When to use Marginal Emissions (and when not to) | A New Shade of Green | Sherry Listgarten | DanvilleSanRamon.com |

Local Blogs

A New Shade of Green

By Sherry Listgarten

E-mail Sherry Listgarten

About this blog: Climate change, despite its outsized impact on the planet, is still an abstract concept to many of us. That needs to change. My hope is that readers of this blog will develop a better understanding of how our climate is evolving a...  (More)

View all posts from Sherry Listgarten

When to use Marginal Emissions (and when not to)

Uploaded: Sep 29, 2019
This week’s post is about electricity emissions. Specifically, I want to talk about a concept called “marginal emissions”. This has come up often in the past few years, particularly in the context of evaluating when to move electric loads from one time to another (aka “load shifting”). I want to talk about what these emissions are, how they are used, when they are useful, and when they are not.

As you may remember from a post earlier this spring, we have a competitive wholesale market for electricity in most of California, administered by CAISO. This market allows new plants to come online and bid for business. CAISO generally brings plants online in order of price. So when they are deciding which plants to run to match supply, every five minutes throughout the day, they typically bring the cheapest plants online first. Since in California lower-emission plants have lower prices (1), this also means that CAISO generally brings on the lowest-emission plants first. This market has been very effective at growing our renewables power sector.

Here is a picture of how this works, from a market operator in the northeast (ISO New England). (1)

Source: ISO New England

In this case, plants A, B, and C are running at full capacity, and D at partial capacity. If more load came online, plant D would handle it. This would be evaluated every five minutes throughout the day, with micro-adjustments in between.

The emissions on the grid at the time shown above would be the average of A, B, C, and part of D. Because plants are chosen in order of lowest price, which is also lowest emissions, the average emissions will be relatively low.

The “marginal emissions” are the emissions that would come online if new load were added. In this case, the marginal emissions are those of power plant D. They are nearly always greater than the average emissions, and often by a considerable amount.

There are some exceptions to this “cheapest runs first” rule, which is why marginal emissions can sometimes be lower than average emissions. For example, some types of plants are less flexible and need to run throughout the day, and that includes many older and less efficient gas plants, as well as nuclear plants.

Average emissions are what we normally talk about and see on graphs like the one below, which shows CAISO’s emissions over time..

Source: CAISO

Marginal emissions are not as easily available for us to look at. They can be negligible on sunny, temperate days when we have more than enough solar (solar is on the margin). Or they can be up to 1400 pounds of CO2 per MWh on hot summer afternoons when we are forced to run our least efficient plants. Christy Lewis at WattTime gave me two data points that are a bit less intuitive. Let’s take them one at a time.

On a cool but sunny Friday afternoon, May 3 2019 at 1:27pm, average emissions were a low 284 lbs/MWh, but marginal emissions were a sky-high 1012 lbs/MWh. The graph below shows the power supply on that day.

Source: CAISO

Why were marginal emissions so high at 1:27pm? It is probably because a fast-acting but very inefficient “peaker plant” was activating. You can see a drop in solar and a rise in gas around that time, perhaps due to some persistent clouds over a large solar installation. That means the “peaker plants” would come online to cover the transient drop in solar.

Here is the second data point. On a recent warm Thursday morning, September 26 2019 at 11:40am, average emissions were 626 lbs CO2 per MWh, while marginal emissions were 702 lbs/MWh. The graph below shows the power supply on that day.

Source: CAISO

You might wonder why the average was so high, given so much solar in the mix and marginal emissions indicating an efficient gas plant on the margin. Something much more inefficient than the marginal gas must be operating. It is possible we were running a less efficient and less flexible gas plant in order to prepare for a potential evening spike (it was a very warm day). Inflexible plants cannot ramp up quickly.

Marginal emissions are volatile and hard to predict. You can imagine that, if you have a flexible load, it is best to avoid times when marginal emissions are high. Since those times can be hard to guess, WattTime provides a platform that companies and organizations can use to schedule their load and lower their emissions. Using the right metrics to optimize load shifting really matters.

