60-Month HadCrut trends as of April month-end 2011
Posted by The Diatribe Guy on June 21, 2011
In my previous post I mentioned that there was a two year stretch of year-over-year increasing anomalies until the latest 9 months.
NOTE: In reviewing the data after I made this post, I realized that the chart presented below is based on 50 months, not 60 months. I have edited the text below to reference 50 months, and the chart is mis-labeled. The slope charts are correct.
On the 50-month chart we can see this detail a bit clearer. A trough occurs about 3 years ago, followed by increasing anomaly readings. The current downtrend is not enough to offset a positive uptrend.
By itself, a raw, fitted 50-month trend in the world of temperature statistics is fairly meaningless. But I’m just showing the charts and letting the reader decide what kind of mental effort is worth putting into it. Probably isn’t worth much more than a noteworthy point of interest.
Of a bit more interest to me (but not much more…) is the current trend in how the 60-month slope values have been changing. This tells us whether or not there is an acceleration or deceleration in the trendline. I find this more interesting because it puts the current trend line in context with previous readings. The actual trend line is probably not as meaningfull as the graph itself, but it’s a nice visual.
What we see here is that the 60-month trend line had achieved a minimum slope value of just under -0.3 about halfway through 2008. We then saw the slope values increase to near 0.15 by the end of 2010. Over the last few months, we’ve seen a fairly sharp reversal, and it looks as if slopes are back on a downturn.
Again, the 60-month slopes are pretty variable, so there’s still no real stunning conclusions to be had here.
Probably the most meaningful and interesting chart regarding 60-month trends is the entire historical view on how the slope values change over time. I find it interesting because one can see that there are a constant series of peaks and valleys in the 60-month slope values. Observation of this chart highlights a couple things of note:
1) The highest 60-month slope occurred back around 1880. Subsequent peaks were of lower magnitude until around 1930, at which point peaks increased in magnitude until the late 1990s.
2) The minimum 60-month trend value followed a series of lower minimums that started in the 1860s and ended around 1905. From that point until around 1940 the minimums were higher.
3) There was a sudden plunge in the 60-month slope that reached a minimum around 1950. This then started an upward trend in minimums until 2000. The chart of minimums from 1905 – 1945 looks very similar to the stretch from 1950 – 2000.
4) If the current peak fizzles at under a 0.2 maximum following a minimum below -0.2, it would mark the first time since the early 1920s that this would have happened. I have no idea what this means, but I thought I’d point it out.
5) More importantly, if you observe what I’ll call the “major peaks” since 1950, which increase over one another, and then observe the “minor peaks” which increase over one another, then a fizzling at under 0.2 would represent the first instance of a significantly separated “minor peak” from previous peak to (a) not be a “major peak” and (b) to decline from previous “minor peak.” This also hasn’t happened since the 1920s.
6) It is apparent that since the mid-1970s a lot more time has been spent in the “positive slope” territory of the chart than in the “negative slope” area. This shouldn’t be surprising, given the fact that the anomalies saw their most significant increases in the 1980s and 1990s.