Some solar farms, like this Recurrent Energy project in Kern County, are being turned off on sunny days. (Recurrent Energy)

Sponsored by

Become a KQED sponsor

Solar energy records are falling left and right in California these days, as the state steams ahead toward its ambitious renewable energy goals.

But the success of solar has brought about a hidden downside: on some perfectly sunny days, solar farms are being told to turn off.

That’s because in the spring and fall, when Californians aren’t using much air conditioning and demand for electricity is low, the surge of midday solar power is more than the state can use.

It’s becoming a growing concern for those running the grid at the California Independent System Operator. At their Folsom headquarters, a team continually manages the power supply for most of the state, keeping the lights on for some 30 million people.

“It’s constantly solving a constant problem, meaning you’re always trying to balance,” says Nancy Traweek, who directs system operations for the grid.


On March 27, a sunny day, some solar farms had to shut down because there was more power on the grid than Californians were using. 

SOURCE: California ISO
SOURCE: California ISO (Teodros Hailye/KQED)

Ups and Downs of Renewables

In the past, balancing California’s electric was fairly straightforward. The power supply was constant, coming from natural gas and nuclear power plants that put out a steady stream of electricity.

But the growth of solar and wind power has thrown a wild card in the mix. The sun and wind are much less predictable.

“All of a sudden you have a major cloud that comes over a solar field,” Traweek says, and that causes the solar power to drop off.

“That [power] needs to come from somewhere else immediately,” she says.

So grid operators have to keep the natural gas plants running in the background. If they’re turned off, many take four to eight hours start up again.

California’s highest demand for electricity also happens right as the sun goes down, when Californians come home from work and lights turn on. Grid operators need natural gas power plants at the ready to meet that peak and to fill the gap that’s left by solar power.

Listen to the Story:

But add up all those energy sources – solar, wind, natural gas, as well as hydropower, nuclear and others – and on some days, they’re making more electricity than California needs.

“When it gets really bad, now we really got to start cutting as much as we possibly can,” Traweek says. “If that’s not done, then you could have a blackout.”

So, grid operators have to tell solar farms to shut off.

“That’s zero-carbon, clean energy,” says Keith Casey, a vice president at the California Independent System Operator. “It would just be a travesty to curtail large amounts of it.”

Casey says the problem will only get worse as more solar and wind connect to the grid. California plans to hit 50 percent renewable energy by 2030.

Joining Grids Across the West

California’s grid operator is developing a solution, one that is garnering controversy across state lines.

Grid operators manage 80 percent of California’s transmission system at the California Independent System Operator. (Lauren Sommer/KQED)

Right now, California’s grid runs mostly on its own, like an island. But there are power lines reaching across the West.

“You’re operating your little piece of the system,” Casey says, “but if you can operate it as an integrated whole, you can just operate the system more efficiently.”

So, Casey is proposing California join up with its neighbors. Instead of having lots of electric grids across the West, each doing their own thing, there would be a larger regional grid, sharing power across state lines.

When California has too much solar power, neighboring states would buy it, preventing California from having to switch off the solar farms.

“It’s a win-win,” Casey says. “We really think we need to seize the most efficient opportunities that are out there for integrating renewables.”

This marriage of electric grids would start with PacifiCorp, a utility that runs its own grid in Oregon, Utah, Idaho and Wyoming.

A study commissioned by PacifiCorp found that customers across both regions would save $154 to $335 million annually through sharing lower-cost renewable energy and avoiding turning off solar farms.

Coal Power on the Line

But PacifiCorp isn’t a partner everyone wants to get in bed with.

“PacifiCorp is by far the largest owner of coal plants in the Western United States,” says Travis Ritchie, an attorney with the Sierra Club. In 2014, more than 60 percent of PacifiCorp’s electricity came from coal power.

SOURCE: (California Independent System Operator)

“That’s a big problem for California,” he says. “We have put forth a lot of really great policy measures to stop coal for climate reasons, for pollution reasons.”

PacifiCorp’s coal power wouldn’t necessarily flow into California. The state’s climate change law, AB 32, means that electricity from other states has a price tacked on that raises the cost of fossil fuels to account for their carbon emissions.

