ABI Innovation: Quantifying, simulating, and visualizing the tree growth and its antecedent endogenous and climatic predictors

过去的环境条件可能对预测当前的过程很重要，例如生态系统生产力和树木生长。此外，过去的环境条件可能会与过去的树木生长模式相互作用，从而影响当前的树木生长。了解过去条件在预测树木当前和未来生长反应方面的作用，对于预测树木、森林和整个陆地生物圈可能如何受到未来环境变化的影响非常重要，例如那些预计将与气候变化同时发生的环境变化。然而，用于推断过去条件对当前过程(例如，树的生长)的重要性的方法并未得到很好的开发。因此，这项研究开发了统计和计算方法，以量化过去(以前)影响树木生长的因素，特别侧重于确定过去的气候条件和过去的树木生长模式影响当前树木生长的时间尺度。这项研究利用了美国西南部多个树种在线提供的大量现有树木年生长数据(树轮)，并将通过对该地区几个树种进行抽样的重点实地研究来提供更多数据。该项目将培养至少三名来自不同学科(生态学/生物学、建模/统计学、计算/软件开发)的研究生和几名本科生。这项研究产生的结果和数据将用于开发与新的本科课程和旨在培训早期职业科学家的短期课程有关的教育模块。过去几天、几周、几个月、四个季节或几年的平均环境条件是植物和生态系统生产力的重要预测因子。例如，树木年轮研究表明树木生长受到先行外源(例如，过去的气候)和内源(例如，过去的年轮宽度)因素的影响，但现有的分析方法没有明确地评估每个因素的作用。为了满足这一需要，本研究将(1)建立一个随机先行模型(SAM)来量化先行气候和内生条件及其对树木生长的影响，(2)应用随机先行模型来估计先行因素影响西南地区多个地点的多个树种的树木生长的时间尺度，(3)识别潜在的先行影响树木生长的生理机制，以及(4)为SAM的更广泛的应用开发软件，以及用于模拟和可视化树木生长。这项研究结合了大数据集、实地研究、文献数据、贝叶斯综合和基于个体的树木生长和生理模型(IBM)。这项研究有望深入了解树木作为过去环境和生理状态的集成者所扮演的角色，SAM方法应该会提高我们预测植物和生态系统对气候变化、干扰或其他扰动的反应的能力。重要的是，用于量化先行条件的时间特性及其对感兴趣过程的影响的一般SAM框架将广泛适用于广泛的领域。为了实现上述研究目标，该项目将为1名博士后、3名研究生(生态学、统计学、计算机科学)和多名本科生提供跨学科培训。这项研究的成果将被纳入亚利桑那州立大学新的本科生研讨会、为年度生态学会议(ESA)组织的贝叶斯建模研讨会，以及针对早期职业和高级生态科学家的为期两周的贝叶斯方法夏季课程。将开发两个软件(R)包，一个用于SAM，一个用于IBM。这项研究产生的数据将通过数据储存库向公众和全球提供。该项目的结果可以在www.ogle.Lab.asu edu上找到。

Past environmental conditions are likely to be important for predicting current processes such as ecosystem productivity and tree growth. Moreover, past environmental conditions may interact with past tree growth patterns to affect current tree growth. Understanding the role of past conditions for predicting current and future growth responses of trees is expected to be important for predicting how trees, forests, and the terrestrial biosphere in general, may be impacted by future environmental changes, such as those anticipated to occur in tandem with climate change. However, methods for inferring the importance of past conditions for current processes (e.g., tree growth) are not well developed. Thus, this study develops statistical and computing methods for quantifying the past (antecedent) factors governing tree growth, with specific focus on identifying the time scales over which past climate conditions and past tree growth patterns affect current tree growth. This study draws upon large amounts of existing data on annual tree growth ("tree rings") available on-line for multiple tree species across the southwestern US, and it will contribute additional data via a focused field study that will sample several species in this region. The project will train at least three graduate students from diverse disciplines (ecology/biology, modeling/statistics, computing/software development) and several undergraduate students. Results and data generated by this study will be used to develop education modules related to new undergraduate coursework and short courses aimed at training early career scientists. Environmental conditions averaged over past days, weeks, months, seasons, or years are important predictors of plant and ecosystem productivity. For example, tree-ring studies indicate that tree growth is affected by antecedent exogenous (e.g., past climate) and endogenous (e.g., past ring widths) factors, but existing analysis methods do not explicitly evaluate the role of each factor.To address this need, this study will (1) develop a stochastic antecedent model (SAM) for quantifying antecedent climatic and endogenous conditions and their influence on tree growth, (2) apply the SAM to estimate the time-scales over which antecedent factors affect tree growth for multiple species across multiple sites in the Southwest, (3) identify potential physiological mechanisms underlying the antecedent effects on tree growth, and (4) develop software for more general applications of SAM and for simulating and visualizing tree growth. This study combines large datasets, field studies, literature data, Bayesian synthesis, and an individual-based model (IBM) of tree growth and physiology. This study is expected to lend insight into the role trees play as integrators of past environments and physiological states, and the SAM approach should improve our ability to forecast plant and ecosystem responses to climate change, disturbances, or other perturbations. Importantly, the general SAM framework for quantifying the temporal properties of the antecedent conditions and their effects on the process of interest will be broadly applicable to a wide range of fields. To address the aforementioned research objectives, this project will provide interdisciplinary training for a post-doc, 3 graduate students (ecology, statistics, computer science), and multiple undergraduates. Products from this study will be incorporated into a new undergraduate seminar at ASU, a Bayesian modeling workshop organized for annual ecology meetings (ESA), and a 2-week summer course in Bayesian methods aimed at early career and senior ecological scientists. Two software (R) packages will be developed, one for SAM and one for the IBM. Data generated from this study will be made publically and globally available via data repositories. Results from the project can be found at: www.ogle.lab.asu.edu.