throughout the atmosphere is mineral dust
from the world’s deserts, which inject thou-
sands of teragrams (1012 g) of dust into the
atmosphere every year.
Scientists long assumed that clay minerals
were the ice-nucleating component of desert
dust. However, recent studies have shown
that the alkali feldspar component of desert
dust is exceptionally efficient at nucleating
ice (1, 8–10). This discovery has helped to bet-
ter model the global distribution of desert-
dust ice-nucleating particles (11). But it has
remained unclear why certain feldspars are
so good at nucleating ice.
Kiselev et al. now report a crucial step
toward addressing this question. Using an
electron microscope, they recorded movies
of ice crystals growing on feldspar surfaces.
They made the striking observation that
most ice crystals grew with the same crys-
tallographic orientation (see the image).
This observation implies that the feldspar
surface properties direct the nucleation and
growth of the crystals.
Using a combination of computer simulations and analysis of the directional growth
of crystals on feldspar surfaces, the authors
conclude that nucleation occurs on a particular crystallographic face of feldspar,
denoted as the (100) plane. The (100) face
is not typically present on feldspar crystals
because it is unstable. However, Kiselev et
al. suggest that it may be exposed at steps,
cracks, and cavities.
Although the work is a substantial step for-
ward, open questions remain. For example,
the study is directly relevant for clouds where
ice nucleates from water vapor (e.g., cirrus),
but not for those in which dust is immersed
in supercooled water droplets (e.g., stratus).
It remains to be shown whether the same
sites that are responsible for nucleation from
liquid water are also responsible for nucle-
ation from the vapor phase.
Kiselev et al.’s study also advances the
field of heterogeneous nucleation in general. There are very few systems where we
can understand and rationalize why a material is good or bad at nucleating another
material. Feldspars are a family of minerals that have similar crystal structures and
compositions, but that nucleate ice with
very different efficiencies (10, 12). They thus
provide a means of systematically varying material properties and examining the
impact of these changes on ice nucleation.
This offers the possibility of deepening our
understanding of why certain feldspars are
so good at nucleating ice. It may also enable the identification of other materials
that nucleate ice effectively. This opens the
door to the design and manufacture of ice-nucleating materials for a range of applications, including artificial cloud seeding and
1. J. D. Atkinson et al. , Nature 498, 355 (2013).
2. A. Kiselev et al. , Science 355, 367 (2017).
3. B. J. Murray et al., Nat. Geosci .3, 233 (2010).
4. R.J.Herbert et al., Geophys. Res. Lett. 42,1599(2015).
5. A. J. Heymsfield et al., Science 333, 77 (2011).
6. I. Tan, T. Storelvmo, M. D. Zelinka, Science 352, 224 (2016).
7. B. J. Murray et al., Chem. Soc. Rev. 41, 6519 (2012).
8. J.D. Yakobi-Hancock, L.A.Ladino,J.P.D.Abbatt, Atmos.
Chem. Phys. 13, 11175 (2013).
9. D. Niedermeier et al. , J. Geophys. Res.120, 5036 (2015).
10. A. Peckhaus et al., Atmos. Chem. Phys. 16, 11477 (2016).
11. J.Vergara-Temprado et al., Atmos. Chem. Phys. Discuss.
12. A. D. Harrison et al. , Atmos. Chem. Phys.16, 10927 (2016).
Upper tropospheric ice clouds
WI THOU T ICE-NUCLEATING
take up water
to make a dense,
A fe w ice
form on INPs
and grow at the
a cloud with a
identify functional trait
variation in tropical forests
By Valerie Kapos
Recognizing the importance of biodi- versity to human well-being, most nations have committed to the Con- vention on Biological Diversity’s Aichi Biodiversity Targets and the Sustainable Development Goals.
However, the pressures on biodiversity
are increasing, and its status is declining
globally (1), raising concerns that national
plans and targets are not ambitious enough
(2) and showing that new solutions are
needed (3). Recognition of synergies among
different targets and goals (4) has brought
forest to the forefront of national land-use
decision-making, which must balance multiple objectives that all demand land (5). Efforts to support decision-making on forests
have focused on individual (typically vertebrate) species and on carbon and other
ecosystem services. Highly resolved views
of functional trait variation in tropical forests reported by Asner et al. on page 385 of
this issue (6) may provide a further basis
for making such decisions.
Planning for national action on REDD+
(Reducing Emissions from Deforestation and
Forest Degradation plus conservation, management and enhancement of forest carbon
stocks) to mitigate climate change has been
a major impetus for developing approaches
to identify where and how forests can best
be preserved to achieve the greatest overall
benefit (5, 7). Modern conservation planning
approaches that take account of potential
ecosystem service delivery, costs, and factors
affecting the feasibility of conservation interventions (8, 9) are equally relevant to other
land-use decision-making. However, major
challenges remain for ensuring that science
can provide robust support for decisions that
achieve multiple objectives through preserving forests and managing landscapes.
UNEP World Conservation Monitoring Centre,
219 Huntingdon Road, Cambridge CB3 0DL, UK.
Ice nucleation on feldspar
As Kiselev et al. show, ice nucleates and
grows on (100) faces of feldspar, a mineral
that is found in desert dust. The results help
to explain how ice nucleates on desert dust
particles in the atmosphere.