Fire increasingly threatens the integrity of the Amazon forest, particularly along its agricultural frontier. In addition, fire alters nutrient availability and herbivory, two factors hypothesized to affect plant diversity. Studying the diversity of regenerating vegetation as affected by fire, nutrient additions, and herbivory therefore addresses basic ecological questions as well as conservation and management issues. This work took place in a large-scale, long-term controlled burn experiment in the southern Amazon and used a factorial design to test interactions between annual fires, nutrient availability, and herbivory on seedling and resprout diversity over the course of one year. The nutrient treatment was established in subplots with additions of nitrogen (N), phosphorus (P), or nitrogen and phosphorus together (N+P) in burned and unburned forest. Within the nutrient plots, herbivore exclosures were constructed to limit herbivory. In addition, effects of fire frequency were examined by comparing diversity between areas subjected to annual versus triennial burns. It was hypothesized that nutrient availability and herbivory would increase diversity in burned plots but that these factors would be less important in unburned forest.
Results/Conclusions
After six annual burns, species richness and diversity were highest in unburned plots without nutrients. Evenness was greatest in unburned plots with P. Increased herbivory was correlated with increases in diversity in unburned plots and burned plots with nutrients. In contrast, in unfertilized, burned plots, higher herbivory was related to lower diversity. Plots with high herbivory were also less even in the burned area. Herbivory therefore does not have consistent effects on regeneration diversity, and overall plots without nutrients or with N have the highest diversity after fire.
The unburned area and the area burned every three years had similar levels of diversity, but the annually burned area was less than half as species rich. Community composition also differed with fire. In both the annually burned and the triennially burned forest, species that are commonly found in the savanna were present, indicating repeated fires may enable savannization of the southern Amazon.
Overall, seedling survival was negatively related to herbivory, although it did not always increase diversity, and only N was related to improved diversity after fire. Fire has deleterious effects on regeneration, and even infrequent burns open space for the recruitment of savanna species.