Updraft Energy

Perspective

Higher Education and the Scope 2 GHG Protocol Revision

By Grant Jones, Director, Updraft Energy

Proponents of hourly clean energy matching are supporting the proposed updates to the GHG Protocol's Scope 2 requirements1. Universities have contributed much of the research behind this movement but have been almost entirely absent from the buyer conversation on carbon-free energy (CFE) accounting.

University researchers, like Princeton's ZERO Lab2, have been significant contributors to the hourly matching conversation around system-level emissions and cost premiums for voluntary CFE procurement3 and hourly matching CFE for grid-based hydrogen production4. As energy buyers, however, universities have been largely absent. In fact, we're not aware of a single college or university that has committed to tracking its electricity against clean energy supply on an hourly basis or set a 24/7 carbon-free energy target.

Case study: UPenn, all 8,760 hours of 2025 See how a solar PPA performs when clean energy is matched hour by hour. Jump to the heatmap & visuals

What are the proposed Scope 2 updates about?

Under current GHG Protocol guidance for Scope 2, an organization can claim 100% renewable electricity by purchasing enough certificates over a year to equal its annual consumption, regardless of when or where that energy was generated. The proposed revision would require market-based claims to match consumption hourly and come from resources deliverable to the grid region where the load is located. This shifts value toward clean firm generation (geothermal, nuclear, hydropower), energy storage, and demand flexibility that can serve load during hours when wind and solar are not producing.

The current Scope 2 guidance allows intermittent energy generation from wind and solar to be applied against all loads, even if the generation doesn’t occur at the same time as the load it’s being matched against.

  1. January 2026First consultation closed with nearly 1,100 responses, which overall indicated low support for hourly matching and deliverability as proposed5
  2. Coming monthsRevised draft and second consultation
  3. Next 1–2 yearsFinal standard expected

What changes for campuses

Implications for higher education

Colleges and universities have been historical leaders in the voluntary clean electricity procurement market.

Higher education was a significant driver of early clean energy procurement models, though large corporations and data centers have become the primary focus in more recent years.

Many campus climate commitments rely on unbundled Renewable Energy Certificates (RECs), often sourced from distant grids, or solar-only Power Purchase Agreements (PPAs). Under the proposed rule changes, those certificates may no longer support a low market-based Scope 2 figure moving forward. Institutions that have reported progress toward carbon neutrality using this approach may see very different results under the revised method.

Campus reporting frameworks, including STARS and Second Nature's climate commitments, are built on GHG Protocol conventions and are likely to follow any changes to GHG Protocol rules.

It's worth noting that smaller institutions may fall below proposed load thresholds and be exempt from hourly requirements, and most hourly matching proponents are in favor of carving out allowances for existing contracts.

Case study

UPenn's Great Cove Solar PPA

In 2020 the University of Pennsylvania signed a power purchase agreement with Community Energy, later acquired by AES, for two solar facilities in Franklin and Fulton Counties, Pennsylvania. Great Cove I and II total 220 MW across 1,600 acres and reached commercial operation in December 2023.

The UPenn solar PPA is a long-term commitment that caused new clean generation to be built, in-region, in a grid that has historically lagged on solar. Even in 2025, Pennsylvania generated under 1% of its electricity from utility-scale solar. The Great Cove projects would not exist without UPenn's commitment. And given the low penetration of renewables in PJM, UPenn’s solar PPA likely has a much greater emissions impact than a similar project in grid regions facing increasing rates of solar curtailment, like CAISO.

PENNSYLVANIA ~140 miles Great Cove I 70 MW, Fulton County Great Cove II 150 MW, Franklin County UPenn campus Philadelphia
Great Cove I and II and UPenn’s campus are all within the PJM grid region. Facility locations are approximate.

We modeled all 8,760 hours of 2025 using UPenn's reported monthly consumption, the PECO Consolidated Commercial & Industrial class hourly load shape, PJM's actual hourly generation by fuel type, and EIA-923 metered monthly output for both Great Cove facilities.

8,760 Hours of 2025

Modeling every hour of 2025 – Estimating UPenn's energy

0% carbon-free100% Columns are days; rows are hours of the day (Eastern Time)

Hover or tap any cell for that hour's CFE score, UPenn load, Great Cove output and PJM grid mix. Scroll or pinch to zoom, drag to pan.

