1) a) Every single line shown are thin lines except for the 6 lowest average-ice years, which is clearly marked in the legend (2006-7, 11-12, 16-17). That makes for 34 thin lines and 6 thick lines. b) Also, why would that visual distinction be misleading again? All it would emphasize are when the low-ice time periods are - in the recent decades. It's not like there was a low-ice period in 1979 or something that the author intentionally made thin.
2) It is not visually possible to pick 40 colors that are distinct enough in color space that you can consistently tell at a glance (i.e. when you look away and only compare the next color from memory). Go ahead and try it. Visual perception researchers have tried (http://vis.stanford.edu/color-names/) and the best you're going to get is maybe 12 (http://colorbrewer2.org/)
3) This is so wrong, it's not even wrong: a) the data only existed starting in 1978, because the Nimbus 7 launched in 1978, so this is the whole available data (http://psc.apl.washington.edu/zhang/Global_seaice/). b) Global sea ice isn't in perfect lockstep with temperature, but larger temperature trends and currents and other factors, so talking about a handful of warmer or colder years doesn't even make sense. c) Even if it were, I still don't know what you're talking about. See the NASA land-ocean temperature index (http://climate.nasa.gov/vital-signs/global-temperature/). You could have started in 1977, or 76, or 75... you'd have to go all the way back to 1945 to find a warmer year, and that would be missing the whole point of the overwhelming trend.
I'd be very careful to make sure that I'm not the one doing the public a disservice when I openly accuse people of manipulating data. I wish everyone showed a similar courtesy.
It is a graph comparing the current year to the previous minimum area years, with the entire dataset for context. The choices they have made are completely legitimate for this purpose.
You are also obviously showing a bias towards one conclusion, and accusing opposition to this conclusion of being "plain dishonest" is not constructive. This is data interpretation. Even attributing 100% to your prior (ie. ignoring the data) is not considered dishonest.
Furthermore, when push comes to shove, let me just say that as someone holding a masters in statistics, there is zero useful data on that chart. I think the the point is to show that we are more than 2 standard deviations from the norm ... And then we have a long list of issues with that conslusion. In increasing importance :
1) there is an obvious reason to pick the start of the time series, and that reason is NOT independent of the conclusion that is reached here. Needless to say, that is a huge no-no.
2) more generally the data supporting the conclusion is not randomly drawn from the distribution. Again, huge no-no.
3) is the relation between time of year and sea ice extent well established ? For instance, does the minimum drift ? How much ? (Why I pick that one: because if you ignore the time of year, flatten the data, we are suddenly very far removed from a 2 sigma deviation, and arrive at a much more sane conclusion : lowest extent ever, but not hugely different from before. And of course, seeing 2 extremes of a slowly evolving variable close together is not exactly strange. When you're climbing the hill, every step sets a "new height record". Focusing on that is misleading, to say the least).
How often do we hit "lowest ever" because the minimum extent falls at a slightly different point every year ? (looking at the data: quite often, for instance, we hit lowest ever and highest ever extents last year. Lines that hit "lowest ever" at some point: dark red, dirty green, thick bright green, thin dark green, thin light green, purple, orange and another medium-dark green. I can't be bothered to look up the years, but I think the point stands: lowest ever is not a rare occurrence, and therefore "more than 2 sigma below the norm" is not an accurate way to report that).
You are making assumptions just by choosing which 2 variables to graph against eachother. Are these assumptions valid ? Why ? If you don't have a reason, then please, shut up until you've done your homework.
It may make a lot more sense to plot the minima at the same points every year, to prevent normal noise from generating these extremes, or just only compare minima and maxima, and even then, smoothing out over 4-5 years seems necessary. Or perhaps just don't report things like this.
3) more fundamentally, why would the data be normally distributed at all ? This seems to me unlikely in the extreme.
Technically speaking, the data is contradictory: it leads to a conclusion that disproves itself. The steps work like this : if the sea ice extent is indeed dropping over time (the conclusion of the graph), then obviously it's distribution over time cannot be normally distributed. Unfortunately that is a necessary precondition for interpreting the data in the first place. If the data is not normally distributed, then there is nothing strange about a 2 sigma deviation, which at that point is nothing but an arbitrary numerical value that has nothing to do with the distribution.
If you want to prove that the data is indeed less than the data of the previous years, there's tests for that. This is not it.
4) that grey band obviously covers 98% of the data in the way years are reported here. Years are reported as special, when the 2-sigma deviation is based on days. The time series regularly dive outside of that band, and it's easy to see why: over (2016-1978)*365 days, we should expect ~2% 2-sigma events (days). That's a LOT of days.
5) we should keep our heads screwed on and ask ourselves if the data is really showing what we think it's showing. I would argue no. It looks to me like something prevented sea ice from reforming between August and December of last year. It seems to me unlikely in the extreme that such a sudden and big change would result from the minute temperature change that global warming contributed this year. Maybe not entirely out of the question, but that's an extraordinary claim, requiring extraordinary evidence. And frankly, this article is not off to a good start.
