Reading the Science of Human-Caused Climate Change: Part 1, Data


  • I realised my understanding of human effects on climate change was stuck at what I learned in school at age 12.
  • I reviewed climate data on tornados, droughts, temperature, CO2 levels and sea levels.

My 12 year-old understanding of climate change

Back in geography class with Mr de Paor, I learned that carbon dioxide is a greenhouse gas, and greenhouse gases cause warming of the earth. At age 32, my understanding had not evolved further than that. This is despite the fact that reducing carbon dioxide emissions is a key motivation for the work I did at Sandymount (now part of Alfa Laval) to reduce the cost, energy and emissions of beer transport with concentrate-based supply chains.

I’ve done quite a bit of statistics and modelling of systems at UCD, MIT, CERN and Sandymount. So, while not an expert, I have some ability to assess data such as that of the world’s climate. I just never took the time to do it.

Governments and companies are directing a lot of resources towards CO2 emission reduction, so I thought it was time to explore the data first hand for myself. Here’s what I learned in two parts, the first focused on measurement data, and the second focused on explanations (modelling).

Are tornados and droughts going up over time?

While we regularly read or watch news about extreme droughts or hurricanes, single extreme events – or even multiple extreme events – don’t prove that climate change is happening (much less whether that climate change is human caused). To establish a pattern of anomalous climate change, one needs compare the prevalence of extreme weather events over an extended period of time.

One key challenge in measuring extreme events over time is record keeping. Today, in 2021, record keeping is very good, but that wasn’t always the case. This means that it is common to see an increase in reported extreme weather events purely because our record keeping has gotten better over time.

One way to address this issue is to focus on the large events – which are more likely to have been recorded in earlier days (because they are big and obvious). The following two graphs show larger tornados (category EF1 and larger), as well as really large tornados (category EF3 up to EF5).

Tornado count by year of large (top) and very large (bottom) tornados in the contiguous 48 states. Taken from NOAA data available at and presented in Unsettled by Steve Koonin.

The occurrence of large tornados in the US has remained largely similar, if not decreasing slightly, over the past seventy years.

Moving on to drought – typically measuring using the Palmer index which combines rainfall and temperature data – here is a plot of that index over time:

Taken from Unsettled by Steve Koonin. Data origin is

There isn’t a clear pattern indicating an increased prevalence of droughts over the last fifty years.

This data above is for the 48 contiguous states. I haven’t looked at global data, but I’ll note that the IPCC 5 WG1 report indicates low confidence in there being any increase in extreme cyclones or in drought on a global basis since 1950.

Is the rise in sea level since 1950 unusual?

To understand the context for sea levels, we can look at data on historical levels over a very long period of time. You have to read the following chart from left to right, starting with today on the left and moving back hundreds of thousands of years on the right. (“Kyrs BP” means thousands of years before present):

Figure above copied from Bianchi et al. 2011, OR

In the grand scheme of things, the world in 2021 is towards the top of a sea-level cycle (we’re also towards the bottom of an ice cycle). Historically, the level of the sea has varied by about 120,000 millimeters (120 metres) over the last half million years. Very roughly, the chart also shows periods of increase and decrease in level by roughly 10 millimeters per year (100,000 mm over 10,000 years).

Now, let’s turn to more recent trends in sea level. In the fifth IPCC (Intergovernmental Panel on Climate Change) report, the first working group has a figure showing the rate of change in sea level over the last century:

From IPCC Working Group 1 Fifth Annual Report.

The above figure shows the rate of change in sea level in millimeters per year since 1900. It’s clear that the sea level was already rising before heavy industrialisation caused carbon dioxide emissions to start in earnest (you can find that graph below). Whether there is a recent (i.e. since 1950s) acceleration in sea level rise is not very clear. Sea levels are definitely rising, but they have been for a long time, including in the early 1900s.

How much have carbon dioxide levels increased since 1950?

Here is a plot of historic C02 concentration in the atmosphere. Data from 1959 onwards is directly measured from the atmosphere whereas earlier data comes from ice cores. Sorry for the small font on the graph. If you find the graph hard to read, it shows CO2 steady around about 280 ppm from 1000 AD until 1850 AD, and then starting to rise around 1850. There is an increase of about 10 ppm by 1900, then an increase of about 40 ppm by 1950 and then an increase of 80 ppm by 2020.

CO2 data from . Note that the y axis starts at 260 so the increase is more dramatic than it looks (although the increase is still dramatic).

The level of confidence is high that the increased levels of CO2 are human-caused because i) the high levels of CO2 in the last 50 years are anomalous – with high confidence – comparing them with the previous decades in recent history (what the above graph shows), and ii) the rough mass balance between fossil fuels burned and CO2 increase in the atmosphere seems to roughly check out (I haven’t checked this myself but would like to).

How much has the global average temperature increased since 1950?

Recent information on temperatures – of both the ocean and atmosphere – comes from direct thermometer measurements (although other measurements like satellite data are now becoming more common).

In the top subgraph below you can see the temperature anomaly of land air (LAT) in red, seawater surface (SST) in blue and HadCRUT (a combination of air and seawater data taken from a separate dataset of measurements) in black. The “temperature anomaly” (on the y axis) is the difference between a temperature measurement and the average temperature over some reference period (typically a short pre-industrial period).

