Fuel Consumption and CO2 Emission Forecasting for Sustainable Maritime Transportation: IMO and Panama Canal Data Analysis
Year: 2026
Type: article
Abstract: Accurate prediction of vessel fuel consumption and carbon dioxide (CO2) emissions is essential for improving energy efficiency, supporting compliance with international regulations, and accelerating the decarbonization of maritime transport. Existing studies frequently rely on isolated data sources, address fuel consumption and CO2 emissions as independent prediction tasks and evaluate models through random data partitioning. This study proposes an integrated machine learning (ML) framework that harmonizes heterogeneous global and regional maritime data, linking information from the International Maritime Organization Data Collection System (IMO DCS), accessed through the IMO Global Integrated Shipping Information System (GISIS), with updated operational records comprising more than two million Panama Canal observations. Fuel-consumption data were converted into corresponding CO2 emissions using appropriate fuel-specific emission factors. The proposed methodology standardizes vessel classifications, measurement units, and temporal structures while deriving physically meaningful and interpretable indicators of operational activity, energy efficiency, and carbon intensity. The ML methods applied include linear regression, quadratic polynomial regression, naïve persistence, and Holt’s linear trend models, which were evaluated across two complementary tasks: vessel fuel-consumption prediction and CO2-emissions estimation. Model performance was assessed using temporally consistent validation procedures to reduce information leakage and provide more credible estimates of generalization to previously unseen periods. The comparative analysis examines predictive accuracy, temporal trends, and relative emission patterns within each dataset, identifying similarities and differences between global maritime activity and the distinctive operational context of the Panama Canal while recognizing the limitations imposed by their different scales, and emission metrics. The resulting framework is expected to support evidence-based operational planning, emissions management, and maritime decarbonization strategies at both regional and international levels.
Source: Sustainability
DOI: 10.3390/su182211456
Language: English
Subfield: Environmental Engineering
Field: Environmental Science
Domain: Physical Sciences
Open Access status: gold
Estimated APC paid: $2,896
OpenAlex ID: W7225733551