Climate Change Statistical Tricks

[XLFD]

Facts are stubborn, but statistics are more pliable.  -- Mark Twain

One of the best tools of the trade of peddling junk science so as to make it look like authentic, conclusive facts is the branch of mathematics called statistics.  Adapting a set of data deceptively to show a premise in a good light is the mark of a good junk scientist. 

I recently reviewed the presentation that Dr. Jeffrey Andresen made to Resilient Monroe, a presentation 90% preserved for the folks who attended the Resilient Ludington meetings that occurred on May 15 and May 16.  I picked apart the majority of the presentation using actually data and records that refuted what amounted to unfounded crisis mongering to promote the cause of resiliency and sustainability.  These two concepts are a large part of the Land Information Access Association (LIAA) marketing to our public officials the general basis behind adopting Agenda 21 in practice, which will profoundly change how our nation treats private property (see how here). 

Global warming/climate change is a convenient boogeyman, much like the threat of Communist Russia was in the Cold War.  In the Cold War, our nation took the needed steps to confront and control the spread of communism.  In this Warming War, we are told that we will lose the battle to save the planet if we don't adopt the measures to shore up the disasters portended by these climate crises.  But the underlying agenda behind Agenda 21 is to move the world in lockstep into socialism, which is good if you like having the government control your life, liberty, and pursuit of happiness, and would enjoy the resultant despotism and despair among the masses. 

But beyond this proselytization on the ultimate agendas behind it, I want to further discuss why and how smart people like Dr. Jeffrey Andresen, State Climatologist, can make almost any conclusion from a set of data, if they present it in the right way.  This is a caveat emptor to anyone who looks at a chart or graph that seems to show an unescapable truth-- a warning to look deeper into the facts and to notice when the data is being used to skew the actual data set. 

Definition of Statistics: The science of producing unreliable facts from reliable figures.  --Evan Esar

In all of Dr. Andresen's presentations, he led off with a commonly presented graph used by climate scientists to show that global warming is not a myth (the temperature anomaly graph).  These graphs usually start off at around 1900 and end with recent time, and they are somewhat compelling in showing that we seem to be heating up dramatically over the last quarter century, and he states: 

A couple of things have to be noted here.  The blue line is a nine-year running average, meaning that for a given year, its value is the average of that year's annual average temperature reading and the four preceding and the four following years.  The actual readings for individual years are not presented, but their difference from the average of the full period is.  This can be deceptive if you believe those values are equal, so I will present an easier to comprehend data set and proceed to show how it can be interpreted in three different ways.

I can prove anything by statistics except the truth.  --George Canning

Let's say that we have data from 1900 onward about average global temperature (AGT).  Every even year, except 1974, the AGT is 49 (degrees Fahrenheit); in 1974 it was 47 degrees.  Every odd year the AGT was 51, except for 2005 when the AGT was 53.  In the period from 1900-2013, we can easily compute the average temperature to be 50 by pairing every two years and dividing by two (including pairing the two oddball years of 1974 and 2005), getting fifties every time. 

Thus every year other than the two is either 1 below the average (even years) or 1 above (odd years).  The two off years (1974, 2005) are 3 below and 3 above average, respectively.  Looking at the data in full, a scientist may presume there is a definite cycle going on with only two statistical 'anomalies' in the 114 points of data. 

--1)  Neither Global Warming or Cooling is Taking Place

The first impulse would be to check the two years where the temperature varied a lot more than usual.  Maybe there was an inordinate amount of sunspot activity in 1974, maybe the 2005 figure was skewed because of a miscalculation.  Whatever the case for those years, one cannot conclude whether global cooling or warming is taking place; pretty much moderation is in order, where next year's AGT is pretty predictable.  Two outlying anomalies, one each in both directions, in the 114 year period shows no trend either way.

--2)  Global Warming is Taking Place

For this I will present a sequence of graphs representing recent data, which will eventually show that global warming has been taking place since at least the year 2001.  Here is a simple plotting of each year's AGT against the average temperature over the full period (50 degrees, represented by zero on the graph):

Let's figure out the same idea using a running three year average.  Because every year is between two years that are opposite, we find that if we present the data this way, almost every year will actually reflect the opposite of what actually happened.  Thus 2002 was one below AGT, but since 2001 and 2003 were one above AGT, the three year average was:   [(-1) + 1 + 1]/3= +1/3 or 0.3 degrees.  Thus even though 2002 had a temperature below average, the 3 year running graph actually depicts that year as being above average.  This happens for every year except those around the year 2005 (the purple graph plotted vs. the actual yearly data in green):

That warm year actually brings the year 2005 to the positive, so instead of having six years show temperature declines we only have five of the 13.  But that is not good enough for showing trends and as we have seen, it has a tendency to flip the data and make warm years cool and vice versa.  So let's look at the five year running average (yellow):

