Upgrading plant-functional-types with plant trait variability in ecohydrological models: A stochastic parameterization approach

结合碳循环和水循环的生态水文模型可以预测植被生长和土壤蓄水库因气候和土地使用扰动而发生的相关变化。 因此，它们对于帮助我们为生态系统和水资源脆弱性做好准备至关重要。 这种模型在确定植被和土壤的反馈方面也发挥着重要作用，这些反馈可以加速或减缓气候变化。 然而，这些模型中的大部分不确定性是由于它们目前如何表示植物而产生的。 这些模型将植物的巨大多样性简化为易于处理的数量。植物功能型？（PFT），每个PFT都有统一的、固定的参数集应用于它们。 尽管PFT将具有相似特征的相关物种分组在一起，但最近汇编的全球植物数据显示，某些植物特征在这些预先指定的PFT组中的变化与不同组之间的变化一样大。 由于这些植物特性会影响植物对水和CO2的吸收，因此本研究将植物性状变异性纳入生态水文模型，以改善对我们不断变化的生态系统，水资源和气候的未来预测。 生态水文模型中固定参数PFT的当前范例需要升级，以符合植物性状变异生态学的新发现。 拟议的工作提供了一种新的随机方法，模拟植物性状的可塑性，这些是适应，随着时间和空间的推移，在复杂的环境驱动程序，并引起PFT内的变异性。 一个时空随机PFT参数化将开发全球植物性状数据的基础上，捕捉内PFT变异的分布。 重要的是，参数化将进一步以时空生物和非生物数据为条件，以填补植物性状数据集的空白，并严格考虑将稀疏全球数据应用于这些模型的不确定性。 这种方法标志着一个新的出发点，从最近的建模工作，将性状变异性作为随机参数固定在空间和时间或确定性的输入，无法解决的不确定性。 随机PFT参数化将首先开发一个沙漠灌木林设置，这迫切需要一个新的模型表示，可以捕捉温度和水分驯化的植物。 新模型的模拟将揭示沙漠灌丛中植物性状与气候和土壤类型等环境变量之间的关系。 植物性状变异是普遍存在的所有PFT中，随机参数化方法在这项研究中产生的，因此，有利于全球生态水文模拟。

Ecohydrological models that incorporate carbon and water cycles can predict related changes in vegetation growth and soil water reservoirs in response to climate and land-use perturbations. They are thus critical for helping us prepare for ecosystem and water resource vulnerabilities. Such models also play an important role in determining feedbacks from vegetation and soils that can accelerate or slow climate change. However, much of the uncertainty in these models arises because of how they currently represent plants. These models simplify the enormous diversity of plants into a tractable number of ?plant-functional-types? (PFTs), each of which have uniform, fixed sets of parameters applied to them. Although PFTs group together related species with similar characteristics, recently compiled global plant data reveal that certain plant traits can vary just as much within these pre-specified PFT groups as between distinct groups. Because these plant properties can affect plant uptake of water and CO2, this study will incorporate plant trait variability into ecohydrological models in order to improve future predictions about our changing ecosystems, water resources, and climate. The current paradigm of fixed-parameter PFTs in ecohydrological models needs upgrading to align with new findings in ecology on plant trait variability. The proposed work offers a new stochastic approach that simulates plant trait plasticity; these are adaptations that occur over time and space in response to complex environmental drivers and give rise to variability within PFTs. A spatiotemporally stochastic PFT parameterization will be developed based on global plant trait data that capture distributions of intra-PFT variability. Importantly, the parameterization will be further conditioned on spatiotemporal biotic and abiotic data to fill gaps in the plant trait datasets and rigorously account for uncertainties in applying sparse global data to these models. This approach marks a novel departure from recent modeling efforts that incorporate trait variability as random parameters fixed in space and time or as deterministic inputs that fail to address uncertainties. The stochastic PFT parameterization will be first developed for a desert shrubland setting, which critically needs a new model representation that can capture temperature and moisture acclimation by its plants. Simulations with the new model will reveal relationships in desert shrublands between plant traits and environmental variables such as climate and soil type. Plant trait variability is ubiquitous among all PFTs; the stochastic parameterization approach generated in this study will thus benefit ecohydrological modeling globally.