The significance of incorporating unidentified vessels into AIS-based ship emission inventory

Abstract

Automatic Identification System (AIS) data have been widely used to estimate ship emission inventories. However, although the AIS dataset can provide high spatial and temporal ship activities, the estimation of ship emission inventory requires vessel information that may not always be available in the vessel parameter database (e.g. engine power). Hence, earlier AIS-based emission inventory studies mainly included vessels with engine specifications (referred to as identified vessels); the significance of emissions from vessels without engine or other pertinent specifications (referred to as unidentified vessels) has not been studied in detail.In this study, we develop an emission inventory of all vessels with installed AIS instruments, including both identified and unidentified vessels. In particular, missing ship parameters for unidentified vessels are estimated through categorical regression with vessels with similar ship types and sizes in the AIS database. Using this approach, we estimate the ship emission inventory for the Pearl River Delta (PRD) in 2015. Our results show that unidentified vessels contributed significantly, accounting for almost 49% of the overall ship emissions of CO2, NOx, CO, and hydrocarbon in the PRD. Overall, we find that identified vessels contributed more over the open ocean and near port areas, whereas unidentified vessels contributed significantly along inland waterways with clear emission tracks and hot spots.