The extent to which tree-grass coexistence and biomass ratios in savannas is driven by differences in ecohydrological traits remains unclear. Understanding the effects of precipitation variability on savanna tree cover is important considering that water is a primary limiting resource in this biome. Responses to such variability are implicitly a function of the relative ability of trees and grasses to withstand periods of water stress, access deep soil moisture, exploit large pulses of rainfall, and translate water uptake into biomass gain. Currently however, we lack comprehensive datasets that allow us tease apart tree-grass differences. Here, we quantified three key leaf-level and whole-plant traits in representative tree and grass species at a lowveld savanna in South Africa. In the field, we quantified drought tolerance in 34 tree and 19 grass species by measuring leaf osmotic pressure and leaf turgor loss point (ψtlp). We also grew 20 grass and 20 tree species in the US under standardized environmental conditions, and measured maximum whole-plant transpiration rates (Tmax) with soils at field capacity using a scale-based gravimetric method. Finally, we calculated relative growth rates (RGR) under well-watered conditions. We compared these three traits between the two functional groups and examined within-functional group trait correlations.
Results/Conclusions
Overall, we found that grass species were more drought-tolerant and had a higher maximum transpiration rates on a leaf-mass basis than tree species. We found a tradeoff between ψtlp and Tmax in grasses but not in trees. Trees exhibited a positive relationship between RGR and Tmax, but there was no such relationship between these factors in grasses. We found no relationship between ψtlp and RGR in either functional group. Fine-leaved tree species (Acacia spp. and Dichrostachys cinerea) displayed greater drought tolerance and higher RGR than broad-leaved trees, indicating both functional variation within trees and grass-like traits in this subgroup. Our results suggest that grasses tend to have widespread advantages over trees in terms of water acquisition, with important implications for how we model savanna responses to rainfall variability. Furthermore, we usually categorize trees and grasses as relatively monolithic functional groups, but our results suggest there are important functional subtypes within trees that need to be considered in order to parse tree-grass relationships.