Using local insights to improve climate forecasts

Thrilled to unveil our latest work: multi-modal machine learning to forecast localized weather! We construct a graph neural network to learn dynamics at point locations, where typical gridded forecasts miss significant variation. Paper: https://lnkd.in/eBmfsJin Weather dataset: https://lnkd.in/ejCG8bKs Code: https://lnkd.in/eQg-JzQJ AI weather models have made huge strides, but most still emulate products like ERA5, which struggle to capture near-surface wind dynamics. The correlation between ERA5 and ground weather station data is low due to topography, buildings, vegetation, and other local factors. In this work, we forecast near-surface wind at localized off-grid locations using a message-passing graph neural network ("MPNN"). The graph is heterogeneous, integrating both global forecasts (ERA5) and historical local weather station data as different nodes. What do we find? First off, ERA5 interpolation performs poorly, failing to capture local wind variations, especially in coastal and inland regions with complex conditions. An MLP trained on historical data at a location performs better than ERA5 interpolation, as it learns from the station's past observations. However, it struggles with longer lead times and lacks the spatial context necessary to capture weather patterns. Meanwhile, our MPNN dramatically improves performance, reducing the error by over 50% compared to the MLP. This is because the MPNN incorporates spatial information through message passing, allowing it to learn local weather dynamics from both station data and global forecasts. Interestingly, adding ERA5 data to the MLP does not improve its performance significantly. The MLP struggles to integrate spatial information from global forecasts, while the MPNN excels, highlighting the importance of combining global and local data. Large improvements in forecast accuracy occur at both coastal and inland locations. Our model shows a 92% reduction in MSE relative to ERA5 interpolation overall. This work showcases the strength of machine learning in combining multi-modal data. By using a graph to integrate global and local weather data, we were able to generate much more accurate localized weather forecasts! Congrats to Qidong Yang and Jonathan Giezendanner for the great work, and thanks to Campbell Watson, Daniel Salles Chevitarese, Johannes Jakubik, Eric Schmitt, Anirban C., Jeremy Vila, Detlef Hohl, and Chris Hill for a wonderful collaboration. Thanks also to our partners at Amazon Web Services (AWS) for providing cloud computing and technical support!

Every year, natural disasters hit harder and closer to home. But when city leaders ask, "How will rising heat or wildfire smoke impact my home in 5 years?"—our answers are often vague. Traditional climate models give sweeping predictions, but they fall short at the local level. It's like trying to navigate rush hour using a globe instead of a street map. That’s where generative AI comes in. This year, our team at Google Research built a new genAI method to project climate impacts—taking predictions from the size of a small state to the size of a small city. Our approach provides: - Unprecedented detail – in regional environmental risk assessments at a small fraction of the cost of existing techniques - Higher accuracy – reduced fine-scale errors by over 40% for critical weather variables and reduces error in extreme heat and precipitation projections by over 20% and 10% respectively - Better estimates of complex risks – Demonstrates remarkable skill in capturing complex environmental risks due to regional phenomena, such as wildfire risk from Santa Ana winds, which statistical methods often miss Dynamical-generative downscaling process works in two steps: 1) Physics-based first pass: First, a regional climate model downscales global Earth system data to an intermediate resolution (e.g., 50 km) – much cheaper computationally than going straight to very high resolution. 2) AI adds the fine details: Our AI-based Regional Residual Diffusion-based Downscaling model (“R2D2”) adds realistic, fine-scale details to bring it up to the target high resolution (typically less than 10 km), based on its training on high-resolution weather data. Why does this matter? Governments and utilities need these hyperlocal forecasts to prepare emergency response, invest in infrastructure, and protect vulnerable neighborhoods. And this is just one way AI is turbocharging climate resilience. Our teams at Google are already using AI to forecast floods, detect wildfires in real time, and help the UN respond faster after disasters. The next chapter of climate action means giving every city the tools to see—and shape—their own future. Congratulations Ignacio Lopez Gomez, Tyler Russell MBA, PMP, and teams on this important work! Discover the full details of this breakthrough: https://lnkd.in/g5u_WctW  PNAS Paper: https://lnkd.in/gr7Acz25

🌐 Accurate weather forecasts save lives, protect economies, and enhance community resilience. More data means better forecasts. Yet many regions remain “blind spots.”  The Systematic Observations Financing Facility (SOFF) seeks to close these gaps in the global observation system. I'm excited that new impact experiments carried out by European Centre for Medium-Range Weather Forecasts - ECMWF show that investing in basic weather and climate observations in under-resourced countries improves the accuracy of weather forecasts both locally and globally.   ✅ Africa sees the greatest benefits: Forecast uncertainty decreases by more than 30 percent over Africa with new investments.  ✅ Pacific Islands matter: Forecasts uncertainty decreases by up to 20 percent in the Pacific region.  ✅ Upper-air data is crucial: Radiosonde (weather balloon) data has an outsized impact, especially in the tropics.  ✅ Local investment, global impact: While local improvements are observed over short timeframes (12 hours), forecast improvements extend beyond borders, benefiting people around the world.  More details: 📎 https://lnkd.in/dN-x3-jd Florence Rabier Florian Pappenberger Thomas Asare Markus Repnik World Meteorological Organization

Across the Horn of Africa, climate shocks now unfold as compound crises. The 2020–2023 drought left over 46 million people food insecure and eroded their livelihoods. Before recovery could begin, the 2023–2024 El Niño rains triggered widespread flooding, displacing hundreds of thousands of people. Drought–flood whiplash is no longer exceptional; it is the region’s operating climate. My research with the Jameel Observatory for Food Security Early Action in northern Kenya reveals that pastoralist communities are already adapting to these shifts with remarkable flexibility. From star calendars to animal behaviour and vegetation cues, herders read a rich tapestry of indicators and now complement these with radio forecasts and satellite data. They do not wait for a single forecast or a rigid trigger. Instead, they adjust grazing routes, stagger herd movements, and pool resources as signals evolve. This flexible anticipatory action challenges the dominant model of fixed thresholds and single-event triggers. It shows that forecast information only has value if it is trusted, timely, and open to renegotiation on the ground. Climate Information Services (CIS) enable this agility by translating global climate models into local, impact-based advisories. Regional centres, such as ICPAC, provide seasonal outlooks to guide rangeland management and food security planning. Communities use this information to develop innovative solutions by layering these scientific forecasts onto their own adaptive calendars. Formal Anticipatory Action (AA) frameworks can learn from this. Kenya’s 2024–2029 AA Roadmap is vital. Fundamentally, it will deliver more if it incorporates flexibility by allowing rolling triggers, locally defined indicators, and iterative decision-making, rather than treating early action as a one-off release of funds. The cost of inaction rises with every season. Investing in flexible, forecast-driven anticipatory systems is both fiscally prudent and politically essential. For governments, regional bodies, and development partners, the way forward is clear: move beyond crisis response and embed adaptive, plural, and community-grounded anticipatory action at the heart of policy and planning. In the Horn of Africa’s climate future, acting early and being flexible is the most innovative and cost-effective form of adaptation. Photo courtesy of United Nations Office for Disaster Risk Reduction (UNDRR)