevengoddard says:
July 2, 2010 at 3:57 pm
EFS_Junior
Read the top of the article now, and compare notes vs. the hockey stick decade long (and ongoing) stonewall.
WUWT corrects within an hour of publication.
___________________________________________________________
Bit of a straw man, I suppose.
Apples vs Oranges.
Well respected peer reviewed climate science literature vs non-peer reviewed blog.
If you want to argue circa 1998 vs circa 2010, go right ahead, but I think most people have moved on, except for one Steve McIntyre.
I just popped over to roflclimate.org to scroll down the topic list. Precious little climate science going on over there at the moment I can tell you.
Quite a bit of mutual tummy tickling though.
The seasonal signal is throwing you way off. If you want to do the analysis properly, my suggestion is to follow this approach:
1. Compute an “annual average” curve, by averaging the data for all years into one composite curve.
2. Compute anomalies, by subtracting for each day’s data the value of the “annual average curve” on that date.
3. Fit a trend line to that.
Let us know what you find.
When there are strong cyclical components in the data, linear regression gives highly variable results, depending not only on the phase at start- and end-points, but also upon the length of the data series. It is not a stable measure of secular trend, but a rather crude band-pass filter, with a lagging response. That’s why regression is not used by professionals in time-series analysis. But try telling this to climate scientist, so very fond of homogenizing station-records.
“If you want to argue circa 1998 vs circa 2010, go right ahead, but I think most people have moved on, except for one Steve McIntyre.”
Ah, yes, nothing to see there, we should all just MoveOn.
Using start and end dates of the 22nd of every month, this is the trend of the trends (in sq km/day). All positive slopes are in Spring/early summer, while the negative are in fall/winter.
Jan -667
FEB -390
Mar -106
Apr 171
May 342
Jun 184
Jul 123
Aug -74
Sep -350
Oct -557
Nov -638
Dec -597
An Inquirer says:
July 2, 2010 at 4:43 pm
Yes, starting and ending dates for a cyclical activity heavily influence linear trends.
I could be a rich man if I could get a dollar for every time I see CAGW proponents making a linear line out of convenient points of a cyclical trend.
True,Enquirer, you would be well-endowded financially a $ at a time but still a pauper when compared with those who toe the political line that without a tax on energy we’re all going to die!
Gotta admit that, we’re all going to die is a tad tautologic but nonetheless is a certainty, the experts are spot on.
We are, gulp, going to die!
toby says:
July 2, 2010 at 2:41 pm
“I am glad you added the addendum – a linear trend fitted to a cyclical function is mathematically dubious.”
BINGO!!! Toby, you said it all when you said that. Congrats on the epiphany!
EFS_Junior
Why on earth would you waste all of your many, many precious talents writing comments on blog site which you don’t respect?
Clearly your skills are far beyond what we deserve, and you should move on to work with well respected scientists (like Mann) devoted to rewriting history for highly respected organisations like the IPCC. We don’t deserve you, and I am sure everyone would understand if you moved on to bigger and better things.
Though I have to admit that I would miss your paranoid rantings.
Steve, don’t apologize for that graph as you did at the end of the your article, as long a it contains complete cycles and is labeled properly, that’s perfectly correct, to me anyway. Others will always scream anyway, as you are aware. Maybe put regression somewhere to shut them up. For some reason AGWers think any straight line on a graph automatically means a “projection into the future” when it really means a history of what past data has done unless explicitly stated as a projection or extrapolation, that’s a different matter.
Great info, I have messed with that very data and totally missed that one. I’m tracking the derivative and if it will average here out as something like -29500 km2/day, it would match 2006 minimum, but, that’s rather far fetched. Think summer could basically end up at the pole in July?? About -54500 km2/day from here will match 2008 with 2009 & 2005 falling right between.
PJP says:
July 2, 2010 at 2:07 pm
Are you absolutely certain that trend line is correct?
Just eyeballing it, it doesn’t look right …
–––
I have the same JAXA data, it’s right on the mark.
July 1, 2002 – June 30, 2010 daily plot regression line:
+136.5 km2/day * 365.25 days/year = ~50000 km2/year.
stevengoddard says:
July 2, 2010 at 3:12 pm
> Speaking of California, when it gets hot inland, what happens to temperatures on the coast?
Hmm, a good reason not to compare areas to sizes of other things.
I’ve never lived in California, but I believe the answer is that the hot, light air causes low pressure to develop and that sucks in maritime air through the San Francisco and the Golden Gate area. (Mark Twain commented something like the coldest winter he ever experienced was one summer in San Francisco.)
I’m less familiar with weather and geography in Siberia, but I’d expect things to behave more like the plains states in the US. Perhaps the “steppes of central Asia” behave differently.
tallbloke says:
July 2, 2010 at 4:54 pm
I just popped over to roflclimate.org to scroll down the topic list. Precious little climate science going on over there at the moment I can tell you.
