Subtropical and tropical forests account for over 50% of soil CO₂ production, 47% of N₂O fluxes of natural ecosystems, and act as both significant sources and sinks of atmospheric CH₄. However, ecosystem-scale estimates of these fluxes typically do not account for uncertainty that arises from environmental heterogeneity over small spatial scales. To assess the effects of small-scale environmental heterogeneity on GHG fluxes in a tropical forest ecosystem, we measured fluxes of CO₂, CH₄, and N₂O across a topographic gradient and at the base of different tree species. We then used Bayesian linear models together with maps of trees and topography to quantify spatial heterogeneity in ecosystem-scale estimates of GHG emissions. The relationship between GHG fluxes and species and topography varied for each gas type. CO₂ varied strongly by species but was only weakly related to topographic variation. In contrast, CH₄ and N₂O, which are more strongly regulated by soil oxygen, had strong relationships with topography but did not vary across species. Assuming spatial homogeneity and average rainfall conditions, we estimated ecosystem soil CO₂ emissions to be 28.91 kg CO₂-C/ha/day, net CH₄ consumption of − 5.15 g CH₄-C/ha/day, and net N₂O emissions of 1.78 g N₂O-N/ha/day. Including variation caused by tree species decreased ecosystem-level estimates of CO₂ emissions by 8.03%, whereas including topographic variation decreased net CH₄ consumption by 12.98% and increased net N₂O emissions by 1.05%. This translates to a net decrease of 8.32% in estimated CO₂-equivalent emissions. Our findings show that ignoring small-scale environmental heterogeneity has implications for bottom-up estimates of GHG fluxes in tropical forests. Given the increasing availability of fine-scale topographic models, incorporating this source of variation in estimates of ecosystem soil GHG emissions could improve our understanding of the role tropical forests play in global GHG cycles.

亚热带和热带森林占土壤 CO ₂产量的50% 以上，占自然生态系统N ₂ O 通量的47% ，并且是大气 CH _{4 的}重要来源和汇。然而，这些通量的生态系统尺度估计通常不考虑由小空间尺度上的环境异质性引起的不确定性。为了评估小尺度环境异质性对热带森林生态系统温室气体通量的影响，我们测量了 CO ₂、CH ₄和 N _{2 的}通量O 跨越地形梯度和不同树种的底部。然后，我们使用贝叶斯线性模型以及树木和地形图来量化生态系统规模温室气体排放估算的空间异质性。每种气体类型的温室气体通量与物种和地形之间的关系各不相同。CO ₂因物种而异，但与地形变化的相关性较弱。相比之下，CH ₄和N ₂ O 受土壤氧的调节更强，与地形有很强的关系，但在不同物种之间没有变化。假设空间均匀性和平均降雨量条件，我们估计生态系统土壤 CO ₂排放量为 28.91 kg CO ₂ -C/ha/天，净 CH ₄− 5.15 g CH ₄ -C/ha/天的消耗量，以及1.78 g N ₂ O-N/ha/天的N ₂ O净排放量。包括由树种引起的变化，CO ₂排放量的生态系统水平估计值降低了 8.03%，而包括地形变化在内，CH _{4 的}净消耗量减少了 12.98%，N ₂ O 的净排放量增加了 1.05%。这意味着估计的 CO ₂净减少 8.32%- 等效排放。我们的研究结果表明，忽略小规模环境异质性对热带森林温室气体通量的自下而上估计有影响。鉴于精细尺度地形模型的可用性越来越高，将这种变异来源纳入生态系统土壤温室气体排放的估计中，可以提高我们对热带森林在全球温室气体循环中所起作用的理解。