We construct a global, temperature and precipitation dependent, empirical model of soil-biogenic NOx emissions using 6-hour general circulation model forcing. New features of this source relative to the latest published ones by Dignon et al. <1992> and Muller <1992> include synoptic-scale modeling of ''pulsing'' (the emissions burst following the wetting of a dry soil), a biome dependent scheme to estimate canopy recapture of NOx, and an explicit linear dependence of emission on N fertilizer rate for agriculatural soils. Our best estimate for annual above-canopy emissions is 5.5 TG N (NOx) with a range of 3.3--7.7 Tg N. Globally, the strongest emitters are agriculture, grasslands, and tropical rain forests, accounting for 41%, 35%, and 16% of the annual budget, respectively. ''Pulsing'' contributes 1.3 Tg N annually, In temperate regions, agriculature dominates emission, and in tropical regions, grassland dominates. Canopy recapture is significant, consuming, on average, possibly 50% of soil emissions. In temperate regions, periodic temperature changes associated with synoptic-scale disturbances can cause emission fluctuations of up to 20 ng N m-2 s-1, indicating a close correlation between emission and warm weather events favorable to O3/smog formation. By the year 2025, increasing use of nitrogen fertilizer may raise total annual emissions to 6.9 Tg N with agriculture accounting for more than 50% of the global source. Finally, biomass burning may add up to an additional 0.6 Tg N globally by stimulating emissions for a short period after the burn. ¿ American Geophysical Union 1995