Tag Archives: UAH

Statistical Shenanigans — A global climate story

Today I thought I’d offer a primer on statistical manipulation.  I am going to take reputable data, cherry pick information and “prove” that global warming since WW II began has been less than advertised.   Here we go.

Let’s begin with the East Anglia University World Temperature Graph

Compare the warming between 1942 and 1991 (or 1878 if you prefer).  The change has been about .1 degree C increase in 50 years.  The climate change shown in this chart before 1950 is presumed by climate hawks to be natural warming (or cooling).

Now lets look at the latest UAH satellite climate data:

Now compare the data for the period beginning in 1991 and ending with the current data (March 2013) as shown on the UAH data set.   No net change.  So the world warmed about .1 degree C. in the 50 years from 1941 to 1991 and there has been no net change since then.  A world that is supposed to be warming isn’t.


I started at a high point.  The answer would have been quite different if I had chosen 1976 as my starting point.   Or 1910 or 1965.  But if I had chosen to begin in …..say…. 1878, I could have shown even slower warming.  Statistical misrepresentation is all about data selection.

The IPCC knows how to select data and present it to their advantage. My last post quotes the IPCC’s 2007 Climate Synopses.  I’ll repeat one of my favorite paragraphs here:

Since the IPCC first report in 1990, assessed projections have suggested  global temperature increases between 0.15 and 0.3  degree C per decade from 1990 to 2005.  This can now be compared with with observed values of about 0.2 degree C per decade, strengthening confidence in our near-term projections.

The statement appears not to match data.  Conveniently the IPCC doesn’t quote a source.  I wonder what data they used?  Not UAH or East Anglia.  Perhaps they didn’t start in 1990. The only way that I can get the arithmetic to match is to start in 1992.   1992 was a very cold year because of the eruption of Mt Pinatubo.  World temperatures dropped about .2 degree C in 1992 because of the volcano.  If there is any year that should not be used as a starting point when calculating climate data, it is 1992.

When I use 1990 or 1991 or 1993 the increase is much less than the .2 degree C stated.  The increase is between .1 degree C and  .15 degree C, depending on the temperature source used and start date selected.  If I had started in 1998 instead, the data would have shown very slight net cooling.  Clearly the IPCC cherry picked data.

The IPCC appears to have ignored temperature variability shown in the East Anglia data that they call natural climate variation in their 2007 AR4 report.  If natural climate variation can produce lower temperatures (and higher temperatures too) for long periods of time, how can you assume any specific set of values supports any specific position?

The IPCC cherry picked data…and then assumed that data proved they were right.  Sloppy science and arrogance displayed in a single paragraph at the most quoted climate document ever produced, the 2007 AR4 Climate Synopses Report, section 3.2.

A few years ago NASA direct temperature data started in 1860 and it showed rapid warming and rapid cooling in the 1870’s.  Today NASA data begins in 1880, while East Anglia still begins in 1850. Why did the NASA data change.

I suspect data manipulation….and SWAG.


Temperature S.W.A.G.

I have lived most of my life in Hawaii and Alaska.  Talk about  extremes in temperature variation.  People in Hawaii watch the weather to get the surf report…and to get information on the occasional tropical storm.   Day to day, nobody cares.

When I lived there I once saw local TV newscaster Joe Moore fake the weather.  He searched his desk for the paperwork, admitted to not knowing where he had put it and then made up something like high of 87, low of 74 with light trade winds (an average day in summer).  Good call.

Right now, without even bothering to check for local conditions, I can get the high and the low for Honolulu to within 3 degrees F.  High of 80, low of 67.  And I’m 2600 miles away.  That forecast is good for today…and tomorrow…and the day after that.  The weather in the tropics changes very slowly.

Anchorage is another story.   Yesterday I drove from South Anchorage (at sea level) along Turnagain Arm toward Girdwood.  At about McHugh Creek State Park the temperature changed.  In about a minute it went from 23 degrees to 36 degrees.

