Innovations in data integration for modeling populations
Innovations in data integration for modeling populations
复制标题
用于人口建模的数据集成创新
DOI：
10.1002/ecy.2713
复制
发表时间：
2019
期刊：
影响因子：
4.8
通讯作者：
Beissinger, Steven R.
中科院分区：
文献类型：
--
作者：
Zipkin, Elise F.;Inouye, Brian D.;Beissinger, Steven R.
评估人类世物种的状况需要了解影响种群动态的基本生态过程以及持续的气候和环境变化的影响。各种力量及其反馈的组合推动了物种在大空间和时间尺度上的分布和丰富度，这可能是使用传统生态工具进行分析的一个挑战。评估这些力量对人口动态的相对影响往往很困难，因为数据不足和/或由不同的类型组成。种群生态学的一个基本挑战是检测物种趋势，跨时空尺度外推推断，并创建对动态和生存能力的可信预测。因此，需要结合使用不同数据类型的多个交互过程的建模方法，并且确实，此类模型的开发是生态学中一个快速增长的研究领域。我们的专题文章说明了如何使用多种数据类型和建模方法，将它们集成在一起，以实现建立可靠的方法来评估和预测人口分布和丰度的共同目标。例如，对于生态学家来说，公民科学代表着一种相对较新且不断增长的数据来源，但这些项目的数据通常与从有重点的实验或专家监测中收集的数据类型不同。与使用单一数据类型相比，集成使用多种技术和/或在研究系统的不同方面收集的两种或更多种数据类型可以产生关于感兴趣的生态过程的更好信息。多种数据类型的整合还有助于理解影响人口动态和趋势的机制。综合模型的主要优点是：(1)能够通过减少单一数据集中固有的偏差来补偿数据收集中的可变性；(2)与从单独分析中获得的人口比率估计相比，提高了人口比率估计的精确度；(3)通过空间方法和人类相互作用产生适合于适应广泛环境过程的模型。
Assessing the status of species in the Anthropocene requires an understanding of basic ecological processes affecting population dynamics and the impacts of ongoing climate and environmental changes. Various combinations of forces and their feedbacks drive species distributions and abundances across large spatial and temporal scales, which can be a challenge to analyze with traditional ecological tools. Evaluating the relative effects of these forces on population dynamics is often difficult because data are insufficient and/or comprised of disparate types. A fundamental challenge of population ecology is to detect species trends, extrapolate inference across spatiotemporal scales, and create credible projections of dynamics and viability. Modeling approaches that incorporate multiple, interacting processes using dissimilar data types are thus needed, and indeed, the development of such models is a rapidly growing area of research within ecology.A meaning of “integrated” is to bring separate or disparate things together into a cohesive whole and this is the definition we use for the Special Feature. Our Special Feature contains papers illustrating ways to use multiple data types and modeling approaches, integrating them for the common goal of building robust methods to assess and project population distribution and abundance. For example, citizen science represents a relatively new and growing source of data for ecologists, but often data from these projects are not the same type of data that are collected from focused experiments or monitoring by experts. Integrating two or more data types, collected with multiple techniques and/or on different aspects of a study system, can yield better information about an ecological process of interest than use of a single data type. Integration of multiple data types can also facilitate understanding of the mechanisms that influence population dynamics and trends. The primary advantages of integrated models are (1) the ability to compensate for variability in data collection by reducing biases inherent in a single data set;(2) improved precision of estimates of demographic rates compared to those obtained from separate analyses; and (3) production of models suited to accommodate a broad suite of environmental processes through spatial methods and human interactions.