Introducing the Open Grid Emissions Initiative
Open Grid Emissions is a Singularity Energy initiative that seeks to fill a critical need for high-quality, publicly-accessible, hourly grid emissions data for greenhouse gas (GHG) accounting, policy-making, energy attribute certificate markets, and academic research. It includes a public dataset of hourly, monthly, and annual U.S. electric grid generation, greenhouse gas, and air pollution data, all calculated using open-source, well-documented, and validated methodologies based on the best available science. We believe this to be the most comprehensive, accurate, and high-resolution public dataset of emissions and generation from the U.S. power sector available to date.
The Open Grid Emissions Initiative furthers Singularity Energy’s mission of data-driven decarbonization, and was motivated by our need for better historical data to validate and improve our real-time and forecasted data offerings. The initiative grew out of an earlier research project proposed by researchers and data scientists at the University of California Davis and Catalyst Cooperative that won the U.S. EPA’s EmPOWER Air Data Challenge. As an open-source research initiative, it will always be free and open, and we invite contributions from others to continue improving it.
Our Vision for the Open Grid Emissions Initiative
Our vision for the Open Grid Emissions Initiative encompasses four primary goals, each aimed at accelerating grid decarbonization.
Make high-quality, hourly, consumed grid emissions data publicly available: The EPA’s eGRID database has served as the primary source of trusted grid emissions data for the U.S. for the past quarter century. However, the eGRID dataset only includes annual-average, generated emission factors. For the purposes of GHG accounting and footprinting, however, recent research has shown that in many cases, annual averages no longer accurately describe the actual hourly emissions intensity of the grid, and that the emissions intensity of the electricity consumed in a region can significantly differ from the emissions intensity of electricity generated in a region due electricity imported from other regions.
The Open Grid Emissions Initiative addresses these needs by providing both generated and consumed emissions data at the plant and regional levels for different temporal resolutions, all built from the same high-quality hourly data.
Validate and improve real-time estimates of grid emissions: There is an increasing demand for near real-time emissions data to enable more timely climate action and next-generation energy attribute markets. There are an increasing number of sources for near real-time grid emissions data (including through Singularity’s platform), but these are all estimated based on real-time generation data. Until now, there has not been a benchmark dataset of primarily measured hourly emissions data against which the real-time estimates could be validated. The Open Grid Emissions initiative provides an exciting opportunity to evaluate the accuracy of real-time estimates and develop new approaches for continuing to improve them.
Build trust in grid emissions data through transparent, documented, and open-source methods: One of the biggest successes of eGRID is the trust that it has earned from its users, partially due to its transparent and well-documented methodology. Our vision is to continue this tradition of trust not only by carefully documenting and explaining the methodologies and assumptions used, but also by integrating community-contributed, peer-reviewed improvements. We released Open Grid Emissions as an open-source initiative because we believe that this fosters greater trust in the end data, and because we believe that this will allow us to more rapidly identify user needs and integrate cutting-edge methodologies to keep the data relevant over time.
Accelerate grid decarbonization research through standardization: The power sector data used to inform emissions research are generally scattered across multiple sources, not standardized, and sometimes messy, which means that researchers and analysts often spend more time cleaning and wrangling the data than analyzing it and generating insights. Inspired by and building on Catalyst Cooperative’s work on the open-source Public Utility Data Liberation (PUDL) Project, we envision the Open Grid Emissions Initiative as the go-to source for ready-to-use grid emissions data. We also envision the Initiative as the main repository of open research questions in this field and of the best available methods for cleaning, imputing, and calculating emissions from the power sector, so that researchers are always building upon the best available science to date.
Our data and methods
The Open Grid Emissions Initiative uses the U.S. EPA’s eGRID annual emissions methodology as its foundation, but integrates key innovations from existing peer-reviewed research (such as these open-source tools from Stanford researchers) as well as certain novel methods to improve the data resolution and refine the emission calculations.
Over 90% of the generation-related carbon emissions in the dataset come from measured, hourly continuous emissions monitoring (CEMS) data reported to EPA’s Clean Air Markets Division. However, most small fossil fuel generators and all non-fossil generators only report monthly data to the U.S. Energy Information Administration (EIA). One important innovation of the Open Grid Emissions Initiative is a method for assigning hourly profiles to month-level data reported to EIA, so that it can be combined with the CEMS data to provide a complete picture of hourly power sector emissions.
Another important innovation of Open Grid Emissions is how it adapts the EPA’s methodologies for cleaning and processing the annually-aggregated CEMS data used for eGRID. Applying these methods to hourly CEMS data allows us to more precisely identify and impute missing or incomplete CEMS data, resulting in a more complete inventory of power sector emissions.
Getting Involved
We are excited to continue improving the quality and relevance of the data produced by the Open Grid Emissions Initiative, but we can’t do it alone! If you use the data, we want to hear how you’re using it and how you think it could be improved. We have already identified many ideas for the next release of the dataset (including additional air pollutants, calculation of grid losses, and consumed fuel mix), but hearing from users will help us prioritize these efforts. If you’re a researcher or subject matter expert, we want to hear how our methodologies can be improved, where our assumptions may be incorrect, and how we can solve any of our open methodological questions. If you’re a professor in search of real-world class projects or a grad student in search of a thesis topic, check out our issue tracker for ideas. If you develop improvements or new features, you can submit a request to have your work added into the project. We hope over time to build an active community of users and contributors working towards a shared mission of data-driven decarbonization.
You can download the data here.
