Effects of Different Treatments of Pasture Restoration on Soil Trace Gas Emissions in the Cerrados of Central Brazil

Planted pastures (mainly Brachiaria spp) are the most extensive land use in the cerrado (savannas of central Brazil) with an area of approximately 50 × 106 ha. The objective of the study was to assess the effects of pasture restoration on the N dynamics (net N mineralization/nitrification, available inorganic N and soil N oxide gas fluxes—NO and N2O), C dynamics (CO2 fluxes and microbial biomass carbon), and diversity of the soil bacterial community using denaturing gradient gel electrophoresis (DGGE) profiles. Sampling was done monthly on a farm in Planaltina, Goiás, Brazil (15°13′S, 47°42′W) from November 2001 to April 2002. Three areas of cerradão (dense cerrado) were converted to pasture (Brachiaria brizantha) in 1991, and after 8 years degradation was evident with the decreasing plant biomass production. Methods to restore these pastures were investigated for their sustainability, principally their effects on trace gas emissions. The pastures have been managed since 1999 as follows: 1) fertilized plot (N = 60 kg ha−1 yr−1, P = 12 kg ha−1 yr−1); 2) grass–legume plot, Brachiaria associated with a legume (Stylosanthes guianensis) with addition of P (12 kg ha−1 yr−1); and 3) a traditional plot without management. A fourth area of cerradão was converted to pasture in 1999 and was not managed (young pasture). Ammonium was the predominant inorganic N form in the soils (∼76 mg N kg−1) for all treatments throughout the study. In December 2001 a reduction in average soil N-NH4+ was observed (∼30 mg N kg−1) compared to November 2001, probably related to plant demand. All plots had high variability of soil N gases emissions, but during the wet season, the NO and N2O soil fluxes were near zero. The results of the water addition experiment made during the dry season (September 2002) indicated that the transition of dry to wet season is an important period for the production of N gases in the fertilized pasture and in the young pasture. Soil CO2 fluxes also increased after the water addition and the grass–legume plot had the highest increase in soil respiration (from ∼2 to 8.3 μmol m−2 s−1). The lowest values of soil respiration and microbial biomass carbon (∼320 mg C kg−1 soil) tended to be observed in the young pasture, because the superficial layer of the soil (0–10 cm) was removed during the conversion to pasture. Trace gas emissions measured after the water addition experiment corresponded to rapid changes in the soil bacterial community. The young pasture sample showed the lowest level of similarity in relation to the others, indicating that the bacterial community is also influenced by the time since conversion. This study indicates that the restoration technique of including Stylosanthes guianensis with B. brizantha increases plant productivity without the peaks of N oxide gas emissions that are often associated with the use of N fertilizers. Additionally, the soil bacterial community structure may be restored to one similar to that of native cerrado grasslands, suggesting that this restoration method may beneficially affect bacterially mediated processes.