But Ritchie warns that joining a regional electric grid, governed across several states with different agendas, could mean giving up some of California’s autonomy.

“Will California actually lose the ability lead on climate issues if it gives up its power to Utah and Wyoming, two states that are actively fighting everything about climate change that California is working to promote?” he says.

Several state legislators have the same concern. In February, they sent a letter to Governor Jerry Brown raising red flags about the merger.

“We have made great strides to make sure that our electricity is as clean as possible,” says California Assembly Speaker Anthony Rendon. “I want to make sure that joining a wider grid doesn’t impact that.”

Storing vs. Sharing

Others point out that joining electric grids isn’t the only way to deal with excess renewable energy.

“Energy storage is probably the biggest hammer in the toolbox,” says Paul Denholm, analyst at the National Renewable Energy Lab in Colorado.

Large batteries and other energy storage technologies could store the excess solar energy California is producing during the day. But until prices come down, storage remains costly.

Denholm says according to his analysis, California will probably need both storage and a regional grid to reach its 50 percent renewable goal in the most cost-effective way.

Other types of energy, apart from solar, could also become more flexible. Many natural gas power plants have signed contracts with utilities that limit how often they can be turned off or ramped up and down.

“We need to make sure we can schedule the power plants to be available when needed,” Denholm says. “Those contracts have to change so the output from the power plants can vary.”

With renewable energy growing at a rapid clip, other states could soon face the same challenges.

“We are looking at a tipping point for wind and solar,” Denholm says. “Wind and solar in many places in the country is actually the cheapest option. And that’s the first time in history where that’s been the case.”

The California Independent System Operator and PacifiCorp will spend the next two years studying their plan to integrate, as well as making their case before regulators in several states. If the plan moves ahead, the two would join up by 2019.

What Will California Do With Too Much Solar? 8 June,2017Lauren Sommer

  • ck

    Just as a consumer, I really smell a rat here both in the nature of this supposed ‘problem’ and the ‘solution’. I think this is one of those issues that would be well-served with some serious investigation and not just taking what advocates say at face value.

    • logic

      Yes. This article was initiated by one of the alternate energies. I have no doubt.

  • Jeff Justice

    Why doesn’t the hydro power shut down? Why don’t we import less?

    • logic

      Exactly. In that chart, I see where we can simple import less during peak solar. I’m betting imports could drop to zero. Bet that scares the person selling the imports. That’s probably who ‘sponsored’ this article.

    • Kevon Martis

      Almost certainly because environmental regulations compel hydro to maintain adequate downstream water flow. And of course there are no environmental benefits to replacing hydro with wind or solar.

    • Frank

      Half of that natural gas supply in dark blue is peaker plant supply that can be pushed up or down at will. The will is simply missing.

  • GM

    I bet Germany has some answers to how they store their solar power.

    • Kevon Martis

      Actually they dump it to neighbors at a loss.

      • jmdesp

        Despite the fact they are much better connected than California. The better connection is much less costly since it’s area is significantly slower, and it’s surrounded by highly industrialized country about everywhere and not the pacific for 50% of it’s perimeter. Even where there’s the sea next to Germany, there’s Denmark and Sweden on the other side, and they are connected :

      • Frank

        A loss? They can sell to nuclear France at peak hours for peak prices.

  • Bill1999

    How about using the excess power to make hydrogen…for cars of the future or for use generating electricity (and it can be stored!). Or, pump water into California’s (central valley) declining groundwater aquifer?

  • grumpy

    Part of the problem is that solar farms are the dumbest generators on the grid, ironically made from the stuff of computers. Fixing that, and making them the smartest things on the grid will allow them to take the place of natural gas plants, filling in when phase or voltage needs adjustment. There is a real time energy auction every 15 minutes that solar is too dumb to respond to. Could make a lot more money if it did, and the grid would get a lot more efficient as other generators Dave fuel.
    A company is fixing this problem.

