The chart is called a "Sankey Diagram", also know as a flow or flows chart. It can measure energy, materials, finances, data, or other numerable quantities. It's useful for visualising where things come from and where they go. Values are based on reports, measurements, and esstimates. The "rejected energy" value mostly refers to inevitable efficiency losses.
Units given are "quads", or "one quadrillion British thermal units" (BTU). One BTU is the energy required to raise the temperature of 1 pound of water 1 degree farenheit (1055 joule, 1.055 kJ, 0.3 Watt-hour (Wh)).
Energy Equivalents: For comparison, 92.9 quad is equivalent to:
- 98 quintillion joule (98 billion billion joule)
- 22.7 thousand TWh electricity
- 16 billion barrels of oil
- 2.34 billion tons of oil
- 3.3 billion tons of coal
Total Energy Use: The most interesting fact is that total energy usage fell, pretty considerably, from 100.2 quads in 2019 to 92.9. I think that's the biggest decrease on record (the 2008-9 recession and 1979 oil embargo would be my first options for remotely comparable declines), and is both huge (never before seen) and small (for all the economic disruption last year, the US saw less than an 8% decrease in total energy usage).
"Rejected Energy": Many people focus on this value and assume it can be easily reduced or lowered. Truth is it's been largely constant through the entire history of the energy flow charts (it's actually increased with time, from 53% in 1976 to 67% in 2020), though again, this may be an artefact of modeling.
For any thermal energy system (most electrical generation, internal combustion motors/transport), there's a maximum possible efficiency, given by Carnot's theorem, and which is entirely dependent on the difference in temperatures between the cold (intake) and hot (exhaust) sides of the engine.
For transportation there are also transmission, parasitic, and drag losses. Again given by formula.
1973 forecast and backcast plots: In 1973 a set of retrospective and predictive charts was made going back to 1950 and forward to 1990. Projections at the time were that the US would be using vastly more energy by 1990 than was actually true at the time, 148 quad, quantities it's still not reached 30 years later.
Having tried to track down the "rejected energy" methodology, the best information seems to be the very-small-print legend on the chart itself. For legibility:
Data is based on DOE/EIA MER (2020). ... Distributed electricity represents only retail electricity sales and does not include self-generation. ... EIA reports consumption of renewable resources (i.e., hydro, wind, geothermal and solar) for electricity in BTU-equivalent values by assumign a typical fossil fuel plant heat rate. End use efficiency is estimated as 65% for the residential sector, 65% for the commercial sector, 21% for the transportation sector and 49% for the industrial sector, which was updated in 2017 to reflect DOE's analysis of manufacturing.
That is: efficiency is estimated, not measured, based on modeling.
The accounting for renewable energy contribution is also an interesting choice.
