1. Benoit Laurent, Béatrice Marticorena, Gilles Bergametti
Frequencies and intensities of mineral dust emissions from Chinese and Mongolian deserts: A modeling approach 
Benoit Laurent, Béatrice Marticorena, Gilles Bergametti
Laboratoire Inter-universitaire des Systèmes Atmosphériques
International Symposium on Sand and Dust Storm, Beijing, Sept. 2004

2. Studied area: the main deserts of eastern Asia
(35.5°N-47°N; 73°E-125°E)

3. Dust emissions are sporadic and spatially heterogeneous
Dust emission processes
Dust emissions are sporadic and spatially heterogeneous
soil particle movement:
wind friction velocity (U*) > threshold friction velocity (U*t)
Saltation
Sandblasting
Vertical flux F
Surface wind
Horizontal flux G
Main processes of dust production

4. Emission processes Model outputs
Dust emission processes
Emission processes
Model outputs
Erosion threshold
Emission frequencies
(location and periods)
Saltation
Emission flux intensities (quantities)
Sand-blasting

5. 1- Dust emission frequencies
- Erosion threshold
Aerodynamic roughness length (Z0)
Particle diameter (Dp)
Soil moisture
Snow cover
Surface wind velocity

6. U*t = f (Dp;Z0) → Determination of Z0
Erosion threshold parameterization
U*t = f (Dp;Z0)
[Marticorena et al., J.G.R., 1997]
For 50 µm < Dp <200 µm (generally always present in arid soils), Z0 is the key parameter to compute dust emission frequencies
→ Determination of Z0

7. What are the required input data ?
To compute the erosion threshold:
- aerodynamic roughness length (Z0)
remote sensing
10 m erosion threshold wind velocity ●
Dp = 210 µm σ = 1.8
- size-distribution
- soil texture: f (depth)
FAO ●
Soil moisture
- precipit., T°, albedo, geopot.
ECMWF
To compute the emission frequencies we also need: ●
Snow cover
- snow depth
ECMWF ●
10 m wind velocity
- surface wind velocity
ECMWF

8. (POLarization and Directionality of the Earth’s Reflectance)
Z0 retrieved from the Protrusion Coefficient (PC)
Remote sensing: POLDER-1
(POLarization and Directionality of the Earth’s Reflectance)
Protrusion Coefficient (PC) derived from POLDER-1 measurements of bidirectional reflectance
Empirical relation
Z0 = a.exp (PC / b)
with a = cm,
and b = is dimensionless
[Marticorena et al., I.J.R.S., 2004] 1 2 3 4 5 6 7 8 9 10 11 12 13
log10(Z0)

9. Z0 and 10 m erosion threshold wind velocities
Z0 map (¼° × ¼°)
U*t = f (Dp;Z0)
and the neutral vertical wind velocity profile
10 m erosion threshold wind velocity map (¼° × ¼°)
m.s-1
Ut(10m)
[Laurent et al., J.G.R., submitted]

10. Z0 and 10 m erosion threshold wind velocities
In the Gobi:
● Our results: median ~15 m.s-1
● Wind velocities associated with dust storms:
11-20 m.s-1 [Natsagdorj et al., Atmos. Env., 2003]
● Wind tunnel and field studies: m.s-1 [Murayama, Met. Satell. Cent. Tech. Note, 1988; Hu and Qu, Chin. Meteo. Press, 1997]
In the Taklimakan:
● Our results: median ~7 m.s-1
● Wind velocities associated with dust storms: 6-8 m.s-1 [Wang et al., Water, Air, and Soil Poll., 2003]
m.s-1
Ut(10m)
[Laurent et al., J.G.R., submitted]

11. Frequent dust emission areas
Simulation of dust emission frequencies ( )
Frequent dust emission areas
Longitude
Latitude %
Dust storm events during
Dust storm occurrences during
[Sun J. et al., J.G.R., 2001]
[Sun L. et al., Water, Air, and Soil Poll., 2003]

12. Seasonal cycle Simulation of dust emission frequencies (1997-1999)
Frequencies computed with soil moisture and snow cover
[Laurent et al., J.G.R., submitted]

13. Seasonal cycle Simulation of dust emission frequencies (1997-1999)
Frequencies computed with soil moisture and snow cover
Frequencies computed with soil moisture and without snow cover
Frequencies computed without soil moisture and snow cover
[Laurent et al., J.G.R., submitted]

