news & views
and reformed several times since its
initiation. Between 44 and 36 million
years ago, ice cover fluctuated between
seasonal and perennial, indicating climate
instability during this time. Intriguingly,
seasonal or perennial sea ice may have
persisted through at least part of the Middle
Eocene Climatic Optimum 40 million
years ago. This was a warming event that
lasted approximately 500,000 years and was
characterized by an increase in atmospheric
CO2 concentrations, a gradual increase
in ocean temperatures by about 5° C, and
rapid biotic change8. Results from the Arctic
are lacking, but would significantly aid in
understanding the causes and effects of this
enigmatic warming event. The perennial
ice cover finally became a stable feature
about 36 million years ago. Uncertainties
in the age model for the core9,10 add
some ambiguity to the exact timing,
but the switch seems to have pre-dated
the glaciation of Antarctica 34 million
years ago11.
Most records of sea ice cover from the
Antarctic region are limited to the past
few million years, but the available data
suggest that a stable Antarctic sea ice
regime probably did not develop until at
least early Oligocene times, as a response
to the growth of grounded ice sheets12.
Marine sedimentary evidence indicates
that this early sea ice was dominantly
seasonal13, with uncertain coverage of
perennial sea ice. Arguably, perennial
sea ice was not a dominant feature of the
circum-Antarctic ocean before 23 million
years ago. It would therefore seem that
the initiation of perennial sea ice cover
in the Arctic predated both seasonal
Antarctic sea ice cover and significant polar
glaciation in either hemisphere11. This
timing is significant to our understanding
of cryospheric evolution and bipolar
linkages, pointing to the Arctic as a
potentially important driver of global
cooling, probably through a long-term
ice-albedo effect.
Darby 3 has shown that following
an initial establishment of year-round
Arctic sea ice cover 44 million years
ago, subsequent intervals experienced
ice-free summers. Exploring the transition
between perennial and seasonal sea ice
may help us to understand what we can
expect when the Arctic becomes ice-free in
future summers.
❐
Catherine E. Stickley is co-founder of the
Biostratigraphic Consultancy Company Evolution
Applied Ltd., UK.
e-mail: catherine.stickley@gmail.com
References
St. John, K. Paleoceanography 23, PA1S05 (2008).
Stickley, C. E. et al. Nature 460, 376–380, (2009).
Darby, D. Nature Geosci. 7, 210–213 (2014).
Zachos, J. C., Dickens, G. R. & Zeebe, R. E. Nature
451, 279–283 (2008).
5. Sexton, P. et al. Nature 471, 349–352 (2011).
6. Darby, D. A. Paleoceanography 23, PA1S07 (2008).
7. Krylov, A. A. et al. Paleoceanography 23, PA1S06 (2008).
8. Sluijs, A. et al. Nature Geosci. 6, 429–434 (2013).
9. Backman, J. et al. Paleoceanography 23, PA1S07 (2008).
10.Poirier, A. & Hillaire-Marcel, C. Geophy. Res. Lett.
38, L14607 (2011).
11.DeConto, R. et al. Nature 455, 652–656 (2008).
12.DeConto, R., Pollard, D. & Harwood, D. Paleoceanography
22, PA3214 (2007).
13.Houben, A. J. P. et al. Science 340, 341–344 (2013).
1.
2.
3.
4.
Published online: 26 January 2014
CLIMATE CHANGE
Impacts in the third dimension
Despite reports of no trends in snow- and rainfall, rivers in the northwest USA have run lower and lower in recent
decades. A closer look at high- and low-altitude precipitation suggests that observational networks have missed a
decline in mountain rain and snow that can explain the discrepancy.
