Publisher’s note: this comment was edited on 22 May 2023. The following text is not identical to the original comment, but the adjustments were minor without effect on the scientific meaning.

This manuscript describes the results of real-world emissions measurements made for ~4700 inland ship passages in the Upper Rhine valley for over a year. This work presents NOx and PM fuel-based emissions factors measured above the ships and on the riverbank and relates them to AIS-obtained ship data. The authors found that the shipping emissions increased NOx and PM for those localized areas from 3-50%. Additionally, their calculated emission factors showed correlation to engine characteristic such as RPM and engine age/category. This work showed the great importance of real-world measurements to validate existing emissions standards and inventory and demonstrate correctly functioning or malfunctioning emission control technologies. This reviewer agrees with the authors that this long-term measurement work is critical to creating reliable estimates of shipping emissions contributions and could be applied to multiple emission source sectors. However, the authors do not consider other related work done outside of Europe on understanding the emissions and engine load relationship (see recommendations below). Additionally, there could be additional forecasting or recommendations provided by the author on how to transition the older shipping fleet to newer engines and what the emission and exposure benefits may be from this transition. Overall, this work provides well-described and illustrated novel scientific contribution and is recommended for publication with revision. Please review the specific line comments and recommendations below.

P10 L19-21 In the latter case, which was true for ~ 5 % of all ship plumes, the BC peak height was manually set to zero, marginally reducing the calculated mean emission factor.

• Integrating near-zero values is still recommended rather than manually setting to 0 in order to detect potential failing in the filter technologies on-board (especially if tracked over time).

P12 L9-12 This value is also typical for diesel engines without exhaust gas aftertreatment systems, assumed to constitute the largest portion of the inland fleet (see Sect. 2.2.3), whereas the use of catalysts and diesel particulate filters (DPF) usually results in a somewhat higher ratio of 0.25– 0.30 (Kurtenbach et al., 2016).

• Additional sources exist detailing the use of onboard emission control technologies that may be cited here (see suggestions below)

P13 L12-14 Monthly average concentrations of NOx were increased by 6.9–14.3 μg m-3 at BRI (above the shipping lane) and by 1.0–2.2 μg m-3 at RIV (river bank). This corresponds to an additional burden of 24–68 % respectively 6–11 %, relative to extrapolated background levels without the shipping contribution.

• Though noted later in the manuscript, it is important to clarify that these increases are extremely localized and dependent on meteorology. Additionally, comments regarding the number of people living along the river area could be noted to increase the significance of this finding.

P15 L7-8 This again indicates potential differences in fleet composition, load status and operating conditions between Lower and Upper Rhine.

• This point needed further clarification as the distinction between the lower and upper Rhine is not clear to the reviewer.

P15 L21-23 The BC emission factor derived in this study (EBC = 0.5±0.3 g kg-1) is in the upper range of values reported for sea-going ships (Petzold et al., 2008; Diesch et al., 2013; Celik et al., 2020).

• Additional literature specifically on BC (and NOx) emissions would help to frame this metric in comparison to other ship types or locations (see suggestions below).

P17 L5-6 This could indicate that real-world particle emissions might not as efficiently be reduced as under test bench conditions.

• This is a very important point and could be emphasized more so in the conclusion of this manuscript. Additional literature has commented / proved on this idea (see suggestions below).

P17 L18-20 For inland ships there is little information available in the literature whereas for sea-going ships detailed studies or reviews are e.g. provided by Celik et al. (2020) and Grigoriadis et al. (2021).

• Additional literature is available on this topic (see suggestions below).

P21 L10-12 With a BC fraction of 38 % for upstream ships and 16 % for downstream ships our results indicate that a higher engine load leads to an increase in BC emission. This is in contrast to sea ships where measurements show a decrease in BC with an increase in combustion efficiency (Celik et al., 2020).

• This finding contradicts other literature that has seen a decrease in BC emissions as a function of engine load (as stated in the manuscript) but does not include suggested literature below.

P24 L7-8 For the small number of ships with modern Euro V diesel engines and aftertreatment system passing the station, we derived very low NOx and PM emission factors underlining their emission reduction potential.

• Expanding on this idea and modeling out how the next generation fleet will improve emission and exposure levels would greatly contribute to this manuscript

Figure S1: Correlation plot of CO2 peak areas derived from Licor and ICAD measurements at BRI.

• Units should be ppm*s

Additional literature to be considered.

Buffaloe, G. M.; Lack, D. A.; Williams, E. J.; Coffman, D.; Hayden, K. L.; Lerner, B. M.; Li, S.-M.; Nuaaman, I.; Massoli, P.; Onasch, T. B.; Quinn, P. K.; Cappa, C. D. Black Carbon Emissions from In-Use Ships: A California Regional Assessment. Atmos. Chem. Phys. 2014, 14, 1881−1896.

Gysel, N. R.; Russell, R. L.; Welch, W. A.; Cocker, D. R. Impact of Aftertreatment Technologies on the In-Use Gaseous and Particulate Matter Emissions from a Tugboat. Energy Fuels 2016, 30, 684−689.

