AIS-driven vessel activity and emissions modelling for offshore decommissioning activities in the North Sea

Introduction Maritime transport is a major source of air pollutants yet remains one of the least regulated sectors globally. In the North Sea, vessel activity associated with offshore energy infrastructure decommissioning is expected to increase over the next decade. However, vessel-specific emissions during these activities remain poorly quantified, with existing approaches relying largely on coarse-scale estimates. This study aimed to develop a data-driven, bottom-up methodology for estimating vessel emissions using Automatic Identification System (AIS) data to enable high-resolution spatial and temporal analysis of offshore energy-related shipping emissions. Methods A custom vessel-tracking tool was developed to process large AIS datasets collected between 2015 and 2021 for the northern North Sea. Vessel movements and operational patterns were reconstructed from AIS records, and emissions of carbon dioxide (CO 2 ), nitrogen oxides (NO x ), and sulphur dioxide (SO 2 ) were estimated based on vessel speed, operational activity, and engine characteristics. Emissions were spatially allocated to a 1 km 2 grid. The methodology was applied to three offshore decommissioning case study areas: Scott, Gannet, and Subsea, representing different infrastructure densities and operational stages within the marine energy sector. Sensitivity analyses were conducted to evaluate uncertainties associated with fuel consumption and auxiliary engine assumptions. Results Emissions were found to be highly localised around offshore installations and displayed strong temporal variability associated with episodic operational activity. Among the case study areas, Gannet exhibited the highest emissions intensity, contributing up to 5% of regional CO 2 emissions during peak operational periods. Correlations between vessel transits and emissions varied between sites, indicating that vessel counts alone do not adequately capture emissions behaviour. Operational modes such as idling and dynamic positioning were identified as important contributors to overall emissions. Sensitivity analysis demonstrated that assumptions relating to fuel consumption rates and auxiliary engine usage were key sources of uncertainty in the emissions estimates. Discussion The findings demonstrate that AIS-based bottom-up approaches can provide more accurate, transparent, and spatially resolved emissions estimates for offshore energy operations compared with conventional regional-scale methods. Incorporating operational vessel behaviour substantially improves emissions characterisation and highlights the importance of accounting for non-transit activities in offshore environments. This methodology has potential applications in emissions reporting, regulatory compliance, and decarbonisation planning for offshore energy systems undergoing transition and decommissioning.