Dusty ice clouds over Alaska


Particles lofted into the atmosphere by desert dust storms can disperse widely and affect climate directly through aerosol scattering and absorption. They can also affect it indirectly by changing the scattering properties of clouds and, because desert dusts are particularly active ice-forming agents, by affecting the formation and thermodynamic phase of clouds. Here I show that dust storms that occurred in Asia early in 2004 created unusual ice clouds over Alaska at temperatures far warmer than those expected for normal cirrus-cloud formation.


Although the direct light-scattering effects of dust storms are readily observed from satellites1, their indirect effects on clouds are difficult to quantify. Knowledge about the effects of desert dusts on clouds has come from polarization lidars (detection systems that work on radar principles but use laser light), which can identify both non-spherical aerosols and the phase of clouds2. The evidence suggests that, even after great transport distances, aerosols from both the Asian and Saharan deserts affect cirrus and supercooled water clouds3,4,5,6. It has long been known that desert dusts are highly effective as nuclei on which ice crystals form heterogeneously7.

From late February to mid May 2004, ruby (0.694 µm wavelength) polarization-lidar observations8 from the University of Alaska at Fairbanks (64.86 °N, 147.84 °W) regularly detected elevated dust layers, which were tracked back to a series of Asian dust storms. In many cases, the dust layers sporadically developed embedded ice-crystal clouds. This phenomenon is illustrated in Fig. 1 by the backscattered laser power and linear depolarization ratio (δ, which is the ratio of the powers in the planes of polarization orthogonal and parallel to that transmitted). Figure 1 shows two hours of lidar display over a range of heights. Aerosols are particularly apparent in stratified layers at between about 4.5 and 6.5 km of height because of their enhanced backscattering (an increase in display brightness) and depolarization (values of δ, 0.1–0.2) compared with that produced by molecular scattering. Also prominent is the embedded ice cloud centred at 5.5 km at about 20:30 GMT, which displays even stronger backscattering and depolarization. This cloud had a minimum temperature of −36 °C and a maximum relative humidity (with respect to ice) of 103% (according to the bracketing local radiosonde data).

Figure 1: Polarization-lidar height against time display, showing ice-cloud formation in an elevated desert-dust aerosol layer.

Displays of backscattered laser power (top; logarithmic grey scale) and linear depolarization ratio (colour scale; δ is the ratio of the powers in the planes of polarization orthogonal and parallel to that transmitted) are given for a two-hour period on 29 February 2004, as debris from an Asian dust storm drifted over the interior of Alaska.

Similar mixed dust and ice layers were observed with cloud-top temperatures between about −15 °C and −40 °C: these are abnormally warm compared with those of cirrus clouds, which form by the homogeneous nucleation of haze particles8. The corresponding relative humidities indicate an almost ice-saturated environment, which also contrasts with the high supersaturations needed for cirrus formation9. In laboratory experiments, dust particles needed ice supersaturations of only 10–15% to nucleate ice crystals directly from the vapour10.

Although there have been previous indications that desert dusts alter clouds3,4,5,6, to my knowledge this is the first time that they have been shown to serve as deposition ice nuclei in the atmosphere and to create ice clouds at modest supersaturations and temperatures. Desert dust storms are a natural part of our world, but we must be wary in case human activities in arid lands, or climate change itself, are increasing aerosol amounts and so further altering climate by affecting distant clouds.


  1. 1

    Kaufman, Y. K., Tunré, D. & Boucher, O. Nature 419, 215–223 (2002).

    ADS  CAS  Article  Google Scholar 

  2. 2

    Sassen, K. in Lidar (ed. Weitkamp, C.) 19–42 (Springer, Berlin, 2005).

    Google Scholar 

  3. 3

    Murayama, T. N. et al. J. Geophys. Res. 106, 18345–18360 (2001).

    ADS  CAS  Article  Google Scholar 

  4. 4

    Sassen, K. Geophys. Res. Lett. 29, 1465 doi:10.1029/2001GL014051 (2002).

    ADS  Article  Google Scholar 

  5. 5

    Sassen, K., DeMott, P. J., Prospero, J. M. & Poellot, M. R. Geophys. Res. Lett. 30, 1633 doi:10.1029/2003GL017371 (2003).

    ADS  Article  Google Scholar 

  6. 6

    Sakai, T., Nagai, T., Nakazto, M. & Matsumura, T. Geophys. Res. Lett. 31, L06128 doi:10.1029/2003GL019332 (2004).

    ADS  Article  Google Scholar 

  7. 7

    Isono, K., Komabayasi, M. & Ono, A. J. Meteorol. Soc. Jpn 37, 211–233 (1959).

    Article  Google Scholar 

  8. 8

    Sassen, K. & Campbell, J. R. J. Atmos. Sci. 58, 481–496 (2001).

    ADS  Article  Google Scholar 

  9. 9

    Comstock, J. M., Ackerman, T. P. & Turner, D. D. Geophys. Res. Lett. 31, L11106 doi:10.1029/2004GL019705 (2004).

    ADS  Google Scholar 

  10. 10

    Bailey, M. & Hallett, J. Q. J. R. Meteorol. Soc. 128, 1461–1483 (2002).

    ADS  Google Scholar 

Download references

Author information



Corresponding author

Correspondence to Kenneth Sassen.

Ethics declarations

Competing interests

The author declares no competing financial interests.

Rights and permissions

Reprints and Permissions

About this article

Cite this article

Sassen, K. Dusty ice clouds over Alaska. Nature 434, 456 (2005).

Download citation

Further reading


By submitting a comment you agree to abide by our Terms and Community Guidelines. If you find something abusive or that does not comply with our terms or guidelines please flag it as inappropriate.