AI Extreme Weather and Climate

About AI Extreme Weather and Climate

Brace yourself for a deep dive into the science of how artificial intelligence is revolutionizing our understanding of extreme weather and climate change. Each episode brings you cutting-edge research and insights on how AI-powered tools are being used to predict and mitigate natural disasters like floods, droughts, and wildfires. We'll unravel the complexities of climate models, explore the frontiers of AI-powered early warning systems, and discuss the ethical implications of AI-driven solutions. Join us as we break down the science and uncover the transformative potential of AI in tackling our planet's most pressing challenges.

Publisher

Zhi Li

Category

education · science

Language

en

Explicit

No

First episode

31 Mar 2025

Latest episode

24 Mar 2026

Latest episodes

14 episodes in the feed.

1. 

   24 Mar 2026

   Target Concept Tuning: Solving the AI Blindspot in Extreme Weather Forecasting

   In this episode of AI Extreme Weather and Climate, Allen and Sydney explore a major breakthrough in meteorological AI: predicting rare but high-impact events like typhoons. While foundation models like Pangu-Weather excel at everyday forecasts, they often stumble during extreme anomalies due to severe data imbalance. We dive into a newly proposed framework called Target Concept Tuning (TaCT) which acts like a "specialized meteorologist" inside the neural network. By using Sparse Autoencoders to untangle superposed features, TaCT automatically identifies the exact internal concepts that cause the model to fail during extreme weather. It then selectively fine-tunes only those specific concepts, dramatically improving typhoon forecasting accuracy without causing the model to "forget" how to predict normal weather patterns. Tune in to learn how making AI more interpretable is making our early warning systems safer and more reliable! Paper Discussed in this Episode: Ren, S., Gu, X., Peng, Z., Zhang, H., Niu, P., Wu, B., Wang, X., Sun, L., & Wen, J. (2026). Target Concept Tuning Improves Extreme Weather Forecasting. arXiv preprint arXiv:2603.19325. https://doi.org/10.48550/arXiv.2603.19325 (https://doi.org/10.48550/arXiv.2603.19325)

2. 

   15 Jan 2026

   NeuralGCM: Observation-Based Hybrid Modeling for Global Precipitation Forecasting

   This paper introduces NeuralGCM, a hybrid atmospheric model that integrates machine learning with traditional differentiable physics to improve global precipitation simulations. Unlike older models that rely on high-resolution simulations for training, this framework is trained directly on satellite observations, specifically the IMERG dataset. By leveraging this observational data, the model effectively corrects common biases in extreme weather events and the diurnal cycle of rainfall. In comparative tests, the model outperformed the ECMWF ensemble in mid-range forecasting and showed superior accuracy over CMIP6 climate models. Additionally, the architecture is exceptionally efficient, running simulations at speeds orders of magnitude faster than conventional general circulation models. These findings suggest that hybrid neural models offer a more reliable and computationally accessible path for predicting future climate impacts.

3. 

   9 Dec 2025

   Flow-Matched Neural Operators for Continuous PDE Dynamics

   Researchers introduce the Continuous Flow Operator, a neural framework that learns continuous-time dynamics of Partial Differential Equations, achieving superior stability and data efficiency across multiple benchmarks.

4. 

   5 Nov 2025

   Ep. 11: Principals of Diffusion Models

   Industry experts discuss diffusion models, detailing their foundational principles through variational, score-based, and flow-based views, and explore advanced techniques for accelerating the slow sampling process.

5. 

   9 Oct 2025

   Ep 10. RainSeer: Physics-Guided Fine-Grained Rainfall Reconstruction

   Researchers introduce RainSeer, a new framework that reconstructs high-resolution rainfall fields using radar reflectivity as a structural prior, improving flood forecasting by capturing localized extremes and sharp transitions in rainfall patterns; this is a research discussion.

6. 

   21 Sept 2025

   Ep. 9: FlowCast-ODE Cntinuous Hourly Weather Forecasting with Dynamic Flow Matching and ODE Integration

   This episode explores FlowCast-ODE, a new deep learning framework that achieves accurate, continuous hourly weather forecasting by modeling atmospheric state evolution as a continuous flow, outperforming existing models.

7. 

   3 Sept 2025

   Ep.8 AQUAH: An Automatic Quantification and Unified Agent in Hydrology

   Host Dr. Sarah Chen interviews Dr. John Smith about AQUAH, a new language-based agent automating hydrologic modeling to improve water resource management.

8. 

   5 Aug 2025

   Ep 7. cBottle: Climate in a bottle - foundational AI weather prediction

   NVIDIA's cBottle uses generative diffusion to compress petabyte-scale climate data by 3000x, enabling low-latency generation of realistic kilometer-scale weather data and interactive digital twins of Earth.

9. 

   30 Jun 2025

   Ep.6 How to fine tune a weather foundation model to hydrological variables?

   Researchers evaluate the Aurora weather foundation model's performance using lightweight decoders to predict hydrological and energy variables, revealing that this approach reduces training time while maintaining strong accuracy.

