Spatial and temporal variability of soil Carbon dioxide and Nitrous oxide fluxes in tropical forest soils: The influence of tree species, precipitation, and soil texture

This thesis presents data collected and analyzed to determine the influence of above ground biological factors on the spatial and temporal variability of tropical soil biogeochemical processes. The tropics are the largest natural source of CO2 and N2O to the atmosphere and many tropical forests are changing due to climate or land-use change and forest fragmentation. To understand how these changes are affecting ecosystem feedbacks to climate change we need to understand the plant-soil interactions in tropical forests and how these scale to the whole forest and region. For this thesis, I completed four studies that investigated the interaction of the tree species and forest growth with soil trace gas fluxes within the Amazon basin.;Locally, tree species do influence soil N2O fluxes, with fluxes close to four out of fifteen tree species consistently elevated above the overall mean and consistently low near two species. These results suggest that tree species composition can significantly influence soil biogeochemistry. Experimental sugar additions elevated N2O fluxes, suggesting that N2O cycling is mostly driven by heterotrophic (carbon-limited) denitrifiers, and that carbon transport into the soil by trees may present a mechanism for the observed differences. Alternatively, tree-soil competition for nutrients could explain the tree species related soil N2O flux differences. However, soil nitrate concentrations and low N2O fluxes associated with legumes, suggest that nitrogen is not limiting. This work provides evidence that vegetation species composition can be a controlling factor in overall trace gas emissions in tropical forests, although I found that only rather large (20%) change in species composition would cause an appreciable effect on the landscape-scale forest soil gas flux.;In order to better isolate the effects of species on soil processes, I also investigated soil gas fluxes in a tropical plantation, where trees were grown in monoculture plots. As in the natural forest, different tree species were associated with flux differences, though the species N2 O flux rank order was unlike what I found in the forest, indicating that monoculture plantation plots are not generally informative for species influence on soil processes in diverse forests. Fast tree growth rates and overall lower fluxes of CO2 and N2O from plantation vs. natural forest or agricultural soils suggest that conversion of abandoned farmland to plantation may reduce greenhouse gas fluxes to the atmosphere in tropical systems, a net benefit from a climate change policy perspective.;Finally, motivated by the finding that tree carbon export may influence N2O gas production in soils at the local scale, I investigated the effect of variation in growth rates of forest stands on regional variations in N2O production. I found that site-to-site and monthly flux variability within the Tapajos National Forest (south of Santarem, Brazil) and across the central-eastern Amazon basin is significantly and positively correlated with forest growth rates. I hypothesize that this relationship is a consequence of: (1) the dominance of denitrification (as opposed to nitrification) in driving N2O emissions in wet, clay-rich tropical soils that are conducive to anaerobic metabolism (especially in wet seasons); (2) the requirement of denitrifiers, as obligatory heterotrophs, for labile carbon, which is limited in tropical soil, and (3) the fact that downward sap flow in trees is the main source of carbon for both stem growth and root exudation. The plausibility of this hypothesis is supported by results from a process-based, numerical model (PnETDNDC), which, when trained to climatic and ecological data of our sites, reproduces the wood growth and soil N2O flux patterns we observed. Extrapolating the regionally observed correlation between stand-level tree growth and N2O fluxes, to the Amazon basin as a whole using wood growth measurements from forest inventory plots, yields a mean soil N2O flux of 2.6 kg-N ha-1 y-1 for the whole Amazon basin, higher than most previously published estimates. This suggests that tropical forests may be a bigger contributor to global N2O budget than previously thought, a consequence of accounting for the importance of processes that link vegetation carbon dynamics to soil biogeochemistry.