129 inlet of the trace gas monitors is located at a height of 7 to 11 meters above sea level, depending 
130 on the current tidal level. 
131 Seagoing vessels arriving to and departing from Hamburg pass the MS at a distance of 
132 about 150 to 800 m. The MS was set up in 2014 for research purposes and has been in routine 
133 operation for exhaust gas monitoring since 2017. Since then, the trace gases SO2 (HORIBA APSA- 
134 2370), CO (Licor LI-840A), NO, NO, NOx (HORIBA APNA-370), O3 (APNA-370), as well as 
135 ambient meteorological conditions such as air temperature, humidity, pressure, wind speed and 
136 direction (LUFFT WS700-UMB) are measured in 24/7 operation with a temporal resolution of 5 
137 ss. In addition, the AIS (Automated Identification System) signals, sent by passing vessels every 
138 10 s approximately, are recorded. From the AIS, among others, the ship’s identity, the current 
139 ship’s position and some sailing and ship details such as ship size, speed and destination can be 
140 derived. 
towards North S- 
GER 
towards Hamburg harbor 
141 
142 
143 
144 
145 
146 
147 
148 
149 
150 
151 
152 
153 
154 
155 
156 
157 
158 
a 
Map data: OOpenStreetMap contributors, 
A 
Figure 1: The monitoring station location on maps. Map source: OpenStreetMap. 
The FSC of any pass-by vessel can be derived from the measured ratio of SO2 to CO2 in 
the exhaust plume, according to the method described in detail by Kattner et al. (2015). For each 
pollutant, the average background concentration is detected immediately before and after the 
vessel’s signal and is subtracted from the concentration measured in the plume. In case more 
plumes are detected, e.g. due to multiple ship passing, only readings outside the multiple plumes 
are considered for the calculation of the background. To remove the potential impact of different 
response times of the various analysers — in our case the t90 response time of CO2 and SO2 
analysers are <5 s and 30s, respectively — the integral of the signal is used rather than just the peak 
value. The uncertainty of this method is calculated individually for every single measurement 
according to Gaussian error propagation and is strongly dependent on the height of the detected 
plume versus the background variability (signal-to-noise ratio should be greater than 5 as a 
minimum standard), but also on other factors like measurement conditions, instrument calibration 
and correctness of the assumptions made to convert the measured SO2/CO- ratio into an FSC value. 
A detailed description of the error calculation is included in Mellqvist et al. (2022). However, 
when evaluating the uncertainty of the so far taken approximately 60000 plumes, the uncertainty 
of a single plume measurement is on average about 30% of the determined FSC.