You can’t manage what you don’t measure.

When it comes to carbon reductions from flexibility in demand generated through efficiency, demand response, electrification, and other changes to energy use, if we don’t measure carbon correctly, we will end up with policies and actions that don’t have the intended effect.

With growing numbers of states, cities, and utilities adopting carbon goals, it is increasingly important that we get the incentives and policies aligned with the climate outcomes we are seeking.

Especially when it comes to buildings, the way we measure carbon offsets today is incomplete or simply inaccurate. Basing estimated carbon reductions on monthly consumption or savings, multiplied by a monthly or yearly carbon per kWh average, can give us the entirely wrong answer. For example, interventions that save energy when the grid is awash in solar may reduce consumption overall while cutting very little carbon.

As we focus on decarbonization, it’s critical to account for the fact that emissions vary substantially by both time of day and year, and the source of the power being consumed. A kWh saved or generated during system peak can have a dramatically different carbon content than a kWh during a different season or at a different time of day. In addition, reducing demand impacts marginal carbon rather than the average intensity in a given hour. A recent Stanford study concluded that “current methods of estimating greenhouse gas emissions use yearly averages, even though the carbon content of electricity on the grid can vary a lot over the course of a day in some locations. By 2025, the use of yearly averages in California could overstate the greenhouse gas reductions associated with solar power by more than 50 percent when compared to hourly averages.”

The carbon intensity of the California grid already varies significantly by hour and time of year, as the CEC projections below illustrate. These projections, which include the impacts of solar and wind combined with storage, demand response, and other resources, show the rolled up daily average carbon intensity of kWh by month for 2019 and forecasted to 2030.

Similarly, energy usage patterns in buildings vary substantially. So two buildings that use the same kWh per month can have very different carbon footprints depending when they consume energy and even where they are located on the grid.

At the same time, all behind the meter solutions to demand, including energy efficiency and electrification, create load shape impacts that in turn determine the carbon impact. When it comes to counting carbon, there is no simple “reduction” to be had without taking time and location into account.

Aligning policy, measurement, and incentives with actual hourly carbon impacts will drive optimal behaviors and business models.

A More Accurate Approach to Calculating Carbon Impacts

The top and middle panels of the graphic below show the hourly resource curves (impact on demand) of PG&E’s Advanced Home Upgrade Program (Home Performance) and Commercial Deemed Program (largely lighting and refrigeration). The resource curves were determined at the meter using hourly CalTRACK methods. The bottom panel shows the average hourly marginal GHG savings per MWh. (Values are on an average hourly marginal basis and are computed for 2024 using California Climate Zone 4 as a representative region using the CPUC’s avoided cost calculator). The dramatic dip in avoided GHG emissions corresponds to periods of high solar intensity and overgeneration.

In short, if renewables are being curtailed, electricity savings will have no GHG impact. By combining the hourly resource curve (metered impact to demand on an 8760 basis) with the marginal carbon intensity of a MWh we can accurately calculate carbon reduction.

The figure below shows how 1 MWh of savings through the AHUP and Commercial Deemed programs accrue avoided GHG emissions. Each data point represents the multiplication of 365 hourly kWh savings measurements  with associated marginal avoided GHG emissions. As expected from the previous figure, we see that the AHUP savings concentrated in the GHG-intensive evening peak hours yield high avoided emissions, while both programs yield lower avoided emissions during the mid-day hours in which the CPUC forecast indicates a high propensity for overgeneration.

With the AHUP program generally delivering savings during periods of high marginal GHG intensity, 1 MWh of savings yields 59% more emissions reduction than 1 MWh of savings from the Commercial Deemed program. With other states encountering duck curves and various other animal shapes depending on their mix of renewables and weather, this type of accounting will be essential across the board.

The figure below shows a heat-map representation (24 hours x 12 months) and carbon accounting for a representative program in the Recurve platform. This type of tracking on a real-time basis is now available and is needed to plan and optimize programs for effective decarbonization.

Measuring carbon correctly is critical if we are to put our limited resources to maximal use fighting climate change. If we measure monthly using average values for GHG intensity, we will send the wrong signal, produce inaccurate results, and ultimately do a bunch of stuff that misses the point.

The good news is that we now have methods that can track hourly savings using AMI data. These methods, including CalTRACK, are now being used by multiple utilities and regulators in a number of states. Combining this approach to metering savings with the type of marginal carbon data available in most states from public sources allows us to finally present a real picture of when, where, and how much carbon is actually being saved by efficiency and electrification programs so that we can adjust in order to meet our climate goals.

Measure twice, cut once!

Contact Recurve to learn more about how our platform can help states, utilities, and local communities meet their decarbonization goals.