Or to put it more technically : it looks 2016 is drawn from a different distribution than the other years. That means you should find out what happened using other means, and stop using statistics. It cannot help you in this case.
Conclusion: there is something weird happening, true, but based on intuition alone. And let's be honest : chances are very good it's not global warming. Something changed, and it was not a 0.2 degree temperature change. This is way too big for that to be a valid explanation.
Oh, yes, the Planck satellite was launched because of the problem to be analysed. Feedback loop closed.
Meteo satellites were launched because of the problem to be analysed. Feedback loop closed.
GPS satellites were launched because of the problem to be analysed. Feedback loop closed.
Are you aware that you have arrived to a cnoclusion and you're trying very hard to bent reality to adapt to your conclusion even when your arguments are absurd?
>? Disastrously wrong results to follow.
And exaclty why the results are disatrously?
But pedantically the point stands. The planck satellite studies a phenomenon that exists independently of the satellite or the humans studying it. This satellite was launched to study climate change by people who claim they're causing it. There's a clear difference.
> And exaclty why the results are disatrously?
Because using data about something you're doing yourself to study that phenomenon invalidates the basic axioms of statistics. If you do that, all your conclusions about the data become wrong. Why ? Because you yourself are now a factor in the phenomenon under study. At that point, and I've seen this many times, there is no longer any real hope of being correct, as there either a big gaping hole in the model, or circular reasoning. Both obviously invalidate the model.
You don't have provided any rational argument for you criticism, when you were proved wrong anon when the time series started, you changed the goalposts.
As I'm very tired of deniers like you, have a nice day.
Funny how your criticism of me consists of 2 things :
1) I have a predetermined conclusion
2) and it's the wrong one (obviously you have a predetermined conclusion as well and it doesn't seem to tolerate people of different opinions)
My big argument is this : why don't we look at the main point ? Obviously that last year of data doesn't fit the data series. Obviously it's an outlier (just freaking look at it). That means it should be treated as an outlier : don't base conclusions on it, and figure out what has changed.
And it's an outlier so let me tell you: the effect seen here is NOT (directly) related to a very slow change that started in 1850. It is something else.
This also means, don't use statistics to analyse it. It won't work.
Also, you can certainly study outliers with statistics. In fact, most scientific discoveries (these days) all start out as outliers or statistical anomalies. The Higgs boson, Top quark, J/Psi,... every single particle described by the Standard Model (the most accurate scientific theory by far) all started out as anomalies and outliers. The fact that this years sea ice measurements are outliers doesn't mean they are necessarily a large statistical fluctuation, it could be that there is a new climate phenomenon taking place (look up the runaway greenhouse effect). I'm not saying that is what is happening here but we have no idea a priori which is the truth at the moment: it's a statistical fluctuation or it's a sign of a new climate phenomenon. Clearly all we can do is continue to study it using, among other tools, statistics.
[0] - http://earthobservatory.nasa.gov/Features/CarbonCycle/page2....
More importantly, the graph shows area. It doesn't show thickness, which has been steadily decreasing for decades. Now the ice is quite thin, so it doesn't have to thin much more for the area to drop rapidly.
If you want to read about the physical processes in detail, and find out why a large sudden drop in ice area due to global warming isn't surprising at all, read A Farewell to Ice by arctic researcher Peter Wadham.
As for the color scheme choice -- almost any other option would have been better. If the colors were random, it would have removed the ability to recover the trend in any way except of tedious manual comparison, but at least it would be certainly bias-free. If mapped to some color gradient -- well, fitting 28 points is hard, but at least close colors would have matched close years. But this scheme is different. It _looks_ like the gradient map, and invites the reader to mentally build the year-to-year trend, but some colors are nearly repeated in distant areas! This is a cardinal sin of using color mapping for dimensional representation, and it's hard for me to imagine this could be done unintentionally.
Not to mention that this is not a good choice for representing YTY trend; 2D color map, like in https://xkcd.com/1410/, would be so much better.
And it took just 30 seconds of search to know why the data starts in 1978[0]:
> Useful satellite data concerning sea ice began in late 1978 with the launch of NASA's Scanning Multichannel Microwave Radiometer (SMMR) satellite.
1) Lines for years with low ice area are marked bold, when high-ice years (e.g. 2013 or 2015) are shown with thin lines.
2) Color scheme is deliberately confusing, with some colors are almost identical (1993 vs 2015, anyone?)
3) The data set is started with 1978, one of the colder years. If it would have been 1979 or 1974, the story would have been less immediately visible.
While I am not saying there isn't a trend, and 2016 is clearly an anomalous year (but what about this year _was_ normal?), visible manipulations like these quickly destroy any remaining trust in climate science, advancing the other side's cause.