Interactive comment on “ Technical Note : Minerals in dust productive soils – impacts and global distribution ”

Response: We are aware that there is no dense network of soil sampling in desert regions and the gap in such measurements limits the accuracy of mineralogy databases. We however believe that by combining the existing data on minerals with other relevant data (soil types, land cover), and mapping the mineral fractions into a high-resolution grid a value is an added value to the existing geographical mineralogy distribution. We developed the database on 30sec (~1km) grid so to be compatible with other databases (topography, soil texture, land cover) frequently used in atmospheric modelling. The high-resolution structure of the database permits its use in both highresolution regional models downscaled horizontally up to several kilometers, and in coarser global models as well. To better clarify the requested Referee’s comment, we made the following changes in the article text: At the end of Introduction we added the sentence: Database is applicable for both global and regional scale atmospheric dust models. Furthermore, at 26020,19 we added: The database contains a mean global distribution of soil mineral composition appropriate for implementation in models for global and regional studies. Introducing information on dust mineral composition obtained from the database, in spite of large uncertainties originating from scarce soil sampling is expected to provide additional information for better understanding of the atmospheric dust processes.

Response: Referee's comment relates to : "Gypsum originates mainly from (paleo-) lacustrine sources and has moderate absorption in the infrared spectrum."The importance of lacustrine sources for dust emission was emphasized by Tegen et al. (2002) and in their study lacustrine sources are introduced as a separate soil texture class.For example, the Bodélé depression, one of the largest worldwide dust source consists of fine lacustrine sediments deposited by paleo lake Chad in the Holocene.Tegen et al states that gypsum is mainly existing in lacustrine sediments typical for the Bodele depression, To avoid confusion we decided to remove the following sentence from the paper: "For example, the Bodélé depression, one of the largest worldwide dust source consists of fine lacustrine sediments deposited by paleo lake Chad in the Holocene."Reviewer26018/27: I do not think that this is a good idea, as hematite and goethite (or other iron oxide/hydroxide compounds, which are often not very well distinguished) do have different spectral behavior.If the information on iron speciation is available, I would suggest keeping it a long as possible.
Response: Since we did not have additional data on hematite and goethite as separate fractions, we keep them together under the name of hematite, as was originally done by Claquin et al. (1999).

Reviewer26019/29:
The phosphorus solubility is also depending on atmospheric processing (Nenes et al 2011, doi: 10.5194/acp-11-6265-2011) and in general the P cycle is not well understood (Okin et al. 2011, doi: 10.1029/2010GB003858), so I would suggest not giving a number here would be more favorable.
Response: Accepted; we changed the corresponding sentence as:"Although only small part of phosphorus in mineral dust aerosol is bioavailable (Jickells and Spokes, 2001;Mahowald et al., 2005), it may play an important role in biological response of the marine ecosystem when dust is deposited into the ocean".Reviewer26020/9-10: I understand that there is not more data available, but I have serious doubts whether this is close to reality.Trace contributors to soils like P most probably depend on traces in the geological basement, but soils are a product of basement and climate, so at least a very large uncertainty with respect to the P content should be given as a caveat here.

Response:
In the absence of more detailed sampling data for P, we followed the approach of Okin et al. ( 2004) to downscale spatial distribution of P over arid areas.We therefore changed sentence as:" In the absence of more detailed information, we assumed that all soil types that belong to the same group have the same phosphorus content.The obtained results are given in percentages equal for both clay and silt populations."Reviewer26020/20: Up to here the paper is on soil composition, but now this soil composition is set equal to an atmospheric dust composition.However, the transfer function from soil composition to dust composition is not unity (e. g., Eltayeb et al. 2001, doi: 10.1023/A:1012272208129).In a work like that, I would expect at least some comments on this to aid the user in identifying potential uncertainties.

Response:
We have added a sentence to identify potential uncertainties as required by the Reviewer: "The quantification of minerals in aerosol samples is not easy, mainly because of the very small sample masses involved.Some authors have therefore assumed that the aerosol mineral content is the same as in source soils (Claquin et al., 1999) Reviewer26030/Table 1: Please distinguish carefully between "fractions of clay minerals" (i.e. relative amount of minerals in the clays group) and "clay fraction" (particles with sizes smaller than 2 µm).I suggest -to improve uniqueness -to print explicitly "clay size fraction", if the particle size is referred.Same applies to "fraction of silt".

Response:
Accepted.Titles in Table 1 are changed to:"Table 1. Mineral content in clay and silt size fraction in selected soil types.Mineral content in clay size fraction normalized to 100%Mineral content in silt size fraction normalized to 100%".As also suggested, we improved uniqueness in the paper by introducing "clay size fraction" and "silt size fraction" wherever appropriate in the text.
Reviewer26030/Table 1: Some of these compositions are very similar.I understand that these numbers represent an ideal composition of a particular soil or a composition that represents and average over several samples of that soil.But how variable can be the composition of a single soil type?This would be particularly interesting, when addressing iron and phosphorus contents, which have been most probably not of primary concern when these soil classes have been established.

Response:
We are aware that there is large variability of mineral fractions given in Table1.Claquin et al. report that standard deviations of a mineral content range between 27 and 33% of the average observed value.A corresponding sentence is added in the article text.1: This table implies that clay minerals like kaolinite, illite are never present in grain sizes larger than 2 µm, and that feldspar and gypsum are never present in sub 2 µm particles.This is certainly not true (e. g., Leinen et al. 1994, doi: 10.1029/94JD01735;Reid et al. 2003, 10.1029/2002JD002935;Kandler et al. 2007Kandler et al. , 10.1016Kandler et al. /j.atmosenv.2007.06.047).06.047), but relies on the definition of "clay mineral size fraction" as an a priori assumption.However, for the representation of the true composition, this might introduce a significant error, and should be addressed here.At least, it should be mentioned that this clearcut composition change is artificial.

Reviewer26030/Table
Response: Following the Reviewer's suggestion, we have added in the text (26017/5) a sentence: "However, the value of 2 µm which delineates sizes between clay and silt minerals is arbitrary and used here as a first order approximation." Reviewer26034: This map is in my opinion quite misleading.Looking at the map, I would assume that most of the dust consists of quartz and feldspars.Though this is probably true for the soils, the question of applicability to airborne dust arises again, as that contains considerable amounts of clay minerals.Instead, I would suggest displaying the maps separately for the silt size fraction and the clay size fraction.

Response:
In dust modelling applications, effective percentage of a mineral for a given soil population (clay or silt) is calculated by multiplying fractions of a mineral and a soil population.The mineralogical database actually contains fractions of minerals separately for clay and silt populations, permitting so users to use their own data on clay and silt.To reduce the number of shown maps in the article, we showed the effective percentages of minerals calculated using the hybrid STATSGO-FAO silt/clay fractions.Following the referee suggestion, we will try to provide as well the maps separately for the silt size fraction and the clay size fraction, as a supplementary material following journal rules for supplementary material submission, or we will provide this figures on download web page of database.
Reviewer: Also, there are very large phosphorus deposits in NW Africa (Morocco), which are even mined.But the map doesn't show anything of it, so the relevance of the phosphorus concentrations estimated by that database should be addressed critically.
Response: Like in the approach we used for generating other components of the database, in case of P we have also calculated its fractions using the published evidence on P spatial distribution (Okin et al, 2004) combined with the gridded data on sol types and land cover.Although our approach has limitations resulting from the lack of enough sampling data, we still believe that information on P has a value for modelling research.