Saturday, January 10, 2015

An Exchange on Global Warming

A friend of a friend posted an objection to something I said about global warming in the middle of an email exchange among friends.  I asked him to expand on that, which he did in detail.  It was a great exchange, bereft of the usual rant.  However, he hasn't (yet) given my permission to post it on this blog.   Nevertheless I am posting my response, to which, by the way, he hasn't yet commented on.

First, I don't know of any scientist that believes a simulation proves or disproves anything.  They are only for information because there are always missing factors whose behavior has to be assumed in some fashion.  I do not look at climate simulation for any purpose at all myself.

Next, many of you may be familiar with financial time series which never move linearly but rather bounce up and down as they move.  My background is in statistical models, and any statistical model includes variability that is not part of the underlying trend, called error variability or "random shocks" be economists.  The global temperature includes both an underlying, or "secular", trend and additional variability due to independent factors.  A major independent factor in global temperature is ENSO, the El Nino Southern Oscillation.  The El Nino pushes the global temperatures above the secular trend, and the La Nina down. 

It is common among amateur traders to take a random upward or downward push in a stock price as measuring the underlying trend.  What happens is they end up buying at a high thinking the underlying trend is going up, and then later selling at a low thinking that the stock is tanking.  In other words they are buying high and selling low, a guaranteed method for losing money.  The important thing to understand is that some of the factors affecting the price must be ignored because they are unrelated to the underlying trend in the stock price.

The Cato Institute, a Libertarian/Conservative think tank that I respect in spite of the fact that they are owned by the Koch brothers, brought out an article a few years ago claiming that the global temperatures over the preceding decade are decreasing.  The analysis started with 1998, a year with a very strong El Nino, an independent random shock raising the temperature over the underlying secular trend.  The effect was to bring up the tail end of the linear fit to the data making it look like a negative trend (something I've discussed elsewhere). 

Starting the analysis with 1998, a year with a large positive but irrelevant increase in global temperature, is a classical instance of lying with statistics called "cherry picking", i.e., selecting the data points to get the result you want.  The Cato Institute eventually removed the article and engaged someone else to redo the analysis.  They found a small positive increase in global temperatures. 

Another analysis duplicating the "cherry picking" is now circulating among global warming deniers.  It's amazing to me how global warming deniers can get sucked into bs.

Before talking about the recent slower increase in global temperatures I'd like to bring up another version of "cherry picking", i.e., arctic sea ice decline.  As I mentioned the time series of Arctic sea ice combines an underlying secular trend overlaid by independent factors causing it to go up and down over the trend.  Making a claim about the underlying trend, "the ice cap has expanded since 2012" is cherry picking the data.  Yes, the decrease has slowed the last couple of years, but are you one of those people that sells low because it looks low?  In fact the secular trend is clearly decreasing.  Picking 2012 through 2014 for your measure of trend is cherry picking. 

extent trend graph

If this were a time series of a stock price, I would want additional information to determine whether it had hit a low or not in 2012.  What additional information do the global warming deniers have?  None, they only have the time series.  Climate scientists on the other hand have other reasons for believing that the underlying trend is going to continue downward, trends in greenhouse gases in the atmosphere for which there are strong arguments in physics and chemistry.   If you are a trader, would you bet your farm on the price in 2012 being a bottom where the additional evidence is that the company profits continue to decline, and all else you have is some kind of "gut" feeling about the company that's possibly a fantasy?  I know some people who've lost a lot of money in the market thinking like that.

Now back to the issue of global temperatures.  The later analysis by the Cato Institute shows a small increase in the global temperatures, smaller than the simulations showed.  In fact, climate scientists are wondering about that as well.  For scientists this calls for research.  And there is evidence that the heat is going into the ocean.  Dr. Buckhammer has claimed that the oceans are not heating up, but his data appears to be out of date, seriously so.

Newly published data suggest that a hiatus in rising global air temperatures in the 21st century is the result of heat sinks deep in the Atlantic and Southern oceans. The trend is likely connected to roughly 30-year global warming and cooling cycles, according to researchers.

In other words, the data are hiding an additional independent factor on top of the secular trend in global temperature which very likely continues to rise.

I've discussed this elsewhereResearch at the University of Washington suggests that we may have a lot more to worry about than just the CO2 we're pumping into the atmosphere.