    • Ken Gibson

      Solar generating systems are fixed investments that cost nothing to operate – they don’t burn fuel. Therefore, taking it offline would be the same as running the electricity they generate into the ground. The saving grace of natural gas turbines is that they can ramp down or up quickly – saving the cost of the fuel when the energy is not needed. Modifying our extensive hydro-power facilities to create more, if needed, pumped storage also makes better sense than fission or burning hydrocarbons.

  • normy

    agree with other commenters that have noted we could reduce use of hydro. also as for storing excess solar, we have a proven and already existing technology called pumped storage. using existing hydro plants you pump the water uphill back into the reservoir using access available Electric Power.this technology is already in place and there are hundreds of megawatts of capacity just in California alone already in service for many years.

    • Kevon Martis

      How long can they deliver that 1000MW and how many would you need at what cost?

  • How about sending the excess electricity to the electrical generating sources so they can slow the waterfall out of the dams a bit

  • MITDGreenb

    As a veteran of the Combustion Turbine (i.e., natural gas generator) industry, I wanted to add some perspective on why the contracts for these are “inflexible”:

    CTs are very high power engines, designed to work for a year between maintenance shutdowns. To operate at peak efficiency, they use very high temperatures with blades passing just a fraction of a millimeter from the casing. As you can imagine, starting up and shutting down such a machine is a choreographed dance of heat, metal stress, resonance, and more. When it’s not done correctly, you get a “train wreck” (as we called it) that destroys the machine ($10’s of millions). When it *is* done correctly, the startup creates wear and tear on the machine equivalent to 20 or more hours running at full power.

    So, think about that. If you are going to run the thing for 4 hours for the evening peak and it takes 20 hours of wear and tear to start it up… then you have to charge for a whole day’s use (minus 20 hours of fuel) for those 4 hours. The “inflexibility” in the contracts is a direct consequence: “peaking units” are much more expensive than base load units… because those are the only sustainable business models.

    The answer, as others have pointed out, is storage. As Bill Gates has pointed out, this is a major challenge – storing huge amounts of power and being able to release it when needed is just plain difficult.

    The problem in this article is a well-known issue in renewables. Someone mentioned Germany. Denmark (next door) has a lot of wind power… a lot. But the only reason it can do so is that the Danish grid is tied to the German grid, which provides power on days when the wind is not blowing.

    • Pete Danko

      Aren’t there a new generation of plants — like the “fast-start” Lodi Energy Center that opened in 2012 — designed to work much more efficiently in tandem with variable renewables?

      • MITDGreenb

        This is a good point Pete. This sort of plant does create the impetus for more flexible agreements… which is a good thing… but it’s not 100% rosy in the details. (It’s more good than bad… just not completely good.) Let’s look…

        The plant is combined cycle = an engine and a boiler. The engine has a number of things in common with your internal-combustion automobile. It’s a reasonable amount of energy, and wear and tear, to start the engine and get it warmed up. In my day, though, that was the choice: on or off. That led to two types of power agreements reflecting the very different economics: base load (turn it on once/year) and peak load (turn on as needed, probably once/day).

        Now, with sophisticated controls, the engine no longer needs to run at full power (or not at all) – the engine is more like your car’s. You can start it and warm it up. That still counts. But you don’t have to turn it off. Rather, it can idle, like at a stoplight, producing little power beyond what it needs to keep running and using little fuel… albeit inefficiently. You then “accelerate” to full power (the turbine actually never turns faster… it just turns with more torque, but that’s a detail). Eventually you get up to full power (and peak efficiency), although I surmise that efficiency and pollution increase as you accelerate and decelerate. In any case, it’s not as hard on the engine to accelerate/decelerate than it is to start cold.

        But this is a combined cycle plant. That means there’s a boiler… and that ramps slower than the engine. So even when the engine has reached full power and efficiency, there is a time before the plant as a whole does.

        So, on the whole, sophisticated controls are a good thing. They enable much more flexibility in generation. However, what you trade is efficiency (and pollution) over a constant full-power plant. If there were really good storage, you would not have to trade for that.

        But back to our topic – what you *do* get is an impetus to create power contracts so that they support renewable power instead of standing in its way! That’s a definite good.

