[1] Using closed chamber techniques, soil fluxes of NO, N 2 O, and CO 2 were measured from September 1999 to November 2000 in savanna areas of central Brazil (cerrado) subjected to prescribed fires. Our studies focused on two vegetation types, cerrado stricto sensu (20-50% canopy cover) and campo sujo (open, grass dominated), which were either burned every 2 years or protected from fire. Soil moisture and vegetation type were more important in controlling NO and CO 2 fluxes than fire regime (early dry season, middle dry season or late dry season burning). N 2 O fluxes, however, were very low and below detection limit in any of the vegetation-fire treatments. NO emissions increased after burning (1.0 ng NO-N cm À2 h À1 ), but flux returned quickly to prefire levels and even lower. In comparison, NO emissions increased 100-fold (to 10.5 ng NO-N cm À2 h À1 ) during a water-addition experiment in unburned campo sujo, and to 1.0 ng NO-N cm À2 h À1 in unburned cerrado and 1.9 ng NO-N cm À2 h À1 in burned cerrado with the first rains. Low NO and N 2 O emissions, low nitrification rates, and the majority of inorganic N in the form of NH 4 + all indicate a conservative N cycle in the cerrado. CO 2 fluxes increased with the onset of the rainy season and after artificial water addition. The highest CO 2 measured in the wet season was 6.3 mmol CO 2 m À2 s À1 in burned campo sujo. During the dry season, soil respiration in burned and unburned treatments were similar (average flux = 1.6-2.3 mmol CO 2 m À2 s À1 ). Differences between fire treatments of cerrado and campo sujo CO 2 fluxes are attributed to differences in relative litter production and root activity.

Abstract. We compared fluxes of CO 2 , CO, NO, and N 2 O, soil microbial biomass, and N availability in a 20-yr-old Brachiaria pasture and a native cerrado area (savanna in central Brazil). Availability of N and NO fluxes were lower in the pasture than in the cerrado. N 2 O fluxes were below detection limit at both sites. The CO fluxes showed weak seasonal variation with slightly higher positive fluxes in the dry season and lower fluxes, including net consumption, during the wet season. The cerrado CO fluxes were higher and more variable than the fluxes in the pasture. Both sites showed a seasonal pattern in CO 2 emissions with lower fluxes (ϳ2 mol CO 2 ·m Ϫ2 ·s Ϫ1 ) during the dry season. There were no significant differences in annual CO 2 soil emissions between the cerrado and the pasture, but the temporal trends differed, with higher fluxes in the pasture during the transition from the wet to the dry season. Artificial water addition in the pasture during the dry season resulted in short-lived pulses of NO and CO 2 .

[1] Soil-atmosphere fluxes of carbon monoxide (CO) were measured from September 1999 through November 2000 in savanna areas in central Brazil (Cerrado) under different fire regimes using transparent and opaque static chambers. Studies focused on two vegetation types, cerrado stricto sensu (ss) (20-50% canopy cover) and campo sujo (open, scrubland), which were either burned every 2 years or protected from fire (for 26 years). CO emissions in transparent chambers varied seasonally, with highest fluxes during the late dry season and transition to wet season (August-October) and lowest fluxes late in the wet season (February-April). Daytime fluxes in the transparent chambers were always higher than in the opaque chambers. Similarly, a diurnal study showed negative fluxes for all nighttime measurements and positive measurements for all daytime measurements made with transparent chambers. Deposition velocities observed in the opaque chambers during the night fell in the 0.002-0.0014 cm s À1 range, which is at the lower end of the range that has been observed in tropical, temperate, and high-latitude regions. No significant differences were found between the daytime annual average fluxes from unburned cerrado and unburned campo sujo (160 Â 10 9 and 190 Â 10 9 molecules cm À2 s À1 , respectively). Fire increased soil surface CO emissions significantly in the burned cerrado plot. Measurements made 30 days after the fire showed daytime CO production over 10 times higher than that of the unburned cerrado ss (812.8 Â 10 9 versus 76.8 Â 10 9 molecules cm À2 s À1 ). Postfire CO emissions were greater than prefire emissions for both opaque and transparent chambers. However, the fire-induced increase was greater in the transparent chambers. This suggests that the fire created both photochemically and thermally reactive precursors. Removal of litter and standing, dead plant material from plots in unburned campo sujo and a pasture was shown to dramatically decrease CO emissions. CO production in burned plots (using opaque chambers) was similar to previous measurements from Venezuelan and African savannas.

The objective of this work was to evaluate the effects of fire regimes and vegetation cover on the structure and dynamics of soil microbial communities, through phospholipid fatty acid (PLFA) analysis. Comparisons were made between native areas with different woody covers ("cerrado stricto sensu" and "campo sujo"), under different fire regimes, and a 20-year-old active palisadegrass pasture in the Central Plateau of Brazil. Microbial biomass was higher in the native plots than in the pasture, and the highest monthly values were observed during the rainy season in the native plots. No significant differences were observed between fire regimes or between communities from the two native vegetation types. However, the principal component (PC) analysis separated the microbial communities by vegetation cover (native x pasture) and season (wet x dry), accounting for 45.8% (PC1 and PC3) and 25.6% (PC2 and PC3), respectively, of the total PLFA variability. Changes in land cover and seasonal rainfall in Cerrado ecosystems have significant effects on the total density of soil microorganisms and on the abundance of microbial groups, especially Gram-negative and Gram-positive bacteria.

No abstract