In Fairbanks, temperatures within the city area regularly vary by 30 degrees F or more.   The hills around Fairbanks are warmer than the city.   Go up in altitude and warm up.  Fairbanks gets horrible temperature inversions every winter.   A drive from downtown up Farmers Loop Road (about 5 miles) nets the aforementioned 30 degree shift on just about any cold winter night.

Climate experts tell us the climate will change more in Polar regions than in more temperate zones.   The UN gives very specific predictions for climate.  They make predictions for 10 years from now, for 90 years from now and for 200 years from now.   How do they know?

As we venture from the equator toward the poles, two important statistical problems develop for temperature predictors.  1) The temperature variability increases and 2) the number of weather stations decreases.   Hmmm.

fluctuating data + few data points = wild ass guesses

 The University of Alaska recently conducted a climate study of Alaska showing cooling in the 2000 to 2010 period.  They used 20 reliable test sites for the whole state.   They need hundreds if they have any hope of being accurate.

Siberia and Northern Canada suffer similar problems….and the Arctic and Antarctic….well forget about it.   The Arctic Ocean averages about 10 working sites and those locations are not fixed, they drift with the ice.  Any ground based temperature data for either the Arctic or Antarctic must include more than a little wild ass guessing.

Perhaps a little perspective, here’s a graphic of the Arctic

Ten sites, all that space.  WOW.

That is why I like Satellite data.   That data includes guessing too, but it’s a different kind of guessing, making adjustments for Satellite drift and other indirect conversion problems.   The big problem with Satellite data, it is a very small data set.  Who knows what the data would have looked like 50 or 500 or 5000 years ago.    Let’s take a peek at the most recent Satellite data courtesy of the University of Alabama at Huntsville

Yep, the world’s a bit warmer than it was 34 years ago.  But it’s worth remembering that in 1974 Time Magazine ran a cover article predicting the next ice age based primarily on 30 years of cooling that began around 1940.  Sure it looks like the world has warmed about .3 degree C since 1979, but I wonder what the Satellite data would have shown had it existed in 1940.

World Temperature Data — A Statistical Mess

I prefer  satellite data to land surface  data when discussing global temperatures.  Satellite data has calibration issues but it successfully avoids so many of the surface issues.   Climate scientists using surface data must accurately adjust data to overcome site location deficiencies, poor and incomplete data and a changing environment.  Unfortunately satellite data has only been around since 1979, making it a ridiculously small data set.

The IPCC admits in their 2007 report that the uneven nature of temperature sights are a problem today.  I can only imagine how difficult the problem must have been….say 150 years ago.

That said, there’s lots of temperature data out there….and when I look at all the different pieces of data, I find it difficult to draw the definitive conclusions so many climate experts draw.  When the IPCC says they are 90% certain man has caused most of the warming seen in the last 60 years, as they did in the last Synopses Report in 2007, I wonder what they see that I don’t see.

Let’s begin by looking at the Vostok Antarctic Ice Core Data Set.  The data ends in 1950 and begins 400,000 years earlier.  This chart shows both temperature and carbon dioxide.

Sometimes carbon dioxide and temperature go in the same direction, sometimes not.   Al Gore used only ice core data as his proof (in his documentary film An Inconvenient Truth) that global warming is the serious problem that he says that is.

I find that more than a bit ironic, because ice cores really don’t support his position well. I am particularly fond of the data from about 130,000 years ago.   In one 10,000 year period shortly after the peak temperature dropped 6 degrees C while carbon dioxide  managed to stay relatively steady at about 270 ppm.  Steady carbon dioxide, falling temperatures…and it went on for 10,000 years.

I’m also a fan of 345,000 years ago when carbon dioxide increased 15% in 10,000 years and temperature actually went down.  I suspect volcanic activity, but then that’s for another post.

Let’s move from Antarctica to Greenland and shorten the time span to 4,000 years.

