The situation we’ve found ourselves in 2020 provides a plausible rapid path to eliminating nearly all carbon emissions from electricity generation in the US, and to electrify much of the rest of our energy use. However, these emissions won’t disappear due to business as usual — especially not at the rapid rate needed to address climate change. The US will need a policy of decarbonization. We have the technology and it is cheap enough — now we need to make a plan, elect politicians and push them to carry it out, and spend some money.
In this post series, we’re going to talk about two engineering-centric plans for electricity and energy decarbonization: the 2035 Report from UC Berkeley's school of public policy and the Rewiring America plan1 from an organization led by Saul Griffith, a MacArthur-winning engineer and inventor. These two plans are largely complementary with the several high level proposals on climate put forward by Democratic party (the 2020 DNC platform, the Green New Deal, and the Biden/Harris plan). These two reports contain a lot of the same ideas, but are more focused on power systems and provide a more detailed accounting of the engineering2. These posts are centered around the electrical system — because I know it well and because it’s surprisingly important to decarbonization. Don’t worry if electricity isn’t your obsession — there are lots of other interesting things to talk about along the way.
Climate and energy are important topics, but it can be hard to tell what’s being proposed from the political conversation and news coverage. We’ve known climate change is a problem my whole life, but policy proposals to seriously deal with it are very new. The idea of a Green New Deal is just two years old — and there’s so much to do, so the details are still being filled in. All this can leave decarbonization feeling mysterious and maybe a bit hopeless to many people - including me before starting on this project. In writing this series, I’ve learned a lot about the details of what’s being proposed, how it might work, what are the hard parts, and what it all will look like in real life. More importantly, this understanding has given me some small hope that we can have a livable climate. The more people that understand these systems and the hopeful possible future we can build, the better — so please join me in this exploration of the exciting world of clean energy.
The 2035 Report essentially says that we can take the current US electricity system and remove 90% of carbon emissions over the next 15 years by adding only renewables and storage.
This 90% clean power system would:
Would renewables work everywhere (not just where it’s really sunny or really windy)? Yes: it’s possible to build enough renewables in every region in the US and maintain a dependable grid (aka keep the lights on). This transformation will only happen with federal policy intervention — they recommend a nation-wide 90% clean energy standard.
Ok, those are some big conclusions. There’s a lot to feel excited and hopeful about in those numbers — and the electricity nerd in me is intrigued by the dependability finding. What they did to figure all this out is pretty cool. First, they did a bunch of sensitivity analysis on prices and policies. They looked at wind and solar resource maps, costs to build renewables, interest rates4, natural gas fuel costs and the difference between a federal policy of 90% clean vs no policy. After all of that work, they had a bunch of hypothetical power systems, modeled at a level of detail similar to what we use to run the real system today. (Impressive!) With this, they set out to answer two questions: “what will electricity cost?” and “can we keep the lights on?” with mostly renewables.
To check system dependability5 they run a bunch of grid simulations (called “unit commitments”) for every region in the US6, for seven years worth of different historical weather scenarios. The main goal of unit commitment is to choose the cheapest available sources of power for each hour of grid operation. It also lets you know if your system doesn’t have enough power to meet demand. And electricity costs also get figured out as they do these simulations.
I worked on something roughly similar for my PhD thesis, but just for one region - they did the whole country! I have a lot of respect for how much work this portion of this study involved. Their work is just what I wanted to see7 to be convinced. The conclusion of grid dependability at 90% clean energy - in every region in the country - is a pretty big deal. It’s a dramatic shift from the studies being done in 2010 (when I started work on my PhD). Ten years ago, both utilities and academia were expressing uncertainty about what would happen with even moderate amounts (say 20%) of renewables on the US grid8. We weren’t even thinking about the question “what would happen at 70%”9. The technology and cost structure of renewables and storage in 2020 has come a long way in a decade. (See my decade in review post for more details.) We really do have the technology decarbonize our electricity quickly to address climate change now.
What do we build? What do we do? After all of that modeling, the study suggests that the US power system needs to build out about 1,100GW of new renewable capacity and about 150GW of battery storage. (Not sure what GW means? See my explanation: “What’s a watt”). For this 90% system we need to build NO new fossil fuel power plants. And better yet, we retire all of the US coal plants by 203510. There are about 250 coal plants running in the US today and this would mean shutting them all down!
To put the renewables numbers in context, we currently have about 130GW of renewable capacity. One modern windmill is 2MW, and one solar farm is about 5MW (although a few are much bigger). Building out our new 1,100GW of renewables is going to be a lot: imagine about 300,000 wind turbines and 100,000 solar farms. That’s a lot of good jobs and cleaner air.
