phys.net1610652016.PDF

Response to Nick Schutgens’ comment: A very useful paper, and it will be interesting to see how to reconcile various measurements. Here I just want to mention the issue of temporal sampling, see also: http://www.atmoschem- /16/1065/2016/ While this temporal sampling issue is important for model evaluation, it is equally important in comparing different observational datasets. I see two issues relevant to the current paper: Page 18, Line 732-735: Fig 7 was apparently made with different samplings of the in-situ and AERONET measurements. Is that the case for other figures as well? How might that affect results? Figures 3-5 were made for matched samplings of the in-situ and AERONET measurements where matched means that the measurements were made within 3 h and 15 km of each other. Given the high correlation in AOD we are reasonably confident that with these sampling constraints the in-situ and remote sensing instruments were measuring in the same atmospheric column. The literature studies cited in Table 4 and included in Figure 6 used similar temporal and spatial matching criteria (see comments column in Table 4), with the exception of the DABEX dust/biomass burning flights (Osborne et al 2008; Johnson et al 2009) which were matched in time but less so in distance (flight profiles were within 100 km of AERONET retrievals). In contrast, Figures 7 and 8 utilized the multi-year climatological data sets available for each measurement which have different samplings. Your work (e.g., Schutgens et al., 2016) shows there can be large differences when comparing values obtained with different samplings (more than 100% for AOD), particularly when there are high levels of variability in the data. In our manuscript the different temporal samplings are likely one contributor to the relatively small differences observed between the in-situ (red line) and AERONET 1.5 AOD (black line) although other things (e.g., assumptions about aerosol hygroscopic