Here is a concrete example that shows how marginal emissions can be more accurate than average emissions for scheduling a flexible load. Suppose there are three power plants on the grid: 10 MW of solar (zero emissions), 10 MW of efficient gas (medium emissions), and 4 MW of inefficient gas (high emissions) that is brought on when demand is high. At 1am, demand is low, just 5 MW, so the efficient gas plant runs with room to spare. (Solar is not available at night, and this grid has no storage.) At 4pm, demand is much higher, 21 MW. The solar and medium gas plants are running at capacity, and the inefficient gas plant is at one-quarter capacity.

The average emissions are lower at 4pm, just over halfway between solar and the efficient gas. But the marginal emissions are lower at 1am. At 1am, new loads get served by the efficient gas plant, while at 4pm they get served by the inefficient gas plant. So it’s a good idea to move a 4pm load to 1am, at least as long as the 4pm load is high enough to engage that dirty plant. But it doesn’t make sense to move a 1am load to 4pm. It will just generate more emissions, though average emissions at that time are lower. Marginal emissions are great for optimizing load shifting, and flexible loads that are scheduled smartly like this help to reduce our emissions.

So all of this makes sense so far. But I’ve seen marginal emissions being used when I don’t think they make sense, so I checked in with Lena Perkins, the Manager for the Program for Emerging Technologies and acting Senior Resource Planner at the City of Palo Alto Utility. It turns out she shares that concern and helped me to think through the problem. She pointed out that marginal emissions are often short-lived and volatile, so while they are useful for short-term scheduling of flexible loads, they are not a good fit for evaluating longer-term changes. And because they reflect only the margin, they are also inapplicable for evaluating scenarios with very large loads, which may extend beyond the margin.

So it wouldn’t make sense to use marginal emissions to talk about carbon footprints or to talk about the impact of a new appliance. In fact, I struggle to find many good applications for marginal emissions beyond load shifting, though I’m still thinking this through. Let’s walk through a scenario involving adding a new electric appliance.

Suppose you have a grid with a 50 MW solar plant and a 100 MW gas plant. The typical demand is 100 MW, so the typical power would be half solar, half gas. The marginal power source is gas, meaning that any additional load would be matched with gas.

Now let’s say the homes in the area all have gas heaters. Residents can exchange them for $50 for electric heaters. If the electric heater is plugged into a 100% gas grid, its emissions are the same as the gas heater. But if the heater is plugged into a 50% gas grid, its emissions are one-half that of the gas heater. The electric heater has lower emissions than the gas heater as long as it is not running on 100% gas.

The residents are all concerned about global warming. You might think that they would rush to exchange their gas heaters for electric ones, since the grid is generally only half gas. But the residents are worried about marginal emissions, so they do no such thing. Plugging the heaters into the grid would only fire up the gas plants more, they reason. The heaters would generate the same emissions, plus it would cost them $50. The residents have somehow internalized that, even though the plugs in their houses all connect to the same “half gas” grid, any *new* electric appliance will actually be plugged into an “all gas” grid.

Even as the grid gets cleaner over time -- after 20 years, it is only 30% gas -- the residents cling to their gas heaters because they don’t want to plug into that marginal power source. (Of course by then the electric heater company has gone out of business.) Not until the grid has virtually no gas on it do these residents plan to switch over and reap the benefits of the clean supply.

What is wrong with this picture? It makes little sense imo to reason that the clean parts of the grid are allocated only to certain appliances, or that improvements to the grid accrue only to “old” appliances. Marginal emissions are short-lived and volatile, and should not be used to evaluate longer-term changes like this.

Unfortunately, there is some lack of agreement around which emissions model to use for which purposes, and Lena Perkins says that this can make the metrics vulnerable to “cherry picking”, when people use the metric that makes their initiative look most compelling. Consider efficiency improvements. Suppose you want to insulate your house or buy a new refrigerator. For the same reason that new additions to the grid should use average emissions, I believe that these removals from the grid should use average emissions as well, even though that makes the case for them weaker than if marginal emissions were used. But I’ve seen some policy makers suggest using marginal emissions for these initiatives.

Well, you know better now! From my perspective, anyway, it makes sense to default to average emissions for most things but to use marginal emissions to optimize short-term load-shifting. And hopefully we can drive both metrics down with aggressive policies that encourage more renewables to be built, more flexible demand to be developed, cleaner and cheaper storage technologies to come to market, and smarter load-shifting to eliminate dirty peaks. California is at the forefront of much of this, and it is fantastic to live here and support the people and organizations who are making this happen.

