Research to Operations in Data-driven Hydrologic Forecasting and Decision-making

Water hazards related to droughts, floods, and tropical storms remain amongst the costliest and deadliest natural hazards. For large sections of the U.S., an increase and intensification in these extreme hydroclimate events are predicted. Precise, accurate, early, and actionable forecasts of these hazards are needed to save lives, protect property, and sustain commerce needs.

The hydrologic forecasting research community is responding with new ideas, techniques, and tools driven by recent advances in data science, artificial intelligence (AI), and machine learning (ML). Rapid and effective translation of these research advances are urgently needed in water system operations to help improve water hazard responses and planning.

These advances must be developed and incorporated with the operational weather/water forecasting community and the businesses, industry, and the public that depend on them. Simultaneously, graduate students' training requires greater inter- and transdisciplinary curriculum inclusion. This change in training will allow these future water professionals and leaders in water science to comprehend the practical significance of research advances and develop skills to effectively translate them into forecast and decision-making frameworks used in operational settings.

This NSF Research Traineeship (NRT) project will address the multifaceted and integrated needs of researchers, forecasters, and users of forecasts, by launching a unique hydrologic science program focusing on the critical linkage of research to operations, or "R2O." The program of study will be co-produced with those working in the water prediction community to generate a pipeline of interdisciplinary scientists and engineers capable of diagnosing water-hazard forecasting needs. This co-production effort will result in the design of prediction tools and techniques using the latest advances in AI, ML, and data science, and the dissemination of the forecast products in actionable forms for a wide array of decision-makers.

The project anticipates training a diverse set of 115 master's and Ph.D. students, including 28 funded trainees from civil engineering, geography, and computer science. Student recruitment efforts will focus on groups traditionally underrepresented in their participation in academia and water industries.

This project will expose students to a variety of professional and simulated professional contexts and strengthen student competencies to be facilitators, innovators, and leaders. The training program features unique modalities, content, and delivery to build competency through team science, challenge-based learning, co-production, and iterative self-reflection.

Innovative educational aspects include domestic and international study tours, mock operational forecasting, practical labs, roundtable discussions, mixed mentoring, experiential learning, internships, broad-scale and interdisciplinary team building, and professional development. Graduates of the program will bring to the hydrologic forecasting workforce a unique combination of attributes.

The first two are: (1) deep disciplinary knowledge in hydrologic science coupled with (2) comprehensive skills spanning the cutting edge of AI and ML, and the use of advanced industry-standard software. Third and fourth are: (3) a holistic understanding of the complex hydrologic forecasting and decision-making system paired with (4) the competencies to be thought leaders in the hydrologic forecasting community of practice.

Trainees will accelerate research advances in three focal areas: (1) creating new data science workflows to characterize multi-scale geophysical and climate drivers of hydrologic processes; (2) advancing models and predictive tools to reduce the uncertainty of hydrologic prediction; and (3) improving the communication of forecast products for practical operations. Coordinated internal and external project evaluation from multiple disciplines, open-source software development, data curation, and new pedagogical approaches to train at the interface of computer science, engineering, and geoscience will support the project goals of convergent research and the delivery of broader impacts in academia, government, and the private sector.

This project is jointly funded by the NSF Research Traineeship (NRT) program and the Established Program to Stimulate Competitive Research (EPSCoR).

The NSF Research Traineeship (NRT) Program is designed to encourage the development and implementation of bold, new potentially transformative models for STEM graduate education training. The program is dedicated to effective training of STEM graduate students in high priority interdisciplinary or convergent research areas through comprehensive traineeship models that are innovative, evidence-based, and aligned with changing workforce and research needs.

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.