Date: 8th April 2025, Tuesday (10:00am – 11:00am)
Presenter: David John Gagne (NCAR)
Topic: CREDIT: Community Research Earth Digital Intelligence Twin

Abstract:
The Community Research Earth Digital Intelligence Twin (CREDIT) is a new open foundational software platform for developing and deploying AI weather/atmospheric prediction models at time scales spanning from hours to years. CREDIT enables users to build custom data and modeling pipelines to load data, train configurable AI forward models, and deploy them for real-time forecasting, hindcasting, or projecting scenarios. CREDIT builds on the first generation of AI weather prediction models by enabling more customization throughout the whole modeling pipeline from data processing to choice of architecture and training procedure. The CREDIT development team has produced multiple advances in our development and understanding of AI weather prediction: stabilizing training and predictions with spectral normalization, training a 1-hour timestep model competitive with the ECMWF IFS, creating architecture-independent physical constraint layers that improve precipitation predictions, and implementing an emulator of the Community Atmospheric Model, CAMulator, that can produce stable 30-year rollouts at 1 degree grid spacing while reproducing many of the behaviors of the original model. The presentation will cover the overall CREDIT framework as well as results from a selection of projects utilizing the CREDIT framework for model development. Many of the projects use a new architecture developed for CREDIT called WXFormer, which combines elements of the Crossformer and U-Net architectures to learn relevant features across scales with minimal predictive latency. The presentation will showcase verification results for the different CREDIT models, show how they can conserve mass, energy, and moisture budget, and how those advancements connect with forecast improvements. Finally, progress on newer developments in CREDIT, including ensemble generation and regional modeling, will be discussed.

Speaker Profiles:
David John “DJ” Gagne is the head of the MILES (Machine Integration and Learning for Earth Systems) group at the NSF National Center for Atmospheric Research. His research focuses on developing and using advanced machine learning methods for emulation, uncertainty quantification, and physical understanding of atmospheric phenomena across space and time scales. He has also conducted extensive research into the elements of trustworthy AI. David John began conducting machine learning research as an undergraduate intern in 2007 and received his Ph.D. in meteorology from the University of Oklahoma in 2016. He has worked at NSF NCAR since then first as an Advanced Study Program Postdoctoral Fellow and then as a Machine Learning Scientist. He serves as a co-chair of the WMO AI for Nowcasting Pilot Project and on the WMO WWRP Nowcasting and Mesoscale Research working group.

Date: 15th April 2025, Tuesday (4:00pm – 5:00pm)
Presenter: Camilla Mathison (UK Met Office)
Topic: An introduction to the PRIME Earth system emulator and FASTMIP

Abstract:
We present PRIME, a framework for analysis of scenarios of regional impacts for user-prescribed future emissions. PRIME combines global mean temperature and CO2 concentrations from the emissions driven FaIR simple climate model, as used in the IPCC Sixth Assessment Report, with patterns of climate change from CMIP6 Earth System models to drive the JULES land model. This simulation system projects regional changes to the land surface and carbon cycle. We evaluate PRIME by running it with Shared Socioeconomic Pathways and illustrate its robustness by comparing these known scenarios with ESMs that have also been run for the same scenarios. PRIME correctly represents the climate response for these known scenarios, which gives us confidence that PRIME will be useful for rapidly providing probabilistic spatially resolved information for novel climate scenarios; substantially reducing the time between the scenarios being released and being used in impact assessments. Therefore, PRIME fulfils an important need, providing the capability to include the most recent models, science, and scenarios to run ensemble simulations on multi-centennial timescales and include analysis of many variables that are relevant and important for impact assessments. PRIME is part of the FASTMIP activity led by ETH Zurich with strong contributions of the UK Met Office and PNNL, which aims to provide a coordinated experiment of regional emulators for a wide range of scenarios. We will introduce this activity and discuss how these systems tend to be flexible and fast to run and therefore represent a wealth of future development opportunities. We will focus on how PRIME and similar frameworks will enable rapid probabilistic assessment of novel scenarios emissions scenarios that have not yet been run in ESMs thereby providing a useful insight and the capability to quantify societally relevant climate impacts.

Speaker Profiles:
Dr Camilla Mathison is a climate scientist at the UK Met Office. Her current work focuses on developing capabilities using the FaIR simple climate model for rapid assessment of climate scenarios, providing crucial mitigation advice to government customers and supporting Earth system model development. Camilla joined the Met Office in Forecasting research in 2002, where she worked on both Convective-Scale Data Assimilation and the assimilation of ozone. In 2009, she moved to climate research, contributing to European projects including HELIX, AVOID, CSSP Brazil, WCSSP South Africa, ENSEMBLES and WATCH. Subsequently, Camilla worked on simulating the impacts of climate on rivers, irrigation and crops in South Asia using JULES. In 2020, she was responsible for the Met Office contribution to phase 2b of the Inter-Sectoral Impact Model Intercomparison Project (ISIMIP).

Date: 28th April 2025, Monday (11:00am – 12:00pm)
Presenter: Tim Cowan (BoM)
Topic: Insights from the Northern Australia Climate Program: droughts, floods, and hot cows

Abstract:
The Northern Australia Climate Program (or NACP) is a research, development, extension and adoption project, aimed at helping the northern Australian red meat industry manage climate risk. The NACP originated in 2017 as a collaboration between the University of Southern Queensland, the Bureau of Meteorology, UK Met Office and state agencies across Queensland, Northern Territory and Western Australia. In this talk, I will first discuss the framework of NACP (and why it’s a hugely successful R&D&E program!), and then delve into examples of the innovative research and monitoring/forecast products that have been developed over the past 7 years. I will show how flooding in early 2019 led to the development of a prototype rainfall forecast product that was made operational by the Bureau. Other examples of climate tools include the Flash Drought monitor, Green Break of Season (e.g., onset of growing season) and Australia-wide cattle thermal stress forecasts. In the final part of the talk, I will showcase the extension program of NACP and demonstrate how regional climate champions, called Climate Mates, help deliver NACP’s tailored climate service to graziers, farmers, and others in the red meat supply chain. Such an extension and adoption framework can be implemented in other regions and for other commodities (e.g., grains, sugar, dairy).