Notes and References
0. If you are interested in learning more about fuel switching, Palo Alto Utilities is hosting the Bay Area Electrification Expo on Thursday, October 10 from 2-7pm. There will be speakers and hands-on exhibits for residents and builders.

1. In states that use coal, for example, price and emissions are not aligned since coal is relatively cheap but has very high emissions.

2. If you are interested in what marginal emissions look like, below is a heat map of marginal emissions for 2018 for the northern portion of CAISO (NP15), provided by Christy Lewis of WattTime. (These marginal emissions were calculated using the standard heat rate model for marginal emissions.) The bottom x-axis shows month of the year, from January on the left to December on the right. And the left y-axis shows hour of the day, from midnight and very early morning at the bottom to 11:30pm at the top. You can see that marginal emissions are generally lowest at midday and in spring, and highest in the evenings and late summer.

Current Climate Data (August 2019)

Global impacts, US impacts, CO2 metric, Climate dashboard (updated annually)

Comment Guidelines

I hope that your contributions will be an important part of this blog. To keep the discussion productive, please adhere to these guidelines, or your comment may be moderated:
- Avoid disrespectful, disparaging, snide, angry, or ad hominem comments.
- Stay fact-based, and provide references (esp links) as helpful.
- Stay on topic.
- In general, maintain this as a welcoming space for all readers.
Local Journalism.
What is it worth to you?


Posted by Tom, a resident of Menlo Park,
on Sep 29, 2019 at 11:06 am

I think it's even simpler than you show for making electrification decisions (electric load growth) or electric efficiency decisions (electric load reduction). Instead of using the statistical average of what the mix currently is, I use the Renewable Portfolio Standard (RPS) policy that the utility follows. e.g. Palo Alto and SVCE follow a 100% carbon free resource policy, Peninsula Clean Energy follows a 90% carbon free policy going onto 100% by 2021. PG&E follows a RPS scheduled to get them to 60% qualified renewables by 2030. e.g. to add 100 kWh load by then PG&E needs to get at least 60% of it from renewables and 40% or less from fossil fired resources.
So in Palo Alto and SVCE if someone adds an electric heat pump using 1,000 kWh/year the utility has to go get an additional 1,000 kWh/year from carbon free sources, they don't have policy room in their portfolio to meet the new load with fossil fired resources. Same with PCE in 2 years and almost the same now since they have an almost 100% carbon free policy already. If someone else reduces load by 1,000 kWh then the utility didn't have to do any new procurement. In that case maybe the new load at one customer was met by the declining load at another.
Zero Carbon or low carbon RPS make a clean electric stepping stone for us to conveniently decarbonize by electrifying old fossil uses. Electrification of transport (EVs) and building needs (Heat Pumps) is needed to avoid looting the future from our kids. Lets get on with electrification.
For the next decade we probably will still have a thin band of fossil fired generators on the margin ready to ramp up if needed. Adding load doesn't necessarily make that fossil band thicker, but it does cause any California utility (or load serving entity, LSE) to add renewables. Building dispatch-able storage or building dispatch-ability into load can thin down that band of fossil generation needed to follow load. In the meantime I move my movable loads to the greener mid day hours of the carbon tapestry graph, just to soak up some of my rooftop solar and help smooth things out for the hard working grid.

Posted by Anon, a resident of Another Palo Alto neighborhood,
on Sep 30, 2019 at 1:53 pm

I understand that this is an aside/nit regarding the point you are making about marginal emissions, but, I hope most people reading this understand that electrical watts are not completely interchangeable with thermal watts:

>>Now let's say the homes in the area all have gas heaters. Residents can exchange them for $50 for electric heaters. If the electric heater is plugged into a 100% gas grid, its emissions are the same as the gas heater.

