Savannas tropical grasslands Distribution Climate Controlling factors soils

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 15 00 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. 9 7 1. 0 0. 83 4 5 6 0. 75 Herbs (12 spp. ) 3 0. 83 1. 0 0. 75 4 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 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 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 p. H (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

No. of trees (0. 2 ha plot) Effects of fire protection on tree abundance, SW Nigeria

No. of trees (after 14 yrs) Effects of fire protection on tree abundance in savanna, Zambia

% change (1949 -1963) Effects of fire protection on grass abundance in savanna, Zambia

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 “Oskar. Gulliver” 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.

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

Serengeti dynamics rainfall (mm/month) grass (‘ 00 s tons) adult wildebeest (‘ 000 s) lions (‘ 000 s) 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 1980 s)

“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