Functional traits offer generalizability to the prediction of ecosystem processes such as production, and community-weighted mean trait values are increasingly used for such predictions. However, the underlying causal direction between traits and ecosystem processes are often indirect and sometimes even tenuous. In this study, we aimed to uncover underlying causal mechanisms between traits, habitat adaptation and canopy productivity. We used canopy production data estimated from leaf litter traps, and trait and habitat association data obtained from 40 permanent vegetation plots in the Nee Soon catchment in Singapore, which contains a heterogeneous mix of freshwater swamp and dry-land tropical forests. Mean canopy production across the catchment was estimated to be 768 g m⁻² year⁻¹, which is similar to other tropical dry-land forests in the region. Biweekly per-basal-area canopy production was found to be consistently lower in swamp than non-swamp plots, and positively correlated with monthly mean temperature. Structural equation models fitted to data of canopy production, leaf traits, plot type (swamp versus non-swamp), basal areas and habitat adaptations of 69 tree species–plot combinations suggested that tree species possessing leaf traits associated with more conservative resource acquisition strategies, viz. low specific leaf area, high leaf C:N ratio, and thicker leaves, are better adapted to stressful, waterlogged swamp conditions, but that this adaptation also reduces canopy—and likely total—net primary productivity. These observations suggest that the stressful conditions of waterlogged, anoxic swamp habitats significantly reduce the rate at which nutrients are cycled by communities found in such environments.

功能特征为预测生态系统过程（如生产）提供了普遍性，社区加权平均特征值越来越多地用于此类预测。然而，特征和生态系统过程之间的潜在因果关系通常是间接的，有时甚至是脆弱的。在这项研究中，我们旨在揭示性状、栖息地适应和冠层生产力之间的潜在因果机制。我们使用了从枯枝落叶陷阱估计的冠层生产数据，以及从新加坡 Nee Soon 集水区的 40 个永久植被地块获得的特征和栖息地关联数据，其中包含淡水沼泽和旱地热带森林的异质混合。整个流域的平均树冠产量估计为 768 g m ^-2 年^-1，这类似于该地区的其他热带旱地森林。发现沼泽地每两周一次的基础面积冠层产量始终低于非沼泽地块，并且与月平均温度呈正相关。结构方程模型拟合了 69 种树种-地块组合的冠层产量、叶特征、地块类型（沼泽与非沼泽）、基区和栖息地适应性数据，表明具有叶特征的树种与更保守的资源获取策略相关，即。低比叶面积、高叶 C:N 比和更厚的叶子更能适应压力大、淹水的沼泽条件，但这种适应也降低了冠层——可能还有总的——净初级生产力。这些观察结果表明，淹水的压力条件，