Putting solar and wind power on the grid causes fossil plants to emit more CO2 than they would have otherwise.
Human civilization depends on food, food depends on weather, and weather depends on climate. Every grain and every mammal on the face of the earth – almost everything we eat – evolved under a different climate than the one we're heading toward. We are running toward a cliff, and merely walking toward that cliff isn't a viable strategy. We need to stop. Right now.
If civilization is to survive, we need to get to zero emission of fossil carbon, and we need to get there rapidly. Every ton of carbon we emit stays in the air for centuries, and will continue to warm the planet for centuries.
In this series GETTING TO ZERO we will take a very hard-headed look at current energy policy and energy strategies. We will ask hard questions: does this really get us to zero? How much would it cost? How rapidly can it be deployed? We may find some answers along the way, but don't expect them to be easy.
This diary is part IV of GETTING TO ZERO: Our non-fossil energy future.
Part I of GETTING TO ZERO: The size of the problem.
Part II of GETTING TO ZERO: Is renewable energy economically viable?
Part III of GETTING TO ZERO: Why energy efficiency will not save us.
Figuring out how much CO2 a given power technology emits is a tricky business. Take coal, for instance. It's not that hard to figure out how much energy coal contains, and how much CO2 is emitted when you burn it. But then there's the energy cost of mining the coal. And the energy cost of transporting the coal to the power plant (which can be thousands of miles away). And the energy cost of building the plant itself. And all those energy costs emit more CO2, beyond that emitted just by burning the stuff.
The standard way of handing this is through what's called a "Life Cycle Analysis", or LCA, which tracks all of those emissions and more. Ideally an LCA will give you a "cradle-to-grave" picture of the total emissions from every source and every energy input, including things like waste disposal and plant decommissioning. Over the years a lot of people have done this and come up with answers that are slightly or greatly different, because of different assumptions and different ideas of where to draw system boundaries. But recently, the National Renewable Energy Laboratory has published a series of papers with the intent of harmonizing those assumptions and boundaries, in order to get a realistic picture of what emissions really are for various electrical generation technologies. The emissions data below are drawn from the NREL results.
Technology |
Median LCA emissions
(gCO2e/kWh) |
Hydro |
7 |
Small Hydro |
7 |
Ocean |
8 |
Wind onshore |
11 |
Wind offshore |
11 |
Nuclear |
12 |
Photovoltaic (CdTe) |
14 |
Photovoltaic (a-Si) |
20 |
Geothermal |
25 |
Solar thermal (CSP) |
25 |
Photovoltaic (CIGS) |
26 |
Biomass |
30 |
Enhanced geothermal |
57 |
Natural gas CC |
450 |
Natural gas CT |
670 |
Coal |
980 |
Looking at these numbers, the non-fossil technologies are low, as expected. Solar is higher than wind because it takes a lot of energy to make a solar cell, and that costs emissions.
But both solar and wind have a hidden emissions cost that isn't reflected here, because neither the NREL nor any of the papers they analyzed included it. When the wind starts or stops, wind turbines change their power outputs dramatically over short time periods. And when a cloud drifts over a solar panel, its power output also changes dramatically over short time periods. Either too much or too little power on the grid is a problem, and utilities understand that issue and manage it quite well. The way they do that is by changing the power output of other generators, either up or down, to compensate for the changing outputs of the renewables; this is called "ramping". In fact, some natural gas generators are specifically designed to ramp their output rapidly. They are called open-cycle gas turbines (OCGT), combustion turbines, or conventional turbines (NGCT), but they all mean the same thing.
The issue arises because when the fossil plant is ramping up or down, it's not operating as efficiently as it could if it were operating at a constant speed. The ramping causes the plant to lose efficiency, and therefore emit more CO2 (on a per-kWh basis) than it would have otherwise. A recent paper by Katzenstein and Apt looked at this problem closely. They gathered real-world data from actual natural gas power plants during operations, both at constant speed and while ramping, on a minute-by-minute basis. They also gathered minute-by-minute data on the variability of actual wind turbines and solar farms deployed in the US, in several regions. Then using these data they determined how much ramping a fossil plant would have to do in order to accommodate the variability of renewables, and how much that would cost in terms of extra emissions, compared to a plant running at full speed.
What they found was significant. For both wind and solar, intermittency causes between 21% and 24% of expected emissions reductions to be lost. In physical terms, solar and wind would gain an additional 89 gCO2e/kWh, over and above the numbers listed in Table 1. Considering that natural gas emits 400 to 700, and coal emits 980, even adding in that additional 89 still leaves renewables much, much better than fossil. But it does put wind and solar behind some other low-carbon technologies, and this will be important as we put a price on carbon.
In the corrected table below, I've added that 89 gCO2e/kWh to wind and solar PV. But not to solar thermal, which was not analyzed by Katzenstein and Apt, and which may have reduced ramping issues compared to PV because of thermal inertia (and physical inertia) in the plant.
Technology |
Median LCA emissions
(gCO2e/kWh) |
Hydro |
7 |
Small Hydro |
7 |
Ocean |
8 |
Nuclear |
12 |
Geothermal |
25 |
Solar thermal (CSP) |
25 |
Biomass |
30 |
Enhanced geothermal |
57 |
Wind onshore |
100 |
Wind offshore |
100 |
Photovoltaic (CdTe) |
103 |
Photovoltaic (a-Si) |
109 |
Photovoltaic (CIGS) |
115 |
Natural gas CC |
450 |
Natural gas CT |
670 |
Coal |
980 |
Of course, none of this would matter one bit if we get to a grid where there are no fossil generators. In that world, we would still need fast-ramping generators to handle the intermittent production of renewables, but they would have to be
non-fossil fast-ramping generators. As we will see in future parts of this series, there are ways to do that, but they have issues of their own.
Besides the emissions cost of intermittency, there are financial costs as well, which are also ignored by most sources. In the next part of GETTING TO ZERO we will determine the cost of electricity from various technologies and how much of a carbon tax we really need.
Finally, because this diary is rather short, I'm giving you a video with some really cool visualizations of our fossil fuel use, made by the World Business Council for Sustainable Development. Being CEO types, these guys are pushing their own favorite technology (CCS, or carbon capture and storage, which is both expensive and unproven -- so take the sales pitch with a healthy dose of skepticism). But at least they're also pushing a carbon tax, which everyone can get behind.
References
Katzenstein, W., & Apt, J. (2008). Air emissions due to wind and solar power. Environmental science & technology, 43(2), 253-258.
NREL Life-cycle analysis database. See links at the bottom for details on individual generating technologies.