The critiques I posted of the last two articles apply again here, since the assumptions and conclusions of those two feed into those of this one. I'm sure others will repeat themselves.

The analysis needs some comparative figures. For example,

• the cost of getting the same energy entirely from fossil fuels over the time period envisaged for the conversion to renewables
• the cost in disruptions to economy and society of simply letting the fossil fuels run short and replacing them with nothing

A lack of comparison is the most serious weakness of this paper. For example,

The figure quoted only represents the major energy production plant required. In parallel with this will be a similar cost associated with the changes made by energy users (i.e. electric vehicles, mining and manufacturing equipment, rail lines, power transmission, metal smelters etc. etc....)

You've not given a clear time period for the conversion to renewables. However, it seems fair to assume you're thinking of more than a generation (30 years). Even if coal fell like manna from the heavens and the Earth had a creamy nougat center of oil, over those 30 years much of our vehicles, mining and manufacturing equipment, rail lines, power transmission, metal smelters and the like would be replaced anyway.

So if we have a petrol-burning car now in 2009, and are going to have to replace it in 2020 at the latest, whether we're replacing it with a $10,000 petrol-burning car or a $10,000 electric car doesn't matter economically. It only matters if one is $10,000 and the other $15,000 or $5,000.

Your figures are absolute rather than relative. It's like someone saying to us, "buying your house will cost you $300,000 over 20 years!" Yes, it will. But if we didn't buy the house, we'd have to pay rent anyway, so the question must be what extra cost we have because of our choice.

Just as whether we own or rent, we're not going to let ourselves be homeless over the next 20 years, so too will we as a society not let ourselves be without electricity, transport and so on.

Whatever we do, we're going to spend shitloads of cash on building stuff. The only questions are:

• what will we build?
• how much more will it cost than building something else?
• is it worth it?

Your paper doesn't look at extra costs, only absolute costs. It thus vastly overstates the magnitude of the problem in economic terms. Presumably the aim is to raise the alarm. That's a good thing to do. But when you overstate the problem, people find excuses to dismiss what you say. Between crying wolf! and moderating everything into meaningless bland bureaucratese there's a sensible middle ground where you present the problem and its solutions in a reasonable and fair manner.

This ain't it.

minor typo:
"simple not possible" >> simply not possible

There's an unstated assumption in the figures for nuclear that I would call everyone's attention to:

Fission of 1 tonne of metal yields approximately 1 GW-year of electric power (depending on the thermal efficiency of the rest of the system).  Actual fission of 65,000 tons/year of uranium would indicate a nuclear electric power production of roughly 65 TW, give or take.

Obviously, actual nuclear electric generation is a tiny fraction of this figure.  This shows that the above figures assume (but do not state) a once-through cycle in thermal-neutron reactors.  Fast breeder reactors could increase the available energy by a factor of roughly 100.  (The resource figures may also assume no increase in the price of uranium, which is only a small part of the total price of nuclear power.  The price of uranium could increase tenfold and not increase the cost of power noticeably.  This would obviously increase the amount available.)

Last, thorium is not listed as a resource.  Thorium-232 is 100% convertible to U-233, which is fissile (only 0.7% of natural uranium is fissile).  Also, thorium is about 4x as abundant as uranium.  The addition of thorium to the nuclear energy base increases the total energy supply by a factor of roughly 400 over the once-through thermal-neutron uranium cycle.

Phoenix and Aeldric,
I appreciate the efforts you have made in these posts.
I have some constructive criticisms/comments:

It's completely unrealistic to not assume that nuclear will play a significant role in electricity production using a mix of generation II and III reactors and breeder reactors using both uranium and thorium. Present known reserves will last X100 your estimates.
I would suggest at least 10% of the 198EJ of electricity should be from nuclear.

Secondly you seem to be over-estimating steel production or energy use, since today's steel uses less than 1Billion tonnes coal, less than 25EJ. This includes very high consumption from China with almost no recycling in China. In 40 years the one child policy will mean a lot of recycled steel and a lower steel demand. For example the US now recycles 78% of steel. So even if steel production increases 300% a lot will be from recycled, and direct reduction using electricity will use a lot less energy( you are not heating 2 tonnes air and coal /tonne iron ore). Even less energy if biomass is used for reduction in electric arc furnaces.

