Review of „Last Glacial loess in Europe: luminescence database and chronology of deposition” by Bosq et al.

In the last decade or two, a large amount of new absolute chronological data from European loess sediments has been published, mainly for the last glacial. Most of these are luminescence ages, a smaller part radiocarbon dates. Although some reviews of mass accumulation rates calculated from previously published age data have been published in recent years, the authors have recognised the need to create a single database, which can further be developed in a uniform way. ChronoLoess is a great initiative and the authors have obviously invested a lot of time and effort in the creation of this database of over 1400 entries. They have also partially revamped the previously accepted methodology of extracting useful information from the published absolute ages and in this very exciting manuscript they also present and discuss the spatial and temporal variations of loess accumulation focussing on the North European Loess Belt (NELB) and the Perialpine Loess (PL).

The language of the manuscript is excellent, the structure is logical and the arguments are easy to follow. The methodology is well summarized and seems scientifically correct, with only two comments from my side concerning the 1) calculation of Rb concentrations from the K contents and the 2) treatment of obvious age inversions in chronologies of some of the sites (e.g. Bialy Kosciol). Regarding the first issue: the major problem here is that the equation published by Mejdahl (1987) systemtically overestimates Rb from K and the correlation between these two elements are rather weak in reality (see Fig 3 and discussion in Buylaert et al., 2018, Radiation Measurements 120, 181-187). Please comment on this and the second question on age inversion.

The main conclusions of the authors are correct and I can agree with most of them, although there are some comments that I would suggest for reflection or improvement. One critical issue is how to determine in a statistically exact way when the time-activity curve crosses the uniform random distribution (URD), as the authors have used this as a basis for inferring the initiation of significant dust accumulation and have correlated this with certain Heinrich events (HE). In my opinion, this has considerable uncertainties on the order of 1000-3000 years that need to be taken into account, however, I believe that the dust accumulation peaks are well defined (NELB: 21.8 ka, PL: 23.9 ka) and may indeed be causally related to HE2.

I have further suggestions, especially for certain interpretations, such as the potential causes of the time lag in the aforementioned peak accumulation (NELB vs. PL). This may indeed depend on several factors, as the authors rightly state, but these factors need to be reconsidered/refined. However, these are rather pedantic points and do not affect the main conclusions. I believe that if the authors improve these points in a small revision and the other peer reviews end up favourably, this paper can be published in ESSD.

The study “Last Glacial loess in Europe: luminescence database and chronology of deposition” by Bosq et al., aims to create a unique database of European loess osl dated chronologies and accumulation rates during MIS 1-MIS 3 period. The data appears robust, and the methods used are sound and appropriately applied.

In my opinion this is a comprehensive study which will be a significant contribution to the loess, geochronological and palaeoclimatological community. The manuscript itself is well-structured, language is appropriate, the methods ant data are well presented and the conclusions are sound. The figures are also well designed and complement the data accurately. I would like to highlight that the authors did an excellent job with the manuscript and I very much enjoyed reading it.

I recommend publication, however, I would only like for the authors to take into account a few of my questions.

• Have you considered the dating resolution as a parameter when calculating MARs? In my experience, dating resolution can have a significant impact on the accurate calculation of MAR distribution over the investigated time period. Maybe you should mention that.
• Have you thought on maybe applying an additional age-depth modelling method and compare the results? I acknowledge that this would be a large amount of additional work, but often different modelling software produces different results. It would be interesting to see the discrepancies. This is just a thought, not a critique point.
• Have you thought of expanding the investigation beyond MIS 3? I do know that these chronologies are very scares, but there are some studies from the Carpathian Basin where LPS were dated beyond MIS 5.
• I think it would have been interesting to compare these results with MARs from LPS outside of Europe (e.g. Chinese Loess Plateau). We might get a better insight into the atmospheric dust activity on an intercontinental scale. However, I do recognize that this might be beyond the scope of this paper. Please take this just as an idea and/or a suggestion.