Jiang, Y.; Yang, J.; Gagné, S.; Chan, T. W.; Thomson, K.; Fofie, E.; Cary, R. A.; Rutherford, D.; Comer, B.; Swanson, J.; Lin, Y.; Van Rooy, P.; Asa-Awuku, A.; Jung, H.; Barsanti, K.; Karavalakis, G.; Cocker, D.; Durbin, T. D.; Miller, J. W.; Johnson, K. C. Sources of Variance in BC Mass Measurements from a Small Marine Engine: Influence of the Instruments, Fuels and Loads. Atmos. Environ. 2018, 182, 128−137.

Nuszkowski, J.; Clark, N. N.; Spencer, T. K.; Carder, D. K.; Gautam, M.; Balon, T. H.; Moynihan, P. J. Atmospheric Emissions from a Passenger Ferry with Selective Catalytic Reduction. J. Air Waste Manage. Assoc. 2009, 59, 18−30.

Schlaerth, H.; Ko, J.; Sugrue, R.; Preble, C.; Ban-Weiss, G. Determining Black Carbon Emissions and Activity from In-Use Harbor Craft in Southern California. Atmos. Environ. 2021, 256, 118382

Sugrue R A, Preble C V, Tarplin A G and Kirchstetter T W 2022 In-use passenger vessel emission rates of black carbon and nitrogen oxides Environ. Sci. Technol. 56 7679–86

The authors present the results of a comprehensive study on the differentiated assessment of inland shipping emissions on the Upper Rhine near Worms in Germany. They use a wide range of measurement techniques to detect gaseous (CO₂, NOx and O₃) and particulate (PNSD, PNC, PMx, soot) air pollutants. Two sites have been selected for the measurements, allowing different scenarios to be mapped. One site was located on a bridge in order to record the plumes from passing ships close to the source. The second site was chosen directly on the banks of the Rhine. In this way, it is possible to determine the level of emissions that could affect people living near the Rhine. Particularly noteworthy is the methodology developed to identify individual ship plumes. The algorithm used avoids overlapping plumes, which can be caused by several ships passing at the same time. As a result, only clearly identifiable ship plumes are included in the evaluation. This results in a significantly reduced number of evaluable ship plumes and also reduces the number of individual ships in the composition of the shipping fleet. At the same time, the quality of the subsequent allocation and classification is significantly improved. In particular, the continuous long-term measurements over a period of one year provide a good picture of the emissions of the shipping fleet in this part of the Rhine. In addition, the emission factors can be calculated under real conditions, leading to a better understanding of the impact on inland navigation. This work represents a solid contribution and, in part, a new scientific approach to the measurement and characterisation of emissions from inland navigation under real conditions. The work is recommended for publication by this reviewer. The following suggestions may be incorporated into the authors' opinion.

P3 L21

…high temporal resolution of ~1 s…

Maybe one can mention, that the SMPS has a different and longer temporal resolution for a whole scan of the size range. Additionally, one could also explain the “problem” with scanning devices as a SMPS with a moderate sampling time. The assumption with a scanning device as the SMPS is that the aerosol spectrum does not change much over the time of a scan. However, this can occur with passing ships and short-term increases and thus lead to a distorted PNSD.

P4 L11

…Instrument-specific sampling lines of 4-5 m length…

It seems that the calculated particle loss under 10 percent is relatively low. I would expect a higher particle penetration at this length of the sampling line. Did you use separeted sampling lines or did you use one sampling line with a higher volume flow and a manifold leading to the individual measuring devices?  

P4  L12

…to enable an undisturbed incoming flow.

Doesn't the bridge itself generate turbulence that can contribute to influencing the wind field at the measurement site? Are downwind eddies possible that carry road traffic emissions down to the measurement site and superimpose the ship plumes as well?

P5 L5

…to avoid strong interferences from road traffic.

You have chosen the locations to also avoid the influence of traffic related air pollutants. I am not familiar with the local conditions, but a look at the Nibelungen Bridge shows that this is a double bridge with two lanes each. What traffic volume can be expected there? Is there rush hour and congestion with traffic jams on the bridge? Especially with winds from northern directions, lee vortices could transport the TRAPs to the sampling point.

P6 Table 1

Here the temporal resolution from the AIS signals is 1 s. To the best of my knowledge, an inland vessel sends a data set only every 10 s, depending on the current movement status.

P12 L20-21

…further results […] refer to this instrument.

This sentence is somewhat confusing, since in the coming chapters the results on RIV site will also be reported, which, however, were measured with the SMPS.

P12 L26-27

The study by Pohl et al. was performed in Duesseldorf. So please change Upper to the Lower Rhine.

P14 L16

…as well as modern ships with exhaust after treatment…

With regard to the CLINSH project. Weren't up to 40 ships retrofitted with downstream exhaust aftertreatment systems? Are the data or names of the ships available the authors to specifically read them out in their data set in order to be able to better scale up the positive effect of the emission reduction? This would be a good contribution, especially in view of the continuing increase in shipping traffic in the future.

P21 L10

With a BC fraction of 38 % for...

It is (for me) not clear to which correlation the value is. Can you please more specify this. Is it BC_{880 nm} to total BC?

P21 L14

The proportion coming from biomass burning is mentioned here as about 10 % from biofuel combustion. Could it be a possible reason that the analyzed probe isn’t just from ships because you also measure the background were also particles coming from wood fires, cigarette smoke, etc. Maybe there could be a hint, if the amount of bb is higher during the wintertime due to fireplaces?

P21 L25

…(see methods).

Please refer to the chapter.