You are using equal amounts of wind and solar although wind is cheaper per kWh delivered. Presently wind is X 5 solar, either solar costs will reduce below wind, hence lower overall costs or wind will be at least X2-X5 solar. Either results in lower costs.

Your 25% capacity factor is only appropriate to N Europe( excluding Scotland which is >30% some farms 40-50%). Off-shore capacity factors are higher and in US average is 35% capacity factor for wind. I would suggest 33% a better world figure.
As more remote sites( better wind) are connected to grid capacity factor will increase. You can see this in UK, first farms in England(lower capacity) now a lot more in better wind locations in Scotland and off-shore islands.( see BWEA.com )Improvements in turbine design are increasing capacity 0.5% per year.

Demand for copper assuming 60% of the electricity comes from wind(20,000GW capacity) would require 20million tonnes of copper, while present copper reserves are 550 million tonnes with another 500Million reserve base. Similarly Nickel would be 1-2 tonnes/MW(20-40Million). Very large reserves of laterite ores exist in tropics as well as low grade sulphide ores. Both can be extracted by low energy cost acid leaching.

The major component of wind turbines is steel, 20,000GW would require 2Billion tonnes ( less than 2 years world production or 5% over 40 years, or less if you are assuming steel production will increase X3 fold)

It's worrying when you see statements along the lines of 'algal biodiesel has to work'. US Energy Secretary Chu seems to be saying 'CCS has to work'. What if they don't?

Hydrogen from 4th generation nuclear could solve a lot of problems. It could reduce iron ore or upgrade low net energy biofuels. I agree we have to put everything on the table like it or not. It may come down to 4th gen nukes vs dieoff.

It is a bit of a nitpick, but thermal storage with solar thermal might decrease overall cost. The basic assumptions are
total cost = C + S + T
Where: C is the cost of the collectors (pipes plus mirrors)
S is the cost of thermal storage.
T is the cost of the turbomachinery and power conditioning.
Clearly; C scales linearly with planned annual output (J/year)
S probably scales with the thermal storage capacity.
T scales with the nameplate capacity.
The idea is to overbuild the collectors, i.e. the turbines are underpowered wrt. the capacity of the collectors, the excess is stored. Then the turbines can continue to run into the night. If the turbines are a significant fraction of the overall cost, then overbuilding the collectors, and adding sufficient thermal storage can cut the cost per joule. In any case, this price optimization is not likely to be a huge factor, but at least it provides an avenue for reducing overall cost while simultaneously smoothing at the supply temporal variations.

Since no one has done any ad hominens, I'll start: "Nanny nanny boo-bo"!

" * Reduction in energy demand through conservation
* reduction in energy demand growth through population limitation/reduction
* continuing the limited access to energy to large portions of the world population

* breakthroughs in energy conversion efficiency
* large new resource identification
* the roll out of renewable energy generation raised to a war footing for all industrialised countries"

I noticed that the last three of these are essentially wishful thinking.

I also noticed that it hasn't taken long for the nuke enthusiasts to come out of the woodwork. So let's again reiterate:
--too dangerous in a world characterized by a global war on terrorism
--too expensive in a world in the throws of The Greater Depression
--too hazardous to current residents near mines, near roadways carrying fuel and waste, near plants...
--too hazardous to posterity
.
.
.

The first, second, third....Nth priority has to be "reduction in demand."

Education, taxes, rationing, anti-consumption ads, curtailment of consumption, intentional recessions and depressions, aggressive and effective population control starting with the biggest consumers, end to corporate (super-) person-hood...Total and rapid powerdown at all levels.

These, rather than focusing on technofixes, are the necessary approaches we have to carefully consider then rapidly implement.

First, an excellent article. My few nits are essentially opinion.

