How Plant Functional-Type, Weather, Seasonal Drought, and Soil Physical Properties Alter Water and Energy Fluxes of an Oak-Grass Savanna and an Annual Grassland

Savannas and open grasslands often co-exist in semi-arid regions. Questions that remain unanswered and are of interest to biometeorologists include: how do these contrasting landscapes affect the exchanges of energy on seasonal and annual time scales; and, do biophysical constraints imposed by water supply and water demand affect whether the land is occupied by open grasslands or savanna? To address these questions, and others, we examine how a number of abiotic, biotic and edaphic factors modulate water and energy flux densities over an oak-grass savanna and an annual grassland that coexist in the same climate but on soils with different hydraulic properties. The net radiation balance was greater over the oak woodland than the grassland, despite the fact that both canopies received similar sums of incoming short and long wave radiation. The lower albedo and lower radiative surface temperature of the transpiring woodland caused it to intercept and retain more long and shortwave energy over the course of the year, and particularly during the summer dry period. The partitioning of available energy into sensible and latent heat exchanged over the two canopies differed markedly. The annual sum of sensible heat exchange over the woodland was 40% greater than that over the grassland (2.05 GJ m{sup -2} per year versus 1.46 GJ m{sup -2} per year). With regards to evaporation, the oak woodland evaporated about 380mm of water per year and the grassland evaporated about 300mm per year. Differences in available energy, canopy roughness, the timing of physiological functioning, water holding capacity of the soil and rooting depth of the vegetation explained the observed differences in sensible and latent heat exchange of the contrasting vegetation surfaces. The response of canopy evaporation to diminishing soil moisture was quantified by comparing normalized evaporation rates (in terms of equilibrium evaporation) with soil water potential and volumetric water content measurements. When soil moisture was ample normalized values of latent heat flux density were greater for the grassland (1.1-1.2) than for the oak savanna (0.7-0.8) and independent of moisture content. Normalized rates of evaporation over the grassland declined as volumetric water content dropped below 0.15 m{sup 3} m{sup -3}, which corresponded with a soil water potential of -1.5MPa. The grassland senesced and quit transpiring when the volumetric water content of the soil dropped below -2.0MPa. The oak trees, on the other hand, were able to transpire, albeit at low rates, under very dry soil conditions (soil water potentials below -4.0MPa). The trees were able to endure such low water potentials and maintain basal levels of metabolism because ecological forcings kept the tree density and leaf area index of the woodland low, physiological factors forced the stomata to close progressively and the trees were able to tap deeper water sources (below 0.6 m) than the grasses.