The manuscript entitled “A Climate Data Record of Stratospheric Aerosols” presents a new climate data record of stratospheric aerosols (CREST) which comes from merging long-term datasets of vertically resolved stratospheric aerosol extinction derived from six limb and occultation satellite instruments such as the SAGE II on ERBS, GOMOS and SCIAMACHY on Envisat, OSIRIS on Odin, OMPS on Suomi-NPP, and the SAGE III on the International Space Station. It is worth mentioning that this new dataset is complementary to NASA-GloSSAC but it provides aerosols at a different wavelength, 750 nm, and it also uses a different collection of individual datasets included in the merged dataset and a different merging approach. For the development of this merged dataset, the authors used aerosol extinction profiles from satellite instruments measuring in the limb-viewing geometry performing a transformation of the aerosol datasets from each individual instrument to the same wavelength 750 nm and their de-biasing and homogenization by adjusting the seasonal cycles. After such homogenization, the data from individual satellite instruments are in good agreement.

In my opinion, the use of the merged CREST dataset is innovative and of high interest to the field. It is a useful tool which can be used in various climate-related studies. I agree with the authors that CREST can be used as a proxy in the regression models for trend analyses, or as a forcing in simulations with chemistry-transport models. At the moment, CREST covers the period from 1984 to 2022, and it is intended to be extended in the future. The use of new instruments could provide added value to the existed dataset. Overall, the manuscript is clear and well-structured giving a detailed analysis of the methodology that has been followed. Below I give some specifying remarks which can improve the manuscript.

Specific comments:

• L85-87: The implementation of the cloud-filtering to the SAGE II aerosol extinction profiles is not well supported. According to the authors, the cloud-affected altitudes are defined at locations where aerosol extinction values at 1020 nm are lower than5 ⋅ 10⁻³ 𝑘𝑚⁻¹ and ratio of extinctions (𝛽_{525 𝑛𝑚}/𝛽_{1020 𝑛𝑚}) is lower than 1.75 without giving more detailed information for the selection of these threshold values.
• L157-204: It is quite difficult to read and understand exactly the implemented methodology for the aerosol dataset merging described in Section 3. There are a lot of phrases and equations involved which cause confusion to some degree. The authors are requested to seek a way to provide a clear description of the Section " Merged dataset of aerosol profiles" which is the key part of the analysis. Moreover, use two subsections instead of one whole section to describe the methodology giving more details in each section. The first one will be focused on the description of the conversion of the extinction coefficients from each instrument’s frequency to the 750nm and the second one will describe the merging of aerosol datasets. 
• In Figures 1 and 2, use the label “Extinction 1/km” instead of “aerosols 1/km”. Add units in the plotted aerosol extinction profiles in Figure 3.
• L259: The authors are requested to give more information regarding the comparison of aerosol extinction time series and profiles between the CREST and GloSSAC datasets since this is the most significant part proving the reliability of the merged dataset.
• As already mentioned in lines 316-324, the comparison of CREST dataset using limb-viewing instruments with aerosol extinction profiles retrieved form active remote sensing instruments such as Aeolus, CALIPSO and the forthcoming EarthCARE satellite mission would provide added value increasing the reliability of the stratospheric aerosol climate data record. This is an interesting and challenging task which can be implemented in a future work (can be mentioned in the discussion or conclusion section).

This dataset provides a very useful summary of global stratospheric aerosol optical depth measurements, created by combining a number of satellite observations. This work clearly parallels the work reported by Kovilakam et al., 2020, but represents an alternate approach. The good agreement between the two datasets strengthens the validity of both (with the caveat that they start with many of the same observation datasets.

My only difficulty with this dataset is that there is a discussion of how uncertainties are calculated, but I cannot find these anywhere in the dataset itself. I had expected each dataset to be matched with an uncertainty estimate. Am I missing something here?  A clarification on this would strengthen the usefulness of the dataset.

There are only a couple of minor technical issues that I noticed that am sure will get corrected.  (Line 167 two full stops; highlighting in the "Data availability" section.