Disentangling dust transport pathways to western Crete using aerosol measurements and HYSPLIT trajectories
Abstract. Mineral dust transported to the eastern Mediterranean affects atmospheric conditions, radiation balance, soil development processes, and ecosystem functioning in deposition areas. While the dominant source regions of deposited dust can be roughly identified with the help of sediment fingerprinting approaches, different methodological challenges hamper the exact identification of aeolian sediment sources and their potential temporal and spatial shifts. This study evaluates whether the combination of ground-based aerosol concentration measurements and backward-trajectory modelling can provide a temporally more resolved perspective on dust-source variability in western Crete. With this aim, aerosol concentrations from a monitoring network in and around the Lefka Ori mountains were used to identify days with elevated surface-near dust loads between October 2023 and September 2025. For these days, air-mass pathways were reconstructed using the atmospheric transport and dispersion modelling system HYSPLIT, and grouped into recurring transport patterns. The resulting clusters show that elevated aerosol concentrations were primarily associated with transport from southern directions representing North Africa, especially Libya and Egypt, but also with frequent northerly pathways connecting Crete to the Balkan, the Black Sea region and eastern Europe. Clusters summarising southerly transport patterns showed the strongest seasonal relevance during the main dust season in spring and were associated with higher aerosol concentrations, whereas clusters associated with northerly transport patterns occurred frequently but were generally associated with lower aerosol concentrations. The applied method enabled a temporally high resolved assessment of dust source regions, thereby revealing subordinate but nevertheless relevant source areas. Future work should combine longer observation periods with chemical, mineralogical, and isotopic analyses to quantify the relative contribution of these source regions to the deposited dust.
Review of manuscript egusphere-2026-5456 entitled: “Disentangling dust transport pathways to western Crete using aeorosl measurements and HYSPLIT trajectories”, by Simon Bitzan, Fabian Kirsten and Kerstin Schepanski.
In this well-structured and easy-to-read manuscript, the authors present a set of almost two years (October 2023 – September 2025) of optical measurements of aeolian dust at eight stations on the western part of the island of Crete. Although the eight stations were set up in a relatively small area (less than 30 x 50km) there is large variability in the amount and timing of the dust events that were registered. Only events that were observed simultaneously by at least three stations were classified as ‘dust events’. The authors then tried to reconstruct the provenance of these dust events by calculating backward trajectories with the Lagrangian model HYSPLIT.
When accepting to assess the manuscript, I had not realized that there would be so much statistics used in it. Clearly, I am not an expert in statistics and therefore, I cannot judge the validity of the used approach. However, I do know a bit about the use of HYSPLIT to reconstruct wind trajectories and its large error bars and uncertainties. Small changes in parameters used for the calculations can lead to large differences in the reconstructed trajectories. After all, it is only a model…. And this is where the weakness in the approach becomes clear, I think; one cannot improve the error bars on a modelled reconstruction just by applying a lot of statistical analyses. The authors state in their conclusions (lines 544-549) that: “The temporally highly resolved assessment of trajectories on days with elevated dust concentrations in western Crete further shows that seasonally dominant source regions, particularly during the main dust season, are accompanied by additional dustbearing trajectories from other source areas. This implies that mixed deposition may occur even over relatively short time intervals. Consequently, dust transport during the main dust season should not be interpreted as a continuous flux from the dominant source region with only varying intensity, but rather as an intermittent process that is frequently interrupted or accompanied by contributions from weaker additional sources.” In other words: no single dust event originates from one single dust source, but throughout the event, many different sources may contribute to it. In itself, this is not a new observation, but it could also be explained by the huge uncertainties in the backward trajectories modelling. The authors conclude that analyses of the actual dust in terms of e.g., radiogenic isotopes would help constrain the provenance of the dust. These data were presented in the 2026 JGR paper by the authors but it would have helped to have a summary of the North African samples here to compare the two independent provenance methods.
IF HYSPLIT is really so reliable, it should be able to show forward modelling towards Crete as well; have the authors checked this?
In addition to my worries about the approach and a too high confidence in the HYSPLIT trajectories, I am also curious how the authors explain such a large variability in the observations of dust in such a relatively small area? As can be seen in figure 4, only a relatively small amount of dust events is registered by all stations. Could it be that the large gradients in topography lead to variability in wind patterns and therefore dust transport? The authors have shown that deposition is fairly uniform across the sampling stations but apparently this is not the same for the amounts of dust observed by the optical instruments? Or could there be another source of variability here? Could smoke or pollution be mistaken for dust?
By the way, the way of presenting these figures (4 and 5) is confusing; rather than ranging from zero to seven, I would think that the stations should be labelled one to eight, right? The same holds for figure 5; at zero stations there were N=715 dust events? How many were actually observed by all N=8?
Minor issues
Figure 1 should zoom out a tick, so that one can see the orientation and distance of the island of Crete relative to North African dust sources. In line 86 you state that the island is situated north of the Lybian coast. Please also show this.
Line 31; the reference to dust particles being <100µm is outdated. In your 2026 JGR paper you present coarser-grained material. Please refer to e.g., Adebiyi et al., 2023.
Check the order of references in your citations, as well as in your reference list.
Line 152 – 157 is unclear to me; you mention that PM10 only represents the smaller fraction and therefore you compare with the PM2.5 fraction? The latter is even finer-grained, right?
Lines 198 – 200; you state that the material presented in your 2026 JGR paper was shown to be fairly uniform in particle size and composition. However, this is deposited dust, correct? The data you present here are optical measurements of suspended dust, correct?
Figure 7 with all possible trajectories is a little disturbing as it shows that winds originate from literally everywhere in the eastern Mediterranean. As was also shown by Van Boxtel et al., 2025, and 2026, there is a weak seasonality in dust transport across the Mediterranean, with a maximum in boreal spring. Would it be possible to colour the trajectories by month or season? This might result in a comparable figure as figure 8a or 10.
I do not see the relevance of figure 12; too many graphs and too small. Would it be possible to show a few typical examples and show the rest in the supplement?
In summary, I think it has become clear that I am not convinced by this manuscript and I am on the fence as for my recommendation. I hope my comments are seen as constructive and I hope the authors can use them to improve their work.
Cheers,
Jan-Berend Stuut
References cited in this review
Adebiyi, A., et al. (2023), A review of coarse mineral dust in the Earth system, Aeolian Research, 60, 100849, doi:https://doi.org/10.1016/j.aeolia.2022.100849.
Van Boxtel, A. P. E., J.-B. W. Stuut, and F. Peterse (2025), Identifying Saharan dust driven export of biogenic material in the ultraoligotrophic eastern Mediterranean Sea, Frontiers in Marine Science, 12, doi:10.3389/fmars.2025.1537028.
Van Boxtel, A. P. E., A. Rice, C. Kurpershoek, T. de Nooij, G. J. de Lange, F. Peterse, and J.-B. W. Stuut (2026), Inter- and intra-annual variations in Saharan dust fluxes in the eastern Mediterranean Sea and their connection to carbon export, Aeolian Research, 77, 101070, doi:https://doi.org/10.1016/j.aeolia.2026.101070.