Smil's point was more that it is possible to control a tremendous amount of energetic capacity and consequent power (industrial, economic, financial, political, ...) that literally comes out of a hole in the ground.
This also has implications when considering EROEI --- energy returned on energy input.
I mentioned the First Oil Well, Bahrain. It produced at 70,000 bbl/day (not 80k as I'd stated earlier). Typical oil well casing pipe is about 12" in diameter, let's call that 30cm, giving a wellhead area of 700 cm^2. Let's call that 0.1m^2 just to make the maths easier.
That is roughly 55GW/m^2 of continuous energy throughput.
Solar power delivers about 1kW/m^2, 10 MW/hectare, or 1 GW/km^2. Effectively, that's typically achieved for only about 1/3 of a day. I'm not factoring that into the discussion which follows.
The 55 GW of the First Oil Well are equivalent to the peak incident sunlight falling on a region 55 km^2 in area --- about 7.5 km (4.6 mi).
The net solar PV generation is a very small fraction of that, typically less than 10% net efficiency, so you'd need to expand that to 550 km^2, or 23 km ( 14 mi) on a side.
(San Francisco is 11km / 7mi square.)
Yes, oil fields are far larger, they're also far less prevalent, and unevenly distributed. Sunlight tends to be much more evenly distributed.
The notion that there's more than adequate solar energy arriving on Earth is attractive, but potentially dangerous given various factors:
- 2/3 of sunlight falls on oceans rather than land. Capturing this would prove extraordinarily difficult.
- Net of conversion, we'd do well to capture 10% of the incident energy.
- Transmission and storage further reduce energy. Transmission is surprisingly efficient. Storage has proved quite challenging.
- PV solar infrastructure has a lifetime of ~20 years. Put another way, 5% of infrastructure assuming no growth would require replacement each year.
- There are ecological impacts to very-large-scale land-use changes.
- The human population is continuing to grow, and is anticipated to do so through the end of this century.
- Roughly 5 billion people would like to achieve at least an approximation of a Western standard of living.
The US standard of living, based on ~100 quads (quadrillion BTUs) per year, and 330 million people, is about 10 kW.
The global standard of living, based on ~500 quad/yr and 8 billion people, is about 2.3kW. That's an average, and for a huge population, the net energy rate is a fraction of this.
At a US standard of living, 10 kW/person translates to 2,400 quad/year, 4.4 times greater than total present consumption.
At 0.1 kW/m^2, this is the equivalent of 802,000 km^2.
Or about 895 km on a side.
Since I mentioned net production is about 1/3 of nameplace capacity, we actually need 3x the area. So: 2,700,000 km^2, or 1,600 km (1,000 mi) on a side.
That's ... large. But not entirely infeasible.
In truth, population is projected to grow through ~11 -- 16 billions by 2100. And there are other factors which would further diminish net delivered energy. I suspect the global average per-capita power consumption by 2100 will be somewhat below 10 kW/person.
Ok, I went looking for a comparison. Looks like cities occupy ~3,500,000km^2. So that's an area similar to every urban development.
I always expected solar farms to make roof-top solar irrelevant in the future, but maybe that's not realistic. But lack of economies from scale is the worst lesson I take from those numbers, they are not very bad.
Anyway, I'm not sure ocean solar farms are that hard to build. There are many reasons to think they aren't.
The figure I recollect is that urbanisation accounts for 1% or 3% of Earth's total land area. That's in line with requirements for direct solar power.
In practice, that's not entirely appropriable (people and urban environments have some demand for sunlight other than solar PV), and you'd want some geographic distribution to account for weather, seasonality, and power demand variability (daily, weekly, monthly, seasonal), as well as other forms of interruption.
Solar is in theory largely sufficient, and with addition of wind, hydro (both power and storage), and geothermal, affords one possible route to a reasonably-sustainable, reasonably technological future, for a largish population. I suspect it still presents challenges and would probably fall below levels presently experienced in the US and Western Europe, especially at higher latitudes.
Got it! Not household use but total national energy usage divided by population.
It’s per capita energy usage but I hesitate to call it per capita consumption. Consumption would probably want take into account the massive trade deficit, and so its even higher!
This also has implications when considering EROEI --- energy returned on energy input.
I mentioned the First Oil Well, Bahrain. It produced at 70,000 bbl/day (not 80k as I'd stated earlier). Typical oil well casing pipe is about 12" in diameter, let's call that 30cm, giving a wellhead area of 700 cm^2. Let's call that 0.1m^2 just to make the maths easier.
That is roughly 55GW/m^2 of continuous energy throughput.
Solar power delivers about 1kW/m^2, 10 MW/hectare, or 1 GW/km^2. Effectively, that's typically achieved for only about 1/3 of a day. I'm not factoring that into the discussion which follows.
The 55 GW of the First Oil Well are equivalent to the peak incident sunlight falling on a region 55 km^2 in area --- about 7.5 km (4.6 mi).
The net solar PV generation is a very small fraction of that, typically less than 10% net efficiency, so you'd need to expand that to 550 km^2, or 23 km ( 14 mi) on a side.
(San Francisco is 11km / 7mi square.)
Yes, oil fields are far larger, they're also far less prevalent, and unevenly distributed. Sunlight tends to be much more evenly distributed.
The notion that there's more than adequate solar energy arriving on Earth is attractive, but potentially dangerous given various factors:
- 2/3 of sunlight falls on oceans rather than land. Capturing this would prove extraordinarily difficult.
- Net of conversion, we'd do well to capture 10% of the incident energy.
- Transmission and storage further reduce energy. Transmission is surprisingly efficient. Storage has proved quite challenging.
- PV solar infrastructure has a lifetime of ~20 years. Put another way, 5% of infrastructure assuming no growth would require replacement each year.
- There are ecological impacts to very-large-scale land-use changes.
- The human population is continuing to grow, and is anticipated to do so through the end of this century.
- Roughly 5 billion people would like to achieve at least an approximation of a Western standard of living.
The US standard of living, based on ~100 quads (quadrillion BTUs) per year, and 330 million people, is about 10 kW.
The global standard of living, based on ~500 quad/yr and 8 billion people, is about 2.3kW. That's an average, and for a huge population, the net energy rate is a fraction of this.
At a US standard of living, 10 kW/person translates to 2,400 quad/year, 4.4 times greater than total present consumption.
At 0.1 kW/m^2, this is the equivalent of 802,000 km^2.
Or about 895 km on a side.
Since I mentioned net production is about 1/3 of nameplace capacity, we actually need 3x the area. So: 2,700,000 km^2, or 1,600 km (1,000 mi) on a side.
That's ... large. But not entirely infeasible.
In truth, population is projected to grow through ~11 -- 16 billions by 2100. And there are other factors which would further diminish net delivered energy. I suspect the global average per-capita power consumption by 2100 will be somewhat below 10 kW/person.