Explaining CO2 emission pathways in G20 nations through SHAP analysis decoupling assessment and deep learning forecasting

This study examines the drivers, decoupling progress, and future trajectories of CO 2 emissions across G20 economies, which collectively account for roughly 80% of global emissions. Using data from 19 countries spanning 1991 to 2022, the research integrates three analytical lenses—drivers, decoupling, and forecasting—that existing literature typically treats separately. Seven indicators are analysed as predictors of emissions: GDP, renewable energy use, total energy consumption, population, trade per capita, and surface air temperature. An XGBoost model with SHAP analysis reveals that energy consumption is the dominant emissions driver in 17 of 19 economies, with the strongest positive effects in China, the United States, and India. GDP and trade per capita are also positively associated with emissions, while renewable energy use correlates with lower emissions. Applying Tapio’s decoupling index across six five-year windows from 1991 to 2022, the study finds substantial heterogeneity: seven countries demonstrate strong decoupling of economic growth from emissions, six show weak or unstable decoupling, and six remain closely coupled. For forecasting 2023–2035, four machine-learning models are compared. The Feed-Forward Neural Network achieves the highest short-run accuracy (R 2 = 0.996; MAPE = 5.44%), while the Long Short-Term Memory network produces the most stable long-run projections. Results indicate that advanced economies will likely stabilise or reduce emissions, whereas emerging economies—particularly China, India, Indonesia, and Türkiye—face continued increases driven by economic growth and rising energy demand. The findings suggest that emission-reduction policies must be differentiated by structural driver and decoupling stage, prioritising renewable-energy expansion and energy-efficiency investment in high-growth, energy-intensive economies while sustaining decarbonisation support in already-decoupled nations.

Authors

Institutions

Publication Details

Journal
Discover Sustainability
Published
2026-09-19
DOI
https://doi.org/10.1007/s43621-026-04471-4
Primary Topic
Environmental Impact and Sustainability
Type
article
Field-Weighted Citation Impact
0.00
Controls
|||
ALL TIME
JAN
FEB
MAR
APR
MAY
JUN
JUL
AUG
SEP
article

Explaining CO2 emission pathways in G20 nations through SHAP analysis decoupling assessment and deep learning forecasting

Sazzad Parwez, Siva Aditya Vaddiraju
Discover Sustainability
Environmental Impact and Sustainability
article

Explaining CO2 emission pathways in G20 nations through SHAP analysis decoupling assessment and deep learning forecasting

Sazzad Parwez, Siva Aditya Vaddiraju
article en

Abstract

This study examines the drivers, decoupling progress, and future trajectories of CO 2 emissions across G20 economies, which collectively account for roughly 80% of global emissions. Using data from 19 countries spanning 1991 to 2022, the research integrates three analytical lenses—drivers, decoupling, and forecasting—that existing literature typically treats separately. Seven indicators are analysed as predictors of emissions: GDP, renewable energy use, total energy consumption, population, trade per capita, and surface air temperature. An XGBoost model with SHAP analysis reveals that energy consumption is the dominant emissions driver in 17 of 19 economies, with the strongest positive effects in China, the United States, and India. GDP and trade per capita are also positively associated with emissions, while renewable energy use correlates with lower emissions. Applying Tapio’s decoupling index across six five-year windows from 1991 to 2022, the study finds substantial heterogeneity: seven countries demonstrate strong decoupling of economic growth from emissions, six show weak or unstable decoupling, and six remain closely coupled. For forecasting 2023–2035, four machine-learning models are compared. The Feed-Forward Neural Network achieves the highest short-run accuracy (R 2 = 0.996; MAPE = 5.44%), while the Long Short-Term Memory network produces the most stable long-run projections. Results indicate that advanced economies will likely stabilise or reduce emissions, whereas emerging economies—particularly China, India, Indonesia, and Türkiye—face continued increases driven by economic growth and rising energy demand. The findings suggest that emission-reduction policies must be differentiated by structural driver and decoupling stage, prioritising renewable-energy expansion and energy-efficiency investment in high-growth, energy-intensive economies while sustaining decarbonisation support in already-decoupled nations.

Discover Sustainability
Woxsen School of Business (IN)
Affordable and clean energy
Openalex Percentile: Top 18%
Environmental Impact and Sustainability
AI Navigator

Ask Laika to Summarize, Analyze, and Connect papers live on the map.

Summarize Papers & Methodologies

Extract key findings, datasets, and comparative methods across publications.

Benchmark Rankings & Visual Analytics

Rank top research institutions, authors, funders, topics, and journals by Field-Weighted Citation Impact (FWCI) and paper volume with instant charts.

Connect Distant Disciplines

Bridge topological clusters on the map to find hidden collaborative intersections.