Environmental Engineering Reference
In-Depth Information
Perhaps most notably, China's ofi cial i gures kept underestimating the country's
arable land, and the FAO kept accepting this well-known misinformation, which
put the country's arable land at no more than 95 Mha in 1995, about 30% less
than the real i gure.
During the 1990s, China's Statistical Yearbook cautioned that “i gures for the
cultivated land are under-estimated and must be further verii ed”—many years after
it was known that the ofi cial total was wrong. The earliest remote sensing studies,
based on imagery with inadequate resolution (Landsat, with a resolution of 80 m),
indicated farmland area as large as 150 Mha, detailed sample surveys of the late
1980s came up with the range of 133-140 Mha (Smil 1993), and a correction using
land survey data for 1985 ended up with 137.1 Mha for the year 1995 (Heilig
1997). The best remote sensing evaluation, using classii ed imagery from the Keyhole
series of intelligence satellites, yielded the total of 133-147 Mha for the year 1997
(MEDEA 1997; Smil 1999a).
Only in the year 2000 did China's State Statistical Bureau i nally relent and
correct the total to 130 Mha (SSB 2000), a i gure that was promptly adopted by
the FAO. How many other underestimates of this kind are hidden behind the ofi cial
i gures submitted to the FAO by many Asian or African countries? And how uncer-
tain are all those estimates made in the FAO's headquarters on the basis of old or
fragmentary information for countries beset by endless violent conl icts, such as
Afghanistan, Sudan, or Somalia? Of course, we now have satellite sensors whose
resolution could distinguish the smallest vegetable i elds in Asia, but their use would
be prohibitively expensive, and hence the global mapping of cropland typically
proceeds at much coarser resolution and, inevitably, its results carry substantial
aggregation or omission errors.
The most accurate data on the totals of arable land as well as for actually planted
and harvested areas are available for all afl uent countries of North America, the
EU, East Asia (Japan, South Korea, Taiwan), and Australia and New Zealand, and
some of these data series are available for the entire twentieth century and part of
the nineteenth century; they were collated in the annual statistics published by the
International Institute of Agriculture, the FAO's Rome-based pre-World War II
precursor, since 1910. The i rst global-scale effort to estimate the history of farmland
expansion going back more than 150 years was offered by Richards (1990), whose
series began with 265 Mha in 1700 and progressed to 537 Mha by 1850, 913 Mha
by 1920, and 1.17 Gha by 1950. A new approach to the reconstruction of past
farmland expansion came during the 1990s, when backcasting models provided
results for a gridded (0.5°
×
0.5°) distribution of farmland.
Search WWH ::




Custom Search