Chapter Five


Whistling in the dark with the lights on

The Roadmap has a slick graph that depicts a completely unrealistic buildout schedule. It calls for more than half of the buildout in the first ten years (2015–2025), and another 25% of the buildout in the following five years.1

That's 15 years to build more than three-quarters of a $15.2 Trillion, nationwide, fuel-free renewables grid.

The 35-year Roadmap would entail manufacturing (or importing) and installing:

  • 496,000 5-MW wind machines
  • 18 billion square meters of PV panels
  • 50,000-plus wind and solar farms
  • 75 million residential rooftop systems
  • 2.7 million commercial rooftop systems

On 131,200 square miles, not counting the rooftops.2

Nuts and bolts

Yes, we stepped
up for
World War II, and yes we can do it again.But can we keep it up for 35 years?

To execute the Roadmap, the entire country would have to shift to a war footing and stay hard on it for over three decades. And like we said, if other countries follow suit and overseas fabricators can't fill our orders, we'll have to make our own gear.

Which is a lot of stuff. Yes, we stepped up for World War II, and yes we can do it again. But can we keep it up for 35 years?
And do some of it two or even three times over? Because remember, the buildout will last longer than the wind turbines, and nearly as long as the
solar panels.

So even when the buildout is complete, it'll never end.

Like our military-industrial complex, born in the cradle of WW II and still going strong, we'll have to keep fabricating, installing, and recycling 1.23 million square meters of PV panels every single day – forever – just to keep the Roadmap working.3

Fabricating and installing that many panels each day would be difficult enough. Recycling the old panels that the new ones replace would become a polluting, resource-intensive industry unto itself, involving a series of mechanical, thermal, and chemical processes, each with its own energy requirement and waste stream.4

And don't forget, we'll also have to refurbish all of the 340,000 onshore turbines every 15–20 years (gearboxes, generators and blades), and do the same with the 156,000 offshore turbines every 10 years because of their harsh marine environment.

The U.S. doesn't have anywhere near the industrial capacity to get this done.

For example, just to stay on-track with the Roadmap's second 5-year portion (the period 2020–2025), we'll have to exceed our best year ever in PV panel production by almost 29X, and our best year ever in wind turbine production by nearly 17X, based on U.S. production totals for 2016.5

Dozens of factories will have to be built overnight, and we'll have to run them three shifts a day. That may seem like a good thing, since it would be a national jobs program that can't be beat.

But it could also amount to biting off more than we can chew. Because if we can't ramp up that much and that fast, we'll find ourselves hemorrhaging money with nothing much to show for it.

Public morale will falter, and the mobilization will seem more like the home front during the Vietnam War than World War II, with all the political turmoil, protests and culture wars that came with it.

And keep in mind, the longer it takes to get mobilized, the more those Xes go up.

So despite the optimistic curves in the Roadmap's graph, the buildout will actually be a constant scramble to catch up for three exhausting decades.

As of this writing (autumn 2017), we're already two years behind schedule.

Low energy? You might have a mineral deficiency

Copper and silver are just two of the critical minerals used to fabricate wind turbines, PV panels, and the parabolic (curved) mirrors for CSP solar.

We currently import a third of our copper and most of our silver. Imports would necessarily skyrocket if we make our own Roadmap gear. And even if we had the equipment made overseas, those countries would still have to mine or import the material themselves.

So how much copper and silver would we need for our nation's Roadmap?

The copper industry says that PV solar needs about 5 tonnes per MW, and wind turbines need about 3 tonnes.6 Panel makers say they'll soon be reducing their silver consumption to 13 mgs (milligrams) per dc watt.7 By pure coincidence, CSP's parabolic mirrors also need 13 mgs per ac watt.8

Doing the math, the U.S. Roadmap will need 24.4 million tonnes of copper9 and 51,300 tonnes of silver.10 And that's not counting all the copper for the tens of thousands of miles of new transmission lines. Or mirrors for CSP backup systems.

We'll assume that all the copper and silver in our worn-out panels and turbines will be recycled for the new panels and turbines needed to maintain the Roadmap.

Regardless, our sudden increase in demand, along with the decline in ore grade that typically occurs with each new dig, would result in rising prices and bottlenecks around the world.

The material, however, does exist, even if it doesn't exist here. So the U.S. Roadmap, in theory, could actually be built. But there's a catch:

If the Roadmap goes global, the worldwide buildout will consume about one-third of the world's proven copper reserves11, along with 90% of proven silver reserves12 – meaning the copper and silver that we know for sure is still in the ground.

New silverware and silver jewelry would have to be banned. And mirror technology would have to be revamped – silver, the best reflector of visible light, has been used for centuries. The list goes on: Electrical contacts, batteries, printed circuits, etc.

