Here is a topic for sharing meetings/conferences/workshops of relevance to this community.
AMOS 2025 (Cairns, June 23-27 2025). This working group has submitted a session proposal titled âAdvancing Weather and Climate Research with Machine Learningâ. If anyone has suggestions for a keynote speaker, please let the co-chairs know.
2 Likes
PyCon AU is coming in Melbourne from 22nd to the 26th of November.
Hereâs a direct link to the first day of the program:
This may be of interest with a broad-based interest in Python programming, data science, tools for data science and open source code.
1 Like
SIAM members have access to the SIAG on Mathematics of Planet Earth Community.
Meeting recordings are publicly available on the SIAM MPE Community Meetings YouTube channel.
The 2025 IEEE International Geoscience and Remote Sensing Symposium, 3 - 8 August 2025, Brisbane, Australia will include many topics related to AI and ML for remote sensing, as well as topics related to climate and weather.
Hi all, see below an advert for the AI session at AGOS:
We would like to draw your attention to the forthcoming Asia Oceania Geosciences Society (AOGS) conference âAOGS2025â (27 July -1 August 2025, Singapore).
Among the many sessions related to meteorology and hydrology, there will be an exciting session focusing on AI in Weather and Climate Prediction. While the session will be made up of research papers as is customary in scientific conferences, the aim is to broaden the scope to include discussions on operational meteorological services. Specifically, the organisers are inviting contributions that address concepts, planning, and strategies for applying AI in both R&D and operational contexts, also beyond completed research and development. Many operational services are grappling with the challenges of incorporating or transforming their processes through AI technology to fulfil their missions effectively. This session offers an opportunity to explore common interests and exchange insights that could benefit the community at large.
A call for Papers for the AOGS2025 Session Application of Artificial Intelligence to Weather and Climate Prediction (see https://www.asiaoceania.org/aogs2025/public.asp?page=sessions_and_conveners.asp, Session AS-58) has been issued. The deadline for abstract submission is 18 February 2025. Please see information under âSubmit Abstract â Guidelines and Requirementsâ at https://www.asiaoceania.org/aogs2025/public.asp?page=submissions.asp#SA. The organisers welcome all topics related to AI in weather and climate prediction, including meteorological, computational, and applied aspects.
**Conveners: Prof Chih-Pei Chang (National Taiwan University), Prof Hock Lim (University of Singapore), Prof Melinda Peng (University of Colorado ), Prof Bin Wang (University of Hawaii)
**Invited Speakers (so far):
- Dale Durran, University of Washington, Deep-Learning Earth-System Model
- Takemasa Miyoshi, RIKEN Center for Computational Science (R-CCS), AI-NWP
- Fenghua Ling, Shanghai AI Lab, FengWu Forecast Model
- Shijin Yuan, Tongji University, TianXing Forecast Model
- Xiaohui Zhong, Fudan University, Fuxi Forecast Model
- Guangjun Zhang, UC San Diego, Application of Deep Learning in Model Physics Parameterization
- Yoo-Geun Ham, Seoul National University, Deep Learning for Seasonal Forecast and ENSO
- Yoon-Young Lee, APCC, AI-Based Postprocessing of Climate Prediction
- Bipin Kumar, Indian Institute for Tropical Meteorology, Deep Learning Platform for Hydrological Weather Forecasting at IITM
Not a conference or workshop, but I think people following this will be interested - Anna Vaughan (Aardvark Weather) will be giving a hybrid seminar at the Bureau tomorrow:
Tuesday, 4 February 2025, 11:00 - 12:00
Melbourne Office 9E conference room + MS Teams Software (use the link below)
AI for end-to-end weather prediction
Anna Vaughan, University of Cambridge
Abstract
Machine learning is on the verge of transforming operational weather forecasting. While physics-based models have long been the foundation of numerical weather prediction, recent advances in data-driven approaches such as Pangu-Weather and GenCast have demonstrated superior performance over certain operational baselines. However, these models replace only a single component of the traditional forecasting pipeline, the numerical solver, and are still reliant on external initialization and post-processing. This seminar will introduce Aardvark Weather, the first end-to-end data-driven weather forecasting model which takes raw observations as input and generates both global gridded forecasts and local station forecasts. Global forecasts surpass an operational NWP baseline across multiple variables and lead times, while local station forecasts perform competitively with a post-processed global NWP baseline and a state-of-the-art end-to-end forecasting system that incorporates input from human forecasters. Results from foundation modeling and vision-language model research will be presented, with integration into the Aardvark Weather model planned for 2025. The seminar will conclude with a discussion on the future of AI-driven weather prediction, exploring key challenges, opportunities, and potential developments in the field over the next year.
