『Abstract
　Savannas are shaped by drought, herbivory, nutrient limitation, and fire. We assessed the interactions between two of these factors - nutrient limitation and fire - across a savanna landscape in Kruger National Park, South Africa, by quantifying increases in plant-available forms of soil phosphorus (P) following experimental or simulated burns. Plant-available pools of P were defined for granitic surface soils subjected to four experimental fire regimes (no fire, triennial fire in the wet season, triennial fire in the dry season, and annual fire in the dry season). To provide context for these results, we also fractionated surface soils from a granitic catena after simulating burning of these soils in a furnace.
　Burned soils showed comparable pyromineralization rates, with the experimental burn plot soils averaging 0.49±0.04 g labile P m^-2 y^-1 and catena soils averaging 0.63±0.12 g labile P m^-2 y^-1. Only soils from subplots burned triennially during the wet summer season with moderate fire intensities (〜1.1 MW m^-1) showed significant increases in labile P relative to control soils. Soils from other burned subplots with greater fire intensities showed smaller gains in labile P, suggesting pyromineralization rates may peak at intermediate fire intensities. We estimated ash contributed up to 33％ of pyromineralized P. For catena soils, simulated burning led to significant increases in pyromineralized P for the relatively P-rich footslope soils and smaller increases for sandy crest and midslope soils. These pyromineralization P fluxes are of the same order of magnitude as plant P demand estimated using foliar P levels, and about one-half microbial mineralization rates. In P-limited ecosystems where chemical weathering rates are slow, moderate-intensity fires could play a critical biogeochemical role in the supply of labile P.

Keywords: Experimental burn plot; Kruger National Park; Granite; Fire; Catena; Foliar』

1. Introduction
2. Methods
　2.1. Soil samples
　　2.1.1. Sourveld EBP soils
　　2.1.2. Catena soils
　2.2. Plant samples for foliar P analysis
　2.3. Laboratory techniques
　　2.3.1. Simulated burning of catena soils
　　2.3.2. Sequential fractionation of experimental burn plot and catena soils
　2.4. Calculations and statistical analyses
3. Results
　3.1. Sourveld EBP soils
　3.2. Catena soils
　3.3. Sourveld EBP foliar P levels
4. Discussion
　4.1. Partitioning pyromineralized soil P into its ash and heat components
　4.2. Comparison with other pyromineralization studies
　4.3. Ecological implications
　4.4. Assumptions
5. Conclusions
Acknowledgements
Appendix A. Supplementary data
References