Wow, there really was a Medieval Warming Period…in Greenland at least.   Of course Dr. Mann (of the Mann Hockey Stick Graph fame) has called the Medieval Warming Period and the Little Ice Age regional issues.  Well, let’s take a look at the oh so controversial Mann Hockey Stick Graph that was critiqued by me in a previous post

Check out the temperature scale in the previous two charts .   The Mann graph has almost no temperature change for the first 900 years of the chart, the Greenland core shows a change of about 3 degrees C in the same period.  Clearly the Greenland Ice Core disagrees with the Mann Hockey Stick Graph.

I’d be the first to admit that both Antarctica and Greenland are poor proxies for the world as they represent individual very cold locations…and the weather in Greenland will not accurately predict climate anywhere else in the world.   The data supports the argument that temperature is difficult to accurately measure….and it casts doubt on the validity of the Mann study.

Now well finish up with my two favorite very recent charts, the East Anglia University chart beginning in 1850 and a satellite data chart beginning in 1979 provided by the University of Alabama at Hunstville.

Look at all that variation in temperature.  Wow.  And carbon dioxide has been steadily rising.

When I look at all this data I wonder how anyone can make statements about global climate and carbon dioxide that are presumed to be uncontrovertible which appears to have become the new buzz word in the climate science game.

Is NASA Cooking the temperature books?

I have long distrusted NASA temperature data.   My distrust begins with Goddard Space Studies Administrator James Hansen.  Dr. Hansen is a smart guy, but he has an agenda. He has been publicly trying to convince the world that global warming is a man caused crime against humanity since the 1980’s when he testified before Al Gore’s committee in Senate hearings.

Anyone with an agenda is going to be scientifically prejudiced towards their preferred position, and Dr. Hansen clearly has a specific prejudice.  Everybody working at NASA knows about the bosses prejudices….and I think that can and probably does impact results.

Let’s look at some temperature data.   First stop, the University of East Anglia, UK.

Second stop NASA

And finally the University of Alabama Huntsville (UAH) Satellite data

The first two are surface temperature data, and the third is satellite data.  All three are recognized worldwide as experts.  Which one is not like the others?  NASA!

NASA data used to begin in 1860, and included the very warm years in 1878 and 1879 that are included in the East Anglia data.  And that same old data had the late 1930’s and early 1940’s a bit warmer too.   All current NASA data begins in 1880 and the links to data that went back to 1860 are no longer active.

Both East Anglia and UAH show a pronounced peaking of data in 1998, NASA shows a much lower 1998 and a much higher 2010.   The NASA data is presented in a way that makes it look like global warming is increasing steadily, the other sites  show a peak in 1998 with slight cooling since then.  NASA has insisted that 2010 was the warmest year on record with 2005 second and 1998 third.  East Anglia and UAH like 1998.

So who’s right?   I don’t know, but I have suspicions.  A second satellite site nicknamed RSS (Remote Sensing Systems), generally agrees with East Anglia and UAH.

I wish NASA Administrator Hansen were less emotionally involved.  He is a global warming true believer, and I fear his attitude may cause NASA to look for a specific outcome.

Global Temperature Data Varies Wildly — We need lots of Data

Who’d have thought something as simple sounding as temperature could become controversial.  Measuring the Earth temperature is actually a fairly difficult task.  The Earth is a big place, with a limited number of measuring points, particularly in the oceans.  Any answer has more than a little SWAG.

I’ve got to say, I like Satellite data better than land based global temperature data.  It covers the whole world and is less likely to be manipulated by adjusting for local conditions.  But what I really like is lots of data.

I am constantly amazed by the amount of variation within data sets.  It is not unusual for data to vary by .2 or .3 degree C in a given month and changes of .4 degree C in a single year is fairly common.  Natural background noise can be a degree or two C per century.  That’s a lot of variation.  Here is the current UAH Satellite chart:

Lots of month to month and year to year changes.  Now lets take a peek at a Land Ocean Index Chart (although Ocean data before 1995 is pretty spotty).  Here the current NASA chart going back to 1880 that I pulled off the NASA website:

NASA data used to go back to 1866, and the period between 1866 and 1880 had two very warm years, 1877 and 1878 and rapid annual cooling (.43 degree C)  between 1878 and 1879.  Imagine  1877 and 1878 being about the same as 1973 on this chart.  And 1964 and 1976 were almost as low as 1890 back then too.  The whole chart looked a bit flatter, particularly before 1980, and a lot less ominous.   The trend was still up, it does make you wonder….what new information about the 19th century emerged since 2009 when I downloaded the old data?