Michael Dettinger
M
ountain catchments are key
sources for river water and water
supply. In the western US, for
example, they contribute as much as
75% of surface-water and groundwater
supplies. High-altitude catchments play
this crucial role as they tend to receive
more precipitation than surrounding
lowlands. And mountain precipitation
often forms thick snowpacks, which hold
water in deep freeze until the following
warm season. As the weather warms, just
in time to meet high water demand, melt
water feeds the streams. When a longterm decline in streamflow trends was
diagnosed in the Pacific Northwest 1,2 in the
late 2000s despite widespread reports that
precipitation showed no corresponding
trend3,4, the discrepancy was cause for
concern. Writing in Science, Luce and
colleagues5 suggest that a decline in highmountain precipitation resulted from a
166
climate-change-induced slow-down in
the westerly winds that normally bring
mountain rain and snow.
Mountains receive more precipitation
than surrounding lowlands, because
winds and storms that encounter high
topography are forced up in altitude so that
they can pass over them. On the way up,
they cool, and copious amounts of water
vapour condense and fall as precipitation
on and around the mountains. This
effect, generally referred to as orographic
enhancement of precipitation, is more or
less additional to any precipitation that
the storms would generate even in the
absence of mountains. This orographic
enhancement is approximately 6,7 the
product of three factors: horizontal wind
speed, the amount of water vapour in
the air and the slope of the mountain
faces (which determines how quickly the
air is lifted).
Luce and colleagues5 find that highaltitude winds over the Pacific Northwest
have declined in the past few decades,
to a degree that would be expected
to reduce the amount of orographic
precipitation in the mountain ranges
there. They note that the winds have
been slowing down just enough to cause
rates of precipitation decline that would
in turn explain the observed decline in
streamflows. Previously, the most common
explanations for the differing precipitation
and streamflow trends were possible
unaccounted-for impacts of land and water
use, or potential increases in evaporation
or sublimation rates in response to rising
temperatures so that less water reaches
the region’s streams and rivers2,8. The
observation of Luce and colleagues offers
a novel, and very plausible, alternative
explanation. Based on radiation-balance
considerations, Luce and colleagues also
NATURE GEOSCIENCE | VOL 7 | MARCH 2014 | www.nature.com/naturegeoscience
© 2014 Macmillan Publishers Limited. All rights reserved
news & views
0.9
° 90
0.5
0°
Correlation between trends in Nov–Mar 700 mb winds
and a linearly increasing number sequence (year number)
0.7
0.3
0.1
–0.1
–0.3
–0.5
° 180
argue that the effects of warming alone
would probably not have been large
enough to explain observed streamflow
declines in the Pacific Northwest,
although this argument does not seem to
exhaust all options for warming-induced
flow reductions.
If Luce and colleagues are correct, then
important (15%) declines in orographic
precipitation must have been missed by
the observational network. This is a very
real possibility. In the western US, as in
mountain regions around the globe, harsh
weather conditions, complex topography
and long distances from larger human
settlements have historically limited the
numbers and often the quality of highaltitude weather stations. As a result,
long-term declines in precipitation at high
altitudes in the Pacific Northwest could
well have gone unnoticed — ironically,
precisely in the areas that provide most of
the water.
The complexity of interactions between
large-scale wind fields, precipitation
patterns and river-based water supplies
in the Pacific Northwest should stand as a
stark warning for designers and operators
of climate networks in mountainous
regions. Declines in mountain precipitation
in response to slackening winds, if
widespread, could threaten the sources of
many river systems globally, and potentially
could change the timing and character of
many downstream flows. Future flooding,
drought, wildfire, ecosystems and other
landscape features might be affected by a
transition from high-altitude snow to lowaltitude rain in ways that would challenge
long-established ecosystems and water
management systems. In this context,
expansion of mountain-based networks
of observations of surface weather and
streamflow, as well as conditions in the
atmosphere above, will be essential for
resource and hazard managers striving to
accommodate uncertain future climate
changes and impacts.
Luce and colleagues take the argument
further by suggesting that the slowing
winds reflect large-scale atmosphericcirculation changes associated with
shifting or slackening mid-latitude zones
of westerly winds and storm tracks.