When, e.g. natural gas is turned into electricity, by the most efficient process known Web Link , it will only turn into something like 56% heat delivered into your house by your electric resistance heater, (and on average, somewhere around 30% realistically) while, in a home gas furnace, while new energy-star furnaces with new, high-efficiency ductwork can deliver 90%+ Web Link

OTOH, if the electricity is used to drive a heat-pump, you can get some or all or more of that efficiency back. And, you can do even better with your natural gas source with a gas-driven heat pump (a so-called "absorption" heat pump), which I used to see advertised when I was young, but, which are not well-known residentially now. Source: Web Link . These usually were an ammonia-based heat-pump, and, as such, can be a maintenance/safety issue. Anyway, I'm not trying to beat a dead horse, -honest-, but, I think it is good to appreciate that heat-pump options provide additional options. As well as complexity when considering BTUs and Watts. (Sorry.)

Back to marginal emissions.

Posted by DebbieMytels, a resident of Midtown,
on Oct 1, 2019 at 12:55 pm

DebbieMytels is a registered user.

Thanks, Sherry, for an effective job of explaining a VERY complex issue about which few people have any understanding. But it's important that more of us learn about these issues, since the future of our children, grandchildren and civilization depends on getting a better understanding of HOW the miraculous energy that drives our complex, multi-dimensional lives is produced.

For me, the "take away" of this article is that it aims to explain why a resident (and the utility engineers that might be offering "cost-effective" rebates to such residents) would decide NOT to invest the $50 in an emission-free, renewables-using heater. As your article explains, the fact that "some" fossil fuel may sometimes be needed to generate the heat and thus produce "marginal" emissions seems to justify the decision not to make any investment in the emission-free technology. In short, "the perfect" is seen as the enemy of "the good."

Tom from Menlo Park, however, is suggesting a better policy FRAMEWORK for looking at such decisions. We need to state as our POLICY that we will have a 100% renewable SOURCE (or whatever level is being set for the "renewable portfolio standard" -- RPS) -- irrespective of the "marginal emissions" that may be created. Having an RPS goal will lead to the development of MORE renewables (and ideally, the storage to support them in dark, non-windy hours). This will make the issue of "marginal" emissions increasingly irrelevant over time.

Yes, it may mean that we will have to pay a little more per unit for the energy infrastructure that produces our amazing way of life -- and that we won't eliminate all emissions immediately. But we need to consider the big picture: we need to change over the systems now in place -- and do so as QUICKLY as possible! The long-term goal AND the short-term one is to get as many renewables operative ASAP. By keeping that goal foremost, and rapidly growing the capacity of renewables (and storage), the issue of "marginal emissions" will no longer matter.

With the growing load of greenhouse gases in our atmosphere, we really don't have a choice.

Posted by Sherry Listgarten, a DanvilleSanRamon.com blogger,
on Oct 1, 2019 at 1:23 pm

Sherry Listgarten is a registered user.

I love the comments on this blog!

@Tom. That’s a great point that any added electric load in our area results in clean energy being added to the grid, because our utilities have set policies to purchase clean energy to match load. So the fastest approach to a cleaner grid would be for homes/businesses in areas with dirtier electricity (per utility policy) to implement efficiency, while homes with cleaner electricity (per utility policy) would implement electrification.

@Anon. Yes! You are anticipating some of my next post, which will talk more about the relative emissions from gas and electricity to the home, and the implications of that for electrification. BTW, interesting about the gas-powered heat pumps, I wasn’t aware of those!

There are a lot of factors to consider when asking “Is it cleaner to do A than to do B?” At some point, you simplify and just tell yourself “Go electric”, but I’m trying to give readers (and myself!) a better understanding of the different aspects that go into it.

@Debbie. I was writing this while you submitted your comment, but please note that this post says that the residents are not correct to hold onto their heaters on the basis of marginal emissions, and average makes more sense. Tom adds an extra wrinkle, asking what policy their utility has for power purchases, which is a good question :)

Posted by DebbieMytels, a resident of Midtown,
on Oct 1, 2019 at 3:37 pm

DebbieMytels is a registered user.

Actually, Sherry, I did read your conclusion correctly: your conclusion is that residents should rely on average emission calculations and go ahead and purchase the renewable-based heater. But my comment is that the article is giving an explanation about why "marginal emissions" are used by some people (and utilities) as justification for not switching away from fossil-fuel based heaters.

I wrote(see above):
" ...this article ... aims to explain why a resident (and the utility engineers that might be offering "cost-effective" rebates to such residents) would decide NOT to invest the $50 in an emission-free, renewables-using heater. . . . In short, 'the perfect' is seen as the enemy of 'the good.' "

And since many people tend to rely upon such experts for advice on making such decisions, your explanation of WHY it's not a good idea to worry about marginal emissions is a helpful addition to the discussion.

