Digital Diatribes

A presentation of data on climate and other stuff

Arctic Temperatures Since the mid-90s – Climate Sanity

Posted by The Diatribe Guy on September 15, 2009

I ran across a recent post on Climate Sanity that I thought was an interesting follow-up to my Arctic update using the RSS anomaly data.

To summarize my post: one can plainly see that since 2001, the RSS anomalies have declined on an overall basis, with the rate of decline being steeper the shorter the time frame. Beyond that, the linear best-fit on all longer periods is increasing. My analysis was done on the RSS data from latitude 70-85, which should cover things pretty well (though there is that pesky 5 degrees that isn’t covered). It is based on a best-fit of the anomalies on a month-by-month basis through July 2009. The satellite data goes back to 1979.

Our friends at Climate Sanity have looked at the DMI data by going through a process of extracting data from graphs.

The overall conclusions seem pretty similar if you read the post (linked above in the first line – or click on the link under my blogroll. I have not asked permission to duplicate the post here, so I’ll provide reference links instead.) Fitting linear trends yields overall warming on longer periods, but the highest slope occurs from the mid-90s to current.

The interesting part of the post that really helps put a picture to the discussion, though, is his analysis on a day-by-day basis. I will link to two charts (click on the blue text).

1 – Chart #1 shows the trend by day over time from 1958 through 2009.

2 – Chart #2 shows the trend by day over time split between the periods from 1958 – 1995, and from 1995 – 2009.

This is an interesting perspective in that you see almost no seasonality in the trend through 1995, but since 1995 you see a very distinct seasonality. Part of this may be noise (fewer data points), part of this may be accuracy in measurement (one presumes we’ve gotten better), but even considering for those elements the differential is striking.

One argument against this being “noise” is that the difference in seasons trends fairly smoothly. The January portion of the graph shows the highest trend, and this drops off fairly smoothly until the summer months, where – interestingly enough – we see almost no trend at all (and if there is one, it’s actually slightly negative). We then see the fall – winter months increase towards the right of the graph. There is a weird, anomalous drop at the end of December which makes one question the data there, but again with only 15 data points for each day, things like that will occur.

There are some interesting questions that come of this. First, why the change? Someone referred to the AMO cycle, which I have looked at considerably (most recently in June). The AMO cycle crossed into the positive territory around that same time (mid-90s), though it has actually been increasing from its trough since the 1970s. Further, I’m not sure why a positive AMO would cause a seasonal trend while a negative AMO would not – perhaps in an opposite fashion. But this is just me musing.

Another point that was made in the comments at the original post was that the increased trend seems to be weighted precisely during the seasons when it would seem to have the least impact on ice melt. Indeed, if one looks at the historical ice extent, the winter months don’t seem to deviate all that much, even though this study might imply otherwise.

While that may be true, it would also be true that colder temperatures in the winter probably thicken the ice more, and still makes the ice colder. Colder ice will require more energy input to get it up to the melting point, so it seems quite reasonable to me that warmer winter temperatures will still have an impact on overall ice melt during the summer months. It’s simple Physics.

It’s important for us to recognize ALL the trends and data with these issues. We all know that conclusions can be drawn that are entirely opposite from one another depending on which data is cherry-picked. Just tell the truth in the matter.

And here’s the truth, as I see it: The Arctic, over time, has most definitely warmed. In fact, it has warmed at times quite steeply over some short periods of time, and overall it has warmed over long time frames. This warming, however, has definitely stalled since 2001 and we’ve seen the anomalies cooling off in the most recent years. The slope in the short-term is quite negative.

We have seen ice extent in the Arctic go through a decline, with the minimum ice level occurring in 2007. The warmer temperatures continued into the early part of this decade, and even after they started to decline, it was from such a high level that – while a bit lower – the temps still encouraged more melting. It wasn’t until 2007 that the temperature decline seemed to finally reach the point where it would now encourage a rebound in ice level. And so we saw in late 2007 a significant rebound in the winter ice, leading to a higher minimum level in the summer of 2008. Again, winter ice rebounded in 2008, and I think we can now definitively say that we will see the 2009 minimum ice level come in at a higher level than 2008. Below is today’s chart from IJIS:


IJIS Arctic Ice Extent as of 9/15/2009

I’m still curious about the seasonality, and I thank the poster at Climate Sanity for presenting that interesting analysis. But I think, all in all, the conclusions on my end remain unaffected.