        • Pete Danko

          I don’t know how much ramping up and down they’re having to do at the Lodi plant, but it does seem to operate at pretty impressive emissions levels — 380 gCO2eq/kWh in 2014 (1,316,266 MWh generation, 500,115 metric tons CO2eq emissions).

    • Bernard Finucane

      Are you talking about shutting down the plant completely, or about reducing output by say 10%?

      • MITDGreenb

        The latter. It’s possible to turn the engine down quite a bit — the control technology grew from the needs of IPPs in the late 1990s. The turbine runs at lower power and burns less fuel. It’s not as efficient, but the economics of full shut downs (and the prize of rapid response to the power markets) make it a valuable feature. From a maintenance point of view, the engine is in base load – running all the time – but that’s a lot better than paying a 20-hour/restart penalty in maintenance.

        The inflexible power agreements referenced in the article predate this to a large extent. However, remember that a large amount of base load is provided by very inflexible sources (coal and nuclear). Without storage technology for enormous amounts of power, you need those for the stability of the grid. (That was my comment about Denmark: without the huge base load capacity of the German grid, the Danes could not have so much wind power.)

        The solution (aside from power storage) is to create regulatory and market advantages for the development of flexible generation. But this has to be done in the context of what is possible with the engines.

    • Frank

      How is storing a lot of electricity “just plain difficult”? Tesla integrated an 80 MWh battery bank into a grid in less than 3 months. There’s a line for aggregated smart charge/discharge powerwalls that stretches around the block. None of this is all that difficult we’re just dealing with a lack of imagination and political will to make the change.

      • MITDGreenb

        I presume you’re referencing this:

        Let’s do some math, Frank. Average power consumption in NYC is roughly 2GW. That’s average. Suppose at night it’s only half that. (I don’t have any basis for that assumption, but it makes the math easy.) So, you need 1 GW for, say, 8 hours of night time. That’s 8 GWh. That’s merely 100X what Tesla created.

        And it’s just for NYC.

        And it would be a whole lot of money for Elon if it happens. That is the difficulty. This sort of infrastructure is expensive.

        And it takes a good amount of resources. That’s a lot of lithium to mine* and a lot of land.

        And batteries are a bad technology. They fatigue. They can only charge/discharge so fast. There are much better storage technologies in development.

        So, it’s a lot more than “lack of imagination and political will”. I agree that grid storage would be a great thing, but the market has not found the right solution yet.

        * Lithium mining and battery manufacture creates a large “debt” that must be repaid by efficiency over time. For instance, it takes a lot more CO2 (from energy production) to build an electric car than a conventional one… but the amount of CO2 per unit of driving is much less. For an average driver, the cross-over is in the years. The “obvious” choice of a Tesla may still be the better one, but it’s less obvious.

        ** My VW TDi was recalled due to the pollution scandal. Obviously, it needed to go. Except, it got 45 MPG on the highway. Its replacement – a similar VW with a gasoline engine – gets 32 MPG on the highway. If I do a lot of highway driving, I will guarantee that the pollution from the latter is more than the former. Hmmm. And I did not even factor in the cost of taking an asset out of service before the end of its useful life.

        Math is hard. Sometimes it yields nonobvious results.

        • Frank

          Re-imagine this post as if you had zero existing infrastructure and were planning an optimal solution from scratch given the available resources.

          Yes we consume a huge amount of energy, but we just clearly demonstrated that 15,000 houses can be fairly easily powered by renewable production plus batteries. Certainly we can scale that up quite a bit. It’s literally Day 1 of the Renewables Era and we’re already hitting home runs with ease.

          The entire passenger vehicle fleet will be EVs within 15-25 years. That could take up a massive chunk of peak excess right there. Pumped water, hydrogen production for fuel cells, smart homes. There’s plenty of tech out there right now that gets us 20% of the way down the road.

          Aquion is taking common materials and making salt water batteries that can store solar/wind scaled up as high as you like. We’re acting like excess solar is a problem, how absurd is that?

          • MITDGreenb

            Your point on auto batteries is a good one. By distributing the cost and the storage, you create a good deal of value…

            except it does not work, at least not right now. For car batteries to effectively work in the scenario, the cars would have to yield power back to the grid when needed. For instance, you’d charge during the day while you’re at work… and then *discharge* to power your home at night. Right now, we mostly charge at night *and* during the day.

