This content requires Macromedia Flash Player version 8 or later. Get Flash.
If you believe you do have Macromedia Flash Player version 8 or later installed, there is a problem with your Flash installation and we were unable to detect it. Please follow the solution in Adobe TechNote 7d1862a to resolve this issue.
DATA ANALYSIS: DESERTS
Just deserts
Felix Grant sifts the sands of analysis in desert environments
A
ccording to Richard Lederer’s unnamed student[1]
, deserts
are defi ned by a climate ‘such that the inhabitants have to
live elsewhere’ and have to be ‘cultivated by irritation’. Say the word desert to most people in the industrialised west, and they will visualise somewhere hot, sandy, and utterly devoid of any life except, possibly, Peter O’Toole and Omar Sharif striding (with or without camels) across photogenic dunes. In the days when I was trying the patience of successive O-level geography teachers, we were given a more seductively precise and objective defi nition: annual mean rainfall of 250mm or less. That fi gure is still around, although it has generally been supplanted by more sophisticated relative concepts like relative moisture economy defi cit (more water lost through evaporation and transpiration than is received in precipitation). Generally speaking, the most widely
used climatological classifi cation systems go by distribution of vegetation, which both serves as an indicator of available water and decides the viability of other life. This approach naturally appeals to statistical data analytic bean counters like myself, though they do complicate matters by including temperature considerations. A desert, for my purposes here, is a region whose available water economy provides marginal (or nonexistent) support for life. It will usually be arid, but we won’t be too picky about semi-aridity distinctions as many of the problems to be studied involve
12
transition. It will often be megathermal, but not necessarily so. Oceans cover about two thirds of
the earth’s surface. Deserts account for roughly two thirds of what is left (and current trends, despite signifi cant efforts at reversal, are towards net increase) so they are no small matter – either in themselves or in the computational challenges that they can present. It is not much of an exaggeration to say that, only since the explosion of generally accessible scientifi c computing resources, can they be realistically confronted at all.
‘Desertifi cation is not the advance of deserts. Rather, it is the persistent degradation of dryland ecosystems by human activities and climatic variations’
Many of the issues, and the data sets,
are global in scope and impact, and of a complexity that requires high performance computing methods to make them tractable. Climatological investigations (such as the reanalysis of prediction models[2]
by Ganguly and others, indicating
that ‘global average temperatures from the middle to end of the 21st century are likely to be higher than previously believed’, particularly ‘in desert areas, where there is some underprediction’) clearly fall into this category. Others are, thanks to the escalating power-to-cost ratio of both
SCIENTIFIC COMPUTING WORLD OCTOBER/NOVEMBER 2010
hardware and software tools, increasingly brought to the small team’s or individual researcher’s desktop. In between are those where data and results move up and down a chain of cooperation between fi eld and institutional or corporate facilities. Throughout that scale, increasing
desertifi cation is a high priority concern. ‘Desertifi cation is not’, as IFAD (the International Fund for Agricultural Development) is at pains to point out, ‘the advance of deserts, though it can include the encroachment of sand dunes on land. Rather, it is the persistent degradation of dryland ecosystems by human activities and climatic variations.’[3]
Over time, though,
distinction between encroachment and advance can be a fi ne and debatable one. There is suggestion from anecdotal historical sources that the world’s largest single megathermal desert, the Sahara, may have historically expanded considerably through exactly the mechanisms mentioned, though objective measurement to check this (using red and infrared satellite imaging to infer a 200mm annual precipitation isoline from vegetation cover) only started in 1991[4]
and no conclusions are yet possible.
Establishment of baseline marker data of this kind is one of the priorities set by the United Nations Convention to Combat Desertifi cation (UNCCD) as a prerequisite for monitoring problems ranging from ‘expanding deserts in China, India, Iran, Mongolia and Pakistan, the sand dunes of Syria, the steeply eroded mountain slopes of Nepal, and the deforested and overgrazed highlands of the Lao People’s Democratic
www.scientifi
c-computing.com
Page 1 |
Page 2 |
Page 3 |
Page 4 |
Page 5 |
Page 6 |
Page 7 |
Page 8 |
Page 9 |
Page 10 |
Page 11 |
Page 12 |
Page 13 |
Page 14 |
Page 15 |
Page 16 |
Page 17 |
Page 18 |
Page 19 |
Page 20 |
Page 21 |
Page 22 |
Page 23 |
Page 24 |
Page 25 |
Page 26 |
Page 27 |
Page 28 |
Page 29 |
Page 30 |
Page 31 |
Page 32 |
Page 33 |
Page 34 |
Page 35 |
Page 36 |
Page 37 |
Page 38 |
Page 39 |
Page 40 |
Page 41 |
Page 42 |
Page 43 |
Page 44 |
Page 45 |
Page 46 |
Page 47 |
Page 48 |
Page 49 |
Page 50 |
Page 51 |
Page 52 |
Page 53 |
Page 54 |
Page 55 |
Page 56