At the current rate of silver consumption for all industrial products that aren't solar panels, the world would blow through the final 10% of global silver reserves in 4 years. Entire product lines would have to be re-thought. Things will change bigly.

Monopolizing one-third of the world's copper would be just as bad, putting a serious kink in global supply chains and jacking up prices around the world. And we haven't even factored in the transmission wires to connect the hundreds of thousands of new wind and solar farms to their respective national grids.

A global Roadmap would quickly become a victim of its own excess – strip-mining the planet, and carpeting it with wind and solar farms, is not going to save the world. Or us.

The 1,591-GW grid*

(*Batteries not included. Backup is optional at extra cost.)

The Roadmap contends that an all-electric grid could power the nation­ ­– electricity, transportation, heating, industrial processes, the works – with an average (not nameplate / peak) capacity of 1,591 GWs.

We'll take the estimate as a given.

If everything goes according to plan, smart grid technology will manage all of this extra juice (about 3.4X of what's now on the national grid) by sending power to wherever it's needed on a second-by-second basis, adroitly balancing our national supply and demand.

The Roadmap also recommends using LoadMatch, a grid integration computer model, for predicting the amount and availability of power every 30 seconds across the entire grid.

Sounds amazing, but we have our doubts, because no matter how precisely the grid is managed, it'll essentially be a fuel-free system with virtually no backup or storage, and entirely dependent on our ever-changing weather.

Even more amazing, the 1,591-GW average was derived by simulations that Dr. Jacobson and his colleagues had LoadMatch perform for the years 2050 through 2055.13

Think that through: A $15.2 Trillion national WWS

Peering into the
future through
a 35-year
fog bank, and
claiming to read the details of a distant shore, takes a certain amount of chutzpah.

buildout, embraced by millions of renewables advocates, was determined with the aid of a computer model that purports to predict the nation's weather, region by region (not the climate, mind you, but the weather), every 30 seconds . . .

For a 6-year period 35 years in the future. 

Future trippin'

Peering into the future through a 35-year fog bank, and claiming to read the details of a distant shore, takes a certain amount of chutzpah.

Nevertheless, the authors of the Roadmap are confident that a fuel-free national grid is not only achievable, but predictable to the gigawatt.

While computer modeling is improving by leaps and bounds, the accuracy of any model's output depends upon the accuracy – and applicability – of the input.

The only way to make accurate long-term weather

...any long-term bets on the weather
are long shots at best.

predictions is by extrapolating historical data, and that
data is proving to be less and less applicable as climate change disrupts our weather patterns.

Which means that any long-term bets on the weather are long shots at best.

Looking into the past to see the future only works if baseline conditions remain largely intact. But global weather conditions are becoming ever more unpredictable, and doing so at an ever-increasing rate.

Smart grid technology and LoadMatch will supposedly enable us to build up to the grid capacity we need, then add on a mere 4.38% overbuild (69.7 extra GWs) and call it a day.

So much better than the 150% overbuild14 the U.S. resorted to back in the dark days of the 20th Century, before computers made everything run like a Swiss watch . . .

We disagree.

If backup is like training wheels, then overbuild is like spare tires. And anyway, a Swiss watch runs like a Swiss watch without any help from a computer.

Overbuild, as distinct from oversize (yes, there is a difference)

Oversize has to do with power plants. Overbuild has to do with the entire grid.

We walked you through oversizing, which is a new thing in the energy business. Before renewables came along, a power plant was expected to produce exactly what its nameplate said when the thing was tuned up and running at full capacity: A 1-GW plant has always been relied upon to crank out one gig, on demand.

Even so, we still built a lot more power plants than we strictly need, just for just in case. Using thousands of "always-on" baseload plants (coal, gas, hydro and nuclear) we built a 1,167-GW national electric grid – not primary energy, mind you, just electricity.

That's an overbuild of 2.5X our annual average electrical demand of 467 GWs.15 Another way of saying it: Our current safety margin is 150% above demand.

That's how we've kept the lights on 99.9% of the time for more than a century.

Call it overkill if you like, but the idea is sound. So is the idea of converting to an all-electric society. Better living through electricity! Go, USA! However . . .

If you're driving into unexplored territory, you're probably going to pop some tires. Reliability rules, and overbuild is a low-tech, nearly foolproof way of getting down the road. Buy the best tires you can afford, but always carry a spare. Or two. (Even armored limos with run-flat tires carry a spare.)

But the Roadmap chucks all of that Nervous Nelly stuff out the window, because LoadMatch. Which is why the Roadmap's total grid overbuild (as distinct from oversizing each farm) amounts to 69.7 GWs, or just 4.38% above and beyond the basic 2050 grid.

