Geoscience Reference
In-Depth Information
proposed that semi-arid rangelands were disequilibrium systems where frequent droughts
aperiodically knocked back animal populations, keeping them below levels that would cause
degradation of vegetation. These ideas re-examined carrying capacity as a variable rather
than a constant feature of savanna ecosystems (Behnke et al. 1993).
The chance to observe predicted cyclical changes in tree and elephant abundance, or
repeated natural die-offs in Tsavo was dramatically curtailed due to rampant elephant poach-
ing in the 1970s and 1980s. A quota system for ivory, implemented in 1986 aimed to reduce
pressure on elephant populations by limiting ivory sales, but in fact had the opposite effect,
driving the value of ivory up and creating an incentive to poach (Gillson and Lindsay 2003). A
complete ban was come into force in 1989, remaining in place until 1997, when limited legal
sales began again.
Poaching destroyed the natural experiment in elephant management that was taking place
in Tsavo, and there was little long-term information on vegetation change, as written records
only became common form the end of the nineteenth century. As savanna trees are long-
lived, data covering many centuries are needed to understand the natural resilience and
dynamism of the system. If Caughley's woodland-grassland hypothesis was correct, then
overpopulation of elephants would only be a transitory phase and woodland-grassland
cycles should be apparent in the palaeoecological record.
Fossil pollen data from Kanderi Swamp, in Tsavo East, showed evidence of dramatic
changes between woodland and grassland phases over the past 1400 years (Gillson 2004b)
(Figure 2.1). These data were the first from that area of East Africa, as most previous work was
from large Rift Valley lakes, or wetter and more mountainous areas, where sediment accumu-
lates more readily and pollen preservation is better. The grassland phases were from c. 1,400-
1,200 years ago, and again from 430-180 years ago, with wooded vegetation dominant for the
rest of the time. Fourier transformation of the pollen data suggested that tree/grass abun-
dance varies cyclically with a periodicity of 250-500 years (Gillson 2004b).
Most ecological findings are complex and palaeoecological data are no exception. Never-
theless, the pollen data confirm that transitions between woodland and grassland phases
can occur over timescales of decades to centuries. The results suggest a resilient tree popula-
tion that can rapidly repopulate grassland landscapes and transform them to woodlands
when conditions are favourable. At larger spatial scales, the landscape could potentially
remain stable, because woodland-grassland cycles in individual patches of vegetation may
be out of phase (Gillson 2004a). Furthermore, the pollen data suggest that in the past two
centuries, tree abundance, has increased beyond any other time in the past 1400 years, pos-
sibly reflecting increasing CO
2
levels, which benefit savanna trees, and possibly further
enhanced by declining herbivory at the end of the nineteenth century due to rinderpest, and
the later removal of people from the park in the 1940s, who may have previously increased
fire frequency (Dublin et al. 1990, Bond and Midgley 2000).
When viewed in a longer-term context, the declines in tree abundance that caused con-
cern in the 1960s may have been a temporary dip in a period of unusually high tree abun-
dance, which did not require any management intervention The fossil pollen data suggest
that declines in tree abundance had happened before the twentieth century, but they did not
Search WWH ::
Custom Search