GBA Logo horizontal Facebook LinkedIn Email Pinterest Twitter Instagram YouTube Icon Navigation Search Icon Main Search Icon Video Play Icon Plus Icon Minus Icon Picture icon Hamburger Icon Close Icon Sorted
Guest Blogs

If a Solar Plant Uses Natural Gas, Is It Still Green?

The Ivanpah solar plant in California uses natural gas to generate power when sunlight alone isn't enough

Mirrors arranged around a central tower use sunlight to create steam that drives a generator. But sometimes the sun isn't enough.
Image Credit: BrightSource

The giant Ivanpah solar power plant in the California Mojave Desert recently detailed how much natural gas it burned to generate power when the sun wasn’t sufficient: the equivalent to 46,000 tons of CO2 emissions in its first year, according to reports.

Along with its impacts on wildlife and its receipt of federal incentives, news of the CO2 emissions has renewed criticism of the 377-megawatt facility, which supplies 140,000 California homes during peak hours of the day.

Why is a solar power plant using natural gas, and does the associated CO2 disqualify it as “green”?

The use of natural gas to complement solar in fact highlights a trend toward what I call “speckled green” electricity generation – approaches that are not completely green, but in which natural gas enables more widespread, reliable, and affordable deployments of renewables.

Power tower

Unlike the photovoltaic (PV) panels that are proliferating on rooftops such as my own, Ivanpah generates its power by angling mirrors to gather the intense sunlight of the Mojave Desert to produce heat. The mirrors reflect sunlight onto three power towers, where steam turns turbines to generate electricity.

This “concentrating” solar thermal approach is even more sensitive to clouds and particles than PV panels, since the mirrors can concentrate only the sunlight that arrives in a direct beam from the sun. Clouds and particulate matter scatter the light into directions that can still be utilized by photovoltaic panels but not the precisely angled mirrors.

Ivanpah was designed and built to burn some natural gas to maintain peak power generation during times of intermittent clouds. Without the natural gas, Ivanpah’s steam turbines could trip off-line, interrupting power generation. The extra energy from natural gas can enable peak power production to continue until sunny conditions resume or the turbines pause for the night.

Ivanpah, which is the largest solar concentrating plant in the world, began operation in early 2014, but unexpectedly cloudy conditions dampened electricity generation to below intended levels. Thus, soon after opening, Ivanpah petitioned the California Energy Commission to increase its use of natural gas. The commission’s staff analysis noted that the request would increase the plant’s carbon dioxide equivalent emissions limit from 62 pounds per megawatt hour (lb/MWh) to 75 lb/MWh. Of course, actual emissions are lower if sunnier conditions negate the need for extra natural gas use and increase solar output.

On a percentage basis, this looks like a big increase. But context is needed. U.S. power plants in 2012 emitted an average 1,137 lb/MWh of CO2 – 15 times Ivanpah’s “dirtier” new limit. In fact, this number understates the gap, since the Ivanpah limit includes associated maintenance vehicle emissions not counted in the national average.

Fox News and others have branded Ivanpah a “carbon polluter” by noting that its CO2 emissions top a federal threshold that requires reporting as a greenhouse gas emitting facility. Not mentioned is that Ivanpah’s emissions are just half the threshold of a “major stationary source,” denoting large facilities such as fossil fueled power plants or factories.

Transitional technology?

Critics rightly note that Ivanpah’s natural gas use and associated emissions are far higher than originally anticipated, with gas-fired auxiliary power now at times needed 4.5 hours per day rather than one as originally expected. Some critics specifically attack the use of natural gas with solar as “dirty power.”

Indeed, natural gas is a non-renewable fossil fuel whose use causes CO2 and methane emissions. However, unless we expect to switch to completely renewable fuels overnight, which is unrealistic, fossil fuels will remain a major part of our energy mix in the decades ahead.

I believe using those fossil fuels synergistically with renewables offers great potential. The joint deployment of natural gas with solar thermal energy, sometimes referred to as a “hybrid solar” plant, is one such approach. At Ivanpah, natural gas provided less than 5% of the energy, yet may have substantially boosted solar power output by keeping the turbines online.

More broadly, the rapidly dispatchable – that is, able to generate power on demand – and adjustable nature of natural gas power generators enable greater penetration of variable wind and solar on the grid while maintaining reliability.

Desert tortoises

There are, however, other arguments that critics could lodge against Ivanpah.

Its electricity output underperformed expectations in its initial year, though output in 2015 has increased sharply. It’s likely that the added flexibility to use more natural gas, together with an extra year of experience, has contributed to the big boost in output.

As a cutting-edge energy project, Ivanpah received $1.6 billion in loan guarantees from the Department of Energy, plus a $660 million tax refund under the 30% Investment Tax Credit. Ivanpah also benefits from the tradeable renewable energy credits it generates under California’s Renewable Portfolio Standard. Critics of government incentives for renewable energy may oppose the policies that enabled those incentives and credits. Also, as costs have fallen faster for utility-scale photovoltaic plants than for solar thermal plants, Ivanpah is no longer seen as representing the most affordable possible future for solar.

