Tree canopies explain fire effects on soil nitrogen, phosphorus and carbon in a savanna ecosystem

Questions: What are the independent and interactive effects of fire and tree canopies on soil nutrient and C pools in savannas? Does fire differentially affect total and labile pools of C and N? How do these effects differ between nutrient-poor, broad-leaved savannas on sandy soils of granitic origin, and nutrient-rich, fine-leaved savannas on clay-enriched basalts? Location: Kruger National Park, South Africa. Methods: We investigated long-term effects of fire and tree canopies on total soil C and N, labile C, plant-available N and P, and acid phosphatase enzymes at two sites in Kruger National Park, one on granitic and one on basaltic soil. We sampled soils from plots that form part of a 50-yr landscape-level controlled burning experiment. In addition to the two soil types, we sampled across three burning treatments (no burns, annual burns and triennial burns) in replicated blocks, both under and between tree canopies. Results: There was little evidence for direct effects of fire on any of the variables tested on either soil type, with the exception of C mineralization rates on basalt, which suggested a smaller pool of labile C in frequently burned than in unburned plots. Tree canopies were positively associated with sizes of total and labile soil pools of N and C, and negatively associated with plant-available P, particularly on nutrient-rich basalts. Conclusions: Our results suggest that variation in tree canopy cover is the dominant biotic driver of soil N, P and C dynamics in these savanna systems, despite the frequent occurrence of fire. Given the negative effects of fire on size of individual trees in this ecosystem, this suggests that the role of fire on nutrient cycling may be mediated primarily through its effects on canopy cover. This effect is likely to be magnified in nutrient-rich savannas where tree canopy effects on soil nutrient dynamics are strongest.