Naoki Mizukami Associate Scientist, National Centre for Atmospheric Research (NCAR), Boulder, United States

Title: CMIP based hydrologic projection in the Pacific Northwest for Hydrologic System Vulnerability Assessments

Abstract: Climate data from the Coupled Model Intercomparison Project-Phase (CMIP) is increasingly used for regional and national water infrastructure planning under future climate scenarios. The climate-hydrologic modeling chain, which produces high-resolution hydrologic projections, begins with selecting multiple Earth System Models (ESMs), followed by downscaling the ESM outputs to kilometer-scale resolution, and using  the downscaled climate data to derive hydrologic model(s). Each step offers various methodological options–such as different ESM selections, downscaling methods, and hydrologic models–enabling producing ensemble, century-long hydro-climate traces for use for water resources impact assessment, such as reservoir simulations. This work aims to produce approximately 30 ensemble hydrologic traces from 1950 to 2099 across river reaches in the US Pacific Northwest. These support vulnerability assessments and resilience planning for many reservoirs of the US Army Corps of Engineers and its regional partners. A key novelty in our modelling chain is the use of a computationally efficient dynamical weather modelling for downscaling, rather than traditional, statistical methods. Additionally, we apply a calibrated, catchment based hydrology-river model to simulate river discharge at ~10 km scale river reaches. This presentation will show evaluation of historical streamflow simulations at ~ 200 reference river flow sites, and characterise future hydrologic variability and extremes. Our main focus is on assessment of projected change in high flow and flow seasonality and its uncertainty at inflow points for major reservoirs in the Pacific Northwest. Our findings suggest that increases in high flow exceed annual or seasonal mean, indicating intensification of extreme flood events under future conditions. We will also demonstrate how these projections are used for reservoir modelling for planning.

Zong-Liang Yang Professor, University of Texas at Austin, United States

Title: Translating Global Climate Models for Local Water Resilience: Insights from Austin’s Water Forward Plan

Abstract: Austin’s Water Forward 2024 is a 100-year integrated water resource plan designed to secure a sustainable water future for a rapidly growing city under the pressures of climate change, drought, and ecological stress. This presentation describes a collaborative framework developed between UT Austin researchers and Austin Water to translate global climate model (GCM) outputs into actionable local water planning strategies. We evaluated 35 CMIP6 GCMs for their ability to represent observed precipitation, temperature, and dry-day characteristics over the Colorado River Basin, selecting five models that best balance performance, independence, and data availability. These models were applied across multiple emission scenarios (SSP1-2.6, SSP2-4.5, SSP5-8.5), then downscaled using neural networks and quantile mapping to estimate streamflow at 45 control points. Results indicate a projected decline in streamflow with increased variability, underscoring risks of extreme events. The Austin Water framework demonstrates how municipalities can integrate climate science into adaptive planning while addressing model uncertainty, resource constraints, and stakeholder needs. This approach provides a replicable model for cities worldwide grappling with water resilience under climate uncertainty.

PLEASE NOTE: Due to technical difficulties, Prof. Chaheds presentation will be re-scheduled to January/February 2026

Jamel Chahed Professor, University of Tunis El Manar, Tunisia

Title: Climate Modelling Beyond Prediction: Reinforcing a Phenomenological Perspective on Uncertainty and Responsibility

Abstract: This presentation offers a reflection on the core arguments and insights developed in the article “Advanced Climate Modeling Frameworks: State-of-the-Art Techniques, Uncertainties, and the Principle of Responsibility” (Modeling Earth Systems and Environment, July 2025). Beyond summarizing technical advancements in GCM-ESM coupling, sub-grid parameterization, downscaling, ensemble models processing, and the integration of data-driven and AI, the presentation will emphasize a key conceptual proposition: the need to reinforce a phenomenological interpretation of climate models. Rather than viewing models merely as predictive machines, they are approached as cognitive instruments shaped by assumptions, simplifications, and epistemological commitments. The talk will explore how this perspective reframes current debates around uncertainty, bias correction, ensemble strategies, and hybrid physical-AI architectures. It argues that phenomenological awareness enhances not only model interpretability but also scientific integrity, especially when model outputs influence policy and risk communication. Several recommendations will be discussed, including: improving transparency in how sub-grid processes are conceptualised, recognising the epistemic status of models as structured approximations rather than empirical realities, and promoting a modelling culture that combines technical rigour with interpretive humility. The presentation aims to stimulate a dialogue on how the CMIP community might further integrate reflexivity, clarity, and responsibility into modelling practices, acknowledging that better science arises not only from better algorithms, but from a deeper understanding of what modelling means.