Tropical forests represent major uncertainties in climate models and have the potential to act as both net carbon sources and sinks in the future. Projections that hurricanes will be an increasingly powerful disturbance in many tropical forests further complicate our ability to predict how these ecosystems will respond to climate change. By understanding how environmental variation at small spatial scales affects ecosystem processes shaping present-day forests, it may be possible to improve our predictions for how these forests will change in the future. This dissertation consists of three chapters examining the spatial patterns of tree species and soil greenhouse gas fluxes in a tropical forest in the Luquillo Experimental Forest, Puerto Rico. Disentangling the forces that drive the spatial distribution of tree species has been a foundational question in ecology and determining the relative importance of these forces is central to understanding spatial variation in soil biogeochemistry. In chapter 1, I use percolation threshold analysis to examine the clustering patterns of simulated and real tree spatial point patterns to understand the role that environmental filtering and density dependent processes play in shaping tree species distributions. I demonstrate that percolation threshold analysis successfully distinguishes thinning by random, environmental filtering, and density dependent processes. Additionally, the relative importance of these thinning processes varies by species’ traits; fast growing species with low LMA and shade intolerance have stronger evidence of density dependent processes compared to species with high LMA and shade tolerance. In chapter 2, I examine the spatial relationships between soil greenhouse gas fluxes and two proximal drivers of soil environmental variation: tree species and topography. I also examine how incorporating small-scale variation in greenhouse fluxes affects our scaled-up estimates of ecosystem greenhouse gas emissions. I show that including species effects improves estimates of soil CO2 fluxes, and including measures of topography improve estimates of CH4 and N2O fluxes. Incorporating spatial variation in GHG fluxes related to tree species and topography into our estimates of ecosystem GHG emissions decreased estimates of the total CO2-equivalent emissions in this forest by 5%. Finally, in chapter 3 I examine how the GHG fluxes in this forest change after an intense hurricane. I demonstrate that GHG emissions shift following a hurricane; this shift is primarily driven by a 176% increase in N2O emissions that represent a significant net loss of gaseous nitrogen from this forest. N2O fluxes accounted for 4.2% of the post-hurricane GHG-induced radiative forcing (compared to 1.8% pre-hurricane) and the combined increase in CO2, CH4, and N2O emissions observed translates to a 25% increase in CO2-equivalent emissions compared to pre-hurricane conditions. This dissertation focuses on the role of small-scale environmental variation in shaping forest communities and spatial patterns of GHG fluxes and aims to highlight how this variation can help us to better understand the role tropical forests play in the biosphere.