5 Responses to “Arctic Temperatures Since the mid-90s – Climate Sanity”

  1. Bob H. said


    If you look at January 1 and December 31 on the Arctic Temperature vs Day of Year graph, there is a huge step between the temperatures. This is probably a processing error, as I cannot believe that there would be a huge step just because of the new year. Also curiously, there is a much wider variance when the temperatures are colder. This suggests to me the Arctic is more sensitive to external sources of heat (PDO, AMO, El Nino, etc) in the winter, which makes some sense because winter weather tends to vary more than summer weather. Your thoughts?

  2. The Diatribe Guy said

    Bob, I noticed that difference at year end, as well as the large dip at the end of the month. Maybe the data gatherers take Christmas and New Years off…

    I’ve never really looked all that closely at the variance in winter weather versus other seasons. I can buy it on an anecdotal basis (I’ve often noted that we can go from a high of -10 degrees on one day to a high of +30 a few days hence and think it not atypical, and realized that a similar jump in the summer would be highly unlikely). This could explain the seasonal aspect to things. What I don’t understand is why this seasonality was absent prior to 1995.

  3. Bob H. said

    Joe, something’s wrong with the two graphs. The 1996 to present trend is almost a magnitude or more larger than the 1958 to 1995 trend. Such a large discrepancy between measurements suggests some sort of artificial amplification. It would be instructive to look at the 1995 data and the 1958 data. For the 1958 to 1995 trend to look like it does would suggest that 1958 should be nearly flat and 1995 would look like the 1995 to present trend line. This doesn’t seem reasonable to me. If the 1995 to present were scaled down, it would look like the 1958 to 1995 trend. Another possibility is they are comparing 13 years of observations against 38 years of observations. The longer series would have a tendency toward a zero value, whereas you only have 13 data points for each day on the shorter series. Without the actual data I’m just speculating, but I would be interested to see what 5 10-year series would look like (1958-1967, 1968-1977, etc.) I suspect a bit of cherry-picking to get the desired (dramatic) result.

  4. The Diatribe Guy said

    Bob, interesting points. I admit that I have not looked at the data source, and since Tom at Climate Sanity does seem to suggest that he had to take the numbers from charts, there could always be room for some error. I did refer to the “noise” potential in the trends produced. Now, what I will say is that the trend line produced is not out of line with RSS since mid-90’s, but since RSS only goes back to 1979 I can’t say how the 1958-1995 numbers look.

    I don’t think Climate Sanity is pro-AGW, so I’m not sure there’s any cherry-picking going on other than to present the point of time in which it appeared there was a change in the seasonality component of the trend, which really to me is the more interesting part of the analysis.

    I’m not sure I’ll have time to dig into his data, but on his blog post it appears he produces his pdf and source charts/links, so it should be verifiable.

  5. Bob H. said

    Joe, I didn’t mean to imply that Climate Sanity was pro-AGW, but if you had your chance of writing the headline, would you say “Packers Demolish Lions, and then provide juicy detail of the fourth quarter, or would you say “Packers Defeat Lions” and then give a synopsis of the whole game. Or let’s try the Broncos vs Cincinnati, I’m sure you heard about the “Immaculate Deflection” but don’t really know that the rest of the game was relatively boring. Or the headline could read “Broncos Defeat Bengels 12-7, and then describe the game in all its boring detail. Just saying flashier story, probably. Have a good one.

Leave a Reply

Fill in your details below or click an icon to log in: Logo

You are commenting using your account. Log Out / Change )

Twitter picture

You are commenting using your Twitter account. Log Out / Change )

Facebook photo

You are commenting using your Facebook account. Log Out / Change )

Google+ photo

You are commenting using your Google+ account. Log Out / Change )

Connecting to %s

%d bloggers like this: