Amazon forests could be globally significant sinks or sources for atmospheric carbon dioxide, but carbon balance of these forests remains poorly quantified. We surveyed 19.75 ha along four 1-km transects of well-drained old-growth upland forest in the Tapajós National Forest near Santarém, Pará, Brazil (2°51′ S, 54°58′ W) in order to assess carbon pool sizes, fluxes, and climatic controls on carbon balance. In 1999 there were, on average, 470 live trees per hectare with diameter at breast height (dbh) ≥10 cm. The mean (and 95% ci) aboveground live biomass was 143.7 ± 5.4 Mg C/ha, with an additional 48.0 ± 5.2 Mg C/ha of coarse woody debris (CWD). The increase of live wood biomass after two years was 1.40 ± 0.62 Mg C·ha−1·yr−1, the net result of growth (3.18 ± 0.20 Mg C·ha−1·yr−1 from mean bole increment of 0.36 cm/yr), recruitment of new trees (0.63 ± 0.09 Mg C·ha−1·yr−1, reflecting a notably high stem recruitment rate of 4.8 ± 0.9%), and mortality (−2.41 ± 0.53 Mg C·ha−1·yr−1 from stem death of 1.7% yr−1). The gain in live wood biomass was exceeded by respiration losses from CWD, resulting in an overall estimated net loss from total aboveground biomass of 1.9 ± 1.0 Mg C·ha−1·yr−1. The presence of large CWD pools, high recruitment rate, and net accumulation of small-tree biomass, suggest that a period of high mortality preceded the initiation of this study, possibly triggered by the strong El Niño Southern Oscillation events of the 1990s. Transfer of carbon between live and dead biomass pools appears to have led to substantial increases in the pool of CWD, causing the observed net carbon release. The data show that biometric studies of tropical forests neglecting CWD are unlikely to accurately determine carbon balance. Furthermore, the hypothesized sequestration flux from CO2 fertilization (<0.5 Mg C·ha−1·yr−1) would be comparatively small and masked for considerable periods by climate-driven shifts in forest structure and associated carbon balance in tropical forests.