Getting to 1,100GW by 2035 means building 70GW of renewables every year. This is a fast pace, but one that seems possible. In 2002, the US added about 65GW of natural gas generation — so adding lots of power to the grid in one year is possible. Our best single year for building renewables was about 30GW — we would need a bit more than double that record, every year for the next fifteen years. The good news is that half of this new renewable capacity is already planned but is waiting to be permitted to connect to the rest of the grid11.
The 2035 Report does keep around many of the existing natural gas plants, but most of them will operate infrequently or seasonally: a part time fleet.
The choice to keep about two thirds (360GW) of the existing natural gas plants around through 2035 is clever, but a little bit non-obvious if you’re just flipping through the report. Keeping the gas plants around eases any issues with intermittent renewables significantly — if a cloudy storm system suddenly rolls in and blocks the sun, or there are a few hot summer days without wind, you’ve got lots of idle gas plants sitting around to pick up the slack. And you can avoid the costs of building lots of transmission lines and/or battery storage. (That’s something that other decarbonization studies wind up doing — making decarbonization seem harder to get started and more expensive.)
How much natural gas will get used seems to depend on the region and how much solar it has compared with how much wind. In many places gas will only be used infrequently — about 20% of the gas plant fleet in 2035 will run for less than 1% of the time12. There will be a few places where we’d lean heavier on gas. In the 2035 data explorer, we can see that CA and FL (lots of solar, little wind) continue to get about 25% of their power from natural gas. This gets balanced out by windier regions to get the whole country down to 10% of power from natural gas.
The end goal is to get to 100% zero carbon electricity within a few decades, which means stopping the use of the last natural gas plants. But in the meantime, we can use them to make the transition cheaper and faster.
The Rewiring America plan is even more ambitious. Rewiring America starts with the question: “what is the fastest, cheapest way the US could decarbonize to meet the IPCC target for warming?” They look at not just at the electric grid, but at all US energy use — including gasoline for cars, natural gas used to heat buildings, and many other categories.
They propose a fast transition to a zero carbon electrical grid (similar to the 2035 Report). Then they propose shifting everything else that uses energy from fossil fuels over to use electricity instead (aka electrification). For example we’d switch from gasoline to electric vehicles and from natural gas furnaces to electric heat pumps.
Transitioning to a decarbonized and electrified energy system would:
They also estimate that we can do all of that with total government spending of $3T over ten years16. And their plan involves using only existing commercial technology that we already have — no breakthroughs in carbon capture17 or nuclear fusion required.
Sign me up, right?
Well, Rewiring America also concludes that this won’t happen without some pretty major policy changes. (You’re probably like: yeah, duh.) And they find that the only sufficient policy at this point is a zero carbon technology standard that requires a 100% replacement at end-of-life. This means whenever a machine that uses fossil fuels wears out, it always gets replaced with a machine that produces zero carbon. But before we get into the “how do we make this happen” part, let’s talk about the story of what this can look like.
Rewiring America starts with a top-down view of all US energy use (electricity, natural gas, gasoline, etc) and asks “how much of this could we replace with zero carbon electricity?” They take a classic engineering approach — look for the biggest and cheapest components to fix first and work your way down the list. Their answer: if we electrified the entire electric grid, all vehicles, and all buildings18 with zero carbon sources we could reduce overall US emissions by 85%. That’s a big drop — enough to meet our climate goals, if done quickly.
Their next question was “what policies would we need to do this in time?” The charts above are an illustration of the policy choices Rewiring America considered, and their climate implications. The chart on the left shows how each policy choice results in some rate of adoption of zero carbon technology. The chart on the right just translates this adoption rate onto the US contribution to climate change. We’re the orange line right now: the “free” market choice we are currently making has slow market adoption of clean energy, and leads to a pretty uninhabitable planet (in the red zone on the graph at the right). The first three policy lines (yellow, green, and light blue) are more for reference — they aren’t things we have the ability to do. (You don’t have a magic wand, right? And we can’t start replacing things at 100% until we can build enough of replacements).
Rewiring America came to the conclusion that a production ramp and 100% replacement policy (the darker blue line) is necessary to stay in the 2°C warming zone (the orange band) only after considering all other possible policies and finding them insufficient. The other policy outcomes are just so bad — all of them would put the US contribution to warming well over 3 degrees, with no future in which the US reached zero emissions. 100% replacement at end-of-life might sound like an impossible policy — the authors acknowledge it will take both a wartime-level mobilization (the “production ramp”) and very good policymaking. But each additional half degree of warming is incredibly bad, so let’s dig a bit deeper and see if this policy could work.