1) I agree with above post, that solar-thermal-with-thermal-storage has greater potential than offered. Serious unit cost reductions will be achieved by volume installation and advancing technology. NREL / Sargent and Lundy's engineering paper proposes wholesale electricity out costing between $0.035 and $0.062 / kwh after the first 8 GW are built.

2) I think solar PV will cost a LOT less than $5,000 / kw once a market the size proposed is formed. eg. Optical Rectenna has been proven at laboratory level, awaits advancements in nano-scale diode fabrication which can be assumed definite within this time frame.

3) Some of the newer wind generation technologies should be capable of significantly reducing those costs / kwhr out, and possibly percent availability eg. IF WindJet's new turbine proves real, they're discussing useful generation with winds from 1 kmph to 100 kmph at 4x power density per unit land area.

4) On the time scale of 40 to 50 years, it can be taken as certain that SOME novelties we don't yet even envision will crop up. That could be Lighter-than-Air wind generators floating in the jetstream, ocean current generators, ocean differential density exploiters, etc.

5) TOU metering + realtime pricing COULD seriously help to enable intermittent generation, eg. by STRONGLY ENCOURAGING / incentivizing optional useages such as PHEV charging, air conditioning with ice storage, etc. etc. to happen only when the wind is blowing and / or the sun is shining. Needs overspec grids, but that can be handled as well by incentivising load levelling the same way. Mainly needs long-term vision and planning.

Overall though, your articles do an excellent job of laying out the scale of the real issues as accurately as can be done from a scientific / egineering perspective. Kudos. Hopefully in your next, you can address the REAL problem, which is "We know it CAN be done, but WILL it be done? What obstacles there?". I'm guessing that the time frame available is somewhere within the period of exhaustion of Natural Gas resources in each of the major markets, N America, EurAsia, M East and Africa and India, S. America. To my knowledge, N America has only about enough gas (including tight excluding clathrates) to last at present rates of consumption about 40 yrs. Assume LNG can bridge Latin america that far as well. That means that we need to be AT LEAST well into this plan before 2050, and better would be to have completed it by then so we can keep some fossil fuels for other uses for posterity. Will incumbent suppliers allow ANY transition to start before ALL N. Gas is gone? I remain sceptical.

Too much negativity.

Australia's electricity consumption = 1 quad = 300 Twh
Assume 50% electricity from hydrogen storage (150 Twh used in the daytime) and the efficiency of hydrogen is 25% per Ulf Bossel so 600 Twh in hydrogen storage.

Australia's oil consumption = 2 quads = 83 million tons of hydrogen.

It takes ~60 Mwh to make by electrolysis and compress a ton of hydrogen or 5000 Twh.

A 25 kwe Stirling Systems Suncatcher gets a daily 2 kwh/m2d or 730 kwh/m2a.
http://www.stirlingenergy.com/advantages.htm

5000 + 600 + 150 = 5750 Twh
5750 Twh/730 kwh/m2= 7.877 billion m2 or 7876 km2.

Area of Sydney Australia = 12058 km2

So cover the dear old outback with .6 Sydneys of Suncatchers and you're done.

You can thank me later.

The infrastructure required for provision of biofuels will be a significant challenge, hence the limitation of this fuel source for all uses except where it is irreplaceable because of energy density. The production of biofuels on the scale required is unprecedented. Production of sugar or grain based ethanol for this volume could not be contemplated. It appears that the only viable biofuel at this level of production will be production of algal based biodiesel. Research has indicated that this form of biodiesel production will involve plant capital costs in the region of US$6.5 Million per megalitre of production capacity. The capital cost of the infrastructure to produce 5,800 GL per year will therefore be approximately US$37,900 Billion.

Whoops???

Let's go through this in detail.

$6.5 Million per megalitre of production capacity.

A megalitre is 1,000,000 liters, or about 5,400 barrels. I presume that we're talking about annual production capacity, so a "megalitre of production capacity", would produce about 54,000 barrels over 10 years (a conservatively long pay-back/capitalization period).

So, we're talking $120 per barrel before any operating or maintenance costs - these would bring it up to at least $150 per barrel.