14. Comparison with TOMS Absorbing Aerosol Index frequencies
Simulation of dust emission frequencies ( )
Comparison with TOMS Absorbing Aerosol Index frequencies
Location of the most frequent areas
[Laurent et al., J.G.R., submitted]
In the Taklimakan:
Monthly average dust event frequency
Monthly average frequency of TOMS AAI > 0.7
r ~ 0.95
slope ~ 0.44
Latitude
Longitude
Frequencies of significant simulated dust emissions (flux > g.cm-2.s-1)
Latitude
Longitude
Frequencies of AAI TOMS > 0.7

15. 2- Dust emission fluxes - Flux parameterization
Soil “dry” granulometry
% erodible surface

16. Parameterization of the saltation flux
Flux parameterization
Parameterization of the saltation flux
F= a.G = a.S Srel(Dp).C.U*2(1+U*/U*t (Dp,Z0))(1-U* ²/U*t²(Dp,Z0))
[Marticorena and Bergametti, J.G.R.,1995]
Parameterization of the sandblasting efficiency
a = f(% clay)
[Marticorena and Bergametti, J.G.R.,1995]
→ Determination of the soil “dry” granulometry

17. What are the required input data ?
- aerodynamic roughness length (Z0)
remote sensing
10 m erosion threshold wind velocity ●
in-situ measurements
- size-distribution
- soil texture: f (depth)
FAO ●
Soil moisture
- precipit., T°, albedo, geopot.
ECMWF ●
Snow cover
- snow depth
ECMWF ●
10 m wind velocity
- surface wind velocity
ECMWF
To compute the emission fluxes we also need:
in-situ measurements
Soil “dry” granulometry ●
- size-distribution
% erodible surface ●
- % no cover surface
f(Z0)

18. Sand land east of yellow river
Soil “dry” granulometry derived from in-situ measurements
Sandy lands
Med1(µm)
P1(%) s1
Med2(µm)
P2(%) s2
Taklimakan
110
100
1,79 -
Gurban Tunggut
166
95,6
1,37
348
4,4
2,00
Badain Jaran
196
84,6
1,40
15,4
Tengger
157
89,6
1,30
349
10,4
1,46
Ulan Buh
192
96,8
1,35
3,2
Qubqi
95,7
1,18
4,3
1,93
Sand land east of yellow river
180
83,7
1,25
248
16,3
1,31
Mu Us
243
1,55
Gobi
Alluvial- Dunhuang 48
71,5
433
28,5
Alluvial -Yumen 81
86,4
1,85
485
13,6
1,24
Deluvial An'Xi 71
78,5
445
21,5
1,50
Deluvial - Yumen 78
83,4
1,80
479
16,6
Loess
Shenmu County, Shaanxi 65
1,28
Sandy Loess
Chuanbey, Yulin 74
1,17
Derived from measurements of Gengsheng et al. [Global Alarm: Dust and Sandstorms from the World’s Drylands, report of United Nations, 2001]

19. Simulated year 1997 1998 1999 Total emissions (106 T) 309 417 426
Simulation of dust emission fluxes ( )
Simulated year
1997
1998
1999
Total emissions (106 T)
309
417
426
106 T
Longitude
Latitude
Mean annual quantity

20. Seasonal cycle Simulation of dust emission fluxes (1997-1999)
In frequency: end of spring
In intensity: beginning of spring in 1998

21. Conclusion
Simulations of dust emissions ( )
● Two main dust emission areas (both in frequency and intensity): the Taklimakan desert in the north western China and the Badain Jaran desert in the northern China
● A pronounced seasonal cycle of dust emissions with a maximum (both in frequency and intensity) in spring
● A weak influence of the soil moisture and the snow cover on simulations of dust emissions
● ~ 400 MT/year dust emitted from eastern Asian deserts

22. Benoit Laurent, Béatrice Marticorena, Gilles Bergametti
Frequencies and intensities of mineral dust emissions from Chinese and Mongolian deserts: A modeling approach 
Benoit Laurent, Béatrice Marticorena, Gilles Bergametti
Laboratoire Inter-universitaire des Systèmes Atmosphériques
International Symposium on Sand and Dust Storm, Beijing, Sept. 2004