Agree to Disagree
Debate in a non-hostile environment
The Science and Mathematics
of Earth’s Temperatures
I have recently downloaded actual absolute temperature data, for over 24,000 real locations on the Earth (it took me about 7 days to download it).
Not just average temperatures, but
- yearly and monthly average temperatures
- yearly and monthly average high temperatures
- yearly and monthly average low temperatures
- some yearly and monthly record high temperatures
- some yearly and monthly record low temperatures
I have analysed this data, and developed 3 or 4 new types of graph, to show people the data.
I have written 4 articles based on this temperature data, so far.
This article will be the 5th article to analyse this temperature data. However, this article will be quite different to the previous 4 articles. This article will use linear regressions, to develop a global temperature model of the Earth. This model will be used to predict:
- the average temperature, of every country on Earth
- the temperature of the coldest month, of every country on Earth
- the temperature of the hottest month, of every country on Earth
The temperature model of the Earth, will be based on only 4 factors:
- the average latitude of the country
- the average longitude of the country
- the average elevation of the country
- the area of the country
I am sure that many people will expect my temperature model of the Earth to be very inaccurate.
Let me give you a quick summary of what I have achieved, using just these 4 factors. I have been able to explain:
- over 94% of the variation in the average temperature, of every country on Earth
- over 90% of the variation in the temperature of the coldest month, of every country on Earth
- about 59% of the variation in the temperature of the hottest month, of every country on Earth
I consider these results to be quite astonishing. When I started, I would have been very happy to explain 60% to 80% of the variation in these temperatures.
Instead, I have achieved R-squared values of about 0.94, 0.90, and 0.59, between the models estimates of the temperatures, and the actual temperature data that I have downloaded.
In case you think that I am just talking about a few countries, let me state clearly, that I am talking about a total of 216 different countries. Big and small, narrow and wide, over all latitudes and longitudes, over all elevations, and of many different areas.
I have included all countries that I could get data for. For some strange reason, there is no data for Poland. I don’t know why. I had to rely on the website that I downloaded the temperature data from.
Because of the size of this project, I expect to write a series of articles, examining different aspects of the temperature model of the Earth.
I will make this article a general introduction, and show you some of the results that I have achieved.
A selection of the results from the temperature model of the Earth.
Please note, that all temperatures are given in degrees Celsius.
I want to be as honest as I can about these results. Listing the results for the 3 different types of temperature measurement, for the 216 different countries, is probably a bit much for most people. I may post the complete results on my website, at a later date.
For now, I will show you 40 results (i.e. 40 countries), for each different type of temperature measurement. I will include:
- 5 estimates which have a large error, above the actual temperature
- 5 estimates which have a large error, below the actual temperature
- and 30 other estimates which are nearer to the actual temperature
So you are going to see the worst results, as well as the better results. Don’t judge the temperature model of the Earth on only this selection. I won’t be showing you most of the results, at the moment.
The average temperature, of a selection of countries.
The coldest month, of a selection of countries.
The hottest month, of a selection of countries.