Here the results more normalize around what that year was, as the two years before and after the year in question cancel each other out, unless 2005 is in that period.  If that's the case, then positive values result.  You will also notice the right side begins to flat-line, since we do not know future results and have to use the values between 2009 and 2013 for the last few years (the left side does not, since we know those values, but they fall outside the graph).  You will notice these two factors cause us to have only three years that show below average temperatures, making this compelling but not good enough to show definite warming, so we go to seven year running averages (red): 

This still has years 2001 and 2009 falling below average.  How nice it would be to have those years influenced by the warm year of 2005 if you were showing global warming.  We can do this by getting the nine year running average (blue):

Finally the left half of the graph is helped up by being in the range of the year 2005's value; the right half also is brought up since 2005 falls within nine years of the ultimate year.  But let's un-congest things and just look at just the original data and the nine-year running average of that data:

Our metamorphosis is complete!  The yearly data now looks as if we are totally within a warming period, but the process is not yet complete.  The change seems relatively small when compared to the yearly fluctuations, so let's drop the chart's scale by a factor of ten and look at fractions of degrees, and let's change the color of the graph's line to a 'warm orange' instead of a cool blue, and consider a few years before (like back to 1994) this 'warming trend', to show it was once very cool not so long ago:

Still not good enough?  Let's color the area below the graph orange to emphasize the heat:

And to the unwary, we have global warming kicking in at the beginning of the century, when all that has happened is we have had one year that was just two degrees above where it should be.

--3)  Global Cooling is Taking Place

Recall that in our fake data set, 1974 was the cold year that cancelled out our warm year statistically to make our AGT (Average Global Temperature) stay at 50.  In a way similar to what we did for global warming we take our data set all the way back to 1900, and let's take that all the way to just 2012, to not have to worry about the 1 degree above AGT of 2013.  It actually is important in this case to start and end with an even year.

Now one year averages don't show any trend so let's do some running averages, and go big, let's use 101 year running averages.  From 1900 up until 1955, the cold year 1974 is part of these averages, the warm year 2005 is not.  In that 101 year span covered in the running average for these years, we have 50 years of 1 degree above AGT, 50 years of 1 degrees below, and one year of 3 degrees below, making each year about 0.03 below average. 

There then arises a small period where both extremes are in the mix, and the years 1955, 1957, 1959, and 1961 actually show positive values of around 0.01.  But after that, the 101 year running average has endpoints of 1912 to 2012 and within that:  the years 1974 and 2005 whose extremes zero out, 49 years where it is 1 degree above AGT, and 50 years where it is 1 degree below AGT. 

Coolness prevails: 1 degree below the AGT for the 101 year period is graphed at the -0.01 mark for 1962 onward.  Color the line blue, choose the right scale, and use the area to show the following result:

Presto!  Global cooling is definitely taking place, by the proper presentation of the given data.

There are three kinds of lies: lies, damned lies, and statistics. --Benjamin Disraeli

I have tried to make the process of statistics-preparing as easy to follow as I can, without resorting to a full lecture and using lingo of the statistician, by simplifying the data set and arriving at three different conclusions based on how you want to look at the facts.  Statisticians with complex data sets often have more options, that may be less easy to detect-- particularly if they withhold back from the observer salient facts (such as if a graph shows running averages).  Raw data is often hard for the reader to digest, but can be used by the scientist, particularly scientists looking to get funded by grants from politically-motivated grantors, to misrepresent what is found. 

It is not good when science, politics, and businesses-geared-to-capitalize-on-that-science conspire to sell the public a flawed bag of goodies dressed up as science fact.  The raw data does not show that global warming, global cooling, or climate change (whatever that is), man-made or not, exists with any certainty.  Let's not fritter away what's left of our inalienable rights in the midst of a crisis that does not exist.

[Willy]

 Amazing X. This is something everyone should learn. I've never thought about how these graphs and statistics can be skewed in such a manner. This would be an excellent article for the paper if they were not a leftist orginization. I'm going to have to reread this a couple of times to fully abosrb it and will make more comments. Excellent article. I give this an A+

[XLFD]

Thanks, Willy, be sure to let me know if there isn't anything that you don't fully understand; I did have to edit some material out for brevity's sake.

Also, for all interested in directly addressing (without likely being able to comment) the LIAA crowd and our public officials, the first meeting of Community Action Teams are forming tonight at 6:00 to 7:30 PM tonight out at Hamlin Town Hall (June 18).

There has been no indication as of yet that any of our local public officials have any sort of skepticism over the LIAA charts and dissertations.  Even if they did, most are so mesmerized by the likelihood of expanding their own power that they really wouldn't care anyway.