A measurement by satellite altimetry shows the Himalayan glaciers to be losing mass at only moderate rates, but raises broader questions about other methods for estimating mass balance. See Letter p.495
The Himalayas continue to be interesting to glaciologists, which is not surprising when one considers the formidable water-resource problems in the glaciers' regional context and the controversy1 over the rate at which the ice is disappearing. The claim that the glaciers are vanishing rapidly is not difficult to discredit. In the Karakoram, at the western end of the Himalayan arc, glacier mass balance — the change of glacier mass over a given time span — has, in fact, been slightly positive over the past decade2. Kääb et al.3 (page 495 of this issue) clarify this matter further by providing an independent measurement of glacier mass balance based on elevation changes detected by the ICESat satellite's GLAS altimeter. Although these data were collected over only five years (2003–08), they cover the entirety of the Himalayas.
At −210±50 kilograms per square metre per year, the new estimate of glacier mass balance in the Himalayas is more negative than a recent gravimetric estimate4, which is based on satellite measurements of Earth's changing gravitational field. But it is less negative than the current estimate5,6 obtained by the more established method of interpolation from sparse field measurements. The new estimate may help to allay panic about regional water resources, but as far as glaciological methodology is concerned it will set the cat among the pigeons.
Glaciers are often likened to water towers. A sustained negative mass balance means that there is progressively less water in the 'tower', but more in transit to the ocean as meltwater. Because the latter is available for exploitation downstream, this equates to short-term gain in a region with water deficits, but big problems later when the water tower runs dry. If, as now, someone presents a less-negative estimate, our assessment changes. The tower is delivering less water for current use, but will hold water for longer. It is not clear whether such a change is bad news or good. Certainly, if the change were in the other direction, to a more-negative balance, questions would arise about equity between successive generations of water users and about possible ways of keeping the water in the tower7.
Kääb and colleagues' methods raise the bar for future altimetric studies. The Supplementary Information to their paper3 comprises 10,000 well-chosen words about the processing and correction of the raw data. This provides a striking illustration of the time it can take to learn to put new instruments to work effectively. ICESat was launched in 2003, and better information is still being squeezed out of it nearly three years after the last of its three lasers failed in October 2009.
This information includes better-quantified errors than those that accompany the recent gravimetric estimate4 from the GRACE satellite mission. In glaciological interpretation of GRACE data, it is customary to correct for non-glacial mass transfers — mainly, of surface water and groundwater, and of rock in Earth's deep interior — by hydrological and geophysical modelling. The outputs of the models are the largest sources of gravimetric uncertainty. One of Kääb and colleagues' more provocative points is that it could be advantageous to reverse this procedure. If a really good measurement of the glacier mass change is available, as is now true of the Himalayas, why not use it to correct the gravimetric data, which would then contain only the hydrological and deep-interior signals?
The most spirited reactions to this study will probably come not from GRACE specialists but from practitioners of more traditional methods. In the Himalayas, it typically takes five days to walk from the road head to the glacier terminus, and the glacier itself, if it is accessible at all, is extremely difficult to work on (Fig. 1). The fieldwork has nevertheless yielded a small but precious body of facts about Himalayan mass balance. However, the arithmetic average5 of the field measurements obtained over the same time span as Kääb and colleagues' estimate is, at −746 kg m−2 yr−1, more than three times as negative.
Why are the field measurements seemingly so unrepresentative? The usual answers appeal to several factors: to the smallness of the glaciers chosen for fieldwork, although nobody has explained why smaller glaciers should lose mass faster than larger ones; to the accessibility of the chosen glaciers, which probably implies lower altitude and therefore more intense melting, although this implication has yet to be investigated systematically; and to unquantified uncertainties in the field measurements themselves. The last objection carries less force now, because recent work tends to address the errors more explicitly, but local factors could play a part. For example, mass gain by snow avalanching (Fig. 1) is practically unmeasurable in the field.
Some other regional-scale studies8,9 have reported less-negative mass balances than those obtained either by fieldwork or by interpolation from fieldwork. Kääb and colleagues' results therefore add to accumulating evidence that increases the number of questions, rather than the number of answers, about the measurement of glacier mass balance. These questions could be of worldwide import.
The new measurement3 agrees with a recent review of the state of Tibetan glaciers10, which noted a trend towards smaller rates of mass loss farther north and presented the first field report of mass gain, measured on a glacier in the Pamir mountains. Kääb and colleagues also find positive mass balance, just to the south of the Pamir. Thus, three independent sources2,3,10 now confirm this 'Karakoram anomaly', with Kääb et al. offering the most geographically extensive picture.
Kääb and colleagues have shown that the state5 of Himalayan glaciers is not as dire as was widely and wrongly assumed until recently, but their short-term observations tell us nothing about the fate5 of Himalayan glaciers. The observations were made by a now-defunct satellite, the successor of which will not be launched until 2016, and they pose searching questions about GRACE results and about longer-established methods. Monitoring the Himalayan glaciers is clearly going to be a continuing practical and interpretative challenge.