EGU21-10848
Variability in modelled
airborne dust mineralogy
derived from
global soil composition
uncertainties
María Gonçalves Ageitos
Matt Dawson, Vincenzo Obiso, Martina Klose,
Ron Miller, Oriol Jorba, and Carlos Pérez
García-Pando
EGU General Assembly 2021
EGU21-10848 Online vPICO: 30.04.2021
EGU21-10848
Relevance of dust mineralogy
• Mineral dust aerosols interact with different elements of the climate system.
• Rather than a homogeneous species, as many Earth System Models (ESMs) assume, dust is a
heterogeneous mixture of different minerals with varying physical and chemical properties, which
ultimately determine dust-climate interactions:
• Iron oxides absorb solar radiation, while calcite and quartz interact mostly with long-wave
radiation.
• K-feldspars are considered efficient ice-nuclei, therefore influencing the formation of mixed-
phase clouds.
• Calcite and other alkaline compounds influence aerosols’ pH, therefore altering atmospheric
chemistry. They also favor coatings of dust by acidic species, which increase dust
hygroscopicity favoring cloud formation processes.
• Iron and phosphorous compounds in dust can act as nutrients for ocean and terrestrial
ecosystems, therefore affecting global biogeochemical cycles.
• Explicit representation of minerals in ESMs is challenging, due to: (1) existing uncertainties on the
mineral composition at dust sources, and on the resulting (2) emitted minerals Particle Size
Distribution (PSD), (3) high computational demand, due to the increased number of tracers in
models.
4
EGU21-10848
Aim
• This work aims to assess the variability of global dust composition
due to uncertainties in the characterization of the parent soil
mineralogy.
• We use two available global Soil Mineralogy Atlases, developed by Claquin
et al. (1999) –C1999- and Journet et al. (2014) –J2014-, which represent
respectively 8 and 12 relevant minerals for climate.
• We perform two 5-year long simulations at the global scale with the fully-
coupled chemical weather prediction model MONARCH, considering C1999
and J2014 to derive the fractional mineralogy at emission.
• Uncertainties in the soil mineralogical information are identified and the
resulting modelled mineral fractions are compared to observations.
5
EGU21-10848
0.0
0.2
0.4
0.6
0.8
1.0
0.1 2 5 10 20
D(mm)
dV/dlnD
Dust mineralogy in the MONARCH model
Soil
mineralogy
atlas
Size
resolved
mineral
fraction at
emission
Flexible
implementation
into MONARCH
dust module
Calcite annual mean surface
concentration (µgm
-3
). Year 2006-2010
HWSD
7
The chemical weather prediction model, MONARCH (Klose et al. 2021,in review, and
references therein), has the capability to explicitly trace minerals:
• Each mineral is defined through a sectional size distribution (8 bins, with diameters from
0.2 to 20 µm)
• The emission flux results from multiplying the normalized mass fraction of each mineral to
the total dust flux calculated by MONARCH (per bin)
Homogeneous dust annual mean surface
concentration (µgm
-3
). Year 2006-2010
EGU21-10848
8
Global soil mineralogy atlases
Claquin et al. (1999)
Journet et al. (2014)
Arid and humid areas
Focus on arid areas
Methodology
Soil types and units (depend on
formation, physic-chemical characteristics,
etc.)
• FAO map of the soils of the world
• FAO90, FAO74 soil units (HWSD)
Soil texture (%clay, %silt,
%sand)
• Topsoil texture data (HWSD)
• Textural triangle (STATSGO-FAO)
Mean mineralogical tables
• Claquin et al. (1999), based on 239 soil
descriptions. 8 minerals
• Journet et al. (2014), based on more
than 700 soil descriptions. 12 minerals
Soil
mineralogical
composition
(%w for selected
minerals)
2 size classes:
clay (ϕ < 2 µm)
and silt (ϕ 2-63
µm)
Spatial
resolution up to
0.00083º
EGU21-10848
9
Minerals considered in clay and silt
fractions
Clay (ϕ < 2 µm) Silt (ϕ 2-63 µm)
C1999
(Claquin et al. 1999)
J2014
(Journet et al. 2014)
C1999
(Claquin et al. 1999)
J2014
(Journet et al. 2014)
Quartz Quartz Quartz Quartz
Feldspars Feldspars Feldspars
Calcite Calcite Calcite Calcite
Gypsum Gypsum
Illite Illite
Kaolinite Kaolinite
Smectite Smectite
Vermiculite
Chlorite Chlorite
Mica
Hematite Hematite (irox)
Goethite Goethite
EGU21-10848
10
From soil mineral fractions to emitted PSD
0.2
0.4
0.6
0.8
1.0
