414 both CO2 and SO2. We calculated the concentration factor (CF) as a function of time for the five 
415 vessels considered in this study which are shown in Figure 7 for the VMS at 9 m above sea level. 
416 The multiplication of CF values with the input concentration values of Table 2 per pollutant leads 
417 to the generation of the modelled CO2 and SO: concentration time-series. The production of MOP 
418 values for each ship case is calculated by summing up the concentration values of the timeline, as 
419 explained in Figure 2. In reality, dispersion of CO2 and SO: will be marginally different due to 
420 differences in their molecular diffusivity (0.14 cm”/s and 0.11 cm*/s, respectively). In addition, 
421 eddy diffusivity is five orders of magnitude larger than molecular diffusivity in the vicinity of each 
422 case study ship, where modelling results are exported. Practically, eddy diffusivity constitutes the 
423 only term that affects the results. Bonifacio et al. (2014) observed the same phenomenon in the 
424 vicinity of buildings in an urban area, concluding that only eddy diffusivity affects dispersion 
425 results in the surrounding area. The CF time-series that are produced by the CFD modelling at the 
426 virtual path of the VMS result to be bell or log-normal shaped, as shown in Figure 7. The horizontal 
427 axis corresponds to the time in seconds, while the vertical axis shows the CF. The time resolution 
428 has been selected to be 5 s, the same as for the measured CO»: concentration. In this example, the 
429 total duration of the plume being detectable over any background concentration appears to be ın 
430 the order of 90-95 s. The period from 0 to 45 s demonstrates the vessel’s approach to the VMS, 
431 while the vessel is moving away from 45 to 110 s. CF time-series for Ship 1 and 5 present smooth 
432 bell-shaped curves differing from each other in duration and values which in the case of Ship 5 
433 both are smaller. Time-series of Ship 3, also, presents a bell-shaped curve, which is not as smooth 
434 as Ship 1 and 5, due to the higher actual speed (Table 5), which results in a smaller signal duration 
435 and consequently fewer data points. The CF time-series for Ship 2 and 4 present log-normal curves 
436 that are not such smooth due to their greater actual speed The effect of downwash phenomena may 
437 lead to the accumulation of gaseous pollutants on one side of the propagating plume at sea level, 
438 as described in section 3.1, producing log-normal time-series. 
173 
«10-4 
CF Time-series - CFD 
5 
+ 
4 - 
DO 
3 
-----Ship 1 
»=-e--- Ship 2 
——e---Ship 3 
:--e--- Ship 4 
-—-e--- Ship 5 
5 
0 
20 40 60 80 100 120 
Time (s} 
AA; 
A 
%C 
Figure 7: Concentration factor time-series produced by CFD modelling for the five ship case studies. 
19