1) The study analyzed satellite data on changes in Earth's gravity field to measure mass changes in glaciers and ice caps globally between 2003-2010. 2) Unexpectedly, the results found the contribution of glaciers and ice caps to sea level rise during this period was less than half of previous estimates. 3) Mass loss in the Himalayan and surrounding regions was insignificant, in contrast to previous estimates that were over ten times larger.

1. RESEARCH NEWS & VIEWS
criteria on the basis of their current data.
Providing answers to all these questions
Early primary cancer will require further biological investigation
with mutations across species, as well as clinical studies. An
Post-dispersion
mutations additional challenge is posed by the fact that
Primary-tumour there are fewer than 700 cases of medullo-
genetic compartment
blastoma per year in Europe. More routine
biopsy and characterization of human meta­
stases will be essential, and the impetus and
ethical justification for such a fundamen-
Metastatic
Post-metastatic
tal change to clinical practice will, at least
Dispersion
cells
mutations in part, come from experimental studies
Metastatic genetic such as those presented here. Time will tell
compartment
whether this tale of Sleeping Beauty and mice
develops into a clinically relevant human
paradigm. ■
Figure 1 | A bi-compartmental genetic model of cancer metastasis. By analysing the tumours from
Steven C. Clifford is at the Northern Institute
two strains of mice that model the brain cancer medulloblastoma, Wu et al.1 found differences in the
DNA-sequence mutations present in primary and metastatic tumours. They propose that rare cells in the for Cancer Research, Newcastle University,
primary tumour that are capable of metastasizing disperse to other sites in the brain, where they form Newcastle upon Tyne NE1 4LP, UK.
metastases. The cells of the primary tumour and the metastases then continue to accumulate mutations, e-mail: steve.clifford@ncl.ac.uk
generating two distinct genetic compartments.
1. Wu, X. et al. Nature 482, 529–533 (2012).
5 2. Collier, L. S., Carlson, C. M., Ravimohan, S., Dupuy,
human disease exhibits more complex them can drive tumour development, which A. J. & Largaespada, D. A. Nature 436, 272–276
patterns of metastases than are observed in are therapeutically relevant, and which occur (2005).
mice, and is classified into four molecular sub- at sufficient frequency in the human disease 3. Copeland, N. G. & Jenkins, N. A. Nature Rev. Cancer
groups (WNT, SHH, Group 3 and Group 4), to warrant their pursuit as potential targets. 10, 696–706 (2010).
which each display distinct biological and The authors justifiably reason that targets 4. Goodrich, L. V., Milenković, L., Higgins, K. M. & Scott,
clinical characteristics6. The Ptch+/− mice used that are common to primary tumours and M. P. Science 277, 1109–1113 (1997).
5. Pizer, B. L. & Clifford, S. C. Br. J. Neurosurg. 23,
by Wu et al.1 develop SHH-associated medullo­ metastases, in both humans and mice, are 364–375 (2009).
blastomas 4; similar mutagenesis-driven those most attractive for further develop- 6. Taylor, M. D. et al. Acta Neuropathol. http://dx.doi.
approaches using existing mouse models ment. However, only one cellular pathway, org/10.1007/s00401-011-0922-z (2011).
of other medulloblastoma disease groups, such insulin-dependent signalling, meets these 7. Gibson, P. et al. Nature 468, 1095–1099 (2010).
as WNT7, might prove informative.
Perhaps the most urgent question arising
from this study1 is whether the genetic differ- C L I M ATE C HA NG E
ences between the two disease compartments
lead to distinct biological features that make
them respond differently to treatment. In
mice, these compartments remain genomically
Shrinking glaciers
under scrutiny
characterized entities, the biological and ther-
apeutic importance of which is untested. In
humans, clinical-trial data show5 that primary
and metastatic sites respond similarly to cur-
rent therapies (with cure achieved at both sites) Melting glaciers contribute to sea-level rise, but measuring their mass loss over
in around 60% of children with metastatic dis- time is difficult. An analysis of satellite data on Earth’s changing gravity field
ease, but a more objective assessment of treat- does just that, and delivers some unexpected results. See Letter p .514
ment response is confounded by the fact that
primary tumours are mostly removed by sur-
gery prior to treatment. Wu et al. provide initial J O N AT H A N B A M B E R extending beyond 30 years. Extrapolating this
evidence that the tumour compartments may tiny sample of observations to all glaciers and
G
respond differently to current therapies in cer- laciers and ice caps are pivotal features ice caps is a challenging task that inevitably
tain patients, but they rightly caution that these of both water resources and tourism. leads to large uncertainties.