Speaker Profiles:
Dr Tim Cowan is a Senior Research Fellow at the University of Southern Queensland’s Centre for Applied Climate Sciences, and also part of the Seasonal Climate Processes team at the Australian Bureau of Meteorology in Melbourne. Tim holds a PhD in climate science from the University of New South Wales, and has over 20 years of experience in climate change research, including long-term projections and subseasonal-seasonal prediction. His current research interests include climate modes of variability within the Indo-Pacific (e.g., El Niño-Southern Oscillation, Madden Julian Oscillation), the Australian monsoon, and quantifying thermal stress in livestock. Tim currently leads the Research & Development arm of the Northern Australia Climate Program, and develops innovative prototype forecast products tailored for the northern red meat sector to help graziers better manage climate risks. In 2022, Tim was part of the Forewarned is Forearmed project team that won a Bureau Excellence Award for “leading significant improvements to the Bureau’s climate outlooks service, (and) providing users with critical information to support decision making”. As of 2025, Tim is a newly minted member of the regional Working Group on Asian Australian Monsoons.

Date: 29th April 2025, Tuesday (11:00am – 12:00pm)
Presenter: Wenjun Zhu (NTU)
Topic: Index-based Insurance Design for Weather Risk Management

Abstract:
Weather risk affects economy, agricultural production in particular. Index insurance has gained significant attention as a risk management tool for vulnerable populations in the face of increasing climate variability and extreme weather events. This talk will introduce index-based financial facilities, focusing on index insurance. It introduces a framework that integrates a neural network-based optimization scheme into an expected utility maximization problem to design index insurance contracts. By leveraging neural networks to model the highly nonlinear relationship between high-dimensional weather variables and production losses, the framework endogenously determines the optimal insurance premium and demand. This approach reduces basis risk, lowers insurance premiums, and enhances farmers’ utility. The talk will also highlight future design pathways, emphasizing the transformative potential of cutting-edge technologies such as artificial intelligence and blockchain to revolutionize risk modeling and claims processing.

Speaker Profiles:
Wenjun Zhu (PhD, FSA, CERA), is a tenured Associate Professor at Nanyang Business School, Nanyang Technological University, Singapore. She is the Deputy Head of Division of Banking & Finance, and Deputy Director of the Insurance Risk and Finance Research Centre (IRFRC) at NBS. She is a winner of the Society of Actuaries James C. Hickman Scholar, and the 2023 Actuarial Science Early Career Award. Wenjun’s current research interests include climate change and its risk management solutions, predictive analytics with machine learning, agricultural insurance, and longevity risk management. Wenjun has been publishing in leading academic journals including Management Science, Insurance: Mathematics and Economics, North American Actuarial Journal, Journal of Risk and Insurance, etc. She is an Associate Editor of Annals of Actuarial Science, and an invited expert at the IPCC AR7 Report Scoping Meeting in December 2024.

Date: 13th May 2025, Tuesday (11:00am – 12:00pm)
Presenter: Oliver Watt-Meyer (Allen Institute for Artificial Intelligence)
Topic: The Ai2 Climate Emulator (ACE): Capabilities, Challenges and Opportunities

Abstract:
Ai2 Climate Emulator (ACE) is a fast machine learning model that simulates global atmospheric variability in a changing climate over time scales ranging from hours to centuries. ACE is trained on either a global atmospheric model (AGCM)’s output or on observational reanalysis. It has a 1° horizontal grid with eight vertical layers and 6-hourly temporal resolution. The choice of variables predicted assists in climate interpretability and enables the enforcement of mass and energy conservation constraints. The most recent version, ACE2, simulates about 1500 years per day on a single NVIDIA H100 GPU. When forced by realistic insolation, atmospheric CO2 concentration and specified sea-surface temperature, ACE2 accurately emulates climate trends and ENSO-related interannual variability over the period 1940-2020. When coupled to a slab ocean model, ACE2 accurately emulates the equilibrium climate sensitivity of a similarly coupled AGCM to CO2 change, including spatial patterns of surface temperature and precipitation response. However, comparing ACE2 with out-of-sample transient climate change simulations exposes remaining challenges with radiative forcing and energy conservation. In addition to ACE’s utility as a fast emulator of existing climate models, it is envisioned to also provide predictions at the km-scale necessary for stakeholders. Initial work has trained diffusion-based downscaling models for surface precipitation rate on a global 3-km resolution simulation with NOAA’s X-SHiELD model. Due to ACE’s speed, it will be feasible to combine global climate simulation with downscaling over a user’s region of interest in a single workflow.

Speaker Profiles:
Dr. Oliver Watt-Meyer is a Lead Research Scientist in the Climate Modeling group at the Allen Institute for Artificial Intelligence (Ai2). His research interests are in atmospheric and climate dynamics as well as the application of machine learning to climate prediction. He actively works on developing fast, accurate and easy-to-use climate model emulators using AI. In 2016 he received his PhD from the University of Toronto, focused on the topic of stratosphere-troposphere coupling. From 2016 to 2019, he was a NOAA Climate and Global Change and NSERC Postdoctoral Fellow at the University of Washington Department of Atmospheric Sciences.