Good science, like life, is nothing if not a learning process. This is especially the case in a contested and controversial scientific discipline like climate change. Early predictions are often wrong, as one should of course expect them to be. Sometimes they are wildly conservative; other times predictions later prove to have been too alarmist.
An example of the former might be the Arctic sea ice, which is vanishing at a faster rate in the real world than can be captured in the climate models and now seems likely to disappear much earlier than has been forecast by scientists. An example of the latter might be the reassessment of doomsday pronouncements of the impact disappearing glaciers will have on populations dependent on downstream rivers for their water supplies. This is worth a detailed examination, not least because my own earlier predictions now need updating.
In 2002 I visited Peru’s Cordillera Blanca, and wrote a report which appeared in my first book High Tide: News from a Warming World two years later. In it I reported meeting glaciologists Georg Kaser and Bryan Mark who were, amongst others, attending a meeting on glaciers and water supplies in the Peruvian highland city Huaraz, on the Rio Santa. In the book I quoted Kaser as saying:
“During the dry season the runoff is almost exclusively due to the glacier melt… [water which] is of essential importance for the highly populated and cultivated valley… particularly during the dery season.”
In my own words, I then continued:
The situation is the same for all of Peru’s irrigated coastal valleys: no glaciers will mean no runoff in the dry season.
In particular, with regard to the Peruvian capital Lima (which depends for water supplies on the mountain-originated River Rimac), I wrote:
Glaciers in the Cordillera Central [origin of the Rimac]… will disappear altogether in just a few decades unless global temperatures stop rising. Consequently, the Rimac River, which through the late twentieth century has been temporarily charged with additional meltwater from the rapidly-retreating ice fields, will suddenly – and disastrously – dry up for half the year… With no water supply for six months every year, life will quickly become impossible. Where will its residents go?
I then expanded the analysis more widely:
Warnings have also been sounded in Asia, where half a billion people in the Indian subcontinent who depend on rivers flowing down from the Himalayas could be left without water – within this century – if the Himalayan glaciers melt. [snip] As with Lima, if the worst happens and many of the crucial Himalayan glaciers disappear, hundreds of millions of people will be faced with moving or dying of thirst. The scale of this threat is so colossal that it almost defies comprehension.
It now seems clear that in hindsight this analysis, if not flat wrong, is certainly over-pessimistic. At the time there was very little scientific work that had been conducted on precisely how much glacial contribution to water supplies comprised downstream river flows in both Asia and Peru. That has now changed, thanks to some important work by the glaciologists I met in Huaraz, and others in Asia.
Let’s first of all go back to Peru. To start with, it is important to point out that no-one disputes the rapid rates of glacial retreat both in the Andes and other mountain ranges – the issue here is not global warming per se, but the impacts on and vulnerability of human populations to the changes brought about by warming.
Second, according to further work done by Bryan Mark and colleagues (published here in PDF in 2010 in the Annals of the Association of American Geographers) the total runoff from the Rio Santa (Huaraz) watershed I visited back in 2002 is much less impacted by glacial runoff than I and they originally thought. In particular, the trend over recent decades has been downward (17% from 1954-1997) – the opposite of what you would expect if the river was flushed with significant amounts of extra meltwater from retreating glaciers. (Those watersheds with a very high proportion of glacial area have seen increases – but overall the impact is minor.)
The third conclusion is that discharge from groundwater reserves (see also this study [PDF] also co-authored by Bryan Mark) is much more important than previously assumed – and possibly more important than glaciers. The record shows great variability between years, but what it reveals is that precipitation (snow, hail and rain) is the most significant control governing groundwater discharge in the following months and years. This in turn means that the Rio Santa valley – which depends on freshwater both for crop-growing and hydroelectricity – may be much more vulnerable to multi-year droughts than to straightforward glacial disappearance, though the latter may reinforce the lack of water in dry years.
As Bryan Mark and colleagues conclude:
Our results demonstrate that groundwater is proportionately at least as important as glacier melt with respect to total current dry season streamﬂows, and as glaciers recede, the inﬂuence of groundwater, and its role as a seasonal buffer, will become increasingly important.
What of Lima? Will its seven million inhabitants have to move as water supplies vanish as I predicted nearly a decade ago? The short answer is ‘no’. There is a dearth of science on the subject, but one World Bank hydrological modelling study does have some interesting conclusions. Firstly, Lima is complicated because there are more than a dozen reservoirs already built, and more planned – plus a water diversion brings additional supplies over from the Amazon side of the continental divide.
