The carbon intensity of our electricity grids (gCO2e/kWh) plays a huge role in both the operational and embodied carbon of our buildings. Here are some graphs I’ve made to visualize the grid carbon intensities for all Canadian Provinces/Territories, US States, and the national averages for many countries around the world, based on the latest data I could find.
These numbers will obviously affect the operational GHGI of buildings. They can also significantly impact the embodied carbon of different materials, such as steel from Electric Arc Furnaces and primary aluminum production.
These graphs can be a helpful resource to understand one of the key factors that lead to variations in operational carbon intensities of buildings in different locations, as well as the embodied carbon of products produced across different regions.
Thanks for this, it’s a great breakdown. It beggars belief why Canada doesn’t have one of the greenest grids when you consider how much hydro we have. Also, who wants to go build offshore wind + geothermal in Hawaii?
This is great. When we are talking to clients another consideration is the future GHG intensity of the grid based on current commitments of the state,city, or utility. Some have committed to carbon neutrality by 2050 or so, meaning that after the first 25-30 years of life, an all-electric buildings will have only minimal energy-related carbon emissions…and during the first 25-30 years it will decrease. Many other have 20, 30, or 50% reductions over the next few decades.
-Kjell
Thanks for the post Anthony, and as there has been some interest in the data, I’ve got a couple of other resources for folks that some may be interested in, and helps explain some of the information that Anthony has presented here.
StatsCan provides monthly breakdown of generation by type and province. Here you can break down data how you want, and download the info to a spreadsheet as well.
This site from CER (Canada Energy Regulator) also provides National and provincial generation information, as well as other energy sources. I’ve not reviewed the data to compare the StatsCan and CER data, but being in a position to need to find this data, I want to share these resources.
Both show that some provinces have a lot of work to do for greening the grid.
Thanks Will. The provinces that have hydro and nuclear are quite low, such as BC and Ontario. Overall, Canada is still relatively low at 150 gCO2e/kWh national average, but obviously we also have provinces like Alberta and Saskatchewan using more coal and natural gas in the grid mix.
That’s a great point Kjell. I would love to see how this gets incorporated into energy modelling results, and how that would in turn influence the balance between embodied and operational carbon over the lifespan of the building. We tend to use a simplifying assumption that the operational energy use of buildings will be supplied by the current grid mix over the entire lifespan of a building, which is a conservative assumption given the anticipated increase in renewables.
Thanks Anthony, this is great. One thing I am wondering about is places like Texas or Pennsylvania that are big on gas. Multiple studies lately have been showing that the GHGs from the gas production process are much higher than we thought, and that makes me think the US would look worse than a country like Germany that has a high wind/solar mix if this were taken into account. Something to consider. More info here https://www.edf.org/climate/methane-studies
While using operational carbon intensities for design decisions, say selecting energy-efficiency measures or building electrification, there are a couple other aspects to consider. These do require more granular data from energy models and somewhat complex analyses, but can can more accurately capture ‘real-world impact’.
One is using hourly emissions data instead of a single annual average. In California, for example, we know that there are enough renewables feeding the grid-during the daytime, so much so that they have to be curtailed at times. In late summer afternoons and early evenings, however, the grid is dominated by dirtier power plans, and using energy at that time has a much worse impact. See the image below showing this variability; green indicates clean power while red indicates higher grid emissions intensities.
The second, related aspect is utilizing marginal emissions factors which measure the actual environmental consequence of an energy-consuming or conserving action. Only some power plants will scale up or down in response to say, your sports area turning on or off its lights. The impact of this light turn on/off action should be measured using the emissions intensity of the marginal power plant that’s responding. WattTime, a company that provides this data, does a good job of explaining this concept. Our firm, Atelier Ten, partnered with them to write a white paper on how we use their data for these calculations.
Using hourly and marginal emissions factors can sometimes give surprising results. For instance, even though electrification makes sense in the long run, with cleaning-the-grid commitments as @Kjell_Anderson mentioned, we’ve found that in the near future something like electrifying food service can still be worse than traditional gas-powered. Electrification combined with on-site PVs, or say of HVAC using heat pumps still comes out on top, but the “2030 is the new 2050” mantra we’ve been following in the embodied carbon world adds some shades of grey.
I second Prateek’s point. Great explanation. My colleague at the University of Pittsburgh studied a LEED Gold (less-energy-efficient) and a Net-Zero (more-energy-efficient + on-site PV) buildings in Pittsburgh, PA, using both approaches, i.e. the typical average approach vs. using hourly energy consumption and power generation data. He found that the typical LCA approach underestimates the impact of the less-efficient building and overestimated the impact of the net-zero building. It should be noted that these were both primarily office buildings, and as Prateek mentioned, the use type will also greatly affect the outcomes. It depends on the building and the grid. The full paper is here.
