Global Warming Arguments

With the amount of evidence and reports that have been done on global warming even in the past couple of years, I’m always surprised at people that still refuse to believe in climate change. Some claim it’s not happening, while others, like Larry Bell at Forbes.com think that climate change is just the ‘natural cycle’ of the earth.

This article, even though it was written in 2012, takes the argument that a lot of climate change deniers take: that any signs of ‘global warming’ stem from natural shifts in earths cycles and those of the solar system overall. It also factors in volcano activity. A primary part of the article is citing a publication which looks at a lot of different variables, including Jupiter’s positioning, which causes oscillations in the Earth’s temperatures. Aside from the fact that both seem to be hunting for correlations rather than explaining causations, the article produces a graph which uses global surface temperatures as its y-axis, yet it fails to say where it got the data from.

When read on its own, or to an unsuspecting reader, the article holds some ground. It has a pretty visual, and the logic written by Mr. Bell makes sense. However, it completely fails to use any sort of inferential statistics, as does the paper on which its based, which immediately makes me skeptical of any paper. Inferential statistics are basically the measurement of correctness or alignment between variables, using predetermined models. The most common measurement, widely known, is the p-value. Without getting into the possible sources of data bias (or other potential sources of bias of the writers), its hard for me to argue with the correlations shown in both the article and the paper: it really does seem like there’s a correlation. However, correlations don’t necessarily imply causation. Correlations can show up wherever, and there often are other compounding factors that go into correlations.

Using Mr. Bell’s article as an example of this argument wasn’t really a choice: it took me around twenty minutes of intense google searching to find any sort of news article saying that rising global temperatures weren’t caused by humans in some way. This surprised me a little, because there are so many people all over the country that have disbelief, despite every article and news journal pointing to the same thing. Often, these beliefs are entrenched, and go off of a gut feeling, however this is not how I like to look at science: I like data.

At this point I decided to go off and do some of my own research. On the World Bank’s website, I found a dataset to play around with.

This dataset gave me around 35 thousand rows of data, with 135 different variables in over 200 countries and areas to look at. Since my interest is in understanding what causes global warming (if it’s real), I (after much time spent reshaping and refitting the data) first plotted out some emissions data.

Undoubtedly, at least our emissions are going up. This is all fine and good, but now I wanted to see if it lines up with rising global temperatures. Next, I found a global temperatures dataset from Berkeley Earth, to compare. Note: I took the rolling mean across the months, since otherwise there were massive fluctuations.

Surprise! Global temperatures at least seem to be rising in the last 100 years. So now we have our emissions and temperatures rising, so I wanted to put them together and see if there was a correlation strong enough to merit a title of causation.

Not only are the numbers right there, my graph also looks way better than the one in the article.

Anyways, remember when I was talking about those inferential tests? I decided to run the numbers in python to see if there was a correlation.

Unsurprisingly, there’s an 87% correlation between the two variables, which means its a contender for causation. It definitely would have helped to have data on emissions before 1970, as the more data you have the more correct your model will be. This was just one example done by a college student looking at emissions vs. temperatures, but there’s plenty of work and data done by almost every science organization that back up my claims that global temperatures are rising in accordance with global emissions. The data is surely on my side, and this is just another example that data can help win arguments.