SWQU: Composable Next Generation Software Framework for Space Weather Data Assimilation and Uncertainty Quantification

该项目旨在通过汇集地球空间科学、不确定性量化以及软件开发、管理和可持续性方面的专家，开发下一代空间天气建模和预测软件。控制现代生活的电子技术深深依赖于全球定位系统（GPS）等卫星技术，GPS帮助我们在世界各地导航城市和管理空中交通。这些卫星实际上也是技术和商业的计时标准，用于同步全球银行交易、我们的智能手机和互联网。然而，其信号的精度和鲁棒性极易受到地球空间干扰。卫星轨道必须精确控制，为此，必须提前预测地球空间的扰动。具有可量化预测能力的空间天气模型是继续推进卫星技术和依赖于它们的一切所需的工具，但目前在很大程度上是缺乏的。在该项目下开发的可组合软件框架将作为一个基础，可以随着时间的推移而扩展和改进，从而增加空间天气预报能力和空间天气建模社区。可组合软件是计算科学的坩埚，它允许科学家在不重复他人工作的情况下为数值领域做出贡献。该项目的目标是建立下一代空间天气不确定性量化和数据同化框架，作为该领域日益增长的计算工具的基础。该框架的“必备条件”将是基于调度的现代可组合性、可再现性、易用性、性能、可移植性，以及对当前和未来的异构和新型体系结构的可扩展性。该项目将为数据驱动的空间天气模型提供计算上可扩展的算法和开源的基于julia的软件框架，具有以下属性：1)可组合性：当功能和行为协同工作时，软件是可组合的。例如，如果不确定度量化可以应用于程序而无需重写，则程序和不确定度量化组成；ii)可持续性：当程序的作者可以离开项目，项目的新成员可以维护软件时，软件是可持续性的；iii)可移植性：当软件可以在具有各种底层架构的异构硬件上运行时，它就是可移植性的；iv)可重复性：当存在方便且向后兼容的路径供用户随时检查、运行、共享和修改代码时，软件在今天和将来都是可重复性的。该合同是NSF-NASA联合试点项目“具有量化不确定性的空间天气数据驱动模型的下一代软件”（SWQU）的一部分。它由美国国家科学基金会大气与地球空间科学部、数学科学部、先进网络基础设施办公室和多学科活动办公室支持。所有因该奖项而开发的软件将由获奖者免费提供，用于非商业用途；软件许可证将允许修改和重新分发软件，免费用于非商业用途。该奖项反映了美国国家科学基金会的法定使命，并通过使用基金会的知识价值和更广泛的影响审查标准进行评估，被认为值得支持。

This project seeks to develop the next generation of software for space weather modeling and prediction by bringing together experts in geospace sciences, uncertainty quantification, and software development, management, and sustainability. The electronic technologies that govern modern life are deeply dependent on satellite technologies such as the Global Positioning System (GPS), which helps us navigate cities and manages air-traffic all over the world. These satellites are also the de facto timing standard of technology and commerce, used to synchronize banking transactions worldwide, our smart-phones, and the internet. Yet, the accuracy and robustness of their signals are highly vulnerable to geospace disturbances. Satellite orbits have to be controlled precisely and to do so, geospace disturbances have to be predicted in advance. Space weather models with quantifiable predictive capability are the tools that are needed, and are presently largely absent, to continue to advance the satellite technologies and everything that depends on them. The composable software framework to be developed under this project will serve as a foundation that can be expanded on and improved over time, growing both the space weather prediction capabilities and the space weather modeling community.Composable software is the crucible of computational science, allowing scientists to add their contribution to the numerical realm without having to repeat the work of others. The goal of this project is to build a next-generation framework for space weather uncertainty quantification and data assimilation, as the foundation of the growing body of computational tools for the field. The "must-haves" for this framework will be modern dispatch-based composability, reproducibility, ease-of-use, performance, portability, and extendability to today's and tomorrow's heterogeneous and novel architectures. This project will produce computationally scalable algorithms and open-source Julia-based software framework for data-driven models of space weather with the following properties: i) Composability: Software is composable when features and behaviors work together. For example, if uncertainty quantification can be applied to a program without a rewrite, the program and the uncertainty quantification compose; ii) Sustainability: Software is sustainable when the author of the program can leave the project, and new members of the project can maintain the software; iii) Portability: Software is portable when it can perform on heterogeneous hardware with a variety of underlying architectures; iv) Reproducibility: Software is reproducible today and into the future when a convenient and backward compatible pathway exists for users to readily examine, run, share, and modify code.This award is made as a part of the joint NSF-NASA pilot program on Next Generation Software for Data-driven Models of Space Weather with Quantified Uncertainties (SWQU). It is supported by NSF Division of Atmospheric and Geospace Sciences, Division of Mathematical Sciences, Office of Advanced Cyberinfrastructure, and Office of Multidisciplinary Activities. All software developed as a result of this award will be made available by the awardee free of charge for non-commercial use; the software license will permit modification and redistribution of the software free of charge for non-commercial use.This award reflects NSF's statutory mission and has been deemed worthy of support through evaluation using the Foundation's intellectual merit and broader impacts review criteria.