Hourly Averages

Average UPenn load vs. Great Cove solar PPA generation

UPenn – Estimated Average Hourly Load

Great Cove I & II – Estimated Average Hourly Generation

Hourly Matching

Estimating UPenn's energy supply by the hour

Solar PPA (used) Grid carbon-free Grid carbon-based Solar PPA (excess, above load) UPenn load

Key takeaways

  • Excess SolarRoughly 30% of the solar generation under the PPA occurred in excess of UPenn's consumption (i.e., ~109 GWh of solar generation counted under an annual matching framework is not included within hourly matching)
  • Nuclear & Grid CFEGrid-supplied CFE in PJM suffers from low penetration of solar, wind, and batteries; however, nuclear's 31% share of generation in 2025 provides meaningful round-the-clock CFE for PJM loads.
64.4%UPenn's 24/7 CFE score
0%100%
39.6%PJM's grid CFE score
0%100%

The visuals above demonstrate the daily and seasonal realities of a solar-only clean energy portfolio. Tracking and reporting clean energy procurement against load on an hourly basis lets your organization identify the gaps and system-level challenges preventing deep grid decarbonization.

Methods

Using the data sources below, we modeled all 8,760 hours of 2025's energy supply vs. demand for UPenn.

For each hour of 2025:

UPenn demand

Estimated hourly UPenn demand was calculated based on an assumed annual value of 600,000 MWh for its academic campus and the UPenn Health System. This annual 600,000 MWh was then split by month to the same proportional share of electricity consumed by month reported by UPenn within its most recent STARS report for its academic campus.

Note: This monthly electricity consumption was reported for only UPenn's academic campus and may be an inaccurate approximation for the Health System.

Finally, for each hour of the month a load profile factor was applied based on PECO's hourly load data for Consolidated Large Commercial & Industrial Customer Group.

Great Cove solar

Estimated hourly Great Cove solar generation was calculated based on the monthly electricity net generation values reported for Great Cove Solar and Great Cove Solar II in Form EIA-923 for 2025. These monthly supply values were then shaped on an hourly basis to match the actual, proportional solar output across PJM by hour.

Grid-supplied CFE

Hourly values for grid-supplied CFE come from real, hourly PJM grid mix values. CFE each hour is the sum of nuclear, wind, solar, and hydropower generation (with the estimated hourly supply of Great Cove I & II removed).

Note: PJM's System Mix incorporates RECs retired for state RPS compliance. No public hourly residual mix exists for PJM or PECO.

Calculating the CFE score

h
Each hour of 2025, 1 through 8,760
UPenn load, Solar PPA, CFE
Energy in MWh for that hour
PJM CFE share
Carbon-free share of PJM generation in that hour, net of Great Cove I & II
Grid-supplied CFE, each hour
Grid-supplied CFEh=(UPenn loadh−Matched solar PPAh)×PJM CFE shareh
Hourly CFE score
Hourly CFE scoreh= min(Solar PPAh, UPenn loadh) + Grid-supplied CFEhUPenn loadh

Note: The MWh of energy from the solar PPA cannot exceed UPenn's load in any given hour, so the matched solar PPA in each hour is min(Solar PPAh, UPenn loadh).

24/7 CFE score for 2025 (load-weighted)
24/7 CFE score2025= 8,760Σh = 1 UPenn loadh × Hourly CFE scorehAnnual UPenn load

Data Sources & Assumptions

Supply

  • PJM Grid Mix: Hourly data for PJM generation by fuel6
  • Solar PPA: Monthly generation data for 2025 from Form EIA-923 for Great Cove I & II7

Demand

  • Estimated University of Pennsylvania Annual Demand: 600,000 MWh8
  • University of Pennsylvania Share of Electricity by Month (FY23)9
  • PECO Hourly Load Data for Consolidated Large Commercial & Industrial Customer Group10

What stands in the way

Challenges

Hourly matching costs more than annual matching.

Research shows a cost premium at high matching levels when portfolios rely only on wind, solar, and short-duration batteries. The premium decreases when clean firm generation and long-duration storage are included. Given current budget pressures in higher education, large new procurement commitments will be difficult in the near term.

Not all regions in the U.S. have access to certificate registries supporting hourly tracking, and not all higher ed campuses have complete interval data. However, institutions can estimate supply through data resources from their local utility, grid operator, or accessible data platforms like Electricity Maps11 and Grid Status12. While granular meter data is often available from utilities, demand can also be estimated using publicly available load profile data.

Hourly matching is also not the only approach under consideration. Some buyers and analysts favor emissions-impact accounting, and the GHG Protocol is evaluating a version of it outside the Scope 2 inventory. Under either approach, however, institutions need to understand when and where they use electricity and the emissions intensity of the grid at those times.

