NSF Convergence Accelerator - Track D: Hidden Water and Hydrologic Extremes: A Groundwater Data Platform for Machine Learning and Water Management

NSF 融合加速器 - 轨道 D：隐藏水和水文极端情况：用于机器学习和水管理的地下水数据平台

• 批准号: 2040542
• 负责人: Laura Condon
• 金额: $ 100万
• 依托单位: University of Arizona
• 依托单位国家: 美国
• 项目类别: Standard Grant
• 财政年份: 2020
• 资助国家: 美国
• 起止时间: 2020-09-15 至 2022-05-31
• 项目状态: 已结题
• 来源: https://www.nsf.gov/awardsearch/showAward?AWD_ID=2040542&HistoricalAwards=false
• 关键词: NSF; Convergence; Accelerator; Track; Hidden

The NSF Convergence Accelerator supports use-inspired, team-based, multidisciplinary efforts that address challenges of national importance and will produce deliverables of value to society in the near future. The broader impact and potential societal benefit of this Convergence Accelerator Phase I project is to utilize artificial intelligence methods such as machine learning (ML) to achieve better water management outcomes that directly benefit society by developing the ability to better plan for and manage extreme events through improved hydrologic forecasting. HydroFrame-ML is motivated by, and structured around, applied solutions for water management planning and decision making. Extreme events like drought and floods have far-reaching societal impacts. They are common, costly and likely to get worse in the future. The project team is partnered with the Bureau of Reclamation, which is the largest wholesale water provider in the country, providing water to more than 31 million people and 10 million acres of farmland. The Bureau of Reclamation will drive use case design and the metrics used to evaluate success in Phase 1, as well as partner in the expansion of the project team for Phase 2. Additionally, the project team will develop hands-on activities and challenges designed to give undergraduates experience in machine learning and data science, in the context of pressing real-world challenges. Aided by the planned addition of a STEM mentorship program partner in Phase 2, the team will build content with the vision of helping to broaden participation of underrepresented students well beyond the timeframe of this project.The proposed project brings together the most physically rigorous national scale groundwater simulations developed through HydroFrame with national leaders in Earth Systems Modeling and water management. By providing end-to-end workflows combining state of groundwater science with operational management tools, HydroFrame-ML will advance both large-scale water management as well as our understanding of how human operations and groundwater interact in extreme events. Their products will provide innovative ways to improve forecasts and in the process will expand our knowledge about the (1) contributions of groundwater to extreme events in managed systems; (2) biases in our current risk-assessment approaches which do not consider groundwater; and (3) potential to improve long-term sustainability by more actively managing groundwater and accounting for groundwater surface water interactions in projections.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.

NSF 融合加速器支持以使用为基础的、基于团队的多学科努力，以解决国家重大挑战，并将在不久的将来为社会提供有价值的成果。该融合加速器第一阶段项目的更广泛影响和潜在社会效益是利用机器学习（ML）等人工智能方法来实现更好的水管理成果，通过发展更好地规划和管理极端事件的能力，直接造福社会。改进水文预报。 HydroFrame-ML 的灵感来自于水管理规划和决策的应用解决方案，并围绕其构建。干旱和洪水等极端事件具有深远的社会影响。它们很常见，成本高昂，而且未来可能会变得更糟。该项目团队与农垦局合作，该局是该国最大的批发水供应商，为超过 3100 万人和 1000 万英亩农田提供水。垦务局将推动用例设计和用于评估第一阶段成功的指标，并与合作伙伴扩大第二阶段的项目团队。此外，项目团队将开发实践活动和挑战，旨在在紧迫的现实挑战的背景下，为本科生提供机器学习和数据科学的经验。在第二阶段计划增加 STEM 辅导计划合作伙伴的帮助下，该团队将构建内容，其愿景是在该项目的时间范围之外帮助扩大代表性不足的学生的参与。拟议的项目汇集了最严格的国家规模通过 HydroFrame 与地球系统建模和水管理领域的国家领导者共同开发地下水模拟。通过提供将地下水科学状况与运营管理工具相结合的端到端工作流程，HydroFrame-ML 将促进大规模水管理以及我​​们对人类活动和地下水在极端事件中如何相互作用的理解。他们的产品将提供创新的方法来改进预测，并在此过程中扩大我们对以下方面的认识：(1)地下水对管理系统中极端事件的贡献； (2) 我们目前的风险评估方法存在偏差，没有考虑地下水； (3) 通过更积极地管理地下水和在预测中考虑地下水与地表水的相互作用来提高长期可持续性的潜力。该奖项反映了 NSF 的法定使命，并通过使用基金会的智力价值和更广泛的影响审查进行评估，被认为值得支持标准。

项目成果

A Physics-Informed, Machine Learning Emulator of a 2D Surface Water Model: What Temporal Networks and Simulation-Based Inference Can Help Us Learn about Hydrologic Processes

• DOI: 10.3390/w13243633
• 发表时间: 2021-12
• 期刊: Water
• 影响因子: 3.4
• 作者: R. Maxwell;L. Condon;Peter Melchior

Development of a Deep Learning Emulator for a Distributed Groundwater–Surface Water Model: ParFlow-ML

• DOI: 10.3390/w13233393
• 发表时间: 2021-12
• 期刊: Water
• 影响因子: 3.4
• 作者: Hoang Tran;E. Leonarduzzi;Luis De la Fuente;R. B. Hull;Vineet Bansal;Calla Chennault;P. Gentine;Peter Melchior;L. Condon;R. Maxwell

Sandtank-ML: An Educational Tool at the Interface of Hydrology and Machine Learning

Sandtank-ML：水文学和机器学习接口的教育工具

• DOI: 10.3390/w13233328
• 发表时间: 2021
• 期刊: Water
• 影响因子: 3.4
• 作者: Gallagher, Lisa K.;Williams, Jill M.;Lazzeri, Drew;Chennault, Calla;Jourdain, Sebastien;O’Leary, Patrick;Condon, Laura E.;Maxwell, Reed M.
• 通讯作者: Maxwell, Reed M.