They hypothesize that these storm-track
changes might be related to differences
between tropical versus polar warming
trends, which might be associated directly
with anthropogenic global warming. But,
when mapped at a hemispheric scale, the
particular patch of declining winds that
Luce and his colleagues studied (Fig. 1)
can be seen to be localized near the Pacific
Northwest. Hemispherically, the trends in
–0.7
0°
27
–0.9
Figure 1 | Changing winds. Between 1950 and 2012, zonal winds at altitudes of about 3 kilometres above
sea level (at atmospheric pressures of 700 mb) have slackened over some mid-latitude land areas (red
shades). Luce and colleagues5 suggest that the slowing winds have led to significantly reduced highaltitude precipitation in the Pacific Northwest, and perhaps elsewhere. Such a reduction could result from
a reduction in the orographic enhancement of precipitation by weakening high-altitude winds, a change
that has not been picked up by the sparse networks of mountain observations.
westerly wind speeds are seen to be quite
complex and variable, and, if anything,
might be more focused (and explained) by
changing continental climate gradients.
Thus, it is not yet clear how much the
decline in high-altitude precipitation
change in the Pacific Northwest can
be generalized.
The kinds of changes reported by Luce
and colleagues, when viewed at global
scales, are complex and uncertain. Global
climate models of the current generation
smooth over mountain topographies to
a great extent; for example, the highest
altitudes in the western US in the majority
of global models are less than half the
height of the highest mountain ranges
there. Thus, orographic enhancements are
dampened considerably in the models,
and consequently the models are unlikely
to faithfully capture the kinds of changes
in precipitation in mountainous regions
reported by Luce and his colleagues.
Luce and colleagues5 identify
potentially very important changes in
high-altitude precipitation and wind
patterns that are already underway in at
least one corner of the world, the Pacific
Northwest. These changes have mostly
been missed by current observational
networks and climate simulations, and
NATURE GEOSCIENCE | VOL 7 | MARCH 2014 | www.nature.com/naturegeoscience
© 2014 Macmillan Publishers Limited. All rights reserved
show no immediate signs of being simple
or uniform at the largest scales. In this
context, more attention to the crucial third
dimension — high altitudes — will be
needed in our climate models, and even
more so in our observational networks, if
we are going to recognize and prepare for
long-term climate changes in mountainous
regions. In the end, we cannot adapt to
something we cannot observe or predict.❐
Michael Dettinger is at the US Geological Survey
and Center for Western Weather and Water
Extremes, Scripps Institution of Oceanography,
La Jolla, California 92093-0224, USA.
e-mail: mddettin@usgs.gov
References
1. Luce, C. H. & Holden, Z. A. Geophys. Res. Lett. 36, L16401 (2009).
2. Fu, G., Barber, M. & Chen, S. Hydrol. Proc. 24, 866–878 (2010).
3. Mote, P. W., Hamlet, A. F., Clark, M. P. & Lettenmaier, D. P.
Bull. Am. Meteorol. Soc. 86, 39–49 (2005).
4. Barnett, T. P. et al. Science 319, 1080–1083 (2008).
5. Luce, C. H., Abatzoglou, J. T. & Holden, Z. A. Science
342, 1360–1364 (2013).
6. Rhea, J. O. & Grant, L. O in Advanced Concepts and Techniques in
the Study of Snow and Ice Resources: Interdisciplinary Symposium
December 1973, Proceedings (eds Santeford, H. S. & Smith, J. L.)
182–192 (1974).
7. Neiman, P. J., White, A. B., Ralph, F. M., Gottas, D. J. &
Gutman, S. I. Water Management 162, 83–94 (2009).
8. Das, T., Pierce, D. W., Cayan, D. R., Vano, J. A. &
Lettenmaier, D. P. Geophys. Res. Lett. 38, L23403 (2011).
Published online: 9 February 2014
167