Nutrients, not water, limit the distribution of forest at a tropical forest-savanna boundary in central Brazil

Background/Question/Methods

Tropical savanna-forest boundaries are often considered sensitive indicators of climate change, but it has not been demonstrated that they are directly responsive to climate. Understanding what limits tropical forest expansion into savanna is needed for predicting the future dynamics of these two biomes. We performed an ecosystem experiment at a savanna-forest boundary to test the relative importance of water and nutrient availability for limiting the expansion of forest. We set up twelve 70m x 10 plots, each extending across the savanna-forest boundary. Water and Nutrient treatments were randomly assigned to these plots in a blocked 2 X 2 factorial experiment. The water treatment consisted of a control (no added water) or irrigation (60 mm per week throughout the dry season), and the nutrient treatment consisted of a control (no added nutrients or a complete NPK + micronutrients added twice per year). Tree growth was monitored annually with band dendrometers. 

Results/Conclusions

We found nutrients to be the primary factor limiting tree growth and forest expansion into savanna in the absence of fire. Nutrient addition more than doubled rates of tree diameter growth over 4 years (2.4 mm/yr versus 1.1 mm/yr) growth. Irrigation had no detectable effect on tree growth (1.9 mm/yr versus 1.7 mm/yr). Long-term fire suppression at the site has allowed forest tree species to establish in savanna, and these had nearly 4-fold greater growth rate than savanna species. Thus nutrient limitation, combined with the inherent low growth rates of savanna tree species, greatly slows canopy closure in this environment, predisposing the savanna to remain in an open state under regular burning.  These results challenge the widespread perception that rainfall is the primary factor limiting the natural distribution of tropical forest.