CAMEO: Multiscale modeling of Hawaii's coral reef communities

有效管理海洋生态系统的一个关键挑战是将小规模的分布和动态研究转化为区域规模的管理行动。在包括夏威夷群岛在内的许多海洋生态系统中，有来自多个调查人员的关于近岸社区的广泛调查数据，这意味着巨大的资源投资。这些数据往往没有得到充分利用，对管理人员的用处仍然有限。在夏威夷群岛，至少有七个不同的实体在不同程度的协调下从事珊瑚礁群落的调查。这些数据的合成需要在多个尺度上的综合建模方法。这项研究建立在现有的数据库，并扩展了两个现有的模型：珊瑚恢复模型（CRM）的随机珊瑚恢复后的干扰和COMBO模型的酸化和温度增加对珊瑚礁的协同影响。扩展从以前的工作是应用两个创新的建模方法（尺度转换理论和基本生态位建模）来预测珊瑚群落的组成和动态在区域尺度上。基本生态位建模使用多种数据拟合方法（回归，机器学习等）来描述物种与其环境之间的关系，使用分割数据集进行训练和验证。这种方法可以从离散的数据点生成预测和验证的空间连续的物种分布模型。尺度转换建模将使用完整的物种分布数据库作为物种相互作用发生的景观。这些相互作用是由一个本地模型描述的，在这里，基于珊瑚恢复模型的招聘，增长和死亡率。在尺度转换理论中，局部模型加上关于生物及其环境的分布和共分布的景观信息，预测物种组合如何响应该项目将产生四个与夏威夷群岛基于生态系统的管理有关的产品，产生研究界以外的重大影响：（1）夏威夷群岛范围内的珊瑚分布、底栖生物群落数据、鱼类调查和其他数据的地理信息系统数据库，（国家公园服务），在NOAA的各个部门，水生资源的夏威夷司，和其他来源到一个单一的地理信息系统数据库;（2）整个群岛的珊瑚物种分布的验证，预测和空间连续的地图;（3）经验证的夏威夷主要岛屿珊瑚礁减缓的珊瑚恢复模型;（4）根据已知和预测的珊瑚分布和COMBO模型，预测整个夏威夷群岛珊瑚群落对气候变化的反应。

A key challenge in the effective management of marine ecosystems is translating from small scale studies of distribution and dynamics to the regional scale of management action. In many marine ecosystems, including the Hawaiian Archipelago, there are extensive survey data of nearshore communities from multiple investigators, representing a huge investment of resources. Often, these data are underutilized and remain of limited use to managers. In the Hawaiian Archipelago, at least seven separate entities are engaged in surveys of coral reef communities, with varying degrees of coordination. The synthesis of these data requires integrated modeling approaches at multiple scales. This study builds on an existing database and extends two existing models: the Coral Recovery Model (CRM) of stochastic coral recovery after disturbance and the COMBO model of the synergistic impacts of increasing acidification and temperature on coral reefs. Extending from this prior work is the application of two innovative modeling approaches (scale transition theory and fundamental niche modeling) to predict coral community composition and dynamics at the regional scale. Fundamental niche modeling uses multiple data fitting approaches (regression, machine learning, etc) to describe the relationship between species and their environments, using a split dataset for training and validation. This approach can generate a predictive and validated spatially continuous model of species distribution from discrete data points. The scale transition modeling will use the completed database of species distributions as the landscape on which species interactions occur. These interactions are described by a local model, here, based on recruitment, growth, and mortality from the Coral Recovery Model. In scale transition theory, the local model plus landscape information on the distribution and co-distribution of organisms and their environments predicts how a species assemblage responds (locally and regionally) to changes in biotic and abiotic factors on the landscape.This project will generate four products relevant to ecosystem-based management of the Hawaiian Archipelago, resulting in significant impacts beyond the research community: (1) A Hawaiian Archipelago-wide GIS database of coral distribution, benthic community data, fish surveys, and other data gathered by CRAMP, NPS (National Park Service), various divisions in NOAA, the Hawaii Division of Aquatic Resources, and other sources into a single GIS database; (2) Validated, predictive, and spatially continuous maps of coral species distribution throughout the Archipelago; (3) A validated Coral Recovery Model for coral reef mitigation in the Main Hawaiian Islands; (4) Prediction of coral community response to climate change throughout the Hawaiian Archipelago, based on known and predicted coral distributions and the COMBO model.