            So even if I imagine zero constraints, as you suggest, I have a problem that is not solved. But perhaps that could lead to policies that do solve it.

            And/or we could all work on cost-effective grid storage.

  • SynerGenetics

    Have the excess solar power go towards powering underground dehumidifiers to create water.

  • rattboo

    Charge more electric cars, that will use that extra energy, and store it for use the next day when people drive to work. Electric cars – the real solution.

  • Frank

    Why is it “excess renewables”? Isn’t it just “excess electricity”?

  • Mike Meagher

    Batteries! Lots and lots of batteries.

  • jon

    Why shut it off send it down the line to the next state.

  • jdeely

    While it is interesting that we have already had a day where we had excess solar – it was only one day. Plus it was a Sunday in March (low energy usage Month) In fact, this day – Mar 27 – had the lowest energy usage for a single day in recent CA history – 503,675 MWH. Record low day.

    Also, check out the hourly numbers for Mar 27 here –
    Note that at 2-3M were down near 3GW of thermal (Nat Gas) . Assuming that half of the renewables came from NW hydro… the means that we got 15GW out of a total of 20GW from her emission sources. Not bad.

    Like the rest of the folks on this thread, it would be great to see more detail on how/why/where we got the fossil fuel generation.

    Also no mention of pumped storage which is available in CA. Was the water being pumped up in afternoon hours. If not, why not?

  • Ken Gibson

    This story presents a distorted view of the ‘typical’ CAISO demand curve. In the summer months, when solar generating systems are most productive, peak demand occurs in the early to mid afternoon. Solar generation is an excellent fit to offset air conditioning energy demand in office buildings and throughout the Central Valley. In winter months, solar generating systems produce less midday power and air cooling is in less demand. The morning and evening demand peaks can be offset by suitable placed wind turbine generation particularly if placed near the ocean-land interface. There the morning sea breeze and the evening land breeze coincide with the slower temperature changes of the ocean heat sink. The CAISO demand/supply curve shown, though not specified as such seems drawn on a winter day.

    • Pete Danko

      Solar meets a lot of air-conditioning-driven need, but peak demand is in the late afternoon, and demand remains high up to 8, 9 pm, when solar is totally disappearing. Yesterday was typical:

  • Richard Rowe

    HYDRO POWER STORAGE might help some of the over production

    The Helms Pumped Storage Facility (Helms) is the largest pumped hydroelectric storage project in California.

    Helms produces and stores up to 1,212 megawatts of electricity by moving water between Courtright Lake at a higher elevation and Lake Wishon at a lower elevation. The Helms facility uses three reversible turbine generators to pump water uphill when the demand for electricity is low. Water is released downhill to regenerate electricity when the demand is high. Pumped hydroelectric storage facilities can increase power output quickly. In fact, Helms can go from a dead stop to full generation in 6.5 minutes.

    The Helms facility is located in the Sierra Nevada Mountains approximately 50 miles northeast of Fresno in Fresno County.

  • agunrunner
  • I’m probably too late to the party for anyone to read this, but at a really simplistic level it seems to me that with a regional grid you can ship power east to places where the sun has gone down and home energy use is ramping up as people come home.

    There was another NPR article somewhere that talked about using solar electricity to fill tanks with compressed air, which was then used later to drive turbines. It was on a small scale, but that seems like something that could be scaled up. I would agree with other commenters that political will is a lacking ingredient. Pit political will against entrenched money, and guess who wins…

  • DS

    And why aren’t the grid operators responding by throttling or taking some of the carbon emitting sources offline, or by reducing some of the imports?

    The headline reads like solar is powering the entire grid…yet it’s far from that. Solar was providing only about 1/4 of demand at that time of the day, and only provided about 9% of the daily generation. As an interesting comparison, on that same day, ONE nuclear plant provided 10.9% of the state’s total generation.

    CaIso should focus on getting carbon emitting sources off the grid to help reach our emissions goals, or at least find another use for that energy, be it pumped storage, desalination, etc.