0.1 2 5 10 20 50
D(mm)
dV/dlnD
PHYLLOSILICATES
(illite, smectite, kaolinite)
CALCITE
QUARTZ
FELDSPAR
GYPSUM
HEMATITE
Dispersed PSD
Reaggregation
and
fragmentation
0.0
0.2
0.4
0.6
0.8
1.0
0.1 2 5 10 20
D(mm)
dV/dlnD
0.00
0.25
0.50
0.75
0.2 2.0 20.0
D(mm)
Normalized dV/dlnD
Disturbed soil PSD
Emitted dust PSD
Soil clay in emitted
silt−sized dust
Perlwitz et al., 2015a,b;
Pérez García-Pando et al.,
2016; Pérez García-Pando
et al., in prep
Emitted PSD
0.0
0.2
0.4
0.6
0.8
1.0
0.1 2 5 10 20
D(mm)
Relative fraction
We get a mass fraction per
mineral and bin in each model
cell, which is applied to the dust
emission flux
EGU21-10848
12
Example of differences in soil datasets:
feldspars and quartz in silt
EGU21-10848
13
MONARCH mineralogy:
from soil to emission
Soil atlas: silt, iron oxides
(dispersed PSD)
Fraction of iron oxides in MONARCH bin6
at emission (over total dust flux)
(emitted PSD)
EGU21-10848
2006-2010 Annual mean surface concentration of
dust
MONARCH model
Global domain at 1x1.4º
resolution
48 vertical layers up to 10 hPa
2 experiments of 5 years (2006-
2010), with 1 year of spin-up
Homogeneous dust traced as a
reference
C1999 experiment: 8 minerals
(illite, montmorillonite, kaolinite,
calcite, qypsum, quartz,
feldspars, hematite)
J2014 experiment: 11 minerals (illite, montmorillonite, kaolinite, vermiculite, chlorite,
calcite, gypsum, quartz, feldspars, iron oxides, mica)
EGU21-10848
16
2006-2010 mean iron oxides and calcite fraction (%w) at surface conc.
EGU21-10848
17
2006-2010 mean illite and montmorillonite fraction (%w) at surface conc.
EGU21-10848
18
2006-2010 mean feldspars and quartz fraction (%w) at surface conc.
EGU21-10848
Methodology for minerals’ evaluation
• Perlwitz et al. (2015b) compiles available
observations of mineral fractions at surface
concentration and dry deposition fields, from
fixed lat-lon sites and cruises.
96 locations with information on mineral fractions at surface
concentration
The evaluation procedure is as follows:
1. We calculate model monthly means collocated in time, space and
considering the size fraction of the observations.
2. We calculate the mass fraction of those minerals that are common between
observations and model for each sample
3. We obtain an observation uncertainty comparable to that of a modelled
monthly mean, as in Perlwitz et al. (2015b) considering:
1. The temporal representativeness of the measurement
2. The sample size
3. Occasionally extending the spread of the model to the observational
value
EGU21-10848
Preliminary evaluation results: mineral fractions at
surface concentration
21
C1999 does not
include chlorite
Illite fraction (%w) Montmorillonite fraction (%w)
Kaolinite fraction (%w) Chlorite fraction (%w)
< 2 µm
< 10 µm
< 20 µm
bulk
2-20 µm
< 2 µm
< 10 µm
< 20 µm
bulk
2-20 µm
EGU21-10848
< 2 µm
< 10 µm
< 20 µm
bulk
2-20 µm
Preliminary evaluation results: mineral fractions at
surface concentration
22
Calcite fraction (%w)
Feldspars fraction (%w) Quartz fraction (%w)
Iron oxides fraction (%w)
< 2 µm
< 10 µm
< 20 µm
bulk
2-20 µm
EGU21-10848
Conclusions
• Currently available Soil Mineralogy Atlases (SMAs) rely on a limited number of
observations, which are extended with empirical relationships and geographically
extrapolated through the application of different hypothesis.
• J2014 considers a larger number of samples than C1999, but those are distributed into a
larger number of soil units. The resulting soil mineral fractions differ over dust sources,
particularly for the silt sized fraction.
• Additional choices, made when implementing the minerals in the model framework (e.g.
mapping of soil units to the model grid, normalization of soil mineral fractions, projection of
the size-distributed soil composition on the emitted size distribution), may result in non-
negligible differences on the final modelled mineralogy.
• Using a common approach for the model implementation and experiment design, we find
differences in modelled mineralogy arising solely from uncertainties in the SMAs.
• Clays (illite, smectite, kaolinite) are the less impacted by the choice on the soil mineralogy
atlas.
• Calcite, iron oxides, and feldspars mass fractions at the surface show large differences
between the C1999 and J2014 MONARCH experiments.
