Competition between trees and grasses for both soil water and mineral nitrogen in dry savannas

• We study the competition for multiple resources in dry savannas.
• We consider as limiting resources: soil water and mineral nitrogen.
• We introduce a simple dynamical resource-competition model for dry savannas.
• We use Kendall's test to identify the superior competitor for each resource.
• We use data of 132 dry savannas in Africa.

The co-existence of trees and grasses in savannas in general can be the result of processes involving competition for resources (e.g. water and nutrients) or differential response to disturbances such as fire, animals and human activities; or a combination of both broad mechanisms. In moist savannas, the tree–grass coexistence is mainly attributed to of disturbances, while in dry savannas, limiting resources are considered the principal mechanism of co-existence.Virtually all theoretical explorations of tree–grass dynamics in dry savannas consider only competition for soil water. Here we investigate whether coexistence could result from a balanced competition for two resources, namely soil water and mineral nitrogen. We introduce a simple dynamical resource-competition model for trees and grasses. We consider two alternative hypotheses: (1) trees are the superior competitors for nitrogen while grasses are superior competitors for water, and (2) vice-versa. We study the model properties under the two hypotheses and test each hypothesis against data from 132 dry savannas in Africa using Kendall's test of independence. We find that Hypothesis 1 gets much more support than Hypothesis 2, and more support than the null hypothesis that neither is operative. We further consider gradients of rainfall and nitrogen availability and find that the Hypothesis 1 model reproduces the observed patterns in nature. We do not consider our results to definitively show that tree–grass coexistence in dry savannas is due to balanced competition for water and nitrogen, but show that this mechanism is a possibility, which cannot be a priori excluded and should thus be considered along with the more traditional explanations.