Speaker Bio
Anna is a PhD student in computer science at the University of Cambridge working on applications of probabilistic machine learning to weather and climate. Prior to this she obtained a masters degree in meteorology with a focus on tropical cyclone forecasting at the University of Melbourne. Her current research explores multiple applications including medium range weather forecasting, tropical cyclone intensity forecasting, on-device learning for satellites and greenhouse gas emission detection using remote sensing. Anna also works part time at the United Nations Environment Programme as a machine learning scientist in the International Methane Observatory team automating detection of super-emitters from satellite imagery.
Important Note:
If you are from outside the Bureau of Meteorology, you will join into the âlobbyâ, a virtual waiting room.
Requests for seminars can be sent to bureau-science-seminars-admin@bom.gov.au.
If you are receiving this email, you are currently subscribed to the BoM Science & Innovation seminar distribution list. If you wish to unsubscribe, please reply to this message with UNSUBSCRIBE in the subject line.
Regards,
Bureau Science Seminar coordinators [Zac Hughes Miller, David Hoffmann, Christian Stassen, and Mohammed Bari]
Microsoft Teams Need help?
Meeting ID: 465 685 074 025
Passcode: nC29KE96
Join on a video conferencing device
Tenant key: bomau@m.webex.com
Video ID: 132 039 768 2
For organizers: Meeting options
Disclaimer: This meeting is hosted by the Bureau of Meteorology. You must join with your full name and follow instruction given by the host. Disclosure or unauthorised use of content from this meeting may be a serious criminal offence.
Wednesday, 12 February 2025, 11:00 - 12:00
MS Teams Software (use the link below)
A High-Resolution AI Weather Model for End-to-End Forecasts from Observations
Haiyu Dong, Microsoft
Abstract
In recent years, Artificial Intelligence Weather Prediction (AIWP) models have achieved performance comparable to, or even surpassing, traditional Numerical Weather Prediction (NWP) models by leveraging reanalysis data. However, a less-explored approach involves training AIWP models directly on observational data, enhancing computational efficiency and improving forecast accuracy by reducing the uncertainties introduced through data assimilation processes. In this study, we propose OMG-HD, a novel AI-based regional high-resolution weather forecasting model designed to make predictions directly from observational data sources, including surface stations, radar, and satellite, thereby removing the need for operational data assimilation. Our evaluation shows that OMG-HD outperforms both the European Centre for Medium-Range Weather Forecasts (ECMWF)'s high-resolution operational forecasting system, IFS-HRES, and the High-Resolution Rapid Refresh (HRRR) model at lead times of up to 12 hours across the contiguous United States (CONUS) region. We achieve up to a 13% improvement on RMSE for 2-meter temperature, 17% on 10-meter wind speed, 48% on 2-meter specific humidity, and 32% on surface pressure compared to HRRR. Our method shows that it is possible to use AI-driven approaches for rapid weather predictions without relying on NWP-derived weather fields as model input. This is a promising step towards using observational data directly to make operational forecasts with AIWP models.
Speaker Bio
Haiyu Dong is Principal Applied Science Manager at Microsoft Weather.
Important Note:
If you are from outside the Bureau of Meteorology, you will join into the âlobbyâ, a virtual waiting room.
Requests for seminars can be sent to bureau-science-seminars-admin@bom.gov.au.
If you are receiving this email, you are currently subscribed to the BoM Science & Innovation seminar distribution list. If you wish to unsubscribe, please reply to this message with UNSUBSCRIBE in the subject line.
Regards,
Bureau Science Seminar coordinators [Zac Hughes Miller, David Hoffmann, Christian Stassen, and Mohammed Bari]
Microsoft Teams Need help?
Meeting ID: 483 337 488 151
Passcode: cT3pH7Uh
Join on a video conferencing device
Tenant key: bomau@m.webex.com
Video ID: 131 359 036 0
For organizers: Meeting options
Disclaimer: This meeting is hosted by the Bureau of Meteorology. You must join with your full name and follow instruction given by the host. Disclosure or unauthorised use of content from this meeting may be a serious criminal offence.