90 the Yantian Port Area, Shenzhen (China) and compare them with real measurements. The results 
91 from the CALPUFF model in that this study underestimated the measured concentrations Toscano 
92 et al. (2021) compared CALPUFF and micro-scale CFD modelling with data from wind tunnel 
93 experiments considering the dispersion of pollutants at the Naples port model where a couple of 
94 SO» detection sensors were installed at port locations and around buildings. The CFD model 
95 showed satisfactory results in terms of predicted concentrations, while the CALPUFF model 
96 systematically underestimated concentrations at ground level, where monitoring stations are 
37 typically located. Click or tap here to enter text.Additionally, Badeke et al. (2021) employed the 
98 micro-scale MITRAS model, which is based on Navier-Stokes equations, to evaluate the 
99 downward dispersion of two scenarios. The first considered the entire vessel's structure, including 
100 the stack, while the second only took into account the stack, representing a point source in a 
101 lagrangian or gaussian model. The researchers observed that accounting for the entire ship’s 
102 geometry with the stack led to a 25% greater downward dispersion compared to considering only 
103 the stack. Micro-scale modelling prevails over lagrangian and gaussian models since it is capable 
104 of considering the vessel’s geometry which is responsible for the plume rise and the downwash 
105 phenomena that occurred due to the turbulence effects on the surfaces of the ship (Badeke et al.. 
L06 2021; Dobrucali & Ergin, 2019; Li et al., 2022; Syms, 2004). The few available studies, therefore, 
107 indicate that micro-scale modelling can provide a reliable pollutant dispersion tool in the 
108 environment, in the vicinity of vessels. However, a comprehensive evaluation of such micro-scale 
109 models is currently lacking in the literature. 
110 The current work presents the development of a method to identify the FSC violations of 
111 vessels by producing CO: and SO: concentration data employing CFD modelling. The efficiency 
112 of CFD modelling was assessed using measured concentration data originating from individual 
113 ships collected via a unique measurement station. The CFD approach constitutes the basic asset of 
114 this work since the discretization resolution is finer in contrast to larger scale models, thus 
115 potentially resulting in more accurate results. The current work suggests that plume dispersion 
116 simulation using CFD offers accuracy that can be useful in a variety of applications, apart from 
117 the FSC violation identification technique, including estimating the impact of vessels to air quality 
118 in coastal areas, identification of the optimum location of monitoring stations, air-water 
119 interaction. etc. 
L20 
121 
122 
123 
124 
125 
126 
127 
128 
2 Methods 
2.1 Monitoring station and experimental setup 
The data used for the validation of the current study were measured at a ship emission 
monitoring station (MS), which is operated by the German Federal Maritime and Hydrographic 
Agency (BSH) for compliance monitoring, according to MARPOL Annex VI (European Union. 
2016). The MS was set up to be used in routine operations for monitoring ship exhaust plumes. 
The MS is located in Wedel near Hamburg on the north-eastern bank of the river Elbe, about 10 
km before the entrance to the port of Hamburg, the third largest port in Europe (Figure 1). The gas