Quite a bit of mutual tummy tickling though.
I have a bit of a soft spot for roflclimate. After all, it was them that persuaded me to become a ‘denier’
thanks Gav. I owe you one! But to Dhogoza, certainty!
EFS_Junior says:
July 2, 2010 at 4:52 pm
If you want to argue circa 1998 vs circa 2010, go right ahead, but I think most people have moved on, except for one Steve McIntyre.
Amazingly enough, a lot of warmists still tout the Mann hockey stick as if it’s climate science. But yes, some of the wiser warmies would just like to bury and forget.
EFS_Junior
Now it is time for you retraction. You wrote:
Everything you said was incorrect. I wrote up the addendum within five minutes of the article being published. The JAXA data does not start at the low as you claimed, it starts at the midpoint and ends at the midpoint eight years later.
You have posted rather large amounts of misinformation on WUWT over the last few days. Now let’s hear your retraction.
Achab says:
July 2, 2010 at 3:14 pm
Given that this post is admittedly wrong it should be promptly retracted and let professionals do the job.
__________________________________________________________________
WRONG. In a lab you put a line through the faulty data in you lab notebook, make the correction and given an explanation of the correction then initial and date the correction. This is a good example on how science is supposed to work .
All the “professionals” have done is shown how well “old boy” networking and money can corrupt science.
Remember the definition of a “professional is someone who participates for money”. The key word of course is MONEY. The word professional does not necessarily mean some one with more expertise or intelligence or integrity please keep that in mind.
Oh yes the website probably represents more Phds, Ms Ss and bachelor degrees in science, math, statistics, computer science and other fields than most University Departments.
bob says:
July 2, 2010 at 3:06 pm
Ian H and R. Gates are correct,
“If you can’t do the math, get out of the kitchen.
Although the decrease in ice extent has slowed the last couple of days, 2010 still has a 482,343 square kilometer lead on 2007.
The year of the record low, and with weather reports out of Siberia predicting temperatures into the 80s F, who knows what’s up
That’s bigger than California.”
Yes and were entering the portion of the year where in 2007 over 900,000 sq km was lost in a single week followed by another week later in July when it lost 775,000 sk km in a week. Lets see how long that gap remains if the slowing of the past few days continues. I still argue that given current temps in the Arctic Basin this type of loss is not going to be repeated this season.
Here’s an interesting comparison for you. We all know 2007 recorded the lowest ice extent in the 8 years of JAXA measurements. Well at 30th June 2007 the extent was 9.43 million sq km and went on to the “record” minimum of 4.27 million sq kilometres. But at 30th June 2006 the extent was only 9.31 million sq km but went on to record a much higher minimum of 5.81 million sq km.
We are now just past 30th June. Comparisons to previous years when trying to determine this years results are fairly pointless outside of saying we’ve probably lost enough ice that its unlikely were going to exceed 6 million sq km.
I think the most telling thing this year was the early season high temps which primed the rather rapid early season loss. The ice lost was always going to melt anyway, it just happened a little faster this year because of the warmer temps at that time.
The temps have now been at or below normal for the past month. This didn’t stop the ice loss outside the Arctic Basin because it was already well advanced. I believe it will be a different story in the Arcitic Basin though. I’d be very much doubting were going to get anywhere close to the 2007 minimum but it wouldnt surprise me to see us below the 2009 result.
Sounds like some here want some of the other side:
Dec 1, 2002 – No 30, 2009 daily plot regression line:
-638.5 km2/day * 365.25 days/year = ~-233000 km2/year.
Both are facts. No predictions made, it’s just historic data.
( And I enjoy Steve doing it for everyone so I don’t! So Steve, please continue such posts, intelligent people know exactly what they mean. )
tallbloke says:
July 2, 2010 at 4:54 pm
“I just popped over to roflclimate.org to scroll down the topic list. Precious little climate science going on over there at the moment I can tell you.
Quite a bit of mutual tummy tickling though.”
My impression was it was always that way there. Some weird mutual adoration web site for some strange club of pessimists that hate reality. I tried to engage in conversation a couple of times there, but dissenting views are forbidden there.
This site presents quite a contrast. They allow for dissenting views, and several different views of the same topic. They admit and correct mistakes when they know they were wrong. They stay current but don’t forget history. Overall, a much more informative and pleasant web experience and nice format for the exchange of thoughts and ideas.
ctm, was that nice enough to get my last comment posted?
wayne
My error is much less egregious than Hansen, who chose a starting point at the bottom of one leg of a half cycle in 1979 to calculate his advertised trend of 1.7C / century.
http://data.giss.nasa.gov/gistemp/graphs/Fig.D.lrg.gif
The full cycle trend for GISS is actually closer to 0.6C/century.
This gives me a great idea for an article.
stevengoddard says:
July 2, 2010 at 3:12 pm
Speaking of California, when it gets hot inland, what happens to temperatures on the coast?
It can be a cold, foggy, and windy gray day.