Fire and Vegetation Effects on Productivity and Nitrogen Cycling across a Forest–grassland Continuum

Mixed tree–grass vegetation is important globally at ecotones between grasslands and forests. To address uncertainties vis-à-vis productivity and nitrogen (N) cycling in such systems we studied 20 mature oak savanna stands, ranging from 90% woody dominated to 80% herbaceous dominated, growing on comparable soils in a 32-yr-old fire frequency experiment in Minnesota, USA. Fire frequencies ranged from almost annual burning to complete fire protection. Across all stands, aboveground net primary productivity (ANPP) ranged from 2 to 12 Mg·ha21·yr21, decreased with fire frequency (r2 5 0.59), increased with woody canopy dominance (r2 5 0.83), and increased with soil net N mineralization rates (r2 5 0.79), which varied from 25 to 150 kg·ha21·yr21. ANPP was positively related to total biomass (r2 5 0.95), total canopy leaf N content (r2 5 0.88), leaf area index (LAI; r2 5 0.87), annual litterfall N cycling (r2 5 0.70), foliage N concentration (r2 5 0.62), and fine root N concentration (r2 5 0.35), all of which also increased with increasing tree canopy cover. ANPP, soil N mineralization, and estimated root turnover rates increased with woody canopy cover even for stands with similar fire frequency. ANPP and N mineralization both decreased with fire frequency for stands having a comparable percentage of woody canopy cover. Fine root standing biomass increased with increasing grass dominance. However, fine root turnover rate estimated with a nitrogen budget technique decreased proportionally more with increasing grass dominance, and hence fine root productivity decreased along the same gradient. Via several direct and indirect and mutually reinforcing (feedback) effects, the combination of low fire frequency and high tree dominance leads to high rates of N cycling, LAI, and productivity; while the opposite, high fire frequency and high grass dominance, leads to low rates of N cycling, LAI, and productivity. Carbon and N cycling were tightly coupled across the fire frequency and vegetation type gradients.