In terms of environmental impacts beyond CO2, Ivanpah has been linked to 3,500 bird deaths. Environmentalists have also criticized Ivanpah for its proximity to some endangered species such as the desert tortoise.

Indeed, Ivanpah does not deserve to be considered the pure electric green energy source that we might seek in an ideal world. But all energy sources have environmental impacts, whether from emissions of combustion or obtaining the fuels or building the power plants and solar panels. The bird deaths from solar are tiny compared to those from fossil fuels.

At least from a climate perspective, I believe this particular case of “speckled green” energy is doing pretty well, reducing CO2 emissions per MWh by over 90% compared to our national average for electricity. For costs and more flexible siting, though, other approaches are likely to displace Ivanpah as exemplars of the path toward greener, more affordable electricity.

Daniel Cohan is an associate professor of Environmental Engineering and Faculty Scholar at Baker Institute for Public Policy at Rice University. This post originally appeared at The Conversation.


  1. Expert Member
    Dana Dorsett | | #1

    Power from remote large solar installations are less valuable.
    Utility companies like dispatchable big-iron, including large CSP & PV solar farms, but the infrastructure and real estate required to support them are orders of magnitude larger than the same amount of solar power in a distributed format, sited at the loads, say, on top of your roof, on either side of the meter rather than on the other side of a mountain range out in the desert.

    Even if the lifecycle cost per TWH out of big concentrating power plants were cheaper than rooftop solar, (and there's reason to believe that it isn't), the additional cost of the transmission grid infrastructure makes it more expensive and less valuable than load-sited PV. Load sited distributed resources lower the peak & average loads on the transmission grid, not so for big iron.

    Dividing up the 377 megawatts into the 140,000 homes it's alleged to support during comes out to 2400 watts per home. Installing 2400 watts of PV on each of those 140,000 homes would cost $8000-8500 per home at 2015 installed prices (pre-subsidy), and would take MUCH less stress off the grid during peak hours than Ivanpah can. That adds up to $1.2 billion, a considerable discount from the $1.6 billion in loan guarantees cited. Spending the difference in up front cost on distributed grid-tied batteries would probably exceed the storage capacity at Ivanpah, and would be far more flexible too.

    As visually interesting as they are, concentrating solar power is too expensive to be competitive without ongoing subsidy. Large-scale photovoltaic power is proving be less expensive and has continued to rapidly fall on cost, and is now cheaper than CSP, even when hardened with battery backup. Abengoa (the world's most experienced and best-regarded CSP developer) is on the brink of bankruptcy:

    Unless concentrating solar power costs can be reduced dramatically (and quickly), it is not likely to become a significant power source in the future. Like other big iron, CSP suffers from limited scalability to the small scale, and has siting requirements far less convenient than the roofs of commercial & residential buildings, and is less valuable to both the ratepayer and the grid operator than distributed PV. As grid battery costs continue to plummet, the thermal storage aspects of CSP aren't valuable enough to compensate for the other high costs.

  2. DEnd2000 | | #2

    I have to somewhat disagree
    I have to somewhat disagree with Dana here. And by that I mean all the points he makes are completely valid. I do believe he is missing an important element however, energy source flexibility. As such CSP should not be evaluated just on cost and efficiency of it's solar power operations vs. the solar power cost of other technologies, but also include costs associated with storage, back-up power and transmission. Ivanpah shows a path of good equipment utilization, and allows its equipment cost to be spread over many more kWh than distributed solar does. The question is: Is Ivanpah's cost per kWh (or MWh) cheaper than distributed solar, plus storage, plus back-up power generation. I suspect that the answer is no, as appropriately sized distributed storage should ultimately reduce peaking plant sizes and grid infrastructure costs. But that's just speculation until it's well studied.

  3. Expert Member
    Dana Dorsett | | #3

    The LCOE of Ivanpah vs. alternatives
    The DOE estimated the lifecycle cost of energy out of Ivanpah to be between 12-17 cents per kwh ($120-170 /Mwh) back in 2013:

    And that is without the burdened cost of the transmission grid capacity needed to deliver that energy to the load. Load-sited small to mid scale commercial rooftop PV has a comparable raw LCOE, but imparts a NEGATIVE cost to the transmission grid, since it lowers both the peak & average load on the transmission & distribution grids. The LCOE of grid scale PV at this year's pricing is half that (or less) of Ivanpah, and getting cheaper every quarter. See page 3:

    The value of Ivanpah's dispatchable output is pretty limited in the era of automated demand response, and becoming less valuable every day as more dispatchable storage goes on the CAISO grid. Distributed dispatchable storage solves substation grid-congestion problems, which is much more valuable to the grid operator, local utility and rate payer than Ivanpah's ramp-able output at a fixed location on the grid. Ivanpah is somewhat emulating a base load generator, at a time when the relevance of base load generation is rapidly fading. Yes, the fact that it's flexible and can deliver power after dark is worth something, but it's not worth 2x that of utility scale PV, nor is it worth as much as the lowered grid load delivered by load-sited PV at the same (or now slightly cheaper) lifecycle energy cost.

