The superlative environmental conditions in Southern Amazonia, i.e. high temperatures and annual rainfall, create ideal conditions for high soil organic matter turnover rates and therewith the soil-atmosphere exchanges of greenhouse gases. In this study, we present daily observations of soil-related carbon dioxide (CO2), nitrous oxide (N2O), and methane (CH4) emissions from the dominant land use types in Mato Grosso (cerrado, gallery forest, cattle pasture, and cropland under soybean) and Pará (rainforest and cattle pastures) during the early rain season. We followed the hypothesis that precipitation events provoke hot moments for CO2 and N2O, and lead to alternating uptake and emission for CH4, respectively. Observed fluxes differed significantly between land use types and underlying soil type. CO2 fluxes from soils under natural vegetation ranged from 101 mg m-2 h-1 in the cerrado to 160 mg C m-2 h-1 in the rainforest. Fluxes from cattle pastures varied between locations and were 79 mg C m-2 h-1 from the pasture in Mato Grosso and between 120 and 180 mg C m-2 h-1 from pastures in Pará. For N2O, fluxes were highest from rainforest (16 µg N m-2 h-1) and lowest from cerrado soil (-0.05 µg N m-2 h-1). Similar to CO2, the pastures in Pará emitted higher fluxes of N2O (2–8 µg N m-2 h-1) compared to the pasture in Mato Grosso (0.09 µg N m-2 h-1).  CH4 fluxes were negative on all sites, except for two pasture sites in Pará, where recorded fluxes amounted to 10 and 53 µg C m-2 h-1, respectively. The dynamic behavior during the measurement period, as determined by an indicator function, was moderate for N2O, low for CO2 and CH4, and depended on the site and measurement chamber, respectively. Nevertheless, N2O fluxes from the cropland increased during the end of the vegetation period of soybean, likely as a result of increased nitrogen availability due to ceasing plant roots. At the same time cattle pastures showed an age-related decrease of N2O emissions, whereas they turned out to being considerable sources for CH4. Generally, the low dynamics and missing reactions to precipitation events demonstrate poor process understanding and need to be further investigated.