TL;DR
This paper introduces a comprehensive global dataset of CO2 emissions for 343 cities, integrating self-reported emissions data with a wide range of socio-economic, traffic, and climate-related ancillary data. This dataset aims to facilitate comparative analyses of urban emission drivers and mitigation strategies, addressing a critical gap in urban climate research.
What Problem Does This Paper Address?
Understanding and mitigating urban greenhouse gas emissions is crucial, as cities are major contributors to the global carbon cycle. While many cities report their emissions, a standardized, globally comparable dataset that also includes key influencing factors has been lacking. Previous efforts have often been regional or lacked comprehensive ancillary data. This paper addresses the need for a unified dataset that allows for cross-city comparisons of CO2 emissions and their drivers, enabling a deeper understanding of urban climate action and policy effectiveness.
Methods
The authors compiled their final dataset, Dfinal, by integrating data from three primary sources for CO2 emissions: CDP (Carbon Disclosure Project), carbonn Climate Registry, and Peking University. They implemented rigorous quality control procedures, including documenting reporting methods and correcting for non-CO2 greenhouse gases. A significant portion of the methodology involved collecting and integrating a diverse array of ancillary data. This included socio-economic indicators (GDP, population, etc.), traffic data (congestion indices), and climate data (heating and cooling degree days), as well as calculated urban expansion metrics. Data handling and processing were primarily conducted using Jupyter notebooks and pandas DataFrames. Validation against independent datasets was also a key part of their quality assurance process.
Datasets
The core datasets used to construct the final Dfinal dataset are:
- CDP (Carbon Disclosure Project): Provided emissions data for 187 cities.
- carbonn Climate Registry: Contributed emissions data for 73 cities, particularly from developing regions.
- Peking University: Supplied emissions data for 83 cities in China.
In addition to these primary emission sources, a multitude of ancillary datasets were integrated, including:
- GEA, UITP, WB: Socio-economic and climate-related attributes.
- INRIX, TomTom: Traffic congestion data.
- IESE: Socio-economic indicators from the Cities in Motion dataset.
- DUEX: Calculated urban area expansion data.
- Dclim: Calculated climate indices (HDD and CDD).
- DOTHERS: Externally sourced ancillary data to fill gaps.
The final dataset, Dfinal, comprises 343 cities with 179 columns of data.
Key Insights and Ideas
- The Power of Ancillary Data: The paper strongly emphasizes that understanding CO2 emissions in isolation is insufficient. Integrating socio-economic, traffic, and climate data is crucial for explaining variations in emissions and identifying effective mitigation strategies.
- Data Harmonization is Key: A significant portion of the work involved meticulous data cleaning, quality control, and harmonization across diverse sources, highlighting the challenges and importance of creating a consistent global dataset.
- Self-Reported Data Challenges: The authors acknowledge the inherent limitations of self-reported city emissions data, necessitating extensive quality control and validation against independent sources.
- Bridging the Data Gap: By incorporating data from developing countries and China, the dataset significantly improves global coverage compared to previous regional datasets.
Key Results
The primary result is the creation of the Dfinal dataset, a global compilation of Scope-1 CO2 emissions for 343 cities, enriched with 179 attributes. This dataset includes:
- Emissions Data: Scope-1 CO2 emissions, with quality flags assigned based on consistency checks.
- Ancillary Data: A broad range of socio-economic (GDP, population, urbanization), traffic (congestion indices), and climate (HDD, CDD) variables.
- Geographic Coverage: Representation from various global regions, with a notable inclusion of cities from developing countries and China.
- Temporal Coverage: The majority of cities reported emissions between 2010-2015.
The paper also presents validation results comparing their dataset with independent sources like VULCAN and Markolf et al., demonstrating reasonable correlations and highlighting areas for further investigation.
Limitations and Future Work
- Self-Reporting Bias: The dataset relies heavily on self-reported emissions data, which can introduce inconsistencies and biases. The authors acknowledge that the choice of cities is based on self-reporting, not a designed sampling procedure, meaning it may not be a representative sample of all cities globally.
- Data Gaps and Inconsistencies: Despite extensive efforts, some data gaps and inconsistencies remain, particularly in ancillary data for certain cities.
- Scope-1 Focus: The primary focus is on Scope-1 emissions, with Scope-2 data provided as additional information. A more comprehensive integration of Scope-2 emissions could enhance future analyses.
- Future Work: The authors suggest that the dataset can be used for various analyses, including understanding emission drivers, evaluating mitigation potentials, and developing urban climate policies. The interactive web-based data visualization tool mentioned also points towards future accessibility and use of the data.
Summary updated