Second, most of the glaciers have already gone, reducing from 113 km2 in 1970 to about 40 km2 in 1997. The remaining glaciers are much smaller than those in the Rio Santa catchment, and therefore – according to the report – their contribution to runoff is “not significant”. Moreover, the models see some increase in annual precipitation in future decades as the climate warms, consistent with general predictions for the deep tropics. The overall conclusion is that though the highly glaciated Rio Santa may lose 20% or so of its mean flow around mid-century, for the Rimac which supplies Lima
…projected conditions indicate no significant changes, and at the most suggest the possibility for a slight reduction during the dry season.
Elsewhere the report even suggests increases in water availability:
…water availability [in the 2050s] is expected to increase, particularly with higher discharges in February and March, while there will be little or no reduction in the dry months from June to November. More water is welcome, especially for hydropower production, agriculture and water supply to Lima.
Of course this is just one model, but either way its results hardly accord with the disaster scenario I portrayed in High Tide. This is of course great news, and important for water managers – understanding the projected changes better is critical for planning infrastructure, such as further reservoirs, agricultural schemes and so on which may be needed for climate adaptation and economic development.
In the general Andean context glaciers are relatively more important than elsewhere because they buffer dry season flows. This is not the case for the Himalayas – an important difference which I failed to spot when I wrote High Tide. This is because in the Indian subcontinent the period of the strongest melt – the summer – coincides with the period of highest precipitation, the monsoon. (The reason from the climatic difference is because Peru is tropical, and has little seasonal temperature change.) On the other hand, during the wintertime dry season, glaciers will remain largely frozen and therefore contribute little to river discharge.
As recently as 2005 however, a group of scientists led by Tim Barnett from the Scripps Institute of Oceanography wrote the following in Nature (sub req’d), under the title ‘Potential impacts of a warming climate on water availability in snow-dominated regions’:
Perhaps the most critical region in which vanishing glaciers will negatively affect water supply in the next few decades will be China and parts of Asia, including India (together forming the Himalaya–Hindu Kush (HKH) region), because of the region’s huge population (about 50–60% of the world’s population). The ice mass over this mountainous region is the third-largest on earth, after the Arctic/Greenland and Antarctic regions.
The hydrological cycle of the region is complicated by the Asian monsoon, but there is little doubt that melting glaciers provide a key source of water for the region in the summer months: as much as 70% of the summer flow in the Ganges and 50–60% of the flow in other major rivers. In China, 23% of the population lives in the western regions, where glacial melt provides the principal dry season water source. [my emphasis]
Now it seems this is quite incorrect – the glaciologist Georg Kaser, who I met in Peru back in 2002, just last year published a landmark paper in PNAS which stated in the summary that:
the seasonally delayed glacier contribution is largest where rivers enter seasonally arid regions and negligible in the lowlands of river basins governed by monsoon climates [my emphasis]
That would seem to include the Ganges and the Brahmaputra, for instance, as well as the Yangtze and Yellow rivers in China. In another 2010 paper, this one published in Science (sub req’d), Immerzeel et al find that
meltwater is extremely important in the Indus basin and important for the Brahmaputra basin, but plays only a modest role for the Ganges, Yangtze, and Yellow rivers.
This ‘modest role’ is quantified at 10% for the Ganges, and 8% for the Yangtze and Yellow rivers – though that includes snowmelt as well; glacial melt alone is only a fraction of that figure. For the Indus however, because the downstream basin is so arid, the impact is much more important. The upshot then is
that Asia’s water towers are threatened by climate change, but that the effects of climate change on water availability and food security in Asia differ substantially among basins and cannot be generalized. The effects in the Indus and Brahmaputra basins are likely to be severe owing to the large population and the high dependence on irrigated agriculture and meltwater.
Another quantified contribution comes from a paper by Donald Alford and Richard Amstrong, also published in 2010, which estimates that
the glacierized catchments of the Nepal Himalaya represents approximately 4% of the total annual stream ﬂow volume of the rivers of Nepal, and thus, is a minor component of the annual ﬂow of the Ganges River. [snip] The models developed for this study indicate that neither stream ﬂow timing nor volume of the rivers ﬂowing into the Ganges Basin from Nepal will be aﬀected materially by a continued retreat of the glaciers of the Nepal Himalaya.
However, as documented by the climate reporter Lisa Friedman in an excellent recent news piece, this message of new data and increased complexity is only just beginning to get through. Take this recent story, just published by UPI:
Global warming is causing glaciers in southwest China, the major source of the country’s largest rivers, to melt faster than ever, researchers said. [snip] The shrinking of the glaciers could lead to a water shortage and even a dry-up of rivers in the long run, leading to ecological disasters such as wetland retreat and desertification, the researchers said.
This is demonstrably untrue. As Friedman’s piece quotes Armstrong as saying (in the context of South Asia):
“Most of the people downstream, they get the water from the monsoon,” Armstrong said. “It doesn’t take away from the importance [of glacier melt], but we need to get the science right for future planning and water resource assessments.”