In trying to underscore your point, I may have obfuscated the situation even more... Alas, such is the challenge of trying to explain such complex issues in a forum such as this!

Posted by Anon, a resident of Another Palo Alto neighborhood,
on Oct 2, 2019 at 12:33 pm

>> BTW, interesting about the gas-powered heat pumps, I wasn't aware of those!

At one time, some gas companies were promoting natural gas absorption heat pumps for residential use, although these days, it seems that these units are mainly for larger buildings. Here is something that I found in production today: Web Link . In theory, these should be significantly more efficient in warmer climates than gas->electricity->(heating/cooling/hot-water) applications. There are also proposed units driven off of solar thermal energy, and other variations using water evaporative cooling (outside - "think chillers", not swamp coolers) that is extremely economical with the water. (Aside: I think I mentioned this before, but, back in the day, ammonia leaks were a serious issue that I believe damaged the reputation of these technologies. TMI: When I was young, my family had an ammonia absorption cycle kitchen refrigerator.) . In all, a lot of ways to get more Joules where they are needed than just using direct electrical Watts. Some of these technology applications might reduce carbon emissions in some cases.

Posted by Alan, a resident of Menlo Park: Belle Haven,
on Oct 4, 2019 at 11:52 am

So - is the implication of this that marginal emissions are lower in the middle of the night - when the overall load on the grid is lower? How universal is the "cheapest is the cleanest" rule?

PG&E is changing their pricing plan for homes with electric cars, so the "Off Peak" hours extend until 3 PM, "Peak" is from 4 PM to 9 PM, and the rest of the times are "Part Peak". Web Link . It used to be that rates went up at 7 a.m. .... I am assuming this drop in price reflects solar power coming online in the morning, before the demand is quite so high.

The interesting thing about electric heat pumps is that it's colder at night, so that's when they should be working the hardest. The "dirtier" sources that come online during max load aren't in play when a heat pump has to work its hardest. A/C is another matter, but you aren't going to use natural gas directly to cool your house.

Posted by Sherry Listgarten, a DanvilleSanRamon.com blogger,
on Oct 4, 2019 at 5:05 pm

Sherry Listgarten is a registered user.

@Anon. Wow. At some point I should do a post on all the kinds of heaters. Or at least some of them. I’m sticking with the basics for now, since that’s enough to keep me busy! Plus I understand it’s hard enough to find people who know how the (new) basics work, let alone everything else… Hopefully I’ll have time to look into more variations at some point. Thanks for sharing all that, super interesting.

@Alan. Thanks for forwarding that! I love that utilities are cleaning up their rate schedules to better match the grid. Not sure if you read the “really matters” link in this blog, but there is a section there titled “Bad timing creates emissions”. The article talks about how when we introduced batteries on the grid, emissions went up rather than down because the time-of-use rates were not set to match emissions. They are fixing that now (aligning them with marginal emissions). See here for more information. And, yes, you are right that space heat pumps have to work much harder at night when it’s cooler. You will be interested in my upcoming blog, though I use average rather than marginal emissions, because this is a fuel-switch context rather than a time-switch context.

Posted by Anon, a resident of Another Palo Alto neighborhood,
on Oct 5, 2019 at 4:39 pm

Posted by Alan, a resident of Menlo Park: Belle Haven,, on Oct 4, 2019 at 11:52 am

>> A/C is another matter, but you aren't going to use natural gas directly to cool your house.

You likely won't in California, but, you could:

Web Link

Web Link

Follow this blogger.
Sign up to be notified of new posts by this blogger.



Post a comment

Sorry, but further commenting on this topic has been closed.

Stay informed.

Get the day's top headlines from DanvilleSanRamon.com sent to your inbox in the Express newsletter.

Premiere! “I Do I Don’t: How to build a better marriage” – Here, a page/weekday
By Chandrama Anderson | 0 comments | 1,637 views

Community foundations want to help local journalism survive
By Tim Hunt | 20 comments | 1,335 views

Pop open the beer at the holiday table
By Deborah Grossman | 1 comment | 482 views