    This stuff IS well studied, having undergone extreme scrutiny by the investment banking sector over the past 3-4 years. The fact that Abengoa can't find investors now that government subsidies for CSP are waning is an indication of the perceived financial viability in the face of relentless cost reduction in the distributed PV and storage markets. In the time it takes to plan, design and finance (let alone build) another big-iron concentrating solar project like that, the competition will have already eaten their lunch.

    In less than five years even residential rooftop (unsubsidized) will have a lower LCOE than Ivanpah, and the last peaking power plant ever built in California will already be on line (and at risk of becoming a stranded asset.)

  4. Dana1 | | #4

    Driving the point home...
    The numbers have just been crunched: Distributed grid storage (on both sides of the meter) in CA has doubled the installed capacity in just 3 months of Q3 2015, and that rolling rock is just getting started:

    The combined output of all grid storage on the CAISO grid is already about 1/5 of Ivanpah's full-output rating, and will surpass it well before 2020. The speed at which the value of the dispatchability factor is falling is breathtaking!

  5. charlie_sullivan | | #5

    Evidence of the value of dispatchablity
    Dana argues that the value of dispatchability is low and falling, and cites the rapid growth in installation of storage as evidence. I think that the installation of storage is evidence of exactly the opposite: If there really was no value or need for dispatchability, there would be no reason to install storage.

    That said, the value of a dispatchable source near a load is higher than a remote dispatchable source, and the LCOE of CSP is not looking so so good with PV costs so low. SO I generally agree with Dana's conclusions, even though I'm quibbling with this detail.

  6. DEnd2000 | | #6

    Response to Dana
    Sorry Dana I just realized Ivanpah has no storage, I thought it did... That takes away a huge reason for it to even exist. In my mind the only good reason for CSP is the ability for it to have some relatively cheap, though less flexible storage.

  7. Expert Member
    Dana Dorsett | | #7

    Ivanpah has dispatchable gas fired resources.
    That's not nothing, but it's not the same as storage, but even big storage located out on the grid is far less valuable than distributed storage.

    Charlie: Dispatchability of large generation sources out on the transmission grid still requires the grid capacity. Distributed storage is far more valuable, since it lowers the load on the transmission & distribution grids, and lowers grid congestion, since the power can be source on the same side of the substation. The grid congestion piece and the controllable load (the battery chargers) offer ancillary voltage and frequency stablization services of much greater precision and speed than any peaker plant or grid-remote power source (even large centralized grid storage) could ever achieve. The in-front of the meter storage isn't being implemented by utilities in large random lumps- it's being targeted and sized in location-specific ways, aimed at relieving grid congestion, and providing frequency and voltage stabilization services (more cheaply than other utility assets, or building out new substation infrastructure.) Dispatchable power plants can't do that level of grid tuning at a competitive cost or precision, and that's at this year's battery costs. That snowball is already headed downhill, along with the battery cost, and the growth isn't linear, it's exponential.

    Most of the behind-the-meter storage is much dumber, being used to limit demand charges which is often the biggest line item of commercial ratepayers, bigger than the energy use portion. This has addition value to other ratepayers of lowering the grid peaks, but none of the ancillary services value. Where the regulatory environment doesn't allow aggregators to bid behind the meter storage into the ancillary services markets it won't be used that way, but that's not to say the control software for behind the meter storage isn't capable of that if/when the regulations change.

    There are smart-charger networks for electric car fleets that are doing much more than just charging the car, providing grid stabilization services more valuable than the energy cost of charging the cars. Where the regulatory framework allows. the utilities are PAYING the charging network operator for those services. One of the first such installations in the CAISO region was at the Los Angeles Air Force base:

    This is a rapidly moving target, but the whole centralized generation model is going away, being undone by cheaper distributed resources not all under direct control & ownership of the grid operators or utilities. Some regulators get that (say, those in NY state), others don't (GA). Ivanpah may have made some sense back when it was first conceived (before it's approval 5 years ago), but it was a bit iffy even then, and wouldn't have been built without $1.6B in Federal loan guarantees (on a $2.2B project.)

    Given the rapid financial learning rate of both (more scalable) PV and storage, there's no point to building another one, unless it can under cut the LCOE of small commercial scale PV by more than the cost of the grid infrastructure necessary to get that power to the load. No analysts seem to be promoting that notion, which is why CSP companies like Abengoa have no future, despite a significant recent past.

  8. AndyKosick | | #8

    I'm going to be optimistic. I think distributed generation and storage is inevitable, but having it all tied together is also very advantageous. I'd say the questions are what is the value of the grid, what should it look like, and how are we going to pay for it?

Log in or create an account to post a comment.



Recent Questions and Replies

  • |
  • |
  • |
  • |