This is the third post in a Wattwatchers blog series about high-resolution carbon accounting. In our first post we introduced the concept and its background. The second post focused on the data processing and modelling required for high-resolution carbon accounting. This post focuses on the superior accuracy of high-resolution carbon accounting and why it matters.
In a nutshell, high-resolution carbon accounting calculates the emissions intensity of the grid over small time intervals (30 minutes in our case). This is in contrast to the current practice of using an annual average value for the emissions intensity of the grid. High-resolution carbon accounting promises to be a more accurate method to estimate carbon emissions.
Why is high-resolution accounting more accurate?
Before discussing why accounting accuracy is important, let’s first review why high-resolution carbon accounting is more accurate than annual average accounting.
Annual average accounting assumes that the emissions intensity associated with electricity imports are constant over time. In reality, the emissions intensity is variable. This is especially true in regions with a lot of renewable energy generation capacity (wind and solar). To illustrate this, Figure 1 depicts the emissions intensity of electricity imports in South Australia over a 2 week period.
The pattern of electricity consumption on a single site (e.g. a single household or commercial premises) is also variable. It often includes clear day/night patterns and seasonal patterns. The combination of variable emissions intensity and variability in import patterns means that the actual emissions over a one year period can significantly deviate from emissions calculated with an annual average emissions factor. Calculating and using the emissions factor for every 30-minute interval will result in calculated emissions much closer to the actual emissions.
Why does accuracy matter?
The objective of carbon accounting is to reduce emissions. It’s also vital for offsetting emissions and for putting a dollar value on reductions. Inaccuracies in accounting lead to overestimated emissions for some and underestimated emissions for others. This has a number of undesirable consequences:
- Entities for which the emissions are overestimated will be offsetting emissions they are not responsible for. In effect they are subsidising entities for which emissions are underestimated.
- Entities for which emissions are underestimated don’t get as clear a signal to reduce emissions as they should. This leads to lower emissions reductions that will be slower to materialise.
- No-one gets any signal about the times at which a reduction in electricity imports would most reduce emissions. (The next blog post in this series will focus on this more).
How much more accurate is high-resolution accounting?
The impact of high-resolution accounting on accuracy depends on the usage patterns of the site under consideration. Different sites can exhibit a wide range of usage patterns which will determine how inaccurate using an annual average emissions factor is.
In the extreme case of year-round constant electricity imports, annual average accounting will be accurate over a full year period. The more usage is skewed towards times that the grid’s emissions factor is either above or below average, the more inaccurate annual average accounting will be.
In the remainder of this blog post we analyse the difference in accuracy between annual average accounting and high-resolution accounting for a specific data set.
Accuracy analysis
To get an idea of how much more accurate high-resolution accounting can be for residential sites, we’ll look at the same dataset we used in the earlier posts in this series. The dataset comprises grid imports for 488 residential sites over a one year period (July 1 2021 – June 30 2022).
In the second blog post in this series we presented a basic state-by-state analysis of the inaccuracy of annual average accounting. Figure 2 recaps the inaccuracies found across the 488 sites in 4 National Electricity Market (NEM) states. We will now dig a little deeper to try to explain the inaccuracies.
emissions for 488 sites across 4 NEM states.
Solar vs no solar
One hunch is that the presence or absence of rooftop solar may influence estimation accuracy. Because households with solar tend to import less energy from the grid when the sun shines—correlating with times at which the emissions factor of the grid is low—, we expect average annual accounting to underestimate emissions for households with solar. Put another way, households with solar generate their own energy when the emissions intensity of the grid is low. And they import from the grid when emissions intensity of the grid is high.*
To test this hunch, we divide the data set into sites with solar and sites without solar. We then analyse these two sets of sites separately. As the histograms in Figure 3 and the summary in Table 1 show, our hunch was mostly correct. With the exception of Victoria, annual average accounting underestimated emissions for sites with solar. Also note that for South Australia, annual average accounting underestimated emissions for sites both with and without solar, but more so for sites with solar.
emissions for sites with and without rooftop solar
| Sites without solar | Sites with solar | |
| NSW | 0.1% | -1.5% |
| QLD | 1.6% | -1.7% |
| SA | -2.5% | -7.2% |
| VIC | 1.2% | 1.3% |
emissions for sites with and without rooftop solar (negative values indicate underestimation)
Single site analysis
To highlight how different grid import patterns lead to over- and underestimation when using annual average accounting, we compare two sites in South Australia; one with rooftop solar and one without. (We opted for South Australian sites because the grid there is most representative of a future with increasing generation through renewables)
To get an intuitive idea of what happens at a single site level, we averaged imports over a year by time-of-day in 30-minute intervals.
Figure 4 shows the resulting average import profile for a site with rooftop solar. It clearly shows that the bulk of grid imports happen in the evening (when the sun doesn’t shine). This correlates with the time of day when on average the emissions intensity of the grid is high (see Figure 6). This results in actual emissions associated with grid imports being 18% higher than what annual average accounting would calculate.
Figure 5 shows the resulting average import profile for a site without rooftop solar. It shows that the bulk of grid imports happen during the daytime (when the sun shines). This correlates with the time of day when on average the emissions intensity of the grid is low (see Figure 6). This results in actual emissions associated with grid imports being 5% lower than what annual average accounting would calculate.
Conclusions
Accuracy in carbon accounting is important. For fairness and to provide the right entities with the right (sized) incentives to decarbonise.
Annual average accounting introduces accounting errors because the emissions intensity of the grid isn’t static and the grid import patterns of sites tend to skew to times with a lower or higher than average emissions factor.
Our analysis showed that specific characteristics (in this case, whether a site has rooftop solar) can have a significant impact on actual emissions from energy imports. This can introduce significant inaccuracies when using the annual average emissions intensity to calculate carbon emissions.
Analysis in this article focused on residential sites, which is not where carbon accounting is currently most prevalent (although this will change when homes with mortgages become part of the Scope 3 emissions of banks). However, the same principles will hold true for commercial and industrial sites as well.
Next in this blog series
In the next blog post in this series we will focus on carbon efficiency. We discuss what it is, and show how high-resolution carbon accounting gives companies an additional—and powerful—lever to reduce carbon emissions.
Adriaan Stellingwerff is a Senior Software and Data Engineer at Wattwatchers.
* This does not mean that the overall emissions from the electricity use of houses with rooftop solar is higher. This analysis only takes into account the emissions from electricity imported from the grid, and thus electricity generated on site (with very low emissions) is not part of the analysis.
ACKNOWLEDGEMENT
The My Energy Marketplace project is receiving funding from ARENA as part of ARENA’s Advancing Renewables Program. The views expressed herein are not necessarily the views of the Australian Government, and the Australian Government does not accept responsibility for any information or advice contained herein.
