Daniel Saedi Nia Post Baccalaureate Intern, Oak Ridge National Laboratory, USA

Title: CMIP Data Discovery and Access through ESGF Using Large Language Models

Abstract: The Earth System Grid Federation (ESGF) provides the primary infrastructure for accessing CMIP and other Earth system model datasets by hosting petabytes of data across a globally distributed network. While ESGF enables standardized access to CMIP data, the complexity of metadata conventions can limit efficient data discovery and scientific use, particularly for researchers unfamiliar with CMIP-related metadata structures and terminology. To improve accessibility to CMIP data through ESGF, the ESGF2-US team is developing a domain-specific Large Language Model (LLM) assistant (ESGF-Assistant) that enables natural-language interaction with ESGF data and existing ESGF tools through an easy-to-use interface. Built on an open-source foundation model (Meta’s Llama 3.1 8B), the system is fine-tuned using ESGF and CMIP-specific instruction-output pairs derived from common scientific workflows and documentation. The model responds to natural-language queries by generating ready-to-use Python code leveraging the intake-esgf library, which provides programmatic access to CMIP data through existing ESGF services. This enables easy discovery of relevant CMIP variables, experiments, and datasets while creating reproducible, programmatic workflows. Preliminary results show improved accuracy and contextual alignment with CMIP metadata conventions, as well as more accurate and up-to-date code generation compared to general-purpose LLMs. This work demonstrates how natural-language interfaces can lower barriers to accessing and using CMIP data, enabling more efficient scientific workflows , with future expansion planned for emerging datasets such as CMIP7.

Tom Bearpark Post-Doctoral Fellow, University of Exeter, UK

Title: The economic geography of climate risk

Abstract: Projected temperature changes are variable in both their magnitude and geography. This paper studies how nonlinear damages and general equilibrium forces filter this climate risk across time and space. To do so, we build a tractable dynamic spatial model linking countries through trade and migration, and derive analytical first- and second-order welfare approximations that decompose the mean and variance of welfare changes into damage function, trade, and migration components. Using an ensemble of temperature projections from CMIP-6 and an empirically estimated damage function, we show that climate change-induced welfare risk is large. The standard deviation of country-level projected welfare loss across temperature projections is on average over 8% — compared to an average welfare loss of 13\% — and is spatially unequal. Climate risk inequality is half as large as global income inequality. We show that spatial linkages reshape not only the level of climate damages, but also the spatial distribution of climate risk: accounting for trade and migration can reduce the standard deviation of welfare changes by up to 40% in low-income, internationally integrated nations that can diversify their exposure to local shocks. 

Anastasia Romanou Physical Research Scientist, NASA, Columbia University, USA

Title: TBC

Abstract: TBC