Overview
Emissions describe the release of greenhouse gases (GHGs) and air pollutants into the atmosphere from human activities such as energy production, industry, transport, and agriculture. Reliable emission inventories are a cornerstone of climate change assessment, air quality modeling, and integrated policy analysis, providing consistent information on the magnitude, sources, and spatial–temporal distribution of emissions.
Within PANTHEON, emissions data are provided through the Global Emission Modeling System (GEMS), which represents a key project outcome and a central data resource for downstream modeling and analysis.
What Is the Global Emission Modeling System (GEMS)?
The Global Emission Modeling System (GEMS) is a comprehensive, high-resolution global emission inventory that provides data on greenhouse gases and key air pollutants from multiple sectors. It was developed to improve the accuracy and accessibility of emissions data for research and policy applications.
GEMS compiles emissions across major pollutant categories such as:
· Greenhouse gases (e.g., CO₂, CH₄, N₂O)
· Air pollutants (e.g., NOₓ, SO₂, CO, PM₂.₅, PM₁₀, black carbon, organic carbon)
These emissions are estimated using bottom-up methodologies that combine activity data (like fuel consumption) with emission factors, and resolved spatially at 0.1°×0.1° to support detailed analysis.
Why This Matters
Understanding emissions is critical for:
· Climate change mitigation – tracking GHG sources and trends
· Air quality assessment – quantifying pollutants that affect health
· Policy support – informing mitigation strategies and compliance evaluation
· Environmental research – providing essential inputs for climate and air quality models
Emission inventories such as GEMS support evidence-based decision-making by revealing who emits what, where, and when, and by enabling comparisons across regions and sectors.
Data Access
The GEMS emission inventory can be accessed through the GEMS data portal, which provides detailed documentation and dataset descriptions:
To facilitate broader dissemination and reuse within the atmospheric science community, GEMS is also available via the GEIA–ECCAD platform, which hosts emission inventories for research and modeling purposes:
Both access points provide consistent GEMS data and documentation.
Data structure description
The inventory provides global emissions of CO₂, CO, SO₂, NOx, PM₂.₅, PM₁₀, TSP, BC, OC, BrC, and PAH (16 U.S. EPA priority parent pollutants and 24 other pollutants) for the period 1700-2021 on a 0.1° spatial resolution (see data description file in GEMS website for more details).
Inventory compilation
Energy data: Data for anthropogenic sources from power plants, industry, residential, commercial, transportation, and agriculture sectors mainly come from the International Energy Agency (https://www.iea.org). Data for natural fires after 1997 are obtained from the Global Fire Emissions Database (http://globalfiredata.org/pages/data/), while data prior to 1997 are obtained from the BB4CMIP database (https://esgf-node.llnl.gov/search/input4mips/). Industrial process data mainly come from the International Energy Agency (https://www.iea.org), World Steel Association (https://worldsteel.org), U.S. Geological Survey (https://www.usgs.gov/centers/national-minerals-information-center/commodity-statistics-and-information), Nation Master (https://www.nationmaster.com), and UNSD (https://unstats.un.org/UNSDWebsite/). For specific sources and the corresponding data collection and compilation methods, please refer to the “Activity Level Methodology” table in the GEMS website.
Emission factor: We collected crude emission factors of difference pollutants for different sectors, fuels, and technologies from existing literature and databases. We also measured emission factors by ourselves for sources lacking literature reports.
Technology division
(1) For stove type
In this inventory, stoves are mainly categorized as open stoves, traditional stoves, improved stoves, and high-efficiency stoves. When compiling the emission factors of various stoves, the following sequence is used to determine the stove type: (1) For Chinese literature, if the stove type is
explicitly stated, it is used directly. (2) For literature without stove type information, if a picture of the stove is provided, it can be categorized based on the comparison with the picture. (3) For literature without pictures, it can be categorized based on the thermal efficiency or the modified combustion efficiency (MCE). (4) If emission factors of other incomplete combustion products (such as PM, CO, OC, BC) are determined, they can be used for categorization. (5) If there is no other way to determine the stove type, it can be categorized based on the magnitude of the emission factor.
(2) For transportation
GAINS (IIASA) provides fuel consumption data for up to six years for five types of vehicles (gasoline light-duty, gasoline heavy-duty, diesel light-duty, diesel heavy-duty, and gasoline motorcycles) for over 100 countries and regions. The data are available for the years 1990, 1995, 2000, 2005, 2010, and 2015. The data were sourced by Dr. Shaohui Zhang of IIASA from global data accessed through the GAINS platform. The platform can be accessed at https://gains.iiasa.ac.at.
(3) For power and industrial sources
The classification of emission reduction technologies for power and industrial sources is divided into two parts: power plants and industry, and four types based on control processes: boiler type, PM, SO2, and NOx. In addition, since the steel and cement industries are heavily polluted industries, their technology classification should not be directly included in the industry category, but instead should be separately classified. The final classification should be based on all dust removal/desulfurization/denitrification technologies used for a certain type of boiler in a certain country in a certain year, and should add up to 100%. For parts with continuous statistical data, actual statistical data should be used without fitting. The beginning and end should be completed with an “S”-shaped curve. For parts with discrete data, an “S”-shaped regression curve should be used for fitting. For parts without data, similar methods should be used for supplementation.
Spatial interpolation: By reviewing a large number of literature and databases, emission data for various sources are obtained. For data that is not included, interpolation based on population spatial distribution is used. For detailed interpolation processes for various emission sources, please contact shenhz@sustech.edu.cn.
Relevance for PANTHEON
The GEMS emission inventory serves as a core input and output of PANTHEON, enabling:
· Integrated analysis of energy, emissions, air quality, and climate interactions
· Evaluation of mitigation pathways and policy scenarios
· Harmonized use of emissions data across modeling frameworks and work packages