Join Us for the HEPEX 2025 Workshop – Celebrating 20 Years of Advancing Hydrological Forecasting

Keynote talks 

Rob Hartman is a consulting hydrologist who retired as the Hydrologist in Charge (HIC) of the NWS California-Nevada River Forecast Center (CNRFC) in 2016. Rob has 35 years of federal flood and water supply forecasting experience. Rob’s experience includes all aspects of RFC operations, development, and management obtained in positions at three NWS River Forecast Centers, the USDA/Soil Conservation Service, and the NWS Office of Hydrologic Development. As a consultant, Rob has been actively involved in the Forecast Informed Reservoir Operations (FIRO) project, research applications, and helping clients appropriately apply forecasts in their reservoir and flood management schemes.

Title: HEPEX: Meeting the promise and challenges of probabilistic streamflow forecasting
Abstract: HEPEX is a “grass roots” community of global scientists and practitioners who organized themselves 20+ years ago to better address the challenges of ensemble-based probabilistic streamflow forecasting. Without formal foundational funding from established scientific organizations or societies, HEPEX has furthered the development and applications of probabilistic streamflow forecasting on the shoulders of visionary individuals and supportive institutions. This presentation will address where we came from, what’s been accomplished, and what challenges remain ahead for all of us.

Dr Hwirin Kim has led the Hydrological and Water Resources Service Section (HWR) since 2019. She specializes in flood forecasting, integrated water resources management, and policy. With extensive international experience, she has served as a member of the WMO Commission for Hydrology’s Advisory Working Group, Theme Leader for the WMO Regional Association II Working Group on Hydrological Services, and focal point for the UNESCO International Hydrological Programme and the ESCAP/WMO Typhoon Committee.

Prof. Stefan Uhlenbrook’ s main expertise includes water management in a changing climate, hydrological processes and modelling across scales (i.e., floods and droughts), and the water-energy-food-ecosystems nexus. Stefan leads the Hydrology, Water and Cryosphere branch of WMO, the UN Specialized Agency for Weather, Climate and Water. He is excited to contribute to the Organisation’s program and, among other things, strengthen the early warning systems for floods and droughts as well as water resources assessment and management worldwide.

Title: Global efforts to enhance hydrological forecasting
Abstract: According to the WMO State of the Climate 2024 report, the global mean surface air temperature from January to September 2024 was 1.54°C above the pre-industrial average, with a margin of uncertainty of ±0.13°C. This record-breaking warming, exacerbated by a strong El Niño event, positions 2023 and 2024 as the two warmest years in the 175-year observational record. Extreme weather events in 2024, including widespread flooding and droughts, have caused significant losses and damages globally. Notable examples include droughts across the Americas (Mexico, Central America, and South America), northwestern and southern Africa, and severe flooding in the Sahel, Europe, and East Africa.

In response, WMO has intensified its support to countries, aiming to minimize these impacts through the development technical guidelines, manuals and implementation of hydrological forecasting systems (usually through extra-budgetary projects). Capacity building and enhanced cooperation among global, regional, transboundary and national stakeholders remain central to these efforts. The Assessment Guidelines for End-to-End (E2E) Flood Forecasting and Early Warning Systems (FF&EWS) provide National Hydrological Services (NHSs) worldwide with a structured framework to evaluate their capabilities and enhance their ability to deliver accurate, timely, and actionable flood forecasts. Additionally, the Inventory of Interoperable Models and Platforms for FF & EWS offers valuable guidance for scientists and engineers in NHSs selecting suitable models and platforms, supported by screening criteria to ensure success. A Community of Practice are for E2E FF&EWS is under development to facilitate the exchange of knowledge and experiences. Multiple hydrological models, platforms and systems have been successfully implemented by the WMO Members, particularly Least Developed Countries, supported by external financial and technical assistance. Emerging technologies, including artificial intelligence (AI), the metaverse, and digital twins, are being actively explored to advance hydrological forecasting capabilities. An ongoing AI pilot study in four volunteer countries is generating valuable insights, while case studies on these innovative technologies have been collected to evaluate their potential for operational implementation. To ensure sustainability, the Research-to-Operation (R2O) plan has been initiated, promoting the seamless integration of research advancements into operational systems. The plan fosters a collaborative development culture, ensuring broader and more inclusive participation form the global hydrological community.

This approach underscores WMO’s commitment to strengthening hydrological forecasting systems, building resilience to extreme weather impacts, and supporting informed decision-making at all levels. Combined with the experiences and successes of HEPEX during the last 20+ years, this provides a unique opportunity for tangible impact in the developing world in the future.

Jonathan Frame is Assistant Professor at the University of Alabama in the Department of Geological Sciences. Jonathan also holds a Faculty Fellowship with the Alabama Water Institute and is an affiliate researcher at the National Water Center. Before embarking on an academic career Jonathan was a practicing civil engineer in California designing water resources infrastructure. His research interests span beyond hydrology into fluid dynamics of geophysical systems.

Title: Machine learning advances, from theory to applications to operations, improved predictions and understanding our physical world
Abstract: Machine learning and Artificial Intelligence (AI/ML) has a long and rich history in hydrology, and as recently begun dominating operational forecasting. In fact, AI/ML has recently changed the whole landscape of hydrology spanning beyond academia, into every facet of water resources science and engineering through government and industry. This talk reviews AI/ML approaches for hydrology including ensemble modeling, model selection, and uncertainty quantification. We discuss recent work on ensemble hydrologic prediction, and how AI/ML took ensemble prediction to a new level. We examine performance across distributed catchments, using distribution-based evaluation. We explore the latest in physics-informed machine learning, differentiable conceptual models (δ-CFE) and their unique advantages over “pure” AI/ML, as well as their disadvantages. Geospatial AI/ML advances are extending hydrologic prediction to new scales and in new directions. Generative AI offers new approaches for uncertainty quantification and probabilistic scenarios for hydrologic forecasting. This talk outlines AI’s potential for ensemble prediction, addressing uncertainty through ML-driven model selection, distribution-based evaluation, and generative AI.

Deadline for registration

24 January 2025