After seeing that we need to electrify everything, Rewiring America set out to figure out just what “everything” involves. They did the bottoms-up calculation of what needs to get electrified at a household level (the furnace, the water heater, the car, etc), and then looked at the zero carbon technologies we have today and figured out how much it would cost to use them instead. They came up with an average additional19 spending for electrification of $40k per US household20. They also make a bunch of estimates for electrification costs in the commercial and industrial sectors.
Who has that much money lying around? Rewiring America proposes financing this additional household cost primarily with low-interest, government-backed loans21, combined with grants for low income households. Structuring these costs mostly as loans means that the total government spending is lower — but you could also design a plan where the government did more grants and fewer loans for household electrification costs. Businesses would get slightly different incentives. The next post in this series will explain these policies and how this money would get spent in more detail.
Specifically for the electric grid, the US will need to produce 3-4x more electricity22, which would require adding about 4TW of new renewable capacity23. This is a massive amount — for comparison, the 2035 report calls for building about 1TW of new renewables. Rewiring America is also excited about distributed generation — calling for about 25% of the amount above (1TW) to be rooftop or community solar.
Would an electricity system like that even work? I have a gut feeling that it would — but we need to do some work to prove it (basically a scaled up 2035 study). The bigger challenge is going to be building it quickly.
The 2035 Report makes it clear that we have the ability to quickly build a clean US electricity system that’s cheap and reliable everywhere. Their research on the power system is detailed and convincing. Rewiring America looks at our options for avoiding catastrophic climate change and comes up with a hopeful high-level plan. They take the 2035 clean electricity system and scale it way up. They find that clean electricity could run almost everything we use energy for today and put us on track for a livable climate. This plan is very ambitious but we have all of the technology we need to get started and it’s surprisingly cheap.
Clean energy — and a liveable climate — won’t happen without policy changes. The next post in this series talks about what those policies are, who can make them happen, and how much they would cost.
Thanks again to Amelia Greenhall for lots of editorial work on this post.
Details on the Rewiring America plan can be found in the organization’s “handbook”, a preprint of a longer book slated for publication next year. They also have a Jobs Report paper which is more technical and economics focused. ↩
Although both reports do not provide details at the level you’d expect to find in an academic paper. This should get fixed with time - both organizations have intentions of publishing more technical details of their plans. They both seem to have released reports with an eye towards influencing the 2020 Democratic policy platforms - and policymakers are unlikely to be interested in the engineering details. ↩
To be fair, they estimate that electricity in 2035 would be about 5% more expensive under the 90% clean case vs the business-as-usual case - but only if you don’t include any environmental or carbon costs. Even with that small difference, it would still be cheaper than electricity today. Even under the most pessimistic scenario (high costs for technology, financing, and natural gas - a pretty unlikely trifecta), a 90% clean energy policy would result in 2035 electricity prices that are just 10% higher than they are today. ↩
Utility companies are traditionally regarded as low risk investments and can borrow at good rates. But they are taking out multi-million (or multi-billion for large fossil fuel plants) dollar loans over 30 year time frames - so even small interest rate changes matter. The 2035 Report’s base case interest rate of 3.7% - set pre-pandemic no doubt - is probably now a high estimate (the federal interest rate is 0.25% as of Sept 2020). ↩
The 2035 authors were careful to always say “dependable”, and not “reliable”. Aren’t they the same - to-mae-to, to-ma-toe, right? In the electricity industry, “reliability study” refers to a specific kind of study that power systems engineers perform. It’s focused on detailed “stress testing” - simulating a power system’s ability to very quickly (seconds) respond if there is a sudden transmission line or generator failure. The goal is to make sure the system can avoid large blackouts. The 2035 Report did not do a reliability study - for good reasons. They require even more information about the power system and are hyper-specific - on a hypothetical system 15 years in the future, a reliability study wouldn't mean much. ↩
More precisely, they model every region in the continental US. Modeling isolated power grids like Hawaii and Alaska is a niche expertise. And it probably doesn’t affect the overall numbers much. ↩
I’ll be excited to read their more detailed peer-reviewed academic papers about it too. ↩