24
EGU21-10848
Conclusions
• When compared to observations, calcite mineral fraction is underestimated at the surface
in MONARCH-C1999, while overestimated in MONARCH-J2014.
• Feldspars fraction at the surface is slightly underestimated both in MONARCH-C1999 and
MONARCH-J2014. However, the hypotheses applied in J2014 to extend the mineralogy of
the silt fraction derive in a lower variability for the fractions at diameters below 10 µm.
• Quartz fraction at the surface is overestimated in both MONARCH-C1999 and MONARCH-
J2014, particularly at sizes below 10 µm.
• New approaches to project soil composition to the emitted PSD will be tested to add
information on the preferential size distribution of the minerals (besides the crude
discrimination used now, which only accounts for two soil size classes: clay and silt).
• Our results support the need for further observational constraints to better characterize the
soil and thus the airborne dust composition. The experimental campaigns on airborne dust
characterization planned in the framework of the FRAGMENT ERC project, as well as the
soil characterization that will result from the EMIT NASA mission are promising initiatives in
this sense.
25
EGU21-10848
Acknowledgments
• We thankfully acknowledge the computer resources at Marenostrum4, granted
through the PRACE project eFRAGMENT2 and the RES project AECT-2020-3-
0020; as well as the technical support provided by the Barcelona
Supercomputing Center and the CES team of the Earth Sciences Department.
• This work was supported by the ERC Consolidator Grant FRAGMENT (grant
agreement No. 773051), and the AXA Chair on Sand and Dust Storms at BSC
funded by the AXA Research Fund both led by Dr. Carlos Pérez García Pando,
who also acknowledges the Ramon y Cajal program (grant RYC-2015-18690) of
the Spanish Ministry of Science, Innovation and Universities and the ICREA
program .
• The research leading to these results has also received funding from the
Spanish Ministerio de Economía y Competitividad as part of the NUTRIENT
project (CGL2017-88911-R) and the H2020 GA 821205 project FORCeS.
• We also thank Dr. Jan Perlwitz for his valuable contribution to the minerals’
evaluation presented here, and Dr. Yves Balkanski and Dr. Emilie Journet for
their help with the definition of the soil mineralogy maps. Finally, we wish to
acknowledge the work of all the members of the BSC Earth Science
Department group who contribute to MONARCH model and infrastructure
developments.
27
EGU21-10848
References
• Claquin, T., Schulz, M., and Balkanski, Y. J.: Modeling the mineralogy of atmospheric dust sources, Journal of Geophysical
Research Atmospheres, https://doi.org/10.1029/1999JD900416, 1999.
• Journet, E., Balkanski, Y., and Harrison, S. P.: A new data set of soil mineralogy for dust-cycle modeling, Atmospheric
Chemistry and Physics, 14, 3801–3816, https://doi.org/10.5194/acp-14-3801-2014, http://www.atmos-chem-
phys.net/14/3801/2014/, 2014.
• Klose, M., Jorba, O., Gonçalves Ageitos, M., Escribano, J., Dawson, M. L., Obiso, V., Di Tomaso, E., Basart, S., Montané
Pinto, G., Macchia, F., Ginoux, P., Guerschman, J., Prigent, C., Huang, Y., Kok, J., Miller, R. L., and Pérez García Pando, C.:
Mineral dust cycle in the Multiscale Online Nonhydrostatic AtmospheRe CHemistry model (MONARCH) Version 2.0,
Geoscientific Model Development, in review.
• Kok, J. F.: Does the size distribution of mineral dust aerosols depend on the wind speed at emission?, Atmospheric Chemistry
and Physics, 11, 10 149–10 156, https://doi.org/10.5194/acp-11-10149-2011, 2011.
• Pérez García-Pando, C., R. L. Miller, J. P. Perlwitz, S. Rodríguez, and J. M. Prospero: Predicting the mineral composition of
dust aerosols: Insights from elemental composition measured at the Izaña Observatory, Geophys. Res. Lett., 43,
doi:10.1002/2016GL069873, 2016.
• Perlwitz, J. P., Pérez García-Pando, C., and Miller, R. L.: Predicting the mineral composition of dust aerosols - Part 1:
Representing key processes, Atmospheric Chemistry and Physics, 15, 11 593–11 627, https://doi.org/10.5194/acp-15-11593-
2015, 2015a.
• Perlwitz, J. P., Pérez García-Pando, C., and Miller, R. L.: Predicting the mineral composition of dust aerosols - Part 2: Model
evaluation and identification of key processes with observations, Atmospheric Chemistry and Physics, 15, 11 629–11 652,
https://doi.org/10.5194/acp-15-11629-2015, 2015b.
28