[Side note: The bottom subgraph just shows the difference between the red and black and the blue and black lines in the top subgraph (because the paper is comparing different models and interested in the degree to which the models agree). Don’t worry too much about the bottom subgraph – it is there because the paper is discussing how the red and blue models compare to HadCRUT, which is a commonly used model that – as I understand – combines air and water temperatures.]

Temperature anomalies take from Rohde et al, 2020 (

Looking at the top subgraph above, it is clear that temperatures have increased since 1850. There is some increase from 1920 – 1940 (a period when carbon dioxide levels had not yet risen much), some flatness from 1940 – 1960 (when carbon emissions became significant relative to natural levels) and more of an increase, particularly in air temperature anomalies, after 1960 (when carbon dioxide levels increased significantly).

Satellite temperature data is also now available for more recent decades. One benefit of satellite data is that it does not suffer as much as thermometer based measurements where there can be local warming effects associated with cities, known as urban heat island effects. I need to do more cross referencing before I put up data here, but I think that cross checking earth based measurements with satellite measurements is a sensible thing to do. If you have a peer reviewed reference cross validating thermometer and satellite data I would be interested.

Are recent global temperature increases anomalous compared to historical data?

On page 491 of the fifth IPCC report: The Physical Science Basis, you can see various temperatures over the last two thousand years. If you read the small text in the caption of the diagram you can see what temperature each line corresponds to (broadly speaking, the coloured lines cover different lattitude and land/sea datasets).

Taken from IPCC Fifth Annual Report: The Physical Science Basis

Here is the comment from the IPCC in the report on the above chart:

“Considering these caveats [fewer and less reliable proxy records for earlier centuries], there is medium confidence that the last 30 years were likely the warmest 30-year period of the last 1400 years.”

I can see how drawing conclusions about a trend in the last 30 years of temperatures is difficult because a) historical temperatures show a lot of scatter and b) the graph (see the caption above) uses 50 year smoothing of the data.

What does it mean for global average temperature to increase?

Global average temperature is an abstract concept and doesn’t reflect the temperature at any single place at any single time. Rather – as the term implies – it aggregates temperatures across the globe and across time (typically one year).

When you have an increasing average temperature, that can mean a lot of things. For example:

  1. there may be one or two places that are dramatically warmer than before while everywhere else stayed the same, or
  2. there might be many places that have become slightly less cold while everywhere else has remained the same, or
  3. Some places became much hotter, while fewer places became much colder, so the net effect is a warming.

It is therefore worth breaking down a bit more the temperature patterns that result in an increasing average. I’ll do this with the example of North America.

The next figure shows the trend in reported maximum high and minimum low temperatures over time in the US:

Taken from Unsettled. by Steve Koonan, 2021 – with an original data source of

The number of record high temperatures in the above charge is fairly steady, but the number of record low temperatures is falling. In short, the really cold days got milder.

A non-summary

I’m not going to summarise because I’d like to encourage you to look at the graphs for yourself and go and find other graphs and data. I think that more people looking at and working to understand the data is a good thing, especially when it’s an important topic.

I plan to write a follow on blog (Part 2) where I’ll talk about different explanatory models and hypotheses for what we are seeing in climate data.

In the meantime, please do write to me if there are further graphs or data that you find interesting and/or important or see errors in what I have posted so far. I’ll plan to make some updates to this blog as I get feedback.

2 thoughts on “Reading the Science of Human-Caused Climate Change: Part 1, Data

  1. Great read Ronan. Regarding your selection process, I haven’t read unsettled but you refer quite often to data presented in this book. I note that the area studied is North America, where the changes in climate may not first manifest itself. It would be interesting to see dara from other areas of the world, desertification in North Africa or loss of ice mass at either pole or mountainous areas – Alps, Himalaya etc.

    The data on extreme weather events is very interesting showing no real change over the time period analysed again for N America.

    The data on CO2 increase (exponential) and the data on temperature increase (linear) is probably the most significant evidence of human-related climate change as there is a negative exponential relation between heat transfer coefficient and temperature change over time.

    Thanks for the article.
    Talk soon over a point 5 beer.


    1. Thanks Darragh! Appreciate the comment.

      Yes, three of the eight plots I took directly from Unsettled. The data from those three plots is from the following sources, respectively:
      I’ve now added these links into the article (the first link was already there).
      It would be a good idea – to be very rigorous – to replicate the charts from the raw data.

      Regarding location. The tornado, drought and record high/low temperature graphs are for the US. The other plots are global. The IPCC reports – such as this one (IPCC 5 WG1 report – – does look at extreme events on a global basis. IPCC finds low confidence in increased tornados and droughts globally. I agree it would be good to look at high/low temperature records for other countries. Desertification would be interesting too. As I understand, the ice mass point is kind of like the inverse of the sea level trend – would be interesting to crosscheck that too.

      I agree it is very likely humans have rapidly increased CO2 concentrations since industrialisation for the two reasons I provided in the blog.

      Support for the hypothesis of a recent temperature increase is medium confidence according to IPCC (and I think I agree based on the chart I showed). Assessing human causation of this trend requires a physic model. As you point out, I think the physics relationship between CO2 and net irradiance is an important one to explore and I hope to cover it in part 2.

      Thanks again for the comments Darragh.


Leave a Reply

Fill in your details below or click an icon to log in: Logo

You are commenting using your account. Log Out /  Change )

Google photo

You are commenting using your Google account. Log Out /  Change )

Twitter picture

You are commenting using your Twitter account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )

Connecting to %s