In many ways this uncertainty about the future came from a 2010 grid where renewables produced only 2% of US electricity - it was just hard to imagine even the ~10% world we’re in today. The grid of 2010 had essentially zero battery storage, and studies on the subject were limited to forward-looking academics or the National Renewable Energy Lab (NREL). The 2035 Report cites an NREL paper from 2012 that found that getting to 80% renewables would about double electricity costs - a reasonable estimate given renewable prices at the time. The report later notes that today’s renewable costs are already lower than where NREL (in 2012) thought costs would be in 2050. Additional caution came from utilities who were not used to thinking about “non-controllable” sources of power. Regions that did have some early experience with renewables were dealing with high concentration in one area (the Columbia River Gorge in WA/OR) and conflict with slow-ramping coal plants (Texas). Dramatically cheaper renewables, the appearance of grid scale storage, the decline of coal, and the construction of more geographically spread-out renewables have eased many of the problems and worries of 2010. ↩
In the clean policy scenario renewables generate 70% of electricity in 2035. Hydro and nuclear plants generate an additional 20%, leaving 10% from gas. ↩
Coal generation might be over by 2035 without any policy changes if natural gas stays as cheap as it is today. The 2035 Report estimates that with cheap gas and no new federal policy, the share of electricity generated by coal will decline to just 3.5%. ↩
This permitting is called an “interconnection queue” and reforming the cost and regulatory inefficiency of the process is one of the wonky policy recommendations the authors have. Regulation and cost in this case are related - society will need to figure out how to allocate the $100B (by their estimate) of building new “spur” transmission lines just to connect these renewables plants to the existing grid. The current system often assigns these costs 100% to the potential wind or solar farm operator, making the plants less likely to get built. If there is a utility on the other end of the line hoping to buy that cheaper renewable power, this transmission cost allocation hurts everyone. ↩
These gas plants may need additional compensation to remain open with such infrequent operation. Imagine running a business that sells stuff on just three days per year, but needs to pay salaries and rent all year round and be oncall 24/7/365. It’s not an engineering problem per-se, but a financial and political one. The 2035 authors also suggest that in the future grid we are likely to find a cheaper way to deal with this problem. ↩
They argue that there is a feasible path from 15% in 2035 to net zero carbon by 2050. It involves some advances in technology in areas like aviation fuels, soil management, and industrial processes like steel and concrete. ↩
If you’re wondering how we could possibly save this much money on energy - remember this amount includes not purchasing gasoline for a car and not buying any natural gas to heat the house (and purchasing cheap electricity instead). Total energy spending in the US averages about $4,000 per household - more than half is spent on gasoline. ↩
during the mobilization period, eventually tapering down to 5M sustained jobs ↩
These amounts might sound like a staggering amount of money for the government to spend - but the March 2020 coronavirus relief bill alone cost the government $2.2T (in one year!) and Republican tax cuts in 2017 will cost the government another $2.3T over ten years. In comparison we could spend that kind of money to do something useful - ensuring the US transition to 100% clean energy happens quickly and cheaply. ↩
Rewiring America makes the alarming point that all of the IPCC’s estimates of warming include the assumption that we will be able to sequester significant amounts of carbon at some point in the not-too-distant future. Rewiring America argues that this is highly unlikely, given the physics and current cost projections. ↩
This might seem like everything - but it leaves out a lot of smaller and also technologically trickier or socially messier parts of US energy use like jet fuel, making concrete, agriculture. Rewiring America has some hopeful technological ideas on this last 15% of energy use, but it’s not covered in detail under the plan. ↩
Note this is not total spending on new cars / water heaters / etc. It’s just the additional spending - the difference between buying a zero carbon machine and one that does produce carbon. As an example, they estimate that the additional cost of an electric car is about $6,000 (that is $6k more than a gasoline car). ↩
Note that number is an average - not every household will need to spend $40k. For instance, if you have zero cars instead of two - the American average - congratulations, your costs would be $14k less than the average. ↩
Conveniently, mortgage payments on a 30 year mortgage for $40k with a 3% interest rate comes out to be about $2,000 a year. That’s in the ballpark $1-2000k/year average household savings on energy that Rewiring America. So, depending on interest rates the average household could save money by electrifying, even with a loan. ↩
The 2035 report also assumes a modest amount of electrification happens, leading to 13% more electricity demand in 2035. That’s much less than the 3-4x proposed under the Rewiring America plan. ↩
They also include 100GW of new nuclear capacity in their plan. Nuclear plants do provide good baseload and 100GW isn’t a small amount (especially if you had to replace it with 3x the amount of renewables or more). But I’m skeptical that the US can build that much nuclear power in the next ten years at a reasonable cost. In interviews, Griffith makes it clear that he isn’t 100% attached to the idea. For the rest of the discussion I’m going to ignore it and assume that it doesn’t impact the conclusion of the report significantly. ↩