Collaborative Research: ABI Development: Creating a generic workflow for scaling up the production of species ranges

预测一个物种可以生活在哪里以及它如何应对气候变化的科学仍处于起步阶段。一个物种的地理范围是一个物种可以被发现的地图。它是了解物种生态和进化的基础，在保护中发挥着越来越重要的作用。涵盖了3万种陆生脊椎动物中的大部分物种范围的藏品已经可供科学分析。然而，地球上其他95%的物种都是罕见的。改变这种状况的时机已经成熟。从生物清单、博物馆、公民科学和以前资助的研究中获得大量数据的新途径意味着可以获得足够的数据来估计更多物种的范围。然而，我们目前缺乏可靠的预测方法和计算工具来产生大量的范围。该项目将开发新的计算方法和算法，以预测成千上万个研究不足的物种范围的现状和未来命运。这些方法将用于预测新世界的10万多种植物如何应对气候变化。研究人员将测试保护生物学中关于物种如何应对气候变化的关键假设，以及北美和南美多样性热点地区的地理稳定性已经/将随着时间的推移而变化。他们工作的最终结果将成为生态社区的一种新工具，具有指导生物采样策略的巨大潜力，特别是在保护和公民科学应用方面。提出的研究将考察生物多样性热点是否随时间保持不变，物种气候生态位是否在系统发育上保守，这两个隐含的假设在保护生物学和基础生态学中具有广泛的意义。本研究将开发一个工作流程，通过结合发生数据和GIS数据来预测任何分类类群的物种范围。这一工作流程将应用于所有新大陆植物研究的基本问题，如物种丰富度如何随时间和空间变化（这一主题几乎只研究脊椎动物和树木）。在计算方面，该项目将解决数据清洗、生态位建模实践、新型生态位建模方法和大系统发育分析方法方面的核心挑战。一个免费提供的通用管道将能够将生物多样性发生数据与物种范围联系起来，并将这些计算扩展到1000或10万种物种。这条综合管道将通过以下方式实施：1)适当地清洗数据以消除分类和地理错误；2)确定适用于不同物种的范围建模的明确最佳实践方法；3)创新范围建模方法，整合不同数据（如仅存在的博物馆藏品和基于丰度的plot数据）；4)在HPC环境中缩放计算密集型范围建模；5)将产品输出置于系统发育背景中。该项目将利用一个新的数据库来开发这样一个管道，该数据库包含了对新世界10万多种植物的2000万次观察。所产生的范围预测将用于测试保护生物学中关于物种气候生态位的系统发育保守性和多样性热点随时间的地理稳定性的关键假设。这项研究将通过开发科学代码库，为从初级生物多样性数据到高质量物种范围的生产做出重大贡献。该项目的结果可以通过以下网站（http://bien.nceas.ucsb.edu/bien/和bien3.org）找到。

The science of forecasting where a species can live and how it responds to climate change is still in its infancy. A species' geographic range is the map of where a species can be found. It is fundamental to understanding species' ecology and evolution and increasingly plays a vital role in conservation. Collections of species ranges covering most of the 30,000 terrestrial vertebrate species are already available for scientific analysis. However, collections of species ranges from the other ~95% of species on the planet are rare. The time is ripe to change this. New access to vast quantities of data from biological inventories, museums, citizen science, and previously funded studies mean that adequate data are available to estimate the ranges of many more species. However, we are currently missing robust forecasting methods and the computational tools to produce large numbers of ranges. This project will develop the novel computational methods and algorithms needed to forecast the current state and future fate of the many thousands of poorly studied species ranges. These methods will be applied to forecast how 100,000+ plant species in the New World will respond to climate change. The researchers will test key assumptions in conservation biology about how species respond to changing climate and the geographic constancy of diversity hotspots across North and South America have/will change over time. The end result of their work will be a novel tool for the ecological community that has tremendous potential to guide biological sampling strategies, particularly in conservation and citizen science applications. The proposed research will examine whether biodiversity hotspots are constant through time and whether species climatic niches are phylogenetically conserved, two implicit assumptions with wide-reaching implications in conservation biology and basic ecology. This research will develop a workflow to predict species ranges for any taxonomic group using by combining occurrence data with GIS data. This workflow will be applied to all New World plants to study basic questions, such as how species richness varies across space and time (a topic studied almost exclusively in vertebrates and trees). Computationally, the project will address core challenges in data scrubbing, niche modeling practices, novel niche modeling methods, and mega-phylogeny analysis methods. A freely available generic pipeline will be capable of linking biodiversity occurrence data to species ranges and scaling these computations to 1000s or 100,000s of species. This integrated pipeline will be implemented by: 1) appropriately scrubbing data to remove taxonomic and geographic errors, 2) identifying clear best practice methods for range modeling applicable across diverse species, 3) innovating range modeling methods that integrate diverse data such as presence only museum collections and abundance-based plot data 4) scaling computationally-intensive range modeling in an HPC environment, and 5) placing the outputs of the products in a phylogenetic context. This project will develop such a pipeline using a novel database of 20,000,000 observations of 100,000+ species of plants in the New World. The range forecasts produced will be used to test key assumptions in conservation biology about the phylogenetic conservatism of species climatic niches and the geographic constancy of diversity hotspots over time. This research will make substantial contributions to scientific infrastructure through the development of a scientific codebase for the production of high-quality species ranges from primary biodiversity data. The results of the project can be found via the following websites (http://bien.nceas.ucsb.edu/bien/ and bien3.org).