  • There is actually a pretty cool program in California to deal with this exact problem. It’s called OhmConnect. They pay households to reduce energy at peak times, when demand exceeds renewable supply and they have to turn on dirty, expensive “peaker plants.” It’s cheaper to give people money to offset their usage than to use these plants.

    Here is the link. They have a temporary $20 signup bonus. Note you have to hook up your utility account to get paid. I’ve been using it for about 6 months and I get $10-20 per month with very little effort:

  • Greg

    Where did the data for the excess renewables come from in the chart?

  • UCSBcpa

    There are some fantasticly bright people writing in this comment section!

    My two cents:

    I have quite a bit of experience pricing all sizes of renewable energy systems. I also have a very good sense of storage. I will probably be one of the few “renewable” guys, who are bearish when it comes to storage. Storage reminds me of solar back in 2000. Maybe it will mature much faster, but right now, it is a mess. At the very most basic concept, storage is an added cost to enjoy solar. If you want to get fired-up, try to call one of these storage companies up and get a quote over the phone. Heck, have them come to your house! What you will quickly surmise: they don’t even know the cost and their agreements will down-right scare you.

    Given the fact that storage is an added cost, the economics have to be that much better – or have solar pricing (and storage) come down dramatically. Let’s assume storage can come in and give you a turnkey storage deal for your 7kW system and they price this at $10,000. You can’t get this now, but let’s just assume you can.

    At the very simplest level, this means your PV system now is $10,000 more and there is a good chance this $10k won’t have a 30% Credit tied to it. – No chance am I getting into this now, but let’s just say this: there is IRC guidance on this.

    7kW system cost (let’s assume very low pricing) is $19,000. Thus, your entire system now costs $29,000. But it gets better! – There is no chance your storage will outlast your modules, therefore, you better have some added cost for year n when your storage fails. Let’s, again, be nice, and say this present valued cost of a new storage system in year 11 costs $5,000 installed (big, big point, installed by a qualified electrician).

    Total cost is now $19k + $10k + $5k = $34,000 for a 7kW system. And I was very kind on pricing.

    Basically, what storage does to residential is take it back to 2008 solar pricing.

    And guess what? – All of these financial and storage acrobatics still can be crushed by the local utility who will (yes, read that again, will) change their tariffs to configure this out so they get their returns.

    All this said:

    I am very, very bullish on the pricing of solar and wind. If you believe in this, and consider California is much larger than the other states they will be connecting with, size matters. California will be able to push their renewable ways onto the smaller states. Why this will be? The smaller states would love to buy $0.05 kWh solar power in the day and $0.04 kWh wind power at night. What happens next? Why surely they will shutter their most inefficient / old under-performing coal plant; then another; then another. All the while, solar and wind continue to get cheaper, and they continue to shutter generating plants (starting with very, very expensive “peaker” plants).

    By allowing California to connect to these smaller states, we are providing a huge a push for solar+wind.

  • Pingback: Solar Power Is Cheaper, but the World Is Still Running on Fossil Fuels()

  • Pingback: Solar Power Is Cheaper, but the World Is Still Running on Fossil Fuels | Auto Test 2()

  • Pingback: Solar Power Is Cheaper, but the World Is Still Running on Fossil Fuels – Futurism()

  • Pingback: Solar Power Is Cheaper, but the World Is Still Running on Fossil Fuels | anSabri()

  • hotstock

    Warren Buffet owns PacifiCorp. He’s a hard negotiator.

  • metalen

    If I were them, I would make a swappeble batterysystem for trucks, then these batteries can charge when there is much electricity and discharge when there’s a lack of energy, also, the trucks can drive their next 500 miles with a 5 minit battery swap. Same thing for a bus, name it.


Lauren Sommer

Lauren is a radio reporter covering environment, water, and energy for KQED Science. As part of her day job, she has scaled Sierra Nevada peaks, run from charging elephant seals, and desperately tried to get her sea legs – all in pursuit of good radio. Her work has appeared on Marketplace, Living on Earth, Science Friday and NPR’s Morning Edition and All Things Considered. You can find her on Twitter at @lesommer.