# Microsoft's Aurora 1.5 Boosts Weather AI _Microsoft's Aurora 1.5 open-source Earth system model now offers 22 more weather variables, hourly forecasts, and ensemble predictions for improved climate and weather applications._ **Updated:** 2026-08-22 **Published:** 2026-07-09 **Source:** https://www.startuphub.ai/ai-news/ai-research/2026/microsoft-s-aurora-1-5-boosts-weather-ai --- Microsoft is pushing its open foundation models for Earth sciences forward with the release of **Aurora 1.5**. This significant update to its Aurora Earth System foundation model introduces 22 additional weather variables, hourly temporal resolution, and a crucial probabilistic ensemble forecasting capability. The model is now accessible to researchers and developers via GitHub and Hugging Face, furthering the goal of [Microsoft Research](https://www.microsoft.com/en-us/research/blog/aurora-1-5-extending-open-foundation-models-for-weather-and-earth-system-applications/)'s open science initiatives. Weather AI NeedsDriverneed for more accurate and detailed weather predictionsFrom the article 9 mentionsThis significant update to its Aurora Earth System foundation model introduces 22 additional weather variables, hourly temporal resolution, and a crucial probabilistic ensemble forecasting capability.addressed byMicrosoft Aurora 1.5Coreopen-source Earth system foundation model updateFrom the article 9+ mentionsMicrosoft is pushing its open foundation models for Earth sciences forward with the release of Aurora 1.5.Expanded VariablesContextadds 22 more weather variables for richer dataFrom the article 2 mentionsThe expanded variable set in Aurora 1.5 now covers a broader spectrum of atmospheric conditions, making it more relevant for industries like energy, agriculture, and transportation, as well as for climate risk assessment.Hourly ForecastsContextprovides forecasts with hourly temporal resolutionFrom the article 6 mentionsThe move to hourly forecasts provides the granular detail needed for precise operational guidance.Ensemble ForecastingCoreprobabilistic predictions showing range of outcomesFrom the article 7 mentionsAurora 1.5 addresses this through its new ensemble forecasting capability, which generates multiple forecast members to better capture the spread of possible future weather events.Open ScienceContextFrom the article 4 mentionsThe model is now accessible to researchers and developers via GitHub and Hugging Face, furthering the goal of Microsoft Research's open science initiatives.Greater ConfidenceEffectenables more informed decisions in various sectorsenablesScalable ApplicationsOutcomeFrom the article 4 mentionsAurora 1.5 aims to make advanced weather forecasting more practical and scalable. Aurora 1.5 aims to make advanced weather forecasting more practical and scalable. The addition of ensemble forecasting, a highly requested feature, allows for multiple forecast simulations to illustrate the range and likelihood of potential outcomes, a critical aspect for applications ranging from energy grids to agriculture. This development underscores the growing importance of robust [weather foundation models](/ai-news/artificial-intelligence/2026/google-deepmind-vp-on-ai-s-future-of-intelligence) in navigating climate-related risks. ## Expanding Weather Intelligence The expanded variable set in Aurora 1.5 now covers a broader spectrum of atmospheric conditions, making it more relevant for industries like energy, agriculture, and transportation, as well as for climate risk assessment. The move to hourly forecasts provides the granular detail needed for precise operational guidance. This iteration builds on the original Aurora model, first introduced in 2024 and detailed in Nature in 2025, which demonstrated a single model's adaptability across various Earth system applications. The open-source release encourages evaluation and further development by the global scientific community. ## Ensemble Forecasting for Greater Confidence Uncertainty is inherent in weather prediction. Aurora 1.5 addresses this through its new ensemble forecasting capability, which generates multiple forecast members to better capture the spread of possible future weather events. This is vital for applications where understanding the probability distribution is as important as the most likely outcome. Microsoft reports that Aurora 1.5's ensemble approach outperforms the state-of-the-art ECMWF ensemble forecast on 88.9% of evaluated targets. The model also showed a substantial reduction in track errors for tropical cyclones, including Hurricane Helene, effectively enveloping verified tracks and demonstrating its skill in predicting high-impact weather. This capability is a meaningful step toward making weather foundation models more practical and useful, offering a clearer path for organizations to evaluate and adapt them. ## Beyond Forecasting: Earth System Modeling Aurora 1.5's utility extends beyond traditional weather forecasting. Companies like Terradot are leveraging Aurora-derived data for carbon dioxide removal estimations, showcasing its potential in climate mitigation and public-interest science. Collaborations with entities such as the UK Met Office are exploring how foundation models can complement existing physics-based systems for faster, more flexible forecasts across weather and climate timescales. This integration of open research with operational use is facilitated through platforms like Microsoft Foundry and Planetary Computer Pro, connecting models with essential data and infrastructure. Early adopters like BKW are already using Aurora 1.5 to support energy operations, managing weather-dependent generation and infrastructure planning. The open-source nature of Aurora 1.5 is intended to empower researchers, agencies, and companies to evaluate and build upon the model, fostering responsible deployment and transparency in forecasting and climate resilience efforts. --- Original analysis from [startuphub.ai](https://www.startuphub.ai), the #1 AI startup directory. © StartupHub.ai. All rights reserved. You may not republish this article in full without a license. Search engines and AI research tools may crawl and summarize for reference. Bulk reproduction or model training on this content requires a license. See https://www.startuphub.ai/terms.