Why would anyone pay that much? Long-haul trucking would be entirely replaced by rail at that price. Water shipping would go to batteries and fuel cells. Personal transportation would go to EVs (and some PHEVs that used liquid fuels for perhaps 5% of their motive power). All that's left is air transportation.

5,800 GL per year? That's 32M barrels per year! I think we'd probably use less than 5M barrels per year at that price.

Does anyone suggest that in 50 years we'll have less than 20M of liquid fuels from remaining conventional crude, NGLs, tar sands and biofuels? They'll all be far less expensive.

•Although the cost has been based on a range of assumptions concerning energy technologies, it is unlikely that a different mix of conversions or replacement technologies would greatly affect the bottom line price.

Moving to a mix of, say, 50% wind, 10% nuclear/wave/geothermal, 20% CSP solar and 20% biomass wouldn't affect the costs?? Wouldn't it cut it by almost half?

•The figure quoted only represents the major energy production plant required. In parallel with this will be a similar cost associated with the changes made by energy users (i.e. electric vehicles, mining and manufacturing equipment, rail lines, power transmission, metal smelters etc. etc....)

Not over 50 years. The costs would be negligible, as things were replaced as needed. The costs would come from replacing things before replacement was needed, which would only come from a much faster deployment. Incidentally, I would strongly advocate a much faster deployment to deal with CO2, but that's different.

•The cost assumes a single transition from the current energy production infrastructure to the final renewable infrastructure. This won’t happen. As successive governments are driven by the need to maintain the power on and the fuel tanks full there will be a staged series of interim technologies implemented. Depending on the quality of vision of political and industry leaders these interim technologies could consume as much or more than the cost indicated for the final conversion.

No. Renewable electrity and efficiency are the best candidates, and just about everyone is clear on that.

Hydrogen is already on the way out. Nuclear is still being debated, but it will work just fine to the extent it's deployed. Algae is the only thing I can think of that might fit this description, and I can't imagine it will go anywhere if it can't prove itself. Other biofuels may cause quite a bit of environmental damage, but their direct costs won't be especially high. I can't imagine what you're thinking of.

Given the above considerations it is likely that the total costs associated with transitioning the world to the fully renewable economy will be in the order of US$ 150 Trillion.

The base figure for generation is way too high. The algae proposal is highly unrealistic. This final figure of roughly $50T is completely unsupported.

I appreciate you're giving this a try, but this is barely a pre-draft outline of an analysis.

I ask you to consider Australian experience with the World Solar Challenge, where I participated in 1996. Solar panels for these race cars have recently been downsized from 8 sq meters to 6 sq meters because they were going too fast!

If humanity can't figure out how to get away from the Internal combustion engine at ~20% efficiency and cars which weigh 2 tonnes to carry a single person, we should give up and go back to the 18th century. Considering that the goal is to move people, not metal, then that 20% * 100 kg / 2 tonne = 1%. It's almost as bad for trains and buses. This is no way to bungle into the future, and these models must incorporate better technology assumptions to overcome the absurdities that result from imagined direct oil substitutes.

The USA uses about 25 Quads = about 25 ExaJoules for transportation. Cutting that back to 5 Quads / Ej is feasible, and obviates the absurd notion of making a replacement fuel that perpetuates the automobile which is effectively 1% efficient, while also being responsible for ~ 1 million deaths per year (40,000 in the USA) plus countless injuries.

In California we are on course to build a pilot 100% solar powered non-linear personalized public transportation system that will pay for its energy source (solar PV) in 4 years without any subsidies. http://www.solarevolution.com/.

Macro analyses must incorporate dramatic reduction in energy usage resulting from disruptive technologies. We have done energy / climate / depletion models for certain countries using methods similar to yours and welcome the opportunity to compare notes. Yes it's going to cost money. But spending it, with due consideration of EROEI and other critical metrics, before BAU destroys our basic infrastructure, will save trillions.

Any combusted liquid (or gas) fuel in a personalized vehicle, though plausible in rural areas, is the worst possible BAU in the urban context.