I ran across this blog post on future energy concerns — Less heat, more light: solving the energy crisis, and while much of it seems solid and there are some very good points in it, there are some things that are very, very wrong. And there’s this whole problem with conclusions drawn from invalid premises — you can’t claim they are valid, even if they happen to be correct; you can’t be sure if the correctness is accidental.

Basically, a discussion of how much energy will we be demanding in the future and where will we be getting it. World-wide we use about 14 TW of power (terawatts, or 10^12 watts) — for an idea of scale, that’s like having fourteen one-terawatt light bulbs — and if one assume a 2% annual increase in use, that will double by 2050.

The first issue I have is that the “let’s get more efficient” isn’t first — if the new real demand isn’t actually going to be 14 TW, then let’s use the real number as our target. So the conclusions about nuclear

A two gigawatt plant needs to be built every month from here to 2050. That will get us all of one (1!) terawatt out of the fourteen needed.

is a little off if fourteen TW isn’t actually needed. Also, the conclusions about how much uranium we have available to us

There’s lots more U in sea water, but if you think we should try the environmental disaster of mining seawater — to get 1TW of radioactive energy — you probably got that idea via the fillings in your teeth.

well, sorry, but snark isn’t science. Since we’re basically talking about filtration (technically adsorption on a polymer), the “disaster” part isn’t leaping out at me.

Now, I agree with the notion about biomass. Having energy production fight with food production is probably a really bad™ idea, especially they way we’re doing it in the US, by using corn which has a fairly small net gain in energy (and there are other studies that peg this at less than break-even), and only serves to drive up food prices. Waste products? Sure. But as far as I can see, corn ethanol is a purely political ploy.

And subsidies: yes. Phase out the subsidies for oil and nuclear (and, I would add, ethanol), and put them into future/alternate tech to help those industries grow. Economies-of-scale will kick in, and likely improvements from R&D, to drive down the price down the road.

On to the bigger fish: first, the solar issue. The big problem here, as I noted in the comments, is the confusion between TW and TW-hr, which makes all of the projections too optimistic by a factor of 24. But solar still can be huge — photovoltaic efficiencies are improving, so almost a factor of two can be regained from the projections if 20% efficient panels can be used. But right now, if demand is going up by 2%, that’s 28 GW of capacity per year. Most solar companies right now are ramping up to produce of order a GW of capacity per year. I don’t have a comprehensive list of solar producers, but while there have been a slew of them becoming publicly-traded in the last few years, my guess is that right now solar is probably not even producing enough to accommodate the increase in demand, much less replace any fossil-fuel production.

Second is the efficiency issue. I think this is completely off-base. The efficiency that is noted

Let’s start with what we know. Historically, energy efficiency has improved about 1% per year without any concerted action. When there has been such action, it improved around 4% in the US after the 1970s oil shocks.

is measured in energy to produce a dollar of gross national product. It’s not the same as wall-plug efficiency, which is capped at something less than 100%. And that’s where I think the conclusion

Two percent compounded over 100 years reduces energy use by a factor of 7.2. By 2100, with a world population of 10 billion people, everyone can be living at the current European standard of living and yet expending half the energy we are using today.

falls apart. I can believe that one might be able to produce 7.2 times as many dollars per unit of energy by 2100, but it’s not going to be because I’m using 1/7 the energy I am today, unless all of my equipment is currently less than ~10% efficient, and I don’t see that being the case. My guess is that there were large improvements in GNP production in the US because when we were largely built on agriculture and manufacturing, there were gains made by adding machinery. Now that we have transitioned away from manufacturing, things like computers replacing manual calculators give you huge efficiencies in making a buck, but they use more energy out of the wall-socket. It’s just that economic growth has outpaced energy growth, not that energy use has decreased. And that’s what I think is in store — economic growth, but not a slashing of energy use.

Here’s an example. You can get 683 lumens of light out of a bulb, per Watt of input. That’s the maximum, period, and you’d better like green — if you want white light, it’s about 3 W, for the light of a 40-W incandescent bulb. That’s just over 4 W for a 60-W bulb equivalent, that my compact fluorescent says draws 13 W. So I got a factor of four when I switched, and I can get an additional factor of three (but probably less) from lighting beyond the CFL’s I’m using today — but it won’t keep compounding at 2% per year. It stops.

So I think that what we have here is an incorrect use of “efficiency” that doesn’t offer a real mitigation. If you start lighting up the dark regions on the globe (literally, not metaphorically), it’s going to cost you more energy, as will electric refrigeration and air conditioning etc.