How to Hide the Decline (from yourself)

After exposing Christopher Monckton’s mendacity regarding global sea ice, a certain blogger called “Albertosaurus” decided to disagree. He’s entitled to his own opinion. But he’s not entitled to his own facts.

He begins by showing this graph (which is from the Cryosphere Today website of the University of Illinois at Urbana-Champaign):

and saying:

This isn’t some data set I made up its from

http://arctic.atmos.uiuc.edu/cryosphere/IMAGES/global.daily.ice.area.withtrend.jpg

Perhaps he’s trying to imply that I “made up” the data I used. You can download it from the National Snow and Ice Data Center (NSIDC) and check for yourself. You can also download the data plotted by Cryosphere Today (which I’ll refer to as “UIUC” for “University of Illinois at Urbana-Champaign”) right here.

Albertosaurus then states:

Yes the ice is a little below average but at at least one point in 1980 there was this much ice on the oceans plus within the last year we have been above the mean. Hardly a catastrophic trend the chicken little’s are calling for.

It would appear that he didn’t do any analysis of the data at all, he just looked at the graph and drew the conclusion he wanted to draw. With no analysis, and apparently without even acquiring the actual numerical data, it’s certainly not appropriate for him to comment on the trend.

Then he reproduces my graph:

and says:

I would say the graphs are not even close, maybe it’s another nature trick…

but if you go there its a series of folders does he use all of them some of them? Who knows they are not very exact. It sure looks like the graphs for the exact same information are not the same. Maybe its the smoothing algorithm used, but really do they look the same?

It’s not the exact same information. The graph from Crysophere Today is of daily data for global sea ice area anomaly, while my analysis was of monthly averages of global sea ice extent anomaly. As we’ll soon see, in spite of this the two different data sets tell the same story. As for the “smoothing algorithm,” no smoothing was done. Neither of those factors is related to the reason the graphs look so different.

They look so different because, in the graph from Cryosphere Today, the y-axis is squashed so that the total range of variation appears tiny. That’s because the good folks at UIUC wanted to plot both the raw data and the anomalies on the same graph. Here’s the anomaly data from UIUC, plotted on approximately the same scale as in the Cryosphere Today graph:

Compare that to the Crysophere Today graph; clearly they’re the same. Now here’s the exact same data, plotted with a zoomed-in y-axis:

They don’t look quite the same, do they? But they’re exactly the same data (feel free to download it and check for yourself). Incidentally, I’ve added a trend line as determined by least-squares regression. The estimated rate of decline for the area anomaly data from UIUC is 36 thousand km^2/yr — which is the same as the decline estimated using the extent anomaly data from NSIDC.

Apparently it never occurred to Albertosaurus that the scale of the y-axis could affect the appearance of the graph. Apparently it never crossed his mind to plot the data himself and check it out. Clearly he never performed any actual analysis of the data. He just assumed what he wanted to assume.

We can plot the two data sets (area anomaly from UIUC and extent anomaly from NSIDC) on the same graph. If we do so using the squashed y-axis as in the Cryosphere Today graph we get this:

They sure do look similar. In fact they’re so similar that it’s a little hard to tell that there are two data sets plotted! Click on the graph for a larger, clearer view and you should be able plainly to see, that the two data sets tell essentially the same story. Of course, we can see this in even more detail by plotting them on the same graph, using a better scale for the y-axis:

What’s the bottom line here? First, the data I used show essentially the same changes over time (and the same trend) as the data referred to by Albertosaurus. Second, Albertosaurus seems to lack the analytical or the graphing skills to know this. Third, in both data sets the decline in global sea ice is overwhelmingly statistically significant, showing how wrong Christopher Monckton’s claim is.

So here’s a challenge to Albertosaurus: prove to us what a good, fair-minded, honest guy you are by admitting that you were wrong. No apology is necessary, just a simple statement that you were completely wrong and I was right.

I don’t expect that to happen. After all, Albertosaurus closes with a thinly veiled threat:

Commenters who threaten anyone while here because they are not smart enough to come up with a better answer will have some due diligence done on them. Foul mouthed lefty posters beware.

I can only assume that foul-mouthed “righty” posters will be welcomed.

UPDATE

Albertosaurus decided that rather than admit his error, he’d embarrass himself.

He responded to one of the comments on his blog post with a lengthy tirade. It consists mostly of idiotic denialist memes, which is an embarrassment all by itself. But perhaps most embarrassing is what he doesn’t say. Namely: anything at all to justify his original idiotic post. In fact, his response is clearly an attempt to deflect attention from his error by changing the subject, as often and as quickly as he can — throwing as much garbage as he can muster at the wall, hoping something will stick. He sure has plenty of garbage.

He rambles (rather incoherently) about the wind chill factor in Winnipeg, the MWP, the Holocene climate optimum, sea level, Al Gore, hurricanes, the USA temperature record, UHI, trees growing in Antarctica, and polar bears. But he doesn’t seem to want to talk about the subject, which is global sea ice. The only refutation he can offer to my demolition of his idiocy is that my analysis is just a “an axis stretching nature trick.”

Pathetic.