Abstract

Shipping plays a vital role in global trade, facilitating over 90% of international trade by volume. However, it also contributes to significant environmental pollution, accounting for around 3% of global greenhouse gas (GHG) emissions and substantial amounts of nitrogen oxides (NOx), sulfur oxides (SOx), particulate matter (PM), black carbon (BC), and methane (CH4). These pollutants negatively impact both the climate and public health, particularly in coastal areas.

In response, the International Maritime Organization (IMO) has set ambitious targets for reducing shipping emissions, with the goal of achieving near-zero GHG emissions by 2050. Interim goals include a 20% reduction by 2030 and 70-80% by 2040. As part of these efforts, the IMO introduced regulations such as the Energy Efficiency Existing Ship Index (EEXI) and the Carbon Intensity Indicator (CII) to monitor energy efficiency and emissions.

To support these regulatory efforts, we propose a machine learning framework to predict GHG emissions from vessels navigating the Saint Lawrence River. By utilizing AIS data, vessel-specific details, and meteorological data, our approach will create a detailed emissions inventory. Using deep learning models, the system will predict emissions based on individual vessel activities. This scalable approach will contribute to more accurate environmental monitoring and support Canada’s broader efforts to reduce emissions from maritime transportation.