Now lets look at a chart that compares satellite and a group of land based data.

The data matches pretty well.  UAH and RSS are the two principal US sites that  do Satellite data. They compensate for Satellite errors in different ways.  RSS is regularly a bit higher than UAH, but they match well.  I am not sure which direct measurements were used in this comparison I pulled off Wikipedia.  Probably the East Anglia, UK data or perhaps NOAA, or perhaps an average of many sites.  It’s not NASA.  NASA insists 2005 was warmer than 1998.

Land based data is subject to wild manipulation…and stories of NASA manipulation seem to be everywhere. My favorite NASA manipulation story – NASA changed the way they calculate temperature in 2005, and some people think their approach might be less than rigorous. Here is a link to the story:


I found the site while reading articles about how 2011 was either the 9th warmest in history, the 3 coldest in the 21st century or just a bit above the average of the last 34 years, depending on the article, or the data source and how the data is presented.

I am reconciled to the notion that I will likely die without knowing whether anthropogenic global warming is a serious problem or a political construct.  That is because the data sets vary wildly and the background has the potential to vary wildly too.   We either need to come up with a much more effective way to measure and understand the past, or we need lots more data.

The data is so scattered,  patterns are not going to be immediately identifiable.  Lots of data will be required for confirmation.  Someone looking at Satellite data in  1998 could easily draw a draconian conclusion (sound familiar), but the next 14 years worth of data failed to show continued warming.  March of 2012 in the Eastern USA  was very warm, but it got colder in April,  And so it goes.

April 1998 — Warmest month in history?

A few years ago I did a bunch of research on global temperatures.  I have this vague memory that April 6th 1998 was the warmest day in the UAH (University of Alabama Huntsville) satellite troposphere data base.  I was going to use that day as the basis of a post on climate data.

Well I couldn’t find the data I was looking for, but I did find multiple discussion items written in  February of 2010 using the UAH website data to tout January of 2010 as the warmest month ever and proof that global warming theory was correct.  People talked about all sorts of things, but the UAH January 2010 data was the source of celebration.

When I first started looking at global temperature data I was surprised by the nature of the data.  Each month can vary quite a bit from the month before.  Annual changes can be very large too.   NASA has global surface temperature records that go back to 1866.  In 1879, the temperature dropped .43 degree C in a single year.

UAH data shows a slower rate of warming than the other three websites that measure world temperatures via satellite.  The difference is small but there is a difference..and the scientists at UAH have been among the more skeptical of scientists that work in the field.  So if they show warming ……

Here is the UAH data that was being used, updated to include February 2012 data.

C’mon guys….a single data point doesn’t prove anything one way or the other….and it looks to me like April of 1998 was warmer than January of 2010 anyway.  But there it was, one opinion piece after another, trying to use a single data point to prove a position.

Climate data has wild fluctuations and the data sets are small.  Too many people are trying to use individual data points in wildly fluctuating small data sets to prove whatever point they want to make.  Always a bad idea.  And just about everybody talking global warming science does it all the time.   Al Gore loves to do it, so does Bill Mahr and Sean Hannity …and they all are making the same mistakes.

Satellite data is only 33 years old, the Ice Age we currently live in is 2.5 million years old.   20,000 years ago it was much colder than it is today (probably 10 degrees C colder) and 130,000 years ago it was much warmer (probably 6 degree C warmer).  And sudden rapid warming and cooling is present all throughout the cycle.

The IPCC (the UN Intergovernmental Panel on Climate Change) loves to use short term data to predict climate years into the future.   In their 2007 Climate Synopses Report, the IPCC states that the warming of the late 20th century gave them greater confidence that global warming theory was correct.  Do they have less confidence now…just 5 years later…when their short term predictions for 2010 turned out to be wrong.