BU case study

BU began buying clean energy from the Triple H wind farm in 2020, committing to buying 205,000 MWh of electricity each year through a 20-year PPA.13

BU evaluated 127 wind and solar project proposals from across the country and selected the South Dakota wind project in the Southwest Power Pool grid region due to the projected greatest impact on global emissions. BU estimated the avoided emissions of this project would be greater than a project in New England based on analysis of the marginal emissions rates for additional clean energy generation by grid region. This is often referred to as an “emissionality” approach for siting new clean energy on the grid.

~1,400 miles Triple H Wind Hyde County, SD SPP grid region Boston University Boston, MA ISO-NE grid region
BU buys 205,000 MWh per year from Triple H, about 1,400 miles from campus, in a different grid region. Triple H location is approximate.

Grid comparison

New England (ISO-NE) vs. Southwest Power Pool (SPP), 2020–2025

Project impact

Avoided CO2 over BU's 20-year PPA, 2021–2040

Using EPA’s Avoided Emissions and Generation Tool (AVERT) and the National Laboratory of the Rockies (NLR) Long-Run Marginal Emission Rates for Electricity Workbook, we calculated the marginal emissions rate and estimated total avoided emissions for the Triple H wind project, a New England wind project, and a New England solar project that would each generate 205 GWh of CFE annually.

Thousand metric tons of CO2 avoided, assuming 205,000 MWh per year for each project

2021–2024, short-run (EPA AVERT) 2025–2040, long-run (NLR Cambium, Mid-case)

Methods

Using the data sources below, we compared the carbon intensity of the New England and Southwest Power Pool grids, then estimated the CO2 avoided by three projects that each generate 205,000 MWh per year over BU’s 20-year PPA (2021–2040).

Grid comparison

Generation by fuel for the ISO-NE and SPP grid regions comes from EIA’s hourly and daily grid data. Emissions intensity is estimated CO2 from coal, natural gas, and oil generation, using EIA emission factors, divided by all generation within each region. Values are 12-month rolling averages.

Note: These are average rates across all generators. They show how carbon-intensive each grid is, not the emissions a new project would displace. Imports and biomass emissions are not included.

2021–2024: short-run

EPA publishes annual avoided CO2 rates from AVERT, which estimates how fossil power plants in each region respond to added wind or solar generation based on their actual historical operations. We used each year’s rate for Central onshore wind (Triple H), New England onshore wind, and New England utility-scale solar.

Note: For years 2021–2024, short-run marginal emissions rates (SRMER) analysis from EPA’s AVERT tool is used to estimate emissions impact in each grid region, while years 2025–2040 utilize long-run marginal emissions rate (LRMER) from NLR’s Cambium 2023 model. SRMER-based analysis is more appropriate for estimating immediate impacts to existing power plants on the grid over the course of hours/days/months, while LRMER-based analysis aims to capture grid impacts over the course of many years. Average short-run rates in both ISO-NE and SPP are roughly 2–4x as high as long-run rates on average in Cambium’s modeling from 2025–2040.

2025–2040: long-run

NLR’s long-run marginal emission rates also account for how new generation changes what gets built and retired over time. We used the Cambium 2023 Mid-case for the SPP North region (which includes Triple H’s location in Hyde County, SD) and ISO-NE: CO2 from combustion, measured at the busbar, averaged over 2025–2040 without discounting. Rates by month and hour of day were weighted by AVERT’s default hourly generation profiles for each resource type and region, not project-specific output data.

Note: Cambium 2023 assumes federal clean electricity tax credits continue through 2050, which 2025 legislation has since curtailed. NLR advises that these estimates are illustrative and not a basis for impact claims.

Calculating avoided CO2

Gen
205,000 MWh per project, each year
p, r
Project; its grid region (SPP North or ISO-NE)
y
Year, 2021 through 2024
m, h
Month and hour of day (288 combinations)
Share
Project’s share of its annual generation in that month and hour
Rates
lb CO2 per MWh, converted to metric tons
Short-run, 2021–2024
Avoided CO2, 2021–2024= 2024Σy = 2021 Gen×AVERT ratep, y
Long-run rate, weighted by generation profile
LRMERp=  Σm, h Sharep, m, h×LRMERr, m, h

LRMERr, m, h is NLR’s levelized 2025–2040 rate for that region, month, and hour.

Long-run, 2025–2040
Avoided CO2, 2025–2040=Gen×16 years×LRMERp

Total avoided CO2 is the sum of the two periods.

Data Sources & Assumptions

Grid comparison

  • Generation by fuel: U.S. EIA Open Data, electricity by grid region14
  • Emission factors: U.S. EIA, CO2 emission factors by fuel15

Project impact

  • Short-run rates: EPA AVERT avoided emission rates, 2021–202416
  • Long-run rates: NLR Long-Run Marginal Emission Rates for Electricity, Cambium 202317
  • Generation: 205,000 MWh per year for each project, BU’s contracted annual quantity13; deliveries began December 2020, so 2021 is the first full year

Why campuses are well placed

Campus advantages

Existing flexibility

Many campuses operate central plants, thermal storage, and microgrids.