10. 

    3 Jun 2025

    Ep.5 What is foundation model - drawing from numerical simulation

    Industry experts discuss foundation models, drawing analogies from numerical solutions in fluid dynamics, and explore challenges in building these models for science and engineering.

11. 

    7 May 2025

    Ep.4 Any-to-any Earth Observation Generation and Thinking - TerraMind

    IBM recently released the first-of-its-kind geospatial intelligence any-to-any model TerraMind. In this podcast, we feature this new generative model and learn its capability of multi-modality. I believe there is a lot of potential with such a model. Jakubik, J., Yang, F., Blumenstiel, B., Scheurer, E., Sedona, R., Maurogiovanni, S., Bosmans, J., Dionelis, N., Marsocci, V., Kopp, N., Ramachandran, R., Fraccaro, P., Brunschwiler, T., Cavallaro, G., & Longépé, N. (2025). TerraMind: Large-Scale Generative Multimodality for Earth Observation. ArXiv. https://arxiv.org/abs/2504.11171 (https://arxiv.org/abs/2504.11171)

12. 

    24 Apr 2025

    Ep.3 Geospatial foundation model - Prithvi

    Today, we are featuring a geospatial foundation model Prithvi, produced by NASA and IBM, one of the first foundation model in this space. Trained on a large global dataset of NASA’s Harmonized Landsat and Sentinel-2 data, Prithvi-EO-2.0 demonstrates significant improvements over its predecessor by incorporating temporal and location embeddings. Through extensive benchmarking using GEO-Bench, it outperforms other prominent GFMs across various remote sensing tasks and resolutions, highlighting its versatility. Furthermore, the model has been successfully applied to real-world downstream tasks led by subject matter experts in areas such as disaster response, land use and crop mapping, and ecosystem dynamics monitoring, showcasing its practical utility. Emphasising a Trusted Open Science approach, Prithvi-EO-2.0 is made available on Hugging Face and IBM TerraTorch to facilitate community adoption and customization, aiming to overcome limitations of previous GFMs related to multi-temporality, validation, and ease of use for non-AI experts.

13. 

    15 Apr 2025

    Ep.2 AI models for flood forecasting - HydrographNet

    This research article introduces HydroGraphNet, a novel physics-informed graph neural network for improved flood forecasting. Traditional hydrodynamic models are computationally expensive, while machine learning alternatives often lack physical accuracy and interpretability. HydroGraphNet integrates the Kolmogorov–Arnold Network (KAN) to enhance model interpretability within an unstructured mesh framework. By embedding mass conservation laws into its training and using a specific architecture, the model achieves more physically consistent and accurate predictions. Validation on real-world flood data demonstrates significant reductions in prediction error and improvements in identifying major flood events compared to standard methods. Taghizadeh, M., Zandsalimi, Z., Nabian, M. A., Shafiee-Jood, M., & Alemazkoor, N. Interpretable physics-informed graph neural networks for flood forecasting. Computer-Aided Civil and Infrastructure Engineering. https://doi.org/10.1111/mice.13484 (https://doi.org/10.1111/mice.13484)

14. 

    31 Mar 2025

    Ep.1 AI models for weather forecasting

    We are featuring three papers: Mardani, M., Brenowitz, N., Cohen, Y., Pathak, J., Chen, C., Liu, C., Vahdat, A., Nabian, M. A., Ge, T., Subramaniam, A., Kashinath, K., Kautz, J., & Pritchard, M. (2025). Residual corrective diffusion modeling for km-scale atmospheric downscaling. Communications Earth & Environment, 6(1), 1-10. https://doi.org/10.1038/s43247-025-02042-5 (https://doi.org/10.1038/s43247-025-02042-5) Price, I., Alet, F., Andersson, T. R., Masters, D., Ewalds, T., Stott, J., Mohamed, S., Battaglia, P., Lam, R., & Willson, M. (2025). Probabilistic weather forecasting with machine learning. Nature, 637(8044), 84-90. https://doi.org/10.1038/s41586-024-08252-9 (https://doi.org/10.1038/s41586-024-08252-9) Lam, R., Sanchez-Gonzalez, A., Willson, M., Wirnsberger, P., Fortunato, M., Alet, F., Ravuri, S., Ewalds, T., Eaton-Rosen, Z., Hu, W., Merose, A., Hoyer, S., Holland, G., Vinyals, O., Stott, J., Pritzel, A., Mohamed, S., & Battaglia, P. (2023). Learning skillful medium-range global weather forecasting. Science. https://doi.org/adi2336 (https://doi.org/adi2336)

Reach and audience

Public platform figures. Ratings count people who left a rating, not total listeners.

Apple Podcasts (US)

5.0 / 5

1 ratings

Podcast Authority Score: 11 / 100

A composite of feed quality, social presence, YouTube performance and engagement. Read the methodology.

Quality

22

Social presence

0

YouTube

0

Engagement

0