That's not an overbuild of 4.38 times, mind you, but 4.38 percent.

For an interdependent – and weather-dependent – fuel-free national grid, into which we'll plug every blessed thing in the country. And all of it load-balanced to a T with a computer program, and a dinky little 69.7-GW spare tire for good luck.

Green elephants with training wheels

Back in the day, before elephants were on the endangered list, they were sometimes used in metaphors for comic effect: When a person was drunk they saw pink elephants.

A white elephant was something you wouldn't dare get rid of, even though it was utterly useless and destroyed your finances. The term comes from a time when Thai royalty would gift the rare creatures to especially annoying patrons. The patrons couldn't refuse a royal gift, even though they knew it would ruin their lives.

Utility-scale wind and solar farms could easily become green elephants, even with an endless supply of free "fuel" gifted to us by Mother Nature.

She's mighty annoyed by what we've done to the planet, so we're atoning for our sins by humbly accepting her bounty – no matter how inefficient and wasteful a national renewables grid may prove to be.

That may seem over the top, but we wanted to get your attention, to emphasize an important distinction between a fueled grid and a fuel-free grid:

  • If we launch a buildout of fuel-powered baseload plants (coal, gas or nuclear) and abandon it halfway through, we would still have a collection of fully functioning, independent power plants.
  • If we launch the Roadmap and abandon it halfway through, we would have a herd
    of green elephants that will always need training wheels.

All or nothing

The Green Elephant Scenario is one of the biggest

The Green Elephant Scenario
is one of the biggest drawbacks of a 100% national
WWS grid:
It's an

drawbacks of a 100% national WWS grid: It's an all-or-nothing proposition. Because that's what interdependency is all about – it only works if all (or nearly all) of the pieces are in place and functioning.

If we start the buildout, we'll need to complete the entire project to ensure that each renewables plant has the best possible chance of having enough fuel-free backup.

For that to happen, tens of thousands of wind and solar farms will have to be placed in the widest possible variety of advantageous weather zones. And they'll all have to be
completed, or alternative sites will have to be found.

And then, even if we do build the whole thing, the Roadmap may still not prove to be fully self-supporting. It's entirely possible that training wheels in the form of traditional fueled power plants will still be needed.

We won't really know if the Roadmap will work as advertised until we actually build it. And once we do, we'll have to make it work. The reason is simple:

We can't afford to waste that much money, time, land, and resources, then change our minds and move on to something else.

Aside from using fast-start gas turbines or traditional baseload plants that can operate 24 / 7, and aside from oversizing every wind and solar plant we build, the only reliable way to back up the inherently unreliable performance of renewables is with mass energy storage.

We walked you through P2G. In the next few pages,

24 hours of energy storage
could easily cost
$7.6 Trillion.

we'll be addressing energy storage in the form of grid-scale batteries and pumped hydro. We'll also cover hydrogen, which is being considered as a carbon-free fuel for heavy transportation and process heat.

The bare-bones Roadmap treats adequate storage as an externality. Which is one way to keep the sticker price down: 24 hours of energy storage could easily cost another $7.6 Trillion. The price chart deserves another look:

Of all the WWS plants called for in the Roadmap's 1,591-GW plan, only 7.3% of them (116 GWs) will have their own on-site storage, and it'll be just enough to get them through the night. If it was a sunny day.

CSP: Sunshine in a straw

Concentrated solar power (CSP) is a clever solar technology with a bit of built-in storage – just for over night and pretty much just for itself, but it's a step in the right direction. (More of a gesture than an actual step, but still . . .)

Instead of photovoltaic solar panels, which convert sunlight to electricity, CSP uses simple curved mirrors to heat a pipe of molten (melted) salt, which is used to boil water to run a turbine to generate power.

Since molten salt retains a tremendous amount of heat, some of the salt can be stored in insulated tanks to produce power when the sun goes down – if it was a sunny day.16

 If not, then the CSP plant has to be backed up by another plant, or plants. And even if it was a sunny day, the stored energy will only last till morning.

Not to change the subject, but a Molten Salt Reactor's fuel salt is totally different: It stays molten because the atomic fuel in the salt is actively producing heat.

Since there's nothing in CSP's molten salt to generate its own heat, storing its solar energy in salt tanks is like storing water in a leaky bucket – it'll probably be gone by morning. But no biggie, you can just heat up the salt again the next day.

If it's sunny.

Chapter Five End Notes

  1. Critique.

    See table 2, row 9: 2,326,000 MWp-ac ÷ 160 Wp-ac / m2 = 14.5e9 m2 of solar panels.
    To calculate PV land area, divide by packing factor PF = 0.40 (40%).
    Obtain 36.3e9 m2 = 36,300 km2 land area, or 14,000 sq mi for utility PV solar farms.