effects could also relate to clinical factors such They are also a significant contribu- On page 514 of this issue, Jacob and col-
as radiotherapy being delivered at different tor to sea-level rise. About 1.4 billion people leagues3 describe a study based on satellite data
intensities to different tumour sites. are dependent on the rivers that flow from the for Earth’s changing gravity field that tackles
Some of the mutations identified by Wu Tibetan plateau and Himalayas1. Yet significant this problem*. Their results have surprising
and colleagues’ experiments may also reveal controversy2 and uncertainty surround the implications for both the global contribu-
biological processes or pathways that could recent past and future behaviour of glaciers in tion of glaciers to sea level and the changes
offer drug targets for the improved treat- this region. This is not so surprising when one occurring in the mountain regions of Asia.
ment of primary tumours, metastases, or considers the problem in hand. There are more Melting glaciers are an iconic symbol of
both. The new mouse strains provide excel- than 160,000 glaciers and ice caps worldwide. climate change. On the basis of the limited
lent models in which to test this possibil- Fewer than 120 (0.075%) have had their mass data mentioned above, they seem to have
ity. The multitude and variety of mutations balance (the sum of the annual mass gains and been receding, largely uninterrupted, almost
described by Wu et al.1 are noteworthy, but losses of the glacier or ice cap) directly meas- *This article and the paper 3 under discussion were
the next challenge is to determine which of ured, and for only 37 of these are there records published online on 8 February 2012.
4 8 2 | N AT U R E | VO L 4 8 2 | 2 3 F E B R UA RY 2 0 1 2
© 2012 Macmillan Publishers Limited. All rights reserved

2. NEWS & VIEWS RESEARCH
everywhere around the world for several dec-
J. BALOG/EXTREME ICE SURVEY
ades4. Scaling up the small sample of ground-
based observations to produce global estimates
is, however, fraught with difficulty. Size, local
topography, altitude range, aspect and micro-
climate all affect the response of individual
glaciers in complex ways. Even the seasonality
of changes in temperature and precipitation
strongly influence the glaciers’ response, and
those that terminate in a lake or ocean behave
differently again.
Nonetheless, until recently there was little
alternative to some form of extrapolation of
the terrestrial observations to large regions
and numbers of glaciers. One such high-
profile assessment5 concluded that, during the
period 1996–2006, the mass loss from glaciers
and ice caps (GICs) increased steadily, contrib-
uting a sea-level rise of 1.1 ± 0.24 milli­metres
per year by 2006. In this study5, the authors
concluded that GICs had been the domi-
nant mass contributor to sea-level rise over
the study period, and they extrapolated their
results forward to argue that this would also be
the case in the future.
Then along came the Gravity Recovery Figure 1 | The Leschaux and Talèfre glaciers in the French Alps. The photograph highlights
and Climate Experiment (GRACE), which the complex and intricate topographic setting of these mountain glaciers and the difficulty in
consists of a pair of satellites that have been extrapolating observations from one glacier to others. Jacob and colleagues3 avoided these difficulties
making global observations of changes by using the area-integrated signal from satellite gravity data.
in Earth’s gravity field since their launch
in 2002. They have been used in various GICs) to sea-level rise was less than half the observation period. Some of the regions,
studies to examine the changing mass of the value of the most recent, comprehensive esti- such as the Gulf of Alaska, experience large
great ice sheets of Antarctica and Greenland6 mate8 obtained from extrapolation of in situ inter-annual variations in mass balance that
and several other large glaciated regions 7. measurements for 2001–05 (0.41 ± 0.08 are mainly due to variability in precipitation7.
