Human language is the first new, arbitrarily expressive medium on Earth since DNA
p.376 of ‘Growth’ by Smil.
Up until the 19th century most societies relied largely on biomass and animal labour for non-human work. Smil estimates that annual use of energy was about 18GJ/capita at the height of imperial rule and claims that it didn’t really change for the next 1000 years.
Growth of energy use from the 18th century
- Global primary energy use 1800 - 1900 was 20 EJ to 44 EJ.
- For the 20th century it grew to 391 EJ with fossil fuels supplying 80% of this.
The growth rates have slowed from about 10% a year in the decades of early fossil fuel extraction to about less than 4%. This implies its a logistic curve, that as the absolute quantity increases its rate of change decreases. The logisitic curve of primary energy use indicates 690EJ by 2100.
Agriculture stops being solar powered.
Smil projected that for the 3.7 fold increase in population in the 20th century total harvested land area only increased by 40%. What Smil refers to as ‘anthropogenic subsidies’ increased by nearly 130-fold, from 0.1 to almost 13EJ.
Synthesis of ammonia requires just 1% of global energy use but the application of nitrogen fertilizers has enabled us to grow about 40% of the current global food supply, and about 50% of China’s crop harvest.
Every second person in China are now adequately fed because of this.
Meat consumption in abundance as a modern phenomenon
Global meat output rose from less than 50Mt in 1950 to 230Mt in the year 2000 and by 2015 it had surpassed 300 Mt.
Logistic Regression
How might I replicate these logistic curves smil has? Smil seems to like the idea of fitting logistic curves to things. It implies that there is feedback going on. So, if we take the primary oil consumption historically and fit it, what do we get.