The Great African Plume flows westward from a wind divergence line over Central Africa to pollute the Equatorial Atlantic Ocean. The plume arises from agricultural burning fumes which mix in a 3--4 km deep boundary layer over Africa, then override cooler rainforest air, and finally swerve westward, where their progress into the Atlantic is gated by southern storm systems. Another prominent elevated Global Burning Plume from Tropical South America flows past South Africa above 8 km altitude. Joining elevated African plumes, it influences the South Indian and Southern Oceans. These are results from our GRACES (Global Regional Atmospheric Chemistry Event Simulator), which was used to study carbon monoxide during an intensive experimental period, September--October 1992. Traces of the plumes are also evident in observations of CO in the Measurement of Air Pollution from Satellites (MAPS) samples of October, 1994, suggesting a general phenomena. To arrive at these conclusions, we used the detailed weather reconstructions afforded by MM5 meteorological model dynamics to drive GRACES. Both statistical and detailed (event) comparisons of CO and observed aboard the NASA DC-8 are good, especially < 6 km altitude. The following model adaptations that we required may inform global analysis and global models: (1) Parameterized vertical transport like planetary boundary layer convection and deep cumulonimbus convection strongly control CO, allowing far more precision than trajectory studies and models using only large-scale motions. (2) CO sources during this period were consistent with the Hao and Liu <1994> report-based carbon-burn rates; (3) deep convection is even more active than as parameterized by MM5's Grell-scheme: on the large scale: mass fluxes effective for CO redistribution may be over twice core-updraft values diagnosed strengths. ¿ 1998 American Geophysical Union