    Use the Roadmap's assumed wind farm density of 0.089 km2 / MWp-ac.
    Table 2, row 1: Wind capacity 1,701,000 MWp X 0.089 km2 / MW = 151,400 km2 land area, or 58,500 sq mi.

    Combined PV & onshore wind = 14,000 + 58,500 = 72,500 sq mi for wind & PV solar.

    Using the CSP land density of 0.039 km2 / MWp that describes the Andasol CSP farm in Spain:

    Andasol's land area is 5.85 km2. Its nominal power rating is 150 MWp. 5.85 ÷ 150 = 0.039 km2 / MWp.)

    In the Roadmap's Table 2, rows 10 and 11, CSP capacity: 227,300 + 136,400 = 363,700 MWp.
    Multiply by 0.039 km2 / MW to obtain 14,200 km2; or 5,500 sq mi for utility CSP farms.

    Total onshore wind and solar: 72,500 + 5,500 = 78,000 sq mi.

  3. 18 billion m2 of panels ÷ 14,600 days in 40 years = 1.23 million m2 / day

  5. Ibid. Chapter 5 End Note #1 Critique.
    Search for "intends to ramp up our solar".

    Refer to the table “Copper usage in renewal energy generation”. Power values are expressed in terms of peak capacity.

    In Photovoltaics row, columns 2 and 4:
    350 kilotonnes Cu ÷ 70 GWp cumulative installed PV solar = 350e3 tonnes ÷ 70e3 MWp = 5 tonnes Cu / MWp

    In Wind row, columns 2 and 4:
    714 kilotonnes Cu ÷ 238 GWp cumulative installed wind = 3 tonnes Cu / MWp


    Refer to Fig. 8 on page 11, Frame 13. The data point for 2016 indicates 95 milligrams of silver per cell (crystalline silicon technology).
    Assuming power rating of 3.1 watts per cell, the usage of silver is 95 mg ÷ 3.1 W = 31 mg /W.

    The data point for 2026 indicates 40 milligrams of silver per PV cell. 40 mg /cell ÷ 3.1 W /cell = 13 mg /W.

    The reference PV unit is SunPower Co. module E20-435, containing 128 cells. Module power rating = 401 Wdc under PTC (Photovoltaics for Utility-Scale Test Conditions, often referred to as Practical Test Conditions).

    401 W ÷ 128 cells = 3.1 Wdc /cell.

    Its 435 W nominal rating refers to STC - Standard Test Conditions (laboratory).


    If the link does not work, copy and paste URL into browser: "Material constraints for concentrating solar thermal power."

    See table 3 on page 5. 13 tonnes silver per GWac = 13e6 g / 1e9 W = 13e–3 g / W = 13 mg / Wac

    Alternate: (pay-wall unless registered)
    See Frame 5.


    Download the sixth file, “Solutions-US-2015-Web”.

    Print that S-curve.
    It represents the building schedule for US wind, water, and solar equipment between years 2015 through 2050.

    On the right-side vertical axis, combined solar furnishes 45.25% of 1591 GW total US power, or 720 GWavg for PV and CSP solar combined.

    This does not count the supplementary CSP backup solar farms, equivalent to about 4% of standard fleet power, 1591 GW.

    With an expected 21% capacity factor for combined land-mount and rooftop solar, total solar capacity in the standard fleet must be 720 GWavg ÷ 0.21 = 3430 GWp.

  10. The Roadmap's split between PV and standard fleet CSP is 84% /16%.

    So the PV /CSP split consists of 0.84 × 3,430 GWp-ac = 2,880 GWp-ac for PV, and 550 GWp-ac for CSP.

    The photovoltaic cells themselves must have dc power capacity greater than 2,880 GW, to allow for 85% conversion efficiency of the electronic inverters that convert dc to grid-compatible ac.
    The U.S. PV infrastructure must have dc capacity given by 2,880 GWac ÷ 0.85 = 3,390 GWdc.

    Silver consumption for U.S. PV solar cells in 2050 is given by 3,390 GWdc × 13 mg / Wdc = 44,100 tonnes.

    For standard fleet CSP, silver amount is 550 GW × 13 mg / Wac = 7,200 t.

    Total silver for combined fleet solar: 44,100 t + 7,200 t = 51,300 tonnes.


    Refer to page 2. Units are thousands of tonnes. In Reserves column, world total = 720,000,000 tonnes of copper.

  13. Ibid. Chapter Five End Note #2. Roadmap. See the Abstract.
  14. Ibid. Chapter 5 End Note #1 Critique. See internal footnotes 9 and 11.
  15. Ibid. Critique. See internal footnotes 9 and 10.