Paper data of "<b>Linking leaf dark respiration to leaf traits and reflectance spectroscopy across diverse forest types"</b> in <i>New Phytologist</i>

Leaf dark respiration (Rdark) is an important but minimally quantified component of carbon cycling in forest ecosystems. In terrestrial biosphere models, it is typically simulated through leaf traits such as maximum carboxylation capacity (Vcmax), leaf mass per unit area (LMA), nitrogen and phosphorus concentrations. However, the effectiveness of these relationships between different forest types still needs to be thoroughly evaluated. We collected canopy leaves of three typical temperate (mountain Changbai, 42°23' N, 128°05' E), subtropical (mountain GT, 29°15' N, 118°07' E) and tropical (Xishuangbanna, 21°37' N, 101°34' E) forests in China through the Chinese Academy of Sciences canopy tower crane observation network. Rdark and Vcmax were measured using a portable photosynthesis measurement system, leaf reflectance spectra were measured using a portable ground spectrometer, and leaf morphology and biochemical traits were measured using conventional laboratory methods. We found that leaf magnesium and calcium concentrations are more important in explaining cross site Rdark than commonly used traits such as LMA, nitrogen and phosphorus concentrations, but the univariate trait Rdark relationship remains weak (r2 ≤ 0.15) and varies depending on the forest. Although the multivariate relationship of leaf traits improves model performance, leaf spectroscopy outperforms trait Rdark relationship, accurately predicting cross site Rdark (r2=0.65) and identifying factors that lead to Rdark variation. Our research findings reveal some new traits with greater cross site scalability, challenging the use of empirical trait Rdark relationships in process models, and emphasizing the potential of leaf spectroscopy as a promising alternative method for estimating Rdark, ultimately improving the modeling of terrestrial plant respiration processes.叶片暗呼吸（Leaf dark respiration, Rdark）是森林生态系统碳循环中一项重要但量化程度极低的组分。在陆地生物圈模型中，其通常通过叶片性状进行模拟，例如最大羧化速率（maximum carboxylation capacity, Vcmax）、比叶质量（leaf mass per unit area, LMA）以及氮、磷浓度。然而，这些关联在不同森林类型间的有效性仍有待全面评估。 本研究依托中国科学院冠层塔吊观测网络，采集了中国境内三类典型森林的冠层叶片：温带森林（长白山，42°23' N，128°05' E）、亚热带森林（GT山，29°15' N，118°07' E）以及热带森林（西双版纳，21°37' N，101°34' E）。采用便携式光合测量系统测定Rdark与Vcmax，使用便携式地面光谱仪采集叶片反射光谱，并通过常规实验室方法测定叶片形态与生化性状。 研究结果显示，相较于LMA、氮磷浓度等常用性状，叶片镁、钙浓度对样点间Rdark变异的解释度更高，但单性状与Rdark的关联依然较弱（决定系数r²≤0.15），且因森林类型而异。尽管多性状关联可提升模型性能，但叶片光谱法的表现优于性状-Rdark关联模型，能够精准预测样点间Rdark（决定系数r²=0.65），并明确驱动Rdark变异的关键因子。 本研究揭示了一批跨样点通用性更强的新性状，对过程模型中经验性性状-Rdark关联的应用提出了挑战，同时强调了叶片光谱法作为估算Rdark的可靠替代方法的潜力，最终可优化陆地植物呼吸过程的模型模拟。