Long planning horizons

Institutions can enter long-term agreements for clean firm resources.

A living laboratory

Vast real-world research, education, and entrepreneurship opportunities for institutions that embrace hourly matching.

Campuses can also use their own energy data for research and instruction. Building an hourly CFE score from interval meter data and hourly grid emissions is a practical project for engineering and business students and a useful dataset for energy researchers. The charts above were built entirely from public data sources.

Recommended steps

  1. Track hourly performance: Calculate an hourly CFE score using existing interval data and hourly grid mixes. This requires no procurement changes and shows how current portfolios would perform under the proposed rules. This data collection process through utility collaboration, data sourcing and analysis, and visualization is a phenomenal exercise for interdisciplinary students to have hands-on learning and treat campus operations as a living lab for next-generation clean energy policy frameworks.

  2. Engage networks: Work with peer institutions and learn how others in higher education are thinking about these issues. At Updraft Energy, we’re evaluating how we can build communities of practice within our partner organizations (AASHE, Second Nature, Carnegie, ACE), please reach out if your institution is interested in collaborating!

  3. Participate in the second consultation as institutions: In addition to individual researchers, feedback from higher education leadership on the GHG Protocol's Scope 2 framework updates will have major implications for the industry, including reasonable pushback for allowances of existing contracts, exemptions for smaller institutions, and the use of alternative methods like BU's emissionality-based procurement strategy.

  4. Set interim targets for hourly, deliverable clean energy, and evaluate consortium or university-system procurement to reduce costs.

Grant Jones is a Director and co-founder of Updraft Energy, which advises colleges and universities on decarbonization strategy, clean energy procurement, and sustainability programs, including STARS and the Carnegie Elective Classification for Sustainability. As Deputy Director for Sustainability at the White House Council on Environmental Quality, he supported strategies to meet the federal government's 50% 24/7 carbon pollution-free electricity target under Executive Order 14057 (since rescinded). He also organized a White House convening on hourly matching for the 45V hydrogen production tax credit and participated in an EPRI workshop series with utilities and large-load customers on rate design and barriers to hourly matched clean energy programs.

Sources

  1. GHG Protocol, Scope 2 Guidance. ghgprotocol.org/scope-2-guidance ↩
  2. Princeton ZERO Lab. zero.lab.princeton.edu ↩
  3. Xu et al., System-level impacts of voluntary carbon-free electricity procurement strategies, Joule (2024). cell.com/joule/fulltext/S2542-4351(23)00499-3 ↩
  4. Minimizing emissions from grid-based hydrogen production in the United States, Environmental Research Letters. iopscience.iop.org/article/10.1088/1748-9326/acacb5 ↩
  5. GHG Protocol, Scope 2 public consultation: executive summary and summary of feedback (July 2026). ghgprotocol.org ↩
  6. PJM Data Miner 2, generation by fuel type. dataminer2.pjm.com/feed/gen_by_fuel ↩
  7. U.S. EIA, Form EIA-923 detailed data. eia.gov/electricity/data/eia923 ↩
  8. UPenn Sustainability, Power Purchase Agreement. sustainability.upenn.edu ↩
  9. AASHE STARS, University of Pennsylvania report. reports.aashe.org ↩
  10. PECO Procurement data room, monthly load data. pecoprocurement.com/index.cfm?s=dataRoom&p=monthly ↩
  11. Electricity Maps. app.electricitymaps.com/map/fifteen_minutes ↩
  12. Grid Status. gridstatus.io/live ↩
  13. Boston University Sustainability, BU Wind. bu.edu/sustainability/projects/bu-wind ↩
  14. U.S. EIA Open Data, electricity by RTO: daily region data. eia.gov/opendata/browser/electricity/rto/daily-region-data ↩
  15. U.S. EIA, emission factors (Electric Power Annual, Table A.3). eia.gov/electricity/annual/html/epa_a_03.html ↩
  16. U.S. EPA, AVERT avoided emission rates, 2021–2024 (v4.4 provisional, May 2025). epa.gov/avert/avoided-emission-rates-generated-avert ↩
  17. National Laboratory of the Rockies (NLR), Long-Run Marginal Emission Rates for Electricity: Workbooks for 2023 Cambium Data (Gagnon, 2024). data.nlr.gov/submissions/230 ↩

Get in touch

Some work can't wait.

Tell us what you're trying to accomplish, and we'll tell you honestly whether we're the right partner.

Start a conversation