But, so far, the data have not been analysed compared with 1.1 mm yr−1). Second, losses This is also true for the High Mountain Asia
simultaneously and consistently for all areas. for the High Mountain Asia region — com- region3, and, as a consequence, a different
The difficulty with doing this is that GRACE prising the Himalayas, Karakoram, Tianshan, measurement period could significantly alter
measures the gravity field of the complete Pamirs and Tibet — were insignificant. Here, the estimated trend for this sector. Further-
Earth system. This includes mass exchange the mass-loss rate was just 4 ± 20 gigatonnes more, some areas, such as the European Alps
and/or mass redistribution in the oceans, per year (corresponding to 0.01 mm yr−1 of sea- and Scandinavia, have been relatively well
atmosphere, solid Earth and land hydrology, level rise), compared with previous estimates monitored, and thus constrained, using other
in addition to any changes in GIC volume. To that were well over ten times larger. By a care- approaches. Nonetheless, Jacob and colleagues
determine the latter, it is clearly essential to be ful analysis, the authors discounted a possible have dramatically altered our understanding of
able to separate it from the other sources of tectonic origin for the huge discrepancy, and recent global GIC volume changes and their
mass movement that affect the gravity field. A it seems that this region is more stable than contribution to sea-level rise. Now we need to
second, related issue is the effective resolution previously believed. work out what this means for estimating their
of the observations. The GRACE satellites are What is the significance of these results3? future response. ■
sensitive to changes in the gravity field over Understanding, and closing, the sea-level
distances of a few hundred kilometres. They budget (the relative contributions of mass Jonathan Bamber is at the Bristol Glaciology
cannot ‘see’ the difference between the signal and thermal expansion to ocean-volume Centre, School of Geographical Sciences,
from one glacier or small ice cap and another. change) is crucial for testing predictions of University of Bristol, Bristol BS8 1SS, UK.
To isolate the GIC signal from others at the future sea-level rise. Estimates of the future e-mail: j.bamber@bristol.ac.uk
surface, Jacob and colleagues defined units of response of GICs to climate change are, in
1. Immerzeel, W. W., van Beek, L. P. H. & Bierkens,
mass change — called mass concentrations, general, based on what we know about how M. F. P. Science 328, 1382–1385 (2010).
or mascons — within each of their 18 GIC they have responded in the past. A better esti- 2. Cogley, J. G., Kargel, J. S., Kaser, G. & van der Veen,
regions (including the European Alps; Fig. 1). mate of past behaviour, such as that obtained C. J. Science 327, 522 (2010).
3. Jacob, T., Wahr, J., Pfeffer, W. T. & Swenson, S.
Each region might have many tens of mascons by Jacob and colleagues, will therefore result Nature 482, 514–518 (2012).
defining the geographic extent of significant in better estimates of future behaviour. 4. Kaser, G., Cogley, J. G., Dyurgerov, M. B., Meier, M. F.
ice volume within the sector3. Combined with Discussion of the demise of the Himalayan & Ohmura, A. Geophys. Res. Lett. 33, L19501
global models of land hydrology and atmos- glaciers has been mired in controversy, partly (2006).
5. Meier, M. F. et al. Science 317, 1064–1067
pheric-moisture content, the authors were able because of basic errors2, but also because (2007).
to isolate the GIC mass trends over the eight- of the dearth of reliable data on past trends. 6. Velicogna, I. Geophys. Res. Lett. 36, L19503
year (2003–10) period of the observations. Given their role as a water supply for so many (2009).
7. Luthcke, S. B., Arendt, A. A., Rowlands, D. D.,
What they found was unexpected. people1, this has been a cause for concern and an McCarthy, J. J. & Larsen, C. F. J. Glaciol. 54,
First, the contribution of GICs (excluding outstanding issue. 767–777 (2008).
the Antarctica and Greenland peripheral Of course, eight years is a relatively short 8. Cogley, J. G. Ann. Glaciol. 50, 96–100 (2009).
2 3 F E B R UA RY 2 0 1 2 | VO L 4 8 2 | N AT U R E | 4 8 3
© 2012 Macmillan Publishers Limited. All rights reserved