Wednesday, October 14, 2020

Making the Most of Distributed Energy Resources - Subregional Estimates of the Environmental Value of Distributed Energy Resources in the United States

Distributed Energy Resources (DERs), like rooftop solar and battery storage, have the potential to generate significant social benefits by displacing pollution-emitting electricity generators. Accurately compensating DERs for this environmental and public health value, which some regulators and experts call the “E-Value,” is imperative for making the most out of DERs’ potential. Doing so will ensure DERs are deployed, and used, when and where they create the most value for society. In practice, however, calculating the E-Value of DERs is difficult without a detailed model of the electric power sector because the benefits of avoided air pollution can vary significantly by location and time of day or time of year.

This report provides a new set of hourly E-Values for the whole United States, broken down into 19 subregions, using an open-source reduced-order dispatch model.1 Critically, these granular estimates are shown to vary considerably by geography, hour, and season. The patterns uncovered by these estimates can help policymakers design economically efficient DER policies to reduce air pollution from electricity generators. Because these results come from an open-source model, they can be particularly useful for regulators with mandates to use publicly available data in their decisionmaking or for those who desire to do their own analysis.

This report reveals three novel insights based on the hourly E-Values generated by the model. 

First, the E-Values of DERs depend crucially on the location of the DER, as some regions have more pollution-intensive electricity generators than others. 
Second, unlike the production cost savings of DERs (which are generally greater during periods of high electricity demand), there is no general and consistent pattern that can effectively characterize the E-Value of DERs throughout the day. 
Finally, the E-Value of DERs can be large – potentially greater than the benefits of avoided electricity production costs, and generally greater than what commonly used heuristics would suggest.

Policymakers can use these estimates and insights to create effective DER policies and programs. These findings highlight the need for more accurate and granular valuation of DERs, without which investments in DER technologies are likely to be either meager or misdirected. Policymakers using E-Value estimates in the design of DER compensation schemes or the assessment of other DER policies can rest assured they are making the most of DERs’ potential to deliver social benefits in their jurisdiction.
Results from the reduced-order dispatch model suggest the E-Values of DERs can vary significantly by subregion. Figure 2 displays hourly maps of the E-Value of DERs for each subregion averaged by season and time of day. This figure shows that the E-Value depends largely on the geographic region, and less so on the time of day and season. This variation is because some regions use more pollution-intensive fuels to generate electricity than others. For example, the Great Lakes and Ohio Valley regions are heavily dependent on coal electricity generators, which emit a large amount of CO2 and SO2 per MWh. The E-Value is relatively small in California where little-to-no electricity is generated by coal electricity generators.

Other than geographic location, population density can be a large determinate of the E-Value of DERs. Densely populated areas experience more damage from a given amount of pollution as more people are exposed. Results in Figure 2 show there are consistently higher E-Values in the Northeast compared to the Rocky Mountains, in part because the former is more densely populated than the latter. Analysis on the electric power sector done by the EPA illustrates this point in the context of PM2.5: a ton of PM2.5 released in the eastern region of the United States causes between $130,000 and $320,000 in damages, whereas the same ton in the western part of the United States causes $24,000 to $60,000 in damage. Generally, the hourly E-Value can vary significantly throughout the day, as the marginal electricity generator, marginal emissions, and associated health benefits change hourly. If the hourly E-Value were to follow a consistent and general daily pattern, policymakers could use this information to better design DER compensation policies by, for example, compensating DERs the most during the time of day they generate the most social value. But, if the E-Value does not follow a consistent and general pattern, policy makers would have to directly observe hourly marginal emissions or model the specific region to accurately compensate DERs for their intra-day variation in E-Value.
The average E-Value of DERs, across all 8760 hours in a year and 19 geographic regions, was $57/MWh (with a median of $54/MWh). This value is nearly twice the average cost of electricity simulated by the reduced-order dispatch model ($27/MWh), and greater than the national average wholesale price of electricity in 2018 ($44/MWh).26 Figure 4, which displays the simulated average production cost and average E-Value of DERs in every subregion, shows this relationship holds for every subregion except one.

Figure 4 – Simulated Energy Cost Compared to the E-Value of DERs

Each subregions color is based on the geography of the corresponding NERC region. The diagonal line represents equality between the two values, and subregions to the lower right of the diagonal lines have an average E-Value greater than average simulated energy cost. Benefits from avoided greenhouse gas emissions make up nearly half of the E-Value. Decomposing the E-Value of DERs, as done Figure 5, shows the avoided CO2 pollution is a large component of a DER’s benefits on average across all regions and hours. By using the Social Cost of Carbon, the E-Values presented in this report capture, at least in part, the large future damages from climate change (including from coastal storms, extreme weather events, and human health impacts, such as mortality from heat-related illnesses induced by the use of fossil-fuels). Ignoring the benefits of avoided greenhouse gas emissions will provide an underestimate of the total benefits of DERs. For example, recent analysis by the EPA evaluating only the public health benefits of DERs, excluding avoided GHG emissions, range from $17 to $40/MWh on average.
Finally, the E-Value of DERs are large relative to the estimated environmental value based on average pollution emissions from electricity production. This means that basing policy decisions on the commonly used heuristic of average pollution emissions instead of marginal pollution emissions leads to inefficient deployment... The output of the reduced-order dispatch model shows an E-Value equivalent based on average emissions is less than the E-Value based on marginal emissions in 75% of hourly subregion observations....

The modeling results make clear four specific policy implications. 
First, the relative magnitude of E-Value can tip the scales in favor of DERs. Accounting for both the E-Value and the benefits of avoided energy of DER deployment nearly doubles the benefits of DERs in comparison to valuing the avoided energy alone. Ignoring the E-Value of DERs will therefore result in the deployment of fewer DERs than what is optimal.
Second, the E-Value can identify where in the country different DER technologies are most effective. For example, investing in energy efficiency lightbulbs create the most value where the nightly E-value is the largest, and likewise, efficient air-conditioning and rooftop solar should be directed towards the regions with a high E-Value during summer’s midday. Policymakers should use the E-Value when deciding which technologies to support and where they should be deployed.
Third, general policies rewarding DERs during certain times of day might not effectively capture the benefits of DERs because there is no general and consistent daily pattern in E-Values across all regions. Instead, policymakers interested in granular time-of-day policies must model the specific benefits of a DER’s deployment within their region.
Finally, policies must account for both the climate and public health benefits of DERs when calculating the E-Value. Each component makes up roughly half of total E-Value's on average. Ignoring the either component of the E-Value will result in inefficient DER deployment even if the other component is accounted for.

This report proceeds by first establishing the elements that determine the E-Value of DERs, including marginal generators, marginal emissions, pollutant type, location, and timing. With that established, this report then briefly characterizes the model used to calculate the subregional E-Values of DERs and finally summarizes the modeling results. Before concluding, this report discusses the important role of E-Value in policymaking and the several policy implications.

Interested readers are directed to Appendix A to learn more about the reduced-order dispatch model, and Appendix B for a table of average E-Values for each subregion, season, and time-of-day.

New York University School of Law Institute for Policy Integrity
September 2020 Matt Butner, Ph.D. Iliana Paul Burcin Unel, Ph.D. 

No comments:

Post a Comment