Transcript 09_Savannas

Savannas
(tropical grasslands)






Distribution
Climate
Controlling factors: soils, fire, grazing
Savanna patchiness
Climate change
Desertification
African climate and savanna
distribution
Savanna ecosystems in W Africa
Vegetation
Precipitation
250
500
1000
1500
2000
desert
Sahel zone
Sudan zone
Guinea zone
savanna-forest
rainforest
Desert
Low grass savanna
Tall grass savanna
Thorn
forest
Rain forest
Climatic control on savanna distribution
and type in West Africa
Synoptic situation
ITCZ = northern edge of rains
25°
20°
15°
rainy season
10°
5°
J F M A M J J A S O N D
Savanna
ecosystems in
South and East
Africa
Low grass savanna
Miombo woodland
Mopane woodland
S American climate and savannas
Llanos
Cerrado
2
3
1
4
5
1.
2.
3.
4.
5.
6.
7.
Honduras
Bolivar
Llanos
Rupununi
Amazonas
Cerrado
S. Brazil
6
7
Floristic similarity in S. American savannas
(with Rupununi)
1
1
0.5
2
2
3
0.9
5
0.75
Herbs
(12 spp.)
3
0.83
1.0
0.75 4
0.9
1.0
0.83
4
5
6
7
0.67
6
0.5
Shrubs
(6 spp.)
0.83
7
Caribbean
pine
savannas
(Belize Nicaragua)
Are South American savannas primarily
products of seasonal drought?
500
Savanna
1000
1500
Forest
2000
2500
“the vegetation is xerophytic in many places because of the dry
season that lasts for months . . . But the xerophylly is also
due to the the dry continental climate in general.”
E. Warming, 1909 (on the southern cerrados of Brazil)
the climate of the Venezuelan Llanos is “hostile to woodland”
Schimper, 1903
Common sclerophyllous shrubs, S.
American savannas
Curatella americana
Byrsonima crassifolia
Alternative (or supplementary)
hypotheses
Savanna
Forest
fire
soil senility
topography
seasonal drought
and inundation
Soil hydrology
Van Donselaar (1969), on the basis of work in
Surinam, commented:
“Savanna communities are primarily correlated with
the hydrology of the soil”
In the wet season the savannas of the Venezuelan
Llanos (and other flat-lying savanna areas) may be
inundated by flood waters for several weeks; in the
dry season the water table may drop to depths of
several metres. Such fluctuations may be too severe
for rainforest trees.
Topography:
Monica Cole’s observations
savanna
forest
savanna
forest
savanna
to Brazilian coast -->
Topographic control on savanna/forest
distribution, Rupununi
Savanna-forest boundary
(Kakadu National Park, NT, Australia)
Frederick
Hardy:
the senile
soil
hypothesis
Characteristics of senile soils
• Low pH (4.4 - 5.2) Low cation exchange
capacity
(clay fraction dominated by kaolinite)
• Very low base saturation
• High soluble aluminium
Soil nutrients inadequate to support forest growth;
only alumino-tolerant trees survive.
Plinthite formation
• Development of iron-rich horizon in zone of
fluctuating seasonal water table.
• Long-term lowering of water table causes
irreversible induration of iron-rich horizon
(plinthite / laterite / ferricrete).
• Plinthite inhibits root penetration and
causes perched water tables and seasonal
inundation.
Fire as an agricultural tool in the Guinea
zone, W. Africa
The role
of fire
“..together with the degradation towards a poor savanna
(following on the use of fire) many other changes occur;
the soil definitely deteriorates and lateritic iron pans are
formed.”
Budowski, 1959 (on savanna
formation in Nicaragua)
Fire-maintained boundary:
Rupununi savannas, Guyana
Dry season fires in the forestsavanna zone, Africa, 1987
Fire temperatures in tall
grass savanna
Low grass
savannas with
sub-shrubs,
Rupununi
savannas; a
cold fire
environment
Effects of fire protection on tree
abundance, SW Nigeria
No. of trees (0.2ha plot)
500
400
300
Fire -tender
All species
200
100
0
Late burn
Early burn
Protected
No. of trees (after 14 yrs)
Effects of fire protection on tree
abundance in savanna, Zambia
Effects of fire protection on grass
abundance in savanna, Zambia
Heteropogon contortus
Themeda trianda
% change (1949-1963)
50
40
30
20
10
0
-10
-20
-30
-40
1-yr
2-yr
3-yr
5-yr
Unburnt
The “Oskar-Gulliver” concept
“repeated fires can keep (trees) small, but (they)
rarely suffer mortality, and large (trees) are virtually
immune from fire damage. This .. has been called
the Oskar syndrome (after Günter Grass’ character
Oskar Matzerath), which emphasizes the potentially
advanced age of a small individual, or the Gulliver
syndrome (after Jonathan swift’s character..), which
emphasizes a tree’s potential to be a giant once it
escapes fire. Simulation model(s) … have shown
that the distribution of fire intensities at a site can
shape the structure of the tree stratum”
“Oskar”:
an
individual of
Palicourea rigida
subject to
frequent, lowintensity grass
fires
[Rupununi savannas]
The “OskarGulliver”
syndrome:
field surveys of four
savanna areas in South
Africa subjected to
fire exclusion
treatments for ~50
years
Higgins et al., 2007 Ecology 88, 1119-1125.
QuickTime™ and a
TIFF (LZW) decompressor
are needed to see this picture.
Fire use by hunter-gatherers
(e.g. northern Australia)
“The fine-scale mosaic of burnt and unburnt
areas created by mid-dry season Aboriginal
landscape burning has clear effects on the
distribution of kangaroos. Kangaroos move
into burnt moist habitats and away from burnt
dry, rocky habitats. Isotopic analysis of scats
suggests that the mechanism driving this
effect is the increased abundance of nitrogen
rich grasses in burnt moist habitats.”
Murphy, B.P. & Bowman, D.M.J.S. 2007. J. Biogeography, 34, 237-
Savanna
patch
dynamics
Acacia patches in savanna:
Serengeti, Tanzania
Herbivore defences in
Acacia species
Herbivore
thorns
companion ants
Acacia patch characteristics
Termitaria in the Rupununi
savannas, Guyana
Termitaria in
burnt
savanna,
Kakadu N.P.
(NT, Australia)
Leaf litter
decomposition
by termites
Termite mounds:
high nutrient patches
Grazing sequence: Serengeti
rains
renewed
grass
growth
zebra …. wildebeest … gazelles
coarse …….
grasses
new
shoots
………..
forbs
Serengeti food chain and interactions
rainfall
(mm/month)
grass
(‘00s tons)
adult wildebeest
(‘000s)
lions
(‘000s)
Wolanski, E., et al., 1999. American
Scientist, 87, 526-531 (Fig. 9)
Lines = simulation model; dots = observed data
Grazing patchiness
e.g. interactions between zebra, Grant’s gazelle and
wildebeest in the Serengeti: patchiness is a product
of direct effects (grazing sequence), and indirect
effects (nutrient cycle shunt = scat production)
What are the results of loss of herbivores?
(e.g. 70% of elephants in Serengeti poached in 1980s)
“The last acacia”:
land management issues in the Serengeti
1890’s - outbreak of rinderpest in East Africa led to mega-death of
cattle
(and wildebeest) and starvation amongst pastoralists. Fewer
people, therefore fewer fires to stimulate grass growth (and kill tree
seedlings);
1930-40’s - reduction in fires led to expansion of acacia woodland;
1960’s - wet; cattle numbers increased, but wildebeest did not; fires
in savanna were hot, killing tree seedlings;
1963 - rinderpest control program; ungulates recovered;
1970’s - Old acacias (which live to 60-70 years) were dying; few
replacements; elephants blamed for destroying young trees;
elephants culled;
1990’s - Numbers of buffalo and elephants are far lower due to
heavy poaching (although elephants have been increasing since
the 1990 ivory ban). The wildebeest population has soared to about
1 million; human-set fires are down to about a quarter of what they
The last acacia?
Views of the Serengeti
1980
1991
Photos: Science 19 December 1997: 278. no. 5346, p. 20
Long-term
climate
change:
Late
Pleistocene
lake levels
in African
savannas
Evidence of
